* [PATCH RFC 01/18] OvmfPkg: Add public headers from Xen Project.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver Anthony PERARD
` (26 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This patch imports publics headers in order to use features from Xen
like XenStore, PV Block... There is only the necessary header files and
there are only a few modifications in order to facilitate future merge of
more recent header (that would be necessary to access new features).
There is little modification compared to the original files:
- Use of ZeroMem() instead of memset()
- Replace types to be more UEFI compliant using a script.
Command to run to change types:
find OvmfPkg/Include/IndustryStandard/Xen -type f -name '*.h' -exec sed
--regexp-extended --file=fix_type_in_xen_includes.sed --in-place {} \;
This line is commented instead of been change as I'm not sure why it
does not compile (when s/char/CHAR8/), and it does not seems necessary
so far.
/* __DEFINE_XEN_GUEST_HANDLE(uchar, unsigned char); */
in OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen.h
Avoid changing the 'long' that is not a type (with the first line).
$ cat fix_type_in_xen_includes.sed
/as long as/b
s/([^a-zA-Z0-9_]|^)uint8_t([^a-zA-Z0-9_]|$)/\1UINT8\2/g
s/([^a-zA-Z0-9_]|^)uint16_t([^a-zA-Z0-9_]|$)/\1UINT16\2/g
s/([^a-zA-Z0-9_]|^)uint32_t([^a-zA-Z0-9_]|$)/\1UINT32\2/g
s/([^a-zA-Z0-9_]|^)uint64_t([^a-zA-Z0-9_]|$)/\1UINT64\2/g
s/([^a-zA-Z0-9_]|^)int8_t([^a-zA-Z0-9_]|$)/\1INT8\2/g
s/([^a-zA-Z0-9_]|^)int16_t([^a-zA-Z0-9_]|$)/\1INT16\2/g
s/([^a-zA-Z0-9_]|^)int32_t([^a-zA-Z0-9_]|$)/\1INT32\2/g
s/([^a-zA-Z0-9_]|^)int64_t([^a-zA-Z0-9_]|$)/\1INT64\2/g
s/([^a-zA-Z0-9_]|^)void([^a-zA-Z0-9_]|$)/\1VOID\2/g
s/([^a-zA-Z0-9_]|^)unsigned int([^a-zA-Z0-9_]|$)/\1UINT32\2/g
s/([^a-zA-Z0-9_]|^)int([^a-zA-Z0-9_]|$)/\1INT32\2/g
s/([^a-zA-Z0-9_]|^)char([^a-zA-Z0-9_]|$)/\1CHAR8\2/g
s/([^a-zA-Z0-9_]|^)unsigned long([^a-zA-Z0-9_]|$)/\1UINTN\2/g
s/([^a-zA-Z0-9_]|^)long([^a-zA-Z0-9_]|$)/\1INTN\2/g
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
.../IndustryStandard/Xen/arch-x86/xen-x86_32.h | 171 ++++
.../IndustryStandard/Xen/arch-x86/xen-x86_64.h | 202 +++++
.../Include/IndustryStandard/Xen/arch-x86/xen.h | 273 +++++++
.../Include/IndustryStandard/Xen/event_channel.h | 381 +++++++++
OvmfPkg/Include/IndustryStandard/Xen/grant_table.h | 662 +++++++++++++++
OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h | 275 +++++++
OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h | 150 ++++
OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h | 608 ++++++++++++++
.../Include/IndustryStandard/Xen/io/protocols.h | 40 +
OvmfPkg/Include/IndustryStandard/Xen/io/ring.h | 312 +++++++
OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h | 80 ++
OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h | 138 ++++
OvmfPkg/Include/IndustryStandard/Xen/memory.h | 480 +++++++++++
OvmfPkg/Include/IndustryStandard/Xen/sched.h | 174 ++++
OvmfPkg/Include/IndustryStandard/Xen/trace.h | 310 +++++++
OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h | 44 +
OvmfPkg/Include/IndustryStandard/Xen/xen.h | 897 +++++++++++++++++++++
17 files changed, 5197 insertions(+)
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_32.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_64.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/event_channel.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/grant_table.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/protocols.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/ring.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/memory.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/sched.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/trace.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h
create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/xen.h
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_32.h b/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_32.h
new file mode 100644
index 0000000..98f2f03
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_32.h
@@ -0,0 +1,171 @@
+/******************************************************************************
+ * xen-x86_32.h
+ *
+ * Guest OS interface to x86 32-bit Xen.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2007, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_XEN_X86_32_H__
+#define __XEN_PUBLIC_ARCH_X86_XEN_X86_32_H__
+
+/*
+ * Hypercall interface:
+ * Input: %ebx, %ecx, %edx, %esi, %edi, %ebp (arguments 1-6)
+ * Output: %eax
+ * Access is via hypercall page (set up by guest loader or via a Xen MSR):
+ * call hypercall_page + hypercall-number * 32
+ * Clobbered: Argument registers (e.g., 2-arg hypercall clobbers %ebx,%ecx)
+ */
+
+/*
+ * These flat segments are in the Xen-private section of every GDT. Since these
+ * are also present in the initial GDT, many OSes will be able to avoid
+ * installing their own GDT.
+ */
+#define FLAT_RING1_CS 0xe019 /* GDT index 259 */
+#define FLAT_RING1_DS 0xe021 /* GDT index 260 */
+#define FLAT_RING1_SS 0xe021 /* GDT index 260 */
+#define FLAT_RING3_CS 0xe02b /* GDT index 261 */
+#define FLAT_RING3_DS 0xe033 /* GDT index 262 */
+#define FLAT_RING3_SS 0xe033 /* GDT index 262 */
+
+#define FLAT_KERNEL_CS FLAT_RING1_CS
+#define FLAT_KERNEL_DS FLAT_RING1_DS
+#define FLAT_KERNEL_SS FLAT_RING1_SS
+#define FLAT_USER_CS FLAT_RING3_CS
+#define FLAT_USER_DS FLAT_RING3_DS
+#define FLAT_USER_SS FLAT_RING3_SS
+
+#define __HYPERVISOR_VIRT_START_PAE 0xF5800000
+#define __MACH2PHYS_VIRT_START_PAE 0xF5800000
+#define __MACH2PHYS_VIRT_END_PAE 0xF6800000
+#define HYPERVISOR_VIRT_START_PAE \
+ mk_unsigned_long(__HYPERVISOR_VIRT_START_PAE)
+#define MACH2PHYS_VIRT_START_PAE \
+ mk_unsigned_long(__MACH2PHYS_VIRT_START_PAE)
+#define MACH2PHYS_VIRT_END_PAE \
+ mk_unsigned_long(__MACH2PHYS_VIRT_END_PAE)
+
+/* Non-PAE bounds are obsolete. */
+#define __HYPERVISOR_VIRT_START_NONPAE 0xFC000000
+#define __MACH2PHYS_VIRT_START_NONPAE 0xFC000000
+#define __MACH2PHYS_VIRT_END_NONPAE 0xFC400000
+#define HYPERVISOR_VIRT_START_NONPAE \
+ mk_unsigned_long(__HYPERVISOR_VIRT_START_NONPAE)
+#define MACH2PHYS_VIRT_START_NONPAE \
+ mk_unsigned_long(__MACH2PHYS_VIRT_START_NONPAE)
+#define MACH2PHYS_VIRT_END_NONPAE \
+ mk_unsigned_long(__MACH2PHYS_VIRT_END_NONPAE)
+
+#define __HYPERVISOR_VIRT_START __HYPERVISOR_VIRT_START_PAE
+#define __MACH2PHYS_VIRT_START __MACH2PHYS_VIRT_START_PAE
+#define __MACH2PHYS_VIRT_END __MACH2PHYS_VIRT_END_PAE
+
+#ifndef HYPERVISOR_VIRT_START
+#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START)
+#endif
+
+#define MACH2PHYS_VIRT_START mk_unsigned_long(__MACH2PHYS_VIRT_START)
+#define MACH2PHYS_VIRT_END mk_unsigned_long(__MACH2PHYS_VIRT_END)
+#define MACH2PHYS_NR_ENTRIES ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>2)
+#ifndef machine_to_phys_mapping
+#define machine_to_phys_mapping ((UINTN *)MACH2PHYS_VIRT_START)
+#endif
+
+/* 32-/64-bit invariability for control interfaces (domctl/sysctl). */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+#undef ___DEFINE_XEN_GUEST_HANDLE
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+ typedef struct { type *p; } \
+ __guest_handle_ ## name; \
+ typedef struct { union { type *p; uint64_aligned_t q; }; } \
+ __guest_handle_64_ ## name
+#undef set_xen_guest_handle_raw
+#define set_xen_guest_handle_raw(hnd, val) \
+ do { if ( sizeof(hnd) == 8 ) *(UINT64 *)&(hnd) = 0; \
+ (hnd).p = val; \
+ } while ( 0 )
+#define uint64_aligned_t UINT64 __attribute__((aligned(8)))
+#define __XEN_GUEST_HANDLE_64(name) __guest_handle_64_ ## name
+#define XEN_GUEST_HANDLE_64(name) __XEN_GUEST_HANDLE_64(name)
+#endif
+
+#ifndef __ASSEMBLY__
+
+struct cpu_user_regs {
+ UINT32 ebx;
+ UINT32 ecx;
+ UINT32 edx;
+ UINT32 esi;
+ UINT32 edi;
+ UINT32 ebp;
+ UINT32 eax;
+ UINT16 error_code; /* private */
+ UINT16 entry_vector; /* private */
+ UINT32 eip;
+ UINT16 cs;
+ UINT8 saved_upcall_mask;
+ UINT8 _pad0;
+ UINT32 eflags; /* eflags.IF == !saved_upcall_mask */
+ UINT32 esp;
+ UINT16 ss, _pad1;
+ UINT16 es, _pad2;
+ UINT16 ds, _pad3;
+ UINT16 fs, _pad4;
+ UINT16 gs, _pad5;
+};
+typedef struct cpu_user_regs cpu_user_regs_t;
+DEFINE_XEN_GUEST_HANDLE(cpu_user_regs_t);
+
+/*
+ * Page-directory addresses above 4GB do not fit into architectural %cr3.
+ * When accessing %cr3, or equivalent field in vcpu_guest_context, guests
+ * must use the following accessor macros to pack/unpack valid MFNs.
+ */
+#define xen_pfn_to_cr3(pfn) (((unsigned)(pfn) << 12) | ((unsigned)(pfn) >> 20))
+#define xen_cr3_to_pfn(cr3) (((unsigned)(cr3) >> 12) | ((unsigned)(cr3) << 20))
+
+struct arch_vcpu_info {
+ UINTN cr2;
+ UINTN pad[5]; /* sizeof(vcpu_info_t) == 64 */
+};
+typedef struct arch_vcpu_info arch_vcpu_info_t;
+
+struct xen_callback {
+ UINTN cs;
+ UINTN eip;
+};
+typedef struct xen_callback xen_callback_t;
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* __XEN_PUBLIC_ARCH_X86_XEN_X86_32_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_64.h b/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_64.h
new file mode 100644
index 0000000..7fcf68d
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_64.h
@@ -0,0 +1,202 @@
+/******************************************************************************
+ * xen-x86_64.h
+ *
+ * Guest OS interface to x86 64-bit Xen.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2006, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_XEN_X86_64_H__
+#define __XEN_PUBLIC_ARCH_X86_XEN_X86_64_H__
+
+/*
+ * Hypercall interface:
+ * Input: %rdi, %rsi, %rdx, %r10, %r8, %r9 (arguments 1-6)
+ * Output: %rax
+ * Access is via hypercall page (set up by guest loader or via a Xen MSR):
+ * call hypercall_page + hypercall-number * 32
+ * Clobbered: argument registers (e.g., 2-arg hypercall clobbers %rdi,%rsi)
+ */
+
+/*
+ * 64-bit segment selectors
+ * These flat segments are in the Xen-private section of every GDT. Since these
+ * are also present in the initial GDT, many OSes will be able to avoid
+ * installing their own GDT.
+ */
+
+#define FLAT_RING3_CS32 0xe023 /* GDT index 260 */
+#define FLAT_RING3_CS64 0xe033 /* GDT index 261 */
+#define FLAT_RING3_DS32 0xe02b /* GDT index 262 */
+#define FLAT_RING3_DS64 0x0000 /* NULL selector */
+#define FLAT_RING3_SS32 0xe02b /* GDT index 262 */
+#define FLAT_RING3_SS64 0xe02b /* GDT index 262 */
+
+#define FLAT_KERNEL_DS64 FLAT_RING3_DS64
+#define FLAT_KERNEL_DS32 FLAT_RING3_DS32
+#define FLAT_KERNEL_DS FLAT_KERNEL_DS64
+#define FLAT_KERNEL_CS64 FLAT_RING3_CS64
+#define FLAT_KERNEL_CS32 FLAT_RING3_CS32
+#define FLAT_KERNEL_CS FLAT_KERNEL_CS64
+#define FLAT_KERNEL_SS64 FLAT_RING3_SS64
+#define FLAT_KERNEL_SS32 FLAT_RING3_SS32
+#define FLAT_KERNEL_SS FLAT_KERNEL_SS64
+
+#define FLAT_USER_DS64 FLAT_RING3_DS64
+#define FLAT_USER_DS32 FLAT_RING3_DS32
+#define FLAT_USER_DS FLAT_USER_DS64
+#define FLAT_USER_CS64 FLAT_RING3_CS64
+#define FLAT_USER_CS32 FLAT_RING3_CS32
+#define FLAT_USER_CS FLAT_USER_CS64
+#define FLAT_USER_SS64 FLAT_RING3_SS64
+#define FLAT_USER_SS32 FLAT_RING3_SS32
+#define FLAT_USER_SS FLAT_USER_SS64
+
+#define __HYPERVISOR_VIRT_START 0xFFFF800000000000
+#define __HYPERVISOR_VIRT_END 0xFFFF880000000000
+#define __MACH2PHYS_VIRT_START 0xFFFF800000000000
+#define __MACH2PHYS_VIRT_END 0xFFFF804000000000
+
+#ifndef HYPERVISOR_VIRT_START
+#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START)
+#define HYPERVISOR_VIRT_END mk_unsigned_long(__HYPERVISOR_VIRT_END)
+#endif
+
+#define MACH2PHYS_VIRT_START mk_unsigned_long(__MACH2PHYS_VIRT_START)
+#define MACH2PHYS_VIRT_END mk_unsigned_long(__MACH2PHYS_VIRT_END)
+#define MACH2PHYS_NR_ENTRIES ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>3)
+#ifndef machine_to_phys_mapping
+#define machine_to_phys_mapping ((UINTN *)HYPERVISOR_VIRT_START)
+#endif
+
+/*
+ * INT32 HYPERVISOR_set_segment_base(UINT32 which, UINTN base)
+ * @which == SEGBASE_* ; @base == 64-bit base address
+ * Returns 0 on success.
+ */
+#define SEGBASE_FS 0
+#define SEGBASE_GS_USER 1
+#define SEGBASE_GS_KERNEL 2
+#define SEGBASE_GS_USER_SEL 3 /* Set user %gs specified in base[15:0] */
+
+/*
+ * INT32 HYPERVISOR_iret(VOID)
+ * All arguments are on the kernel stack, in the following format.
+ * Never returns if successful. Current kernel context is lost.
+ * The saved CS is mapped as follows:
+ * RING0 -> RING3 kernel mode.
+ * RING1 -> RING3 kernel mode.
+ * RING2 -> RING3 kernel mode.
+ * RING3 -> RING3 user mode.
+ * However RING0 indicates that the guest kernel should return to iteself
+ * directly with
+ * orb $3,1*8(%rsp)
+ * iretq
+ * If flags contains VGCF_in_syscall:
+ * Restore RAX, RIP, RFLAGS, RSP.
+ * Discard R11, RCX, CS, SS.
+ * Otherwise:
+ * Restore RAX, R11, RCX, CS:RIP, RFLAGS, SS:RSP.
+ * All other registers are saved on hypercall entry and restored to user.
+ */
+/* Guest exited in SYSCALL context? Return to guest with SYSRET? */
+#define _VGCF_in_syscall 8
+#define VGCF_in_syscall (1<<_VGCF_in_syscall)
+#define VGCF_IN_SYSCALL VGCF_in_syscall
+
+#ifndef __ASSEMBLY__
+
+struct iret_context {
+ /* Top of stack (%rsp at point of hypercall). */
+ UINT64 rax, r11, rcx, flags, rip, cs, rflags, rsp, ss;
+ /* Bottom of iret stack frame. */
+};
+
+#if defined(__GNUC__) && !defined(__STRICT_ANSI__)
+/* Anonymous union includes both 32- and 64-bit names (e.g., eax/rax). */
+#define __DECL_REG(name) union { \
+ UINT64 r ## name, e ## name; \
+ UINT32 _e ## name; \
+}
+#else
+/* Non-gcc sources must always use the proper 64-bit name (e.g., rax). */
+#define __DECL_REG(name) UINT64 r ## name
+#endif
+
+struct cpu_user_regs {
+ UINT64 r15;
+ UINT64 r14;
+ UINT64 r13;
+ UINT64 r12;
+ __DECL_REG(bp);
+ __DECL_REG(bx);
+ UINT64 r11;
+ UINT64 r10;
+ UINT64 r9;
+ UINT64 r8;
+ __DECL_REG(ax);
+ __DECL_REG(cx);
+ __DECL_REG(dx);
+ __DECL_REG(si);
+ __DECL_REG(di);
+ UINT32 error_code; /* private */
+ UINT32 entry_vector; /* private */
+ __DECL_REG(ip);
+ UINT16 cs, _pad0[1];
+ UINT8 saved_upcall_mask;
+ UINT8 _pad1[3];
+ __DECL_REG(flags); /* rflags.IF == !saved_upcall_mask */
+ __DECL_REG(sp);
+ UINT16 ss, _pad2[3];
+ UINT16 es, _pad3[3];
+ UINT16 ds, _pad4[3];
+ UINT16 fs, _pad5[3]; /* Non-zero => takes precedence over fs_base. */
+ UINT16 gs, _pad6[3]; /* Non-zero => takes precedence over gs_base_usr. */
+};
+typedef struct cpu_user_regs cpu_user_regs_t;
+DEFINE_XEN_GUEST_HANDLE(cpu_user_regs_t);
+
+#undef __DECL_REG
+
+#define xen_pfn_to_cr3(pfn) ((UINTN)(pfn) << 12)
+#define xen_cr3_to_pfn(cr3) ((UINTN)(cr3) >> 12)
+
+struct arch_vcpu_info {
+ UINTN cr2;
+ UINTN pad; /* sizeof(vcpu_info_t) == 64 */
+};
+typedef struct arch_vcpu_info arch_vcpu_info_t;
+
+typedef UINTN xen_callback_t;
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* __XEN_PUBLIC_ARCH_X86_XEN_X86_64_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen.h b/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen.h
new file mode 100644
index 0000000..e2906f0
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen.h
@@ -0,0 +1,273 @@
+/******************************************************************************
+ * arch-x86/xen.h
+ *
+ * Guest OS interface to x86 Xen.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2006, K A Fraser
+ */
+
+#include "../xen.h"
+
+#ifndef __XEN_PUBLIC_ARCH_X86_XEN_H__
+#define __XEN_PUBLIC_ARCH_X86_XEN_H__
+
+/* Structural guest handles introduced in 0x00030201. */
+#if __XEN_INTERFACE_VERSION__ >= 0x00030201
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+ typedef struct { type *p; } __guest_handle_ ## name
+#else
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+ typedef type * __guest_handle_ ## name
+#endif
+
+/*
+ * XEN_GUEST_HANDLE represents a guest pointer, when passed as a field
+ * in a struct in memory.
+ * XEN_GUEST_HANDLE_PARAM represent a guest pointer, when passed as an
+ * hypercall argument.
+ * XEN_GUEST_HANDLE_PARAM and XEN_GUEST_HANDLE are the same on X86 but
+ * they might not be on other architectures.
+ */
+#define __DEFINE_XEN_GUEST_HANDLE(name, type) \
+ ___DEFINE_XEN_GUEST_HANDLE(name, type); \
+ ___DEFINE_XEN_GUEST_HANDLE(const_##name, const type)
+#define DEFINE_XEN_GUEST_HANDLE(name) __DEFINE_XEN_GUEST_HANDLE(name, name)
+#define __XEN_GUEST_HANDLE(name) __guest_handle_ ## name
+#define XEN_GUEST_HANDLE(name) __XEN_GUEST_HANDLE(name)
+#define XEN_GUEST_HANDLE_PARAM(name) XEN_GUEST_HANDLE(name)
+#define set_xen_guest_handle_raw(hnd, val) do { (hnd).p = val; } while (0)
+#ifdef __XEN_TOOLS__
+#define get_xen_guest_handle(val, hnd) do { val = (hnd).p; } while (0)
+#endif
+#define set_xen_guest_handle(hnd, val) set_xen_guest_handle_raw(hnd, val)
+
+#if defined(__i386__)
+#include "xen-x86_32.h"
+#elif defined(__x86_64__)
+#include "xen-x86_64.h"
+#endif
+
+#ifndef __ASSEMBLY__
+typedef UINTN xen_pfn_t;
+#define PRI_xen_pfn "lx"
+#endif
+
+#define XEN_HAVE_PV_GUEST_ENTRY 1
+
+#define XEN_HAVE_PV_UPCALL_MASK 1
+
+/*
+ * `incontents 200 segdesc Segment Descriptor Tables
+ */
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_set_gdt(const xen_pfn_t frames[], UINT32 entries);
+ * `
+ */
+/*
+ * A number of GDT entries are reserved by Xen. These are not situated at the
+ * start of the GDT because some stupid OSes export hard-coded selector values
+ * in their ABI. These hard-coded values are always near the start of the GDT,
+ * so Xen places itself out of the way, at the far end of the GDT.
+ *
+ * NB The LDT is set using the MMUEXT_SET_LDT op of HYPERVISOR_mmuext_op
+ */
+#define FIRST_RESERVED_GDT_PAGE 14
+#define FIRST_RESERVED_GDT_BYTE (FIRST_RESERVED_GDT_PAGE * 4096)
+#define FIRST_RESERVED_GDT_ENTRY (FIRST_RESERVED_GDT_BYTE / 8)
+
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_update_descriptor(u64 pa, u64 desc);
+ * `
+ * ` @pa The machine physical address of the descriptor to
+ * ` update. Must be either a descriptor page or writable.
+ * ` @desc The descriptor value to update, in the same format as a
+ * ` native descriptor table entry.
+ */
+
+/* Maximum number of virtual CPUs in legacy multi-processor guests. */
+#define XEN_LEGACY_MAX_VCPUS 32
+
+#ifndef __ASSEMBLY__
+
+typedef UINTN xen_ulong_t;
+#define PRI_xen_ulong "lx"
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_stack_switch(UINTN ss, UINTN esp);
+ * `
+ * Sets the stack segment and pointer for the current vcpu.
+ */
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_set_trap_table(const struct trap_info traps[]);
+ * `
+ */
+/*
+ * Send an array of these to HYPERVISOR_set_trap_table().
+ * Terminate the array with a sentinel entry, with traps[].address==0.
+ * The privilege level specifies which modes may enter a trap via a software
+ * interrupt. On x86/64, since rings 1 and 2 are unavailable, we allocate
+ * privilege levels as follows:
+ * Level == 0: Noone may enter
+ * Level == 1: Kernel may enter
+ * Level == 2: Kernel may enter
+ * Level == 3: Everyone may enter
+ */
+#define TI_GET_DPL(_ti) ((_ti)->flags & 3)
+#define TI_GET_IF(_ti) ((_ti)->flags & 4)
+#define TI_SET_DPL(_ti,_dpl) ((_ti)->flags |= (_dpl))
+#define TI_SET_IF(_ti,_if) ((_ti)->flags |= ((!!(_if))<<2))
+struct trap_info {
+ UINT8 vector; /* exception vector */
+ UINT8 flags; /* 0-3: privilege level; 4: clear event enable? */
+ UINT16 cs; /* code selector */
+ UINTN address; /* code offset */
+};
+typedef struct trap_info trap_info_t;
+DEFINE_XEN_GUEST_HANDLE(trap_info_t);
+
+typedef UINT64 tsc_timestamp_t; /* RDTSC timestamp */
+
+/*
+ * The following is all CPU context. Note that the fpu_ctxt block is filled
+ * in by FXSAVE if the CPU has feature FXSR; otherwise FSAVE is used.
+ *
+ * Also note that when calling DOMCTL_setvcpucontext and VCPU_initialise
+ * for HVM and PVH guests, not all information in this structure is updated:
+ *
+ * - For HVM guests, the structures read include: fpu_ctxt (if
+ * VGCT_I387_VALID is set), flags, user_regs, debugreg[*]
+ *
+ * - PVH guests are the same as HVM guests, but additionally use ctrlreg[3] to
+ * set cr3. All other fields not used should be set to 0.
+ */
+struct vcpu_guest_context {
+ /* FPU registers come first so they can be aligned for FXSAVE/FXRSTOR. */
+ struct { CHAR8 x[512]; } fpu_ctxt; /* User-level FPU registers */
+#define VGCF_I387_VALID (1<<0)
+#define VGCF_IN_KERNEL (1<<2)
+#define _VGCF_i387_valid 0
+#define VGCF_i387_valid (1<<_VGCF_i387_valid)
+#define _VGCF_in_kernel 2
+#define VGCF_in_kernel (1<<_VGCF_in_kernel)
+#define _VGCF_failsafe_disables_events 3
+#define VGCF_failsafe_disables_events (1<<_VGCF_failsafe_disables_events)
+#define _VGCF_syscall_disables_events 4
+#define VGCF_syscall_disables_events (1<<_VGCF_syscall_disables_events)
+#define _VGCF_online 5
+#define VGCF_online (1<<_VGCF_online)
+ UINTN flags; /* VGCF_* flags */
+ struct cpu_user_regs user_regs; /* User-level CPU registers */
+ struct trap_info trap_ctxt[256]; /* Virtual IDT */
+ UINTN ldt_base, ldt_ents; /* LDT (linear address, # ents) */
+ UINTN gdt_frames[16], gdt_ents; /* GDT (machine frames, # ents) */
+ UINTN kernel_ss, kernel_sp; /* Virtual TSS (only SS1/SP1) */
+ /* NB. User pagetable on x86/64 is placed in ctrlreg[1]. */
+ UINTN ctrlreg[8]; /* CR0-CR7 (control registers) */
+ UINTN debugreg[8]; /* DB0-DB7 (debug registers) */
+#ifdef __i386__
+ UINTN event_callback_cs; /* CS:EIP of event callback */
+ UINTN event_callback_eip;
+ UINTN failsafe_callback_cs; /* CS:EIP of failsafe callback */
+ UINTN failsafe_callback_eip;
+#else
+ UINTN event_callback_eip;
+ UINTN failsafe_callback_eip;
+#ifdef __XEN__
+ union {
+ UINTN syscall_callback_eip;
+ struct {
+ UINT32 event_callback_cs; /* compat CS of event cb */
+ UINT32 failsafe_callback_cs; /* compat CS of failsafe cb */
+ };
+ };
+#else
+ UINTN syscall_callback_eip;
+#endif
+#endif
+ UINTN vm_assist; /* VMASST_TYPE_* bitmap */
+#ifdef __x86_64__
+ /* Segment base addresses. */
+ UINT64 fs_base;
+ UINT64 gs_base_kernel;
+ UINT64 gs_base_user;
+#endif
+};
+typedef struct vcpu_guest_context vcpu_guest_context_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_guest_context_t);
+
+struct arch_shared_info {
+ UINTN max_pfn; /* max pfn that appears in table */
+ /* Frame containing list of mfns containing list of mfns containing p2m. */
+ xen_pfn_t pfn_to_mfn_frame_list_list;
+ UINTN nmi_reason;
+ UINT64 pad[32];
+};
+typedef struct arch_shared_info arch_shared_info_t;
+
+#endif /* !__ASSEMBLY__ */
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_fpu_taskswitch(INT32 set);
+ * `
+ * Sets (if set!=0) or clears (if set==0) CR0.TS.
+ */
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_set_debugreg(INT32 regno, UINTN value);
+ *
+ * ` UINTN
+ * ` HYPERVISOR_get_debugreg(INT32 regno);
+ * For 0<=reg<=7, returns the debug register value.
+ * For other values of reg, returns ((UINTN)-EINVAL).
+ * (Unfortunately, this interface is defective.)
+ */
+
+/*
+ * Prefix forces emulation of some non-trapping instructions.
+ * Currently only CPUID.
+ */
+#ifdef __ASSEMBLY__
+#define XEN_EMULATE_PREFIX .byte 0x0f,0x0b,0x78,0x65,0x6e ;
+#define XEN_CPUID XEN_EMULATE_PREFIX cpuid
+#else
+#define XEN_EMULATE_PREFIX ".byte 0x0f,0x0b,0x78,0x65,0x6e ; "
+#define XEN_CPUID XEN_EMULATE_PREFIX "cpuid"
+#endif
+
+#endif /* __XEN_PUBLIC_ARCH_X86_XEN_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/event_channel.h b/OvmfPkg/Include/IndustryStandard/Xen/event_channel.h
new file mode 100644
index 0000000..a45c65c
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/event_channel.h
@@ -0,0 +1,381 @@
+/******************************************************************************
+ * event_channel.h
+ *
+ * Event channels between domains.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2003-2004, K A Fraser.
+ */
+
+#ifndef __XEN_PUBLIC_EVENT_CHANNEL_H__
+#define __XEN_PUBLIC_EVENT_CHANNEL_H__
+
+#include "xen.h"
+
+/*
+ * `incontents 150 evtchn Event Channels
+ *
+ * Event channels are the basic primitive provided by Xen for event
+ * notifications. An event is the Xen equivalent of a hardware
+ * interrupt. They essentially store one bit of information, the event
+ * of interest is signalled by transitioning this bit from 0 to 1.
+ *
+ * Notifications are received by a guest via an upcall from Xen,
+ * indicating when an event arrives (setting the bit). Further
+ * notifications are masked until the bit is cleared again (therefore,
+ * guests must check the value of the bit after re-enabling event
+ * delivery to ensure no missed notifications).
+ *
+ * Event notifications can be masked by setting a flag; this is
+ * equivalent to disabling interrupts and can be used to ensure
+ * atomicity of certain operations in the guest kernel.
+ *
+ * Event channels are represented by the evtchn_* fields in
+ * struct shared_info and struct vcpu_info.
+ */
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_event_channel_op(enum event_channel_op cmd, VOID *args)
+ * `
+ * @cmd == EVTCHNOP_* (event-channel operation).
+ * @args == struct evtchn_* Operation-specific extra arguments (NULL if none).
+ */
+
+/* ` enum event_channel_op { // EVTCHNOP_* => struct evtchn_* */
+#define EVTCHNOP_bind_interdomain 0
+#define EVTCHNOP_bind_virq 1
+#define EVTCHNOP_bind_pirq 2
+#define EVTCHNOP_close 3
+#define EVTCHNOP_send 4
+#define EVTCHNOP_status 5
+#define EVTCHNOP_alloc_unbound 6
+#define EVTCHNOP_bind_ipi 7
+#define EVTCHNOP_bind_vcpu 8
+#define EVTCHNOP_unmask 9
+#define EVTCHNOP_reset 10
+#define EVTCHNOP_init_control 11
+#define EVTCHNOP_expand_array 12
+#define EVTCHNOP_set_priority 13
+/* ` } */
+
+typedef UINT32 evtchn_port_t;
+DEFINE_XEN_GUEST_HANDLE(evtchn_port_t);
+
+/*
+ * EVTCHNOP_alloc_unbound: Allocate a port in domain <dom> and mark as
+ * accepting interdomain bindings from domain <remote_dom>. A fresh port
+ * is allocated in <dom> and returned as <port>.
+ * NOTES:
+ * 1. If the caller is unprivileged then <dom> must be DOMID_SELF.
+ * 2. <rdom> may be DOMID_SELF, allowing loopback connections.
+ */
+struct evtchn_alloc_unbound {
+ /* IN parameters */
+ domid_t dom, remote_dom;
+ /* OUT parameters */
+ evtchn_port_t port;
+};
+typedef struct evtchn_alloc_unbound evtchn_alloc_unbound_t;
+
+/*
+ * EVTCHNOP_bind_interdomain: Construct an interdomain event channel between
+ * the calling domain and <remote_dom>. <remote_dom,remote_port> must identify
+ * a port that is unbound and marked as accepting bindings from the calling
+ * domain. A fresh port is allocated in the calling domain and returned as
+ * <local_port>.
+ *
+ * In case the peer domain has already tried to set our event channel
+ * pending, before it was bound, EVTCHNOP_bind_interdomain always sets
+ * the local event channel pending.
+ *
+ * The usual pattern of use, in the guest's upcall (or subsequent
+ * handler) is as follows: (Re-enable the event channel for subsequent
+ * signalling and then) check for the existence of whatever condition
+ * is being waited for by other means, and take whatever action is
+ * needed (if any).
+ *
+ * NOTES:
+ * 1. <remote_dom> may be DOMID_SELF, allowing loopback connections.
+ */
+struct evtchn_bind_interdomain {
+ /* IN parameters. */
+ domid_t remote_dom;
+ evtchn_port_t remote_port;
+ /* OUT parameters. */
+ evtchn_port_t local_port;
+};
+typedef struct evtchn_bind_interdomain evtchn_bind_interdomain_t;
+
+/*
+ * EVTCHNOP_bind_virq: Bind a local event channel to VIRQ <irq> on specified
+ * vcpu.
+ * NOTES:
+ * 1. Virtual IRQs are classified as per-vcpu or global. See the VIRQ list
+ * in xen.h for the classification of each VIRQ.
+ * 2. Global VIRQs must be allocated on VCPU0 but can subsequently be
+ * re-bound via EVTCHNOP_bind_vcpu.
+ * 3. Per-vcpu VIRQs may be bound to at most one event channel per vcpu.
+ * The allocated event channel is bound to the specified vcpu and the
+ * binding cannot be changed.
+ */
+struct evtchn_bind_virq {
+ /* IN parameters. */
+ UINT32 virq; /* enum virq */
+ UINT32 vcpu;
+ /* OUT parameters. */
+ evtchn_port_t port;
+};
+typedef struct evtchn_bind_virq evtchn_bind_virq_t;
+
+/*
+ * EVTCHNOP_bind_pirq: Bind a local event channel to a real IRQ (PIRQ <irq>).
+ * NOTES:
+ * 1. A physical IRQ may be bound to at most one event channel per domain.
+ * 2. Only a sufficiently-privileged domain may bind to a physical IRQ.
+ */
+struct evtchn_bind_pirq {
+ /* IN parameters. */
+ UINT32 pirq;
+#define BIND_PIRQ__WILL_SHARE 1
+ UINT32 flags; /* BIND_PIRQ__* */
+ /* OUT parameters. */
+ evtchn_port_t port;
+};
+typedef struct evtchn_bind_pirq evtchn_bind_pirq_t;
+
+/*
+ * EVTCHNOP_bind_ipi: Bind a local event channel to receive events.
+ * NOTES:
+ * 1. The allocated event channel is bound to the specified vcpu. The binding
+ * may not be changed.
+ */
+struct evtchn_bind_ipi {
+ UINT32 vcpu;
+ /* OUT parameters. */
+ evtchn_port_t port;
+};
+typedef struct evtchn_bind_ipi evtchn_bind_ipi_t;
+
+/*
+ * EVTCHNOP_close: Close a local event channel <port>. If the channel is
+ * interdomain then the remote end is placed in the unbound state
+ * (EVTCHNSTAT_unbound), awaiting a new connection.
+ */
+struct evtchn_close {
+ /* IN parameters. */
+ evtchn_port_t port;
+};
+typedef struct evtchn_close evtchn_close_t;
+
+/*
+ * EVTCHNOP_send: Send an event to the remote end of the channel whose local
+ * endpoint is <port>.
+ */
+struct evtchn_send {
+ /* IN parameters. */
+ evtchn_port_t port;
+};
+typedef struct evtchn_send evtchn_send_t;
+
+/*
+ * EVTCHNOP_status: Get the current status of the communication channel which
+ * has an endpoint at <dom, port>.
+ * NOTES:
+ * 1. <dom> may be specified as DOMID_SELF.
+ * 2. Only a sufficiently-privileged domain may obtain the status of an event
+ * channel for which <dom> is not DOMID_SELF.
+ */
+struct evtchn_status {
+ /* IN parameters */
+ domid_t dom;
+ evtchn_port_t port;
+ /* OUT parameters */
+#define EVTCHNSTAT_closed 0 /* Channel is not in use. */
+#define EVTCHNSTAT_unbound 1 /* Channel is waiting interdom connection.*/
+#define EVTCHNSTAT_interdomain 2 /* Channel is connected to remote domain. */
+#define EVTCHNSTAT_pirq 3 /* Channel is bound to a phys IRQ line. */
+#define EVTCHNSTAT_virq 4 /* Channel is bound to a virtual IRQ line */
+#define EVTCHNSTAT_ipi 5 /* Channel is bound to a virtual IPI line */
+ UINT32 status;
+ UINT32 vcpu; /* VCPU to which this channel is bound. */
+ union {
+ struct {
+ domid_t dom;
+ } unbound; /* EVTCHNSTAT_unbound */
+ struct {
+ domid_t dom;
+ evtchn_port_t port;
+ } interdomain; /* EVTCHNSTAT_interdomain */
+ UINT32 pirq; /* EVTCHNSTAT_pirq */
+ UINT32 virq; /* EVTCHNSTAT_virq */
+ } u;
+};
+typedef struct evtchn_status evtchn_status_t;
+
+/*
+ * EVTCHNOP_bind_vcpu: Specify which vcpu a channel should notify when an
+ * event is pending.
+ * NOTES:
+ * 1. IPI-bound channels always notify the vcpu specified at bind time.
+ * This binding cannot be changed.
+ * 2. Per-VCPU VIRQ channels always notify the vcpu specified at bind time.
+ * This binding cannot be changed.
+ * 3. All other channels notify vcpu0 by default. This default is set when
+ * the channel is allocated (a port that is freed and subsequently reused
+ * has its binding reset to vcpu0).
+ */
+struct evtchn_bind_vcpu {
+ /* IN parameters. */
+ evtchn_port_t port;
+ UINT32 vcpu;
+};
+typedef struct evtchn_bind_vcpu evtchn_bind_vcpu_t;
+
+/*
+ * EVTCHNOP_unmask: Unmask the specified local event-channel port and deliver
+ * a notification to the appropriate VCPU if an event is pending.
+ */
+struct evtchn_unmask {
+ /* IN parameters. */
+ evtchn_port_t port;
+};
+typedef struct evtchn_unmask evtchn_unmask_t;
+
+/*
+ * EVTCHNOP_reset: Close all event channels associated with specified domain.
+ * NOTES:
+ * 1. <dom> may be specified as DOMID_SELF.
+ * 2. Only a sufficiently-privileged domain may specify other than DOMID_SELF.
+ */
+struct evtchn_reset {
+ /* IN parameters. */
+ domid_t dom;
+};
+typedef struct evtchn_reset evtchn_reset_t;
+
+/*
+ * EVTCHNOP_init_control: initialize the control block for the FIFO ABI.
+ *
+ * Note: any events that are currently pending will not be resent and
+ * will be lost. Guests should call this before binding any event to
+ * avoid losing any events.
+ */
+struct evtchn_init_control {
+ /* IN parameters. */
+ UINT64 control_gfn;
+ UINT32 offset;
+ UINT32 vcpu;
+ /* OUT parameters. */
+ UINT8 link_bits;
+ UINT8 _pad[7];
+};
+typedef struct evtchn_init_control evtchn_init_control_t;
+
+/*
+ * EVTCHNOP_expand_array: add an additional page to the event array.
+ */
+struct evtchn_expand_array {
+ /* IN parameters. */
+ UINT64 array_gfn;
+};
+typedef struct evtchn_expand_array evtchn_expand_array_t;
+
+/*
+ * EVTCHNOP_set_priority: set the priority for an event channel.
+ */
+struct evtchn_set_priority {
+ /* IN parameters. */
+ UINT32 port;
+ UINT32 priority;
+};
+typedef struct evtchn_set_priority evtchn_set_priority_t;
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_event_channel_op_compat(struct evtchn_op *op)
+ * `
+ * Superceded by new event_channel_op() hypercall since 0x00030202.
+ */
+struct evtchn_op {
+ UINT32 cmd; /* enum event_channel_op */
+ union {
+ struct evtchn_alloc_unbound alloc_unbound;
+ struct evtchn_bind_interdomain bind_interdomain;
+ struct evtchn_bind_virq bind_virq;
+ struct evtchn_bind_pirq bind_pirq;
+ struct evtchn_bind_ipi bind_ipi;
+ struct evtchn_close close;
+ struct evtchn_send send;
+ struct evtchn_status status;
+ struct evtchn_bind_vcpu bind_vcpu;
+ struct evtchn_unmask unmask;
+ } u;
+};
+typedef struct evtchn_op evtchn_op_t;
+DEFINE_XEN_GUEST_HANDLE(evtchn_op_t);
+
+/*
+ * 2-level ABI
+ */
+
+#define EVTCHN_2L_NR_CHANNELS (sizeof(xen_ulong_t) * sizeof(xen_ulong_t) * 64)
+
+/*
+ * FIFO ABI
+ */
+
+/* Events may have priorities from 0 (highest) to 15 (lowest). */
+#define EVTCHN_FIFO_PRIORITY_MAX 0
+#define EVTCHN_FIFO_PRIORITY_DEFAULT 7
+#define EVTCHN_FIFO_PRIORITY_MIN 15
+
+#define EVTCHN_FIFO_MAX_QUEUES (EVTCHN_FIFO_PRIORITY_MIN + 1)
+
+typedef UINT32 event_word_t;
+
+#define EVTCHN_FIFO_PENDING 31
+#define EVTCHN_FIFO_MASKED 30
+#define EVTCHN_FIFO_LINKED 29
+#define EVTCHN_FIFO_BUSY 28
+
+#define EVTCHN_FIFO_LINK_BITS 17
+#define EVTCHN_FIFO_LINK_MASK ((1 << EVTCHN_FIFO_LINK_BITS) - 1)
+
+#define EVTCHN_FIFO_NR_CHANNELS (1 << EVTCHN_FIFO_LINK_BITS)
+
+struct evtchn_fifo_control_block {
+ UINT32 ready;
+ UINT32 _rsvd;
+ UINT32 head[EVTCHN_FIFO_MAX_QUEUES];
+};
+typedef struct evtchn_fifo_control_block evtchn_fifo_control_block_t;
+
+#endif /* __XEN_PUBLIC_EVENT_CHANNEL_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/grant_table.h b/OvmfPkg/Include/IndustryStandard/Xen/grant_table.h
new file mode 100644
index 0000000..f46b48d
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/grant_table.h
@@ -0,0 +1,662 @@
+/******************************************************************************
+ * grant_table.h
+ *
+ * Interface for granting foreign access to page frames, and receiving
+ * page-ownership transfers.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_GRANT_TABLE_H__
+#define __XEN_PUBLIC_GRANT_TABLE_H__
+
+#include "xen.h"
+
+/*
+ * `incontents 150 gnttab Grant Tables
+ *
+ * Xen's grant tables provide a generic mechanism to memory sharing
+ * between domains. This shared memory interface underpins the split
+ * device drivers for block and network IO.
+ *
+ * Each domain has its own grant table. This is a data structure that
+ * is shared with Xen; it allows the domain to tell Xen what kind of
+ * permissions other domains have on its pages. Entries in the grant
+ * table are identified by grant references. A grant reference is an
+ * integer, which indexes into the grant table. It acts as a
+ * capability which the grantee can use to perform operations on the
+ * granter’s memory.
+ *
+ * This capability-based system allows shared-memory communications
+ * between unprivileged domains. A grant reference also encapsulates
+ * the details of a shared page, removing the need for a domain to
+ * know the real machine address of a page it is sharing. This makes
+ * it possible to share memory correctly with domains running in
+ * fully virtualised memory.
+ */
+
+/***********************************
+ * GRANT TABLE REPRESENTATION
+ */
+
+/* Some rough guidelines on accessing and updating grant-table entries
+ * in a concurrency-safe manner. For more information, Linux contains a
+ * reference implementation for guest OSes (drivers/xen/grant_table.c, see
+ * http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=blob;f=drivers/xen/grant-table.c;hb=HEAD
+ *
+ * NB. WMB is a no-op on current-generation x86 processors. However, a
+ * compiler barrier will still be required.
+ *
+ * Introducing a valid entry into the grant table:
+ * 1. Write ent->domid.
+ * 2. Write ent->frame:
+ * GTF_permit_access: Frame to which access is permitted.
+ * GTF_accept_transfer: Pseudo-phys frame slot being filled by new
+ * frame, or zero if none.
+ * 3. Write memory barrier (WMB).
+ * 4. Write ent->flags, inc. valid type.
+ *
+ * Invalidating an unused GTF_permit_access entry:
+ * 1. flags = ent->flags.
+ * 2. Observe that !(flags & (GTF_reading|GTF_writing)).
+ * 3. Check result of SMP-safe CMPXCHG(&ent->flags, flags, 0).
+ * NB. No need for WMB as reuse of entry is control-dependent on success of
+ * step 3, and all architectures guarantee ordering of ctrl-dep writes.
+ *
+ * Invalidating an in-use GTF_permit_access entry:
+ * This cannot be done directly. Request assistance from the domain controller
+ * which can set a timeout on the use of a grant entry and take necessary
+ * action. (NB. This is not yet implemented!).
+ *
+ * Invalidating an unused GTF_accept_transfer entry:
+ * 1. flags = ent->flags.
+ * 2. Observe that !(flags & GTF_transfer_committed). [*]
+ * 3. Check result of SMP-safe CMPXCHG(&ent->flags, flags, 0).
+ * NB. No need for WMB as reuse of entry is control-dependent on success of
+ * step 3, and all architectures guarantee ordering of ctrl-dep writes.
+ * [*] If GTF_transfer_committed is set then the grant entry is 'committed'.
+ * The guest must /not/ modify the grant entry until the address of the
+ * transferred frame is written. It is safe for the guest to spin waiting
+ * for this to occur (detect by observing GTF_transfer_completed in
+ * ent->flags).
+ *
+ * Invalidating a committed GTF_accept_transfer entry:
+ * 1. Wait for (ent->flags & GTF_transfer_completed).
+ *
+ * Changing a GTF_permit_access from writable to read-only:
+ * Use SMP-safe CMPXCHG to set GTF_readonly, while checking !GTF_writing.
+ *
+ * Changing a GTF_permit_access from read-only to writable:
+ * Use SMP-safe bit-setting instruction.
+ */
+
+/*
+ * Reference to a grant entry in a specified domain's grant table.
+ */
+typedef UINT32 grant_ref_t;
+
+/*
+ * A grant table comprises a packed array of grant entries in one or more
+ * page frames shared between Xen and a guest.
+ * [XEN]: This field is written by Xen and read by the sharing guest.
+ * [GST]: This field is written by the guest and read by Xen.
+ */
+
+/*
+ * Version 1 of the grant table entry structure is maintained purely
+ * for backwards compatibility. New guests should use version 2.
+ */
+#if __XEN_INTERFACE_VERSION__ < 0x0003020a
+#define grant_entry_v1 grant_entry
+#define grant_entry_v1_t grant_entry_t
+#endif
+struct grant_entry_v1 {
+ /* GTF_xxx: various type and flag information. [XEN,GST] */
+ UINT16 flags;
+ /* The domain being granted foreign privileges. [GST] */
+ domid_t domid;
+ /*
+ * GTF_permit_access: Frame that @domid is allowed to map and access. [GST]
+ * GTF_accept_transfer: Frame whose ownership transferred by @domid. [XEN]
+ */
+ UINT32 frame;
+};
+typedef struct grant_entry_v1 grant_entry_v1_t;
+
+/* The first few grant table entries will be preserved across grant table
+ * version changes and may be pre-populated at domain creation by tools.
+ */
+#define GNTTAB_NR_RESERVED_ENTRIES 8
+#define GNTTAB_RESERVED_CONSOLE 0
+#define GNTTAB_RESERVED_XENSTORE 1
+
+/*
+ * Type of grant entry.
+ * GTF_invalid: This grant entry grants no privileges.
+ * GTF_permit_access: Allow @domid to map/access @frame.
+ * GTF_accept_transfer: Allow @domid to transfer ownership of one page frame
+ * to this guest. Xen writes the page number to @frame.
+ * GTF_transitive: Allow @domid to transitively access a subrange of
+ * @trans_grant in @trans_domid. No mappings are allowed.
+ */
+#define GTF_invalid (0U<<0)
+#define GTF_permit_access (1U<<0)
+#define GTF_accept_transfer (2U<<0)
+#define GTF_transitive (3U<<0)
+#define GTF_type_mask (3U<<0)
+
+/*
+ * Subflags for GTF_permit_access.
+ * GTF_readonly: Restrict @domid to read-only mappings and accesses. [GST]
+ * GTF_reading: Grant entry is currently mapped for reading by @domid. [XEN]
+ * GTF_writing: Grant entry is currently mapped for writing by @domid. [XEN]
+ * GTF_PAT, GTF_PWT, GTF_PCD: (x86) cache attribute flags for the grant [GST]
+ * GTF_sub_page: Grant access to only a subrange of the page. @domid
+ * will only be allowed to copy from the grant, and not
+ * map it. [GST]
+ */
+#define _GTF_readonly (2)
+#define GTF_readonly (1U<<_GTF_readonly)
+#define _GTF_reading (3)
+#define GTF_reading (1U<<_GTF_reading)
+#define _GTF_writing (4)
+#define GTF_writing (1U<<_GTF_writing)
+#define _GTF_PWT (5)
+#define GTF_PWT (1U<<_GTF_PWT)
+#define _GTF_PCD (6)
+#define GTF_PCD (1U<<_GTF_PCD)
+#define _GTF_PAT (7)
+#define GTF_PAT (1U<<_GTF_PAT)
+#define _GTF_sub_page (8)
+#define GTF_sub_page (1U<<_GTF_sub_page)
+
+/*
+ * Subflags for GTF_accept_transfer:
+ * GTF_transfer_committed: Xen sets this flag to indicate that it is committed
+ * to transferring ownership of a page frame. When a guest sees this flag
+ * it must /not/ modify the grant entry until GTF_transfer_completed is
+ * set by Xen.
+ * GTF_transfer_completed: It is safe for the guest to spin-wait on this flag
+ * after reading GTF_transfer_committed. Xen will always write the frame
+ * address, followed by ORing this flag, in a timely manner.
+ */
+#define _GTF_transfer_committed (2)
+#define GTF_transfer_committed (1U<<_GTF_transfer_committed)
+#define _GTF_transfer_completed (3)
+#define GTF_transfer_completed (1U<<_GTF_transfer_completed)
+
+/*
+ * Version 2 grant table entries. These fulfil the same role as
+ * version 1 entries, but can represent more complicated operations.
+ * Any given domain will have either a version 1 or a version 2 table,
+ * and every entry in the table will be the same version.
+ *
+ * The interface by which domains use grant references does not depend
+ * on the grant table version in use by the other domain.
+ */
+#if __XEN_INTERFACE_VERSION__ >= 0x0003020a
+/*
+ * Version 1 and version 2 grant entries share a common prefix. The
+ * fields of the prefix are documented as part of struct
+ * grant_entry_v1.
+ */
+struct grant_entry_header {
+ UINT16 flags;
+ domid_t domid;
+};
+typedef struct grant_entry_header grant_entry_header_t;
+
+/*
+ * Version 2 of the grant entry structure.
+ */
+union grant_entry_v2 {
+ grant_entry_header_t hdr;
+
+ /*
+ * This member is used for V1-style full page grants, where either:
+ *
+ * -- hdr.type is GTF_accept_transfer, or
+ * -- hdr.type is GTF_permit_access and GTF_sub_page is not set.
+ *
+ * In that case, the frame field has the same semantics as the
+ * field of the same name in the V1 entry structure.
+ */
+ struct {
+ grant_entry_header_t hdr;
+ UINT32 pad0;
+ UINT64 frame;
+ } full_page;
+
+ /*
+ * If the grant type is GTF_grant_access and GTF_sub_page is set,
+ * @domid is allowed to access bytes [@page_off,@page_off+@length)
+ * in frame @frame.
+ */
+ struct {
+ grant_entry_header_t hdr;
+ UINT16 page_off;
+ UINT16 length;
+ UINT64 frame;
+ } sub_page;
+
+ /*
+ * If the grant is GTF_transitive, @domid is allowed to use the
+ * grant @gref in domain @trans_domid, as if it was the local
+ * domain. Obviously, the transitive access must be compatible
+ * with the original grant.
+ *
+ * The current version of Xen does not allow transitive grants
+ * to be mapped.
+ */
+ struct {
+ grant_entry_header_t hdr;
+ domid_t trans_domid;
+ UINT16 pad0;
+ grant_ref_t gref;
+ } transitive;
+
+ UINT32 __spacer[4]; /* Pad to a power of two */
+};
+typedef union grant_entry_v2 grant_entry_v2_t;
+
+typedef UINT16 grant_status_t;
+
+#endif /* __XEN_INTERFACE_VERSION__ */
+
+/***********************************
+ * GRANT TABLE QUERIES AND USES
+ */
+
+/* ` enum neg_errnoval
+ * ` HYPERVISOR_grant_table_op(enum grant_table_op cmd,
+ * ` VOID *args,
+ * ` UINT32 count)
+ * `
+ *
+ * @args points to an array of a per-command data structure. The array
+ * has @count members
+ */
+
+/* ` enum grant_table_op { // GNTTABOP_* => struct gnttab_* */
+#define GNTTABOP_map_grant_ref 0
+#define GNTTABOP_unmap_grant_ref 1
+#define GNTTABOP_setup_table 2
+#define GNTTABOP_dump_table 3
+#define GNTTABOP_transfer 4
+#define GNTTABOP_copy 5
+#define GNTTABOP_query_size 6
+#define GNTTABOP_unmap_and_replace 7
+#if __XEN_INTERFACE_VERSION__ >= 0x0003020a
+#define GNTTABOP_set_version 8
+#define GNTTABOP_get_status_frames 9
+#define GNTTABOP_get_version 10
+#define GNTTABOP_swap_grant_ref 11
+#endif /* __XEN_INTERFACE_VERSION__ */
+/* ` } */
+
+/*
+ * Handle to track a mapping created via a grant reference.
+ */
+typedef UINT32 grant_handle_t;
+
+/*
+ * GNTTABOP_map_grant_ref: Map the grant entry (<dom>,<ref>) for access
+ * by devices and/or host CPUs. If successful, <handle> is a tracking number
+ * that must be presented later to destroy the mapping(s). On error, <handle>
+ * is a negative status code.
+ * NOTES:
+ * 1. If GNTMAP_device_map is specified then <dev_bus_addr> is the address
+ * via which I/O devices may access the granted frame.
+ * 2. If GNTMAP_host_map is specified then a mapping will be added at
+ * either a host virtual address in the current address space, or at
+ * a PTE at the specified machine address. The type of mapping to
+ * perform is selected through the GNTMAP_contains_pte flag, and the
+ * address is specified in <host_addr>.
+ * 3. Mappings should only be destroyed via GNTTABOP_unmap_grant_ref. If a
+ * host mapping is destroyed by other means then it is *NOT* guaranteed
+ * to be accounted to the correct grant reference!
+ */
+struct gnttab_map_grant_ref {
+ /* IN parameters. */
+ UINT64 host_addr;
+ UINT32 flags; /* GNTMAP_* */
+ grant_ref_t ref;
+ domid_t dom;
+ /* OUT parameters. */
+ INT16 status; /* => enum grant_status */
+ grant_handle_t handle;
+ UINT64 dev_bus_addr;
+};
+typedef struct gnttab_map_grant_ref gnttab_map_grant_ref_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_map_grant_ref_t);
+
+/*
+ * GNTTABOP_unmap_grant_ref: Destroy one or more grant-reference mappings
+ * tracked by <handle>. If <host_addr> or <dev_bus_addr> is zero, that
+ * field is ignored. If non-zero, they must refer to a device/host mapping
+ * that is tracked by <handle>
+ * NOTES:
+ * 1. The call may fail in an undefined manner if either mapping is not
+ * tracked by <handle>.
+ * 3. After executing a batch of unmaps, it is guaranteed that no stale
+ * mappings will remain in the device or host TLBs.
+ */
+struct gnttab_unmap_grant_ref {
+ /* IN parameters. */
+ UINT64 host_addr;
+ UINT64 dev_bus_addr;
+ grant_handle_t handle;
+ /* OUT parameters. */
+ INT16 status; /* => enum grant_status */
+};
+typedef struct gnttab_unmap_grant_ref gnttab_unmap_grant_ref_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_unmap_grant_ref_t);
+
+/*
+ * GNTTABOP_setup_table: Set up a grant table for <dom> comprising at least
+ * <nr_frames> pages. The frame addresses are written to the <frame_list>.
+ * Only <nr_frames> addresses are written, even if the table is larger.
+ * NOTES:
+ * 1. <dom> may be specified as DOMID_SELF.
+ * 2. Only a sufficiently-privileged domain may specify <dom> != DOMID_SELF.
+ * 3. Xen may not support more than a single grant-table page per domain.
+ */
+struct gnttab_setup_table {
+ /* IN parameters. */
+ domid_t dom;
+ UINT32 nr_frames;
+ /* OUT parameters. */
+ INT16 status; /* => enum grant_status */
+#if __XEN_INTERFACE_VERSION__ < 0x00040300
+ XEN_GUEST_HANDLE(ulong) frame_list;
+#else
+ XEN_GUEST_HANDLE(xen_pfn_t) frame_list;
+#endif
+};
+typedef struct gnttab_setup_table gnttab_setup_table_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_setup_table_t);
+
+/*
+ * GNTTABOP_dump_table: Dump the contents of the grant table to the
+ * xen console. Debugging use only.
+ */
+struct gnttab_dump_table {
+ /* IN parameters. */
+ domid_t dom;
+ /* OUT parameters. */
+ INT16 status; /* => enum grant_status */
+};
+typedef struct gnttab_dump_table gnttab_dump_table_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_dump_table_t);
+
+/*
+ * GNTTABOP_transfer_grant_ref: Transfer <frame> to a foreign domain. The
+ * foreign domain has previously registered its interest in the transfer via
+ * <domid, ref>.
+ *
+ * Note that, even if the transfer fails, the specified page no longer belongs
+ * to the calling domain *unless* the error is GNTST_bad_page.
+ */
+struct gnttab_transfer {
+ /* IN parameters. */
+ xen_pfn_t mfn;
+ domid_t domid;
+ grant_ref_t ref;
+ /* OUT parameters. */
+ INT16 status;
+};
+typedef struct gnttab_transfer gnttab_transfer_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_transfer_t);
+
+
+/*
+ * GNTTABOP_copy: Hypervisor based copy
+ * source and destinations can be eithers MFNs or, for foreign domains,
+ * grant references. the foreign domain has to grant read/write access
+ * in its grant table.
+ *
+ * The flags specify what type source and destinations are (either MFN
+ * or grant reference).
+ *
+ * Note that this can also be used to copy data between two domains
+ * via a third party if the source and destination domains had previously
+ * grant appropriate access to their pages to the third party.
+ *
+ * source_offset specifies an offset in the source frame, dest_offset
+ * the offset in the target frame and len specifies the number of
+ * bytes to be copied.
+ */
+
+#define _GNTCOPY_source_gref (0)
+#define GNTCOPY_source_gref (1<<_GNTCOPY_source_gref)
+#define _GNTCOPY_dest_gref (1)
+#define GNTCOPY_dest_gref (1<<_GNTCOPY_dest_gref)
+
+struct gnttab_copy {
+ /* IN parameters. */
+ struct {
+ union {
+ grant_ref_t ref;
+ xen_pfn_t gmfn;
+ } u;
+ domid_t domid;
+ UINT16 offset;
+ } source, dest;
+ UINT16 len;
+ UINT16 flags; /* GNTCOPY_* */
+ /* OUT parameters. */
+ INT16 status;
+};
+typedef struct gnttab_copy gnttab_copy_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_copy_t);
+
+/*
+ * GNTTABOP_query_size: Query the current and maximum sizes of the shared
+ * grant table.
+ * NOTES:
+ * 1. <dom> may be specified as DOMID_SELF.
+ * 2. Only a sufficiently-privileged domain may specify <dom> != DOMID_SELF.
+ */
+struct gnttab_query_size {
+ /* IN parameters. */
+ domid_t dom;
+ /* OUT parameters. */
+ UINT32 nr_frames;
+ UINT32 max_nr_frames;
+ INT16 status; /* => enum grant_status */
+};
+typedef struct gnttab_query_size gnttab_query_size_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_query_size_t);
+
+/*
+ * GNTTABOP_unmap_and_replace: Destroy one or more grant-reference mappings
+ * tracked by <handle> but atomically replace the page table entry with one
+ * pointing to the machine address under <new_addr>. <new_addr> will be
+ * redirected to the null entry.
+ * NOTES:
+ * 1. The call may fail in an undefined manner if either mapping is not
+ * tracked by <handle>.
+ * 2. After executing a batch of unmaps, it is guaranteed that no stale
+ * mappings will remain in the device or host TLBs.
+ */
+struct gnttab_unmap_and_replace {
+ /* IN parameters. */
+ UINT64 host_addr;
+ UINT64 new_addr;
+ grant_handle_t handle;
+ /* OUT parameters. */
+ INT16 status; /* => enum grant_status */
+};
+typedef struct gnttab_unmap_and_replace gnttab_unmap_and_replace_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_unmap_and_replace_t);
+
+#if __XEN_INTERFACE_VERSION__ >= 0x0003020a
+/*
+ * GNTTABOP_set_version: Request a particular version of the grant
+ * table shared table structure. This operation can only be performed
+ * once in any given domain. It must be performed before any grants
+ * are activated; otherwise, the domain will be stuck with version 1.
+ * The only defined versions are 1 and 2.
+ */
+struct gnttab_set_version {
+ /* IN/OUT parameters */
+ UINT32 version;
+};
+typedef struct gnttab_set_version gnttab_set_version_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_set_version_t);
+
+
+/*
+ * GNTTABOP_get_status_frames: Get the list of frames used to store grant
+ * status for <dom>. In grant format version 2, the status is separated
+ * from the other shared grant fields to allow more efficient synchronization
+ * using barriers instead of atomic cmpexch operations.
+ * <nr_frames> specify the size of vector <frame_list>.
+ * The frame addresses are returned in the <frame_list>.
+ * Only <nr_frames> addresses are returned, even if the table is larger.
+ * NOTES:
+ * 1. <dom> may be specified as DOMID_SELF.
+ * 2. Only a sufficiently-privileged domain may specify <dom> != DOMID_SELF.
+ */
+struct gnttab_get_status_frames {
+ /* IN parameters. */
+ UINT32 nr_frames;
+ domid_t dom;
+ /* OUT parameters. */
+ INT16 status; /* => enum grant_status */
+ XEN_GUEST_HANDLE(UINT64) frame_list;
+};
+typedef struct gnttab_get_status_frames gnttab_get_status_frames_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_get_status_frames_t);
+
+/*
+ * GNTTABOP_get_version: Get the grant table version which is in
+ * effect for domain <dom>.
+ */
+struct gnttab_get_version {
+ /* IN parameters */
+ domid_t dom;
+ UINT16 pad;
+ /* OUT parameters */
+ UINT32 version;
+};
+typedef struct gnttab_get_version gnttab_get_version_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_get_version_t);
+
+/*
+ * GNTTABOP_swap_grant_ref: Swap the contents of two grant entries.
+ */
+struct gnttab_swap_grant_ref {
+ /* IN parameters */
+ grant_ref_t ref_a;
+ grant_ref_t ref_b;
+ /* OUT parameters */
+ INT16 status; /* => enum grant_status */
+};
+typedef struct gnttab_swap_grant_ref gnttab_swap_grant_ref_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_swap_grant_ref_t);
+
+#endif /* __XEN_INTERFACE_VERSION__ */
+
+/*
+ * Bitfield values for gnttab_map_grant_ref.flags.
+ */
+ /* Map the grant entry for access by I/O devices. */
+#define _GNTMAP_device_map (0)
+#define GNTMAP_device_map (1<<_GNTMAP_device_map)
+ /* Map the grant entry for access by host CPUs. */
+#define _GNTMAP_host_map (1)
+#define GNTMAP_host_map (1<<_GNTMAP_host_map)
+ /* Accesses to the granted frame will be restricted to read-only access. */
+#define _GNTMAP_readonly (2)
+#define GNTMAP_readonly (1<<_GNTMAP_readonly)
+ /*
+ * GNTMAP_host_map subflag:
+ * 0 => The host mapping is usable only by the guest OS.
+ * 1 => The host mapping is usable by guest OS + current application.
+ */
+#define _GNTMAP_application_map (3)
+#define GNTMAP_application_map (1<<_GNTMAP_application_map)
+
+ /*
+ * GNTMAP_contains_pte subflag:
+ * 0 => This map request contains a host virtual address.
+ * 1 => This map request contains the machine addess of the PTE to update.
+ */
+#define _GNTMAP_contains_pte (4)
+#define GNTMAP_contains_pte (1<<_GNTMAP_contains_pte)
+
+#define _GNTMAP_can_fail (5)
+#define GNTMAP_can_fail (1<<_GNTMAP_can_fail)
+
+/*
+ * Bits to be placed in guest kernel available PTE bits (architecture
+ * dependent; only supported when XENFEAT_gnttab_map_avail_bits is set).
+ */
+#define _GNTMAP_guest_avail0 (16)
+#define GNTMAP_guest_avail_mask ((UINT32)~0 << _GNTMAP_guest_avail0)
+
+/*
+ * Values for error status returns. All errors are -ve.
+ */
+/* ` enum grant_status { */
+#define GNTST_okay (0) /* Normal return. */
+#define GNTST_general_error (-1) /* General undefined error. */
+#define GNTST_bad_domain (-2) /* Unrecognsed domain id. */
+#define GNTST_bad_gntref (-3) /* Unrecognised or inappropriate gntref. */
+#define GNTST_bad_handle (-4) /* Unrecognised or inappropriate handle. */
+#define GNTST_bad_virt_addr (-5) /* Inappropriate virtual address to map. */
+#define GNTST_bad_dev_addr (-6) /* Inappropriate device address to unmap.*/
+#define GNTST_no_device_space (-7) /* Out of space in I/O MMU. */
+#define GNTST_permission_denied (-8) /* Not enough privilege for operation. */
+#define GNTST_bad_page (-9) /* Specified page was invalid for op. */
+#define GNTST_bad_copy_arg (-10) /* copy arguments cross page boundary. */
+#define GNTST_address_too_big (-11) /* transfer page address too large. */
+#define GNTST_eagain (-12) /* Operation not done; try again. */
+/* ` } */
+
+#define GNTTABOP_error_msgs { \
+ "okay", \
+ "undefined error", \
+ "unrecognised domain id", \
+ "invalid grant reference", \
+ "invalid mapping handle", \
+ "invalid virtual address", \
+ "invalid device address", \
+ "no spare translation slot in the I/O MMU", \
+ "permission denied", \
+ "bad page", \
+ "copy arguments cross page boundary", \
+ "page address size too large", \
+ "operation not done; try again" \
+}
+
+#endif /* __XEN_PUBLIC_GRANT_TABLE_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h b/OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h
new file mode 100644
index 0000000..6bfe115
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h
@@ -0,0 +1,275 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_HVM_OP_H__
+#define __XEN_PUBLIC_HVM_HVM_OP_H__
+
+#include "../xen.h"
+#include "../trace.h"
+
+/* Get/set subcommands: extra argument == pointer to xen_hvm_param struct. */
+#define HVMOP_set_param 0
+#define HVMOP_get_param 1
+struct xen_hvm_param {
+ domid_t domid; /* IN */
+ UINT32 index; /* IN */
+ UINT64 value; /* IN/OUT */
+};
+typedef struct xen_hvm_param xen_hvm_param_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_param_t);
+
+/* Set the logical level of one of a domain's PCI INTx wires. */
+#define HVMOP_set_pci_intx_level 2
+struct xen_hvm_set_pci_intx_level {
+ /* Domain to be updated. */
+ domid_t domid;
+ /* PCI INTx identification in PCI topology (domain:bus:device:intx). */
+ UINT8 domain, bus, device, intx;
+ /* Assertion level (0 = unasserted, 1 = asserted). */
+ UINT8 level;
+};
+typedef struct xen_hvm_set_pci_intx_level xen_hvm_set_pci_intx_level_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_pci_intx_level_t);
+
+/* Set the logical level of one of a domain's ISA IRQ wires. */
+#define HVMOP_set_isa_irq_level 3
+struct xen_hvm_set_isa_irq_level {
+ /* Domain to be updated. */
+ domid_t domid;
+ /* ISA device identification, by ISA IRQ (0-15). */
+ UINT8 isa_irq;
+ /* Assertion level (0 = unasserted, 1 = asserted). */
+ UINT8 level;
+};
+typedef struct xen_hvm_set_isa_irq_level xen_hvm_set_isa_irq_level_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_isa_irq_level_t);
+
+#define HVMOP_set_pci_link_route 4
+struct xen_hvm_set_pci_link_route {
+ /* Domain to be updated. */
+ domid_t domid;
+ /* PCI link identifier (0-3). */
+ UINT8 link;
+ /* ISA IRQ (1-15), or 0 (disable link). */
+ UINT8 isa_irq;
+};
+typedef struct xen_hvm_set_pci_link_route xen_hvm_set_pci_link_route_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_pci_link_route_t);
+
+/* Flushes all VCPU TLBs: @arg must be NULL. */
+#define HVMOP_flush_tlbs 5
+
+typedef enum {
+ HVMMEM_ram_rw, /* Normal read/write guest RAM */
+ HVMMEM_ram_ro, /* Read-only; writes are discarded */
+ HVMMEM_mmio_dm, /* Reads and write go to the device model */
+} hvmmem_type_t;
+
+/* Following tools-only interfaces may change in future. */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+
+/* Track dirty VRAM. */
+#define HVMOP_track_dirty_vram 6
+struct xen_hvm_track_dirty_vram {
+ /* Domain to be tracked. */
+ domid_t domid;
+ /* First pfn to track. */
+ uint64_aligned_t first_pfn;
+ /* Number of pages to track. */
+ uint64_aligned_t nr;
+ /* OUT variable. */
+ /* Dirty bitmap buffer. */
+ XEN_GUEST_HANDLE_64(uint8) dirty_bitmap;
+};
+typedef struct xen_hvm_track_dirty_vram xen_hvm_track_dirty_vram_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_track_dirty_vram_t);
+
+/* Notify that some pages got modified by the Device Model. */
+#define HVMOP_modified_memory 7
+struct xen_hvm_modified_memory {
+ /* Domain to be updated. */
+ domid_t domid;
+ /* First pfn. */
+ uint64_aligned_t first_pfn;
+ /* Number of pages. */
+ uint64_aligned_t nr;
+};
+typedef struct xen_hvm_modified_memory xen_hvm_modified_memory_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_modified_memory_t);
+
+#define HVMOP_set_mem_type 8
+/* Notify that a region of memory is to be treated in a specific way. */
+struct xen_hvm_set_mem_type {
+ /* Domain to be updated. */
+ domid_t domid;
+ /* Memory type */
+ UINT16 hvmmem_type;
+ /* Number of pages. */
+ UINT32 nr;
+ /* First pfn. */
+ uint64_aligned_t first_pfn;
+};
+typedef struct xen_hvm_set_mem_type xen_hvm_set_mem_type_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_mem_type_t);
+
+#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
+
+/* Hint from PV drivers for pagetable destruction. */
+#define HVMOP_pagetable_dying 9
+struct xen_hvm_pagetable_dying {
+ /* Domain with a pagetable about to be destroyed. */
+ domid_t domid;
+ UINT16 pad[3]; /* align next field on 8-byte boundary */
+ /* guest physical address of the toplevel pagetable dying */
+ UINT64 gpa;
+};
+typedef struct xen_hvm_pagetable_dying xen_hvm_pagetable_dying_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_pagetable_dying_t);
+
+/* Get the current Xen time, in nanoseconds since system boot. */
+#define HVMOP_get_time 10
+struct xen_hvm_get_time {
+ UINT64 now; /* OUT */
+};
+typedef struct xen_hvm_get_time xen_hvm_get_time_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_get_time_t);
+
+#define HVMOP_xentrace 11
+struct xen_hvm_xentrace {
+ UINT16 event, extra_bytes;
+ UINT8 extra[TRACE_EXTRA_MAX * sizeof(UINT32)];
+};
+typedef struct xen_hvm_xentrace xen_hvm_xentrace_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_xentrace_t);
+
+/* Following tools-only interfaces may change in future. */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+
+#define HVMOP_set_mem_access 12
+typedef enum {
+ HVMMEM_access_n,
+ HVMMEM_access_r,
+ HVMMEM_access_w,
+ HVMMEM_access_rw,
+ HVMMEM_access_x,
+ HVMMEM_access_rx,
+ HVMMEM_access_wx,
+ HVMMEM_access_rwx,
+ HVMMEM_access_rx2rw, /* Page starts off as r-x, but automatically
+ * change to r-w on a write */
+ HVMMEM_access_n2rwx, /* Log access: starts off as n, automatically
+ * goes to rwx, generating an event without
+ * pausing the vcpu */
+ HVMMEM_access_default /* Take the domain default */
+} hvmmem_access_t;
+/* Notify that a region of memory is to have specific access types */
+struct xen_hvm_set_mem_access {
+ /* Domain to be updated. */
+ domid_t domid;
+ /* Memory type */
+ UINT16 hvmmem_access; /* hvm_access_t */
+ /* Number of pages, ignored on setting default access */
+ UINT32 nr;
+ /* First pfn, or ~0ull to set the default access for new pages */
+ uint64_aligned_t first_pfn;
+};
+typedef struct xen_hvm_set_mem_access xen_hvm_set_mem_access_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_mem_access_t);
+
+#define HVMOP_get_mem_access 13
+/* Get the specific access type for that region of memory */
+struct xen_hvm_get_mem_access {
+ /* Domain to be queried. */
+ domid_t domid;
+ /* Memory type: OUT */
+ UINT16 hvmmem_access; /* hvm_access_t */
+ /* pfn, or ~0ull for default access for new pages. IN */
+ uint64_aligned_t pfn;
+};
+typedef struct xen_hvm_get_mem_access xen_hvm_get_mem_access_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_get_mem_access_t);
+
+#define HVMOP_inject_trap 14
+/* Inject a trap into a VCPU, which will get taken up on the next
+ * scheduling of it. Note that the caller should know enough of the
+ * state of the CPU before injecting, to know what the effect of
+ * injecting the trap will be.
+ */
+struct xen_hvm_inject_trap {
+ /* Domain to be queried. */
+ domid_t domid;
+ /* VCPU */
+ UINT32 vcpuid;
+ /* Vector number */
+ UINT32 vector;
+ /* Trap type (HVMOP_TRAP_*) */
+ UINT32 type;
+/* NB. This enumeration precisely matches hvm.h:X86_EVENTTYPE_* */
+# define HVMOP_TRAP_ext_int 0 /* external interrupt */
+# define HVMOP_TRAP_nmi 2 /* nmi */
+# define HVMOP_TRAP_hw_exc 3 /* hardware exception */
+# define HVMOP_TRAP_sw_int 4 /* software interrupt (CD nn) */
+# define HVMOP_TRAP_pri_sw_exc 5 /* ICEBP (F1) */
+# define HVMOP_TRAP_sw_exc 6 /* INT3 (CC), INTO (CE) */
+ /* Error code, or ~0u to skip */
+ UINT32 error_code;
+ /* Intruction length */
+ UINT32 insn_len;
+ /* CR2 for page faults */
+ uint64_aligned_t cr2;
+};
+typedef struct xen_hvm_inject_trap xen_hvm_inject_trap_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_inject_trap_t);
+
+#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
+
+#define HVMOP_get_mem_type 15
+/* Return hvmmem_type_t for the specified pfn. */
+struct xen_hvm_get_mem_type {
+ /* Domain to be queried. */
+ domid_t domid;
+ /* OUT variable. */
+ UINT16 mem_type;
+ UINT16 pad[2]; /* align next field on 8-byte boundary */
+ /* IN variable. */
+ UINT64 pfn;
+};
+typedef struct xen_hvm_get_mem_type xen_hvm_get_mem_type_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_get_mem_type_t);
+
+/* Following tools-only interfaces may change in future. */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+
+/* MSI injection for emulated devices */
+#define HVMOP_inject_msi 16
+struct xen_hvm_inject_msi {
+ /* Domain to be injected */
+ domid_t domid;
+ /* Data -- lower 32 bits */
+ UINT32 data;
+ /* Address (0xfeexxxxx) */
+ UINT64 addr;
+};
+typedef struct xen_hvm_inject_msi xen_hvm_inject_msi_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_inject_msi_t);
+
+#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
+
+#endif /* __XEN_PUBLIC_HVM_HVM_OP_H__ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h b/OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h
new file mode 100644
index 0000000..517a184
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h
@@ -0,0 +1,150 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_PARAMS_H__
+#define __XEN_PUBLIC_HVM_PARAMS_H__
+
+#include "hvm_op.h"
+
+/*
+ * Parameter space for HVMOP_{set,get}_param.
+ */
+
+/*
+ * How should CPU0 event-channel notifications be delivered?
+ * val[63:56] == 0: val[55:0] is a delivery GSI (Global System Interrupt).
+ * val[63:56] == 1: val[55:0] is a delivery PCI INTx line, as follows:
+ * Domain = val[47:32], Bus = val[31:16],
+ * DevFn = val[15: 8], IntX = val[ 1: 0]
+ * val[63:56] == 2: val[7:0] is a vector number, check for
+ * XENFEAT_hvm_callback_vector to know if this delivery
+ * method is available.
+ * If val == 0 then CPU0 event-channel notifications are not delivered.
+ */
+#define HVM_PARAM_CALLBACK_IRQ 0
+
+/*
+ * These are not used by Xen. They are here for convenience of HVM-guest
+ * xenbus implementations.
+ */
+#define HVM_PARAM_STORE_PFN 1
+#define HVM_PARAM_STORE_EVTCHN 2
+
+#define HVM_PARAM_PAE_ENABLED 4
+
+#define HVM_PARAM_IOREQ_PFN 5
+
+#define HVM_PARAM_BUFIOREQ_PFN 6
+#define HVM_PARAM_BUFIOREQ_EVTCHN 26
+
+#if defined(__i386__) || defined(__x86_64__)
+
+/* Expose Viridian interfaces to this HVM guest? */
+#define HVM_PARAM_VIRIDIAN 9
+
+#endif
+
+/*
+ * Set mode for virtual timers (currently x86 only):
+ * delay_for_missed_ticks (default):
+ * Do not advance a vcpu's time beyond the correct delivery time for
+ * interrupts that have been missed due to preemption. Deliver missed
+ * interrupts when the vcpu is rescheduled and advance the vcpu's virtual
+ * time stepwise for each one.
+ * no_delay_for_missed_ticks:
+ * As above, missed interrupts are delivered, but guest time always tracks
+ * wallclock (i.e., real) time while doing so.
+ * no_missed_ticks_pending:
+ * No missed interrupts are held pending. Instead, to ensure ticks are
+ * delivered at some non-zero rate, if we detect missed ticks then the
+ * internal tick alarm is not disabled if the VCPU is preempted during the
+ * next tick period.
+ * one_missed_tick_pending:
+ * Missed interrupts are collapsed together and delivered as one 'late tick'.
+ * Guest time always tracks wallclock (i.e., real) time.
+ */
+#define HVM_PARAM_TIMER_MODE 10
+#define HVMPTM_delay_for_missed_ticks 0
+#define HVMPTM_no_delay_for_missed_ticks 1
+#define HVMPTM_no_missed_ticks_pending 2
+#define HVMPTM_one_missed_tick_pending 3
+
+/* Boolean: Enable virtual HPET (high-precision event timer)? (x86-only) */
+#define HVM_PARAM_HPET_ENABLED 11
+
+/* Identity-map page directory used by Intel EPT when CR0.PG=0. */
+#define HVM_PARAM_IDENT_PT 12
+
+/* Device Model domain, defaults to 0. */
+#define HVM_PARAM_DM_DOMAIN 13
+
+/* ACPI S state: currently support S0 and S3 on x86. */
+#define HVM_PARAM_ACPI_S_STATE 14
+
+/* TSS used on Intel when CR0.PE=0. */
+#define HVM_PARAM_VM86_TSS 15
+
+/* Boolean: Enable aligning all periodic vpts to reduce interrupts */
+#define HVM_PARAM_VPT_ALIGN 16
+
+/* Console debug shared memory ring and event channel */
+#define HVM_PARAM_CONSOLE_PFN 17
+#define HVM_PARAM_CONSOLE_EVTCHN 18
+
+/*
+ * Select location of ACPI PM1a and TMR control blocks. Currently two locations
+ * are supported, specified by version 0 or 1 in this parameter:
+ * - 0: default, use the old addresses
+ * PM1A_EVT == 0x1f40; PM1A_CNT == 0x1f44; PM_TMR == 0x1f48
+ * - 1: use the new default qemu addresses
+ * PM1A_EVT == 0xb000; PM1A_CNT == 0xb004; PM_TMR == 0xb008
+ * You can find these address definitions in <hvm/ioreq.h>
+ */
+#define HVM_PARAM_ACPI_IOPORTS_LOCATION 19
+
+/* Enable blocking memory events, async or sync (pause vcpu until response)
+ * onchangeonly indicates messages only on a change of value */
+#define HVM_PARAM_MEMORY_EVENT_CR0 20
+#define HVM_PARAM_MEMORY_EVENT_CR3 21
+#define HVM_PARAM_MEMORY_EVENT_CR4 22
+#define HVM_PARAM_MEMORY_EVENT_INT3 23
+#define HVM_PARAM_MEMORY_EVENT_SINGLE_STEP 25
+#define HVM_PARAM_MEMORY_EVENT_MSR 30
+
+#define HVMPME_MODE_MASK (3 << 0)
+#define HVMPME_mode_disabled 0
+#define HVMPME_mode_async 1
+#define HVMPME_mode_sync 2
+#define HVMPME_onchangeonly (1 << 2)
+
+/* Boolean: Enable nestedhvm (hvm only) */
+#define HVM_PARAM_NESTEDHVM 24
+
+/* Params for the mem event rings */
+#define HVM_PARAM_PAGING_RING_PFN 27
+#define HVM_PARAM_ACCESS_RING_PFN 28
+#define HVM_PARAM_SHARING_RING_PFN 29
+
+/* SHUTDOWN_* action in case of a triple fault */
+#define HVM_PARAM_TRIPLE_FAULT_REASON 31
+
+#define HVM_NR_PARAMS 32
+
+#endif /* __XEN_PUBLIC_HVM_PARAMS_H__ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h b/OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h
new file mode 100644
index 0000000..2f0bb8b
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h
@@ -0,0 +1,608 @@
+/******************************************************************************
+ * blkif.h
+ *
+ * Unified block-device I/O interface for Xen guest OSes.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2003-2004, Keir Fraser
+ * Copyright (c) 2012, Spectra Logic Corporation
+ */
+
+#ifndef __XEN_PUBLIC_IO_BLKIF_H__
+#define __XEN_PUBLIC_IO_BLKIF_H__
+
+#include "ring.h"
+#include "../grant_table.h"
+
+/*
+ * Front->back notifications: When enqueuing a new request, sending a
+ * notification can be made conditional on req_event (i.e., the generic
+ * hold-off mechanism provided by the ring macros). Backends must set
+ * req_event appropriately (e.g., using RING_FINAL_CHECK_FOR_REQUESTS()).
+ *
+ * Back->front notifications: When enqueuing a new response, sending a
+ * notification can be made conditional on rsp_event (i.e., the generic
+ * hold-off mechanism provided by the ring macros). Frontends must set
+ * rsp_event appropriately (e.g., using RING_FINAL_CHECK_FOR_RESPONSES()).
+ */
+
+#ifndef blkif_vdev_t
+#define blkif_vdev_t UINT16
+#endif
+#define blkif_sector_t UINT64
+
+/*
+ * Feature and Parameter Negotiation
+ * =================================
+ * The two halves of a Xen block driver utilize nodes within the XenStore to
+ * communicate capabilities and to negotiate operating parameters. This
+ * section enumerates these nodes which reside in the respective front and
+ * backend portions of the XenStore, following the XenBus convention.
+ *
+ * All data in the XenStore is stored as strings. Nodes specifying numeric
+ * values are encoded in decimal. Integer value ranges listed below are
+ * expressed as fixed sized integer types capable of storing the conversion
+ * of a properly formated node string, without loss of information.
+ *
+ * Any specified default value is in effect if the corresponding XenBus node
+ * is not present in the XenStore.
+ *
+ * XenStore nodes in sections marked "PRIVATE" are solely for use by the
+ * driver side whose XenBus tree contains them.
+ *
+ * XenStore nodes marked "DEPRECATED" in their notes section should only be
+ * used to provide interoperability with legacy implementations.
+ *
+ * See the XenBus state transition diagram below for details on when XenBus
+ * nodes must be published and when they can be queried.
+ *
+ *****************************************************************************
+ * Backend XenBus Nodes
+ *****************************************************************************
+ *
+ *------------------ Backend Device Identification (PRIVATE) ------------------
+ *
+ * mode
+ * Values: "r" (read only), "w" (writable)
+ *
+ * The read or write access permissions to the backing store to be
+ * granted to the frontend.
+ *
+ * params
+ * Values: string
+ *
+ * A free formatted string providing sufficient information for the
+ * backend driver to open the backing device. (e.g. the path to the
+ * file or block device representing the backing store.)
+ *
+ * type
+ * Values: "file", "phy", "tap"
+ *
+ * The type of the backing device/object.
+ *
+ *--------------------------------- Features ---------------------------------
+ *
+ * feature-barrier
+ * Values: 0/1 (boolean)
+ * Default Value: 0
+ *
+ * A value of "1" indicates that the backend can process requests
+ * containing the BLKIF_OP_WRITE_BARRIER request opcode. Requests
+ * of this type may still be returned at any time with the
+ * BLKIF_RSP_EOPNOTSUPP result code.
+ *
+ * feature-flush-cache
+ * Values: 0/1 (boolean)
+ * Default Value: 0
+ *
+ * A value of "1" indicates that the backend can process requests
+ * containing the BLKIF_OP_FLUSH_DISKCACHE request opcode. Requests
+ * of this type may still be returned at any time with the
+ * BLKIF_RSP_EOPNOTSUPP result code.
+ *
+ * feature-discard
+ * Values: 0/1 (boolean)
+ * Default Value: 0
+ *
+ * A value of "1" indicates that the backend can process requests
+ * containing the BLKIF_OP_DISCARD request opcode. Requests
+ * of this type may still be returned at any time with the
+ * BLKIF_RSP_EOPNOTSUPP result code.
+ *
+ * feature-persistent
+ * Values: 0/1 (boolean)
+ * Default Value: 0
+ * Notes: 7
+ *
+ * A value of "1" indicates that the backend can keep the grants used
+ * by the frontend driver mapped, so the same set of grants should be
+ * used in all transactions. The maximum number of grants the backend
+ * can map persistently depends on the implementation, but ideally it
+ * should be RING_SIZE * BLKIF_MAX_SEGMENTS_PER_REQUEST. Using this
+ * feature the backend doesn't need to unmap each grant, preventing
+ * costly TLB flushes. The backend driver should only map grants
+ * persistently if the frontend supports it. If a backend driver chooses
+ * to use the persistent protocol when the frontend doesn't support it,
+ * it will probably hit the maximum number of persistently mapped grants
+ * (due to the fact that the frontend won't be reusing the same grants),
+ * and fall back to non-persistent mode. Backend implementations may
+ * shrink or expand the number of persistently mapped grants without
+ * notifying the frontend depending on memory constraints (this might
+ * cause a performance degradation).
+ *
+ * If a backend driver wants to limit the maximum number of persistently
+ * mapped grants to a value less than RING_SIZE *
+ * BLKIF_MAX_SEGMENTS_PER_REQUEST a LRU strategy should be used to
+ * discard the grants that are less commonly used. Using a LRU in the
+ * backend driver paired with a LIFO queue in the frontend will
+ * allow us to have better performance in this scenario.
+ *
+ *----------------------- Request Transport Parameters ------------------------
+ *
+ * max-ring-page-order
+ * Values: <UINT32>
+ * Default Value: 0
+ * Notes: 1, 3
+ *
+ * The maximum supported size of the request ring buffer in units of
+ * lb(machine pages). (e.g. 0 == 1 page, 1 = 2 pages, 2 == 4 pages,
+ * etc.).
+ *
+ * max-ring-pages
+ * Values: <UINT32>
+ * Default Value: 1
+ * Notes: DEPRECATED, 2, 3
+ *
+ * The maximum supported size of the request ring buffer in units of
+ * machine pages. The value must be a power of 2.
+ *
+ *------------------------- Backend Device Properties -------------------------
+ *
+ * discard-alignment
+ * Values: <UINT32>
+ * Default Value: 0
+ * Notes: 4, 5
+ *
+ * The offset, in bytes from the beginning of the virtual block device,
+ * to the first, addressable, discard extent on the underlying device.
+ *
+ * discard-granularity
+ * Values: <UINT32>
+ * Default Value: <"sector-size">
+ * Notes: 4
+ *
+ * The size, in bytes, of the individually addressable discard extents
+ * of the underlying device.
+ *
+ * discard-secure
+ * Values: 0/1 (boolean)
+ * Default Value: 0
+ * Notes: 10
+ *
+ * A value of "1" indicates that the backend can process BLKIF_OP_DISCARD
+ * requests with the BLKIF_DISCARD_SECURE flag set.
+ *
+ * info
+ * Values: <UINT32> (bitmap)
+ *
+ * A collection of bit flags describing attributes of the backing
+ * device. The VDISK_* macros define the meaning of each bit
+ * location.
+ *
+ * sector-size
+ * Values: <UINT32>
+ *
+ * The logical sector size, in bytes, of the backend device.
+ *
+ * physical-sector-size
+ * Values: <UINT32>
+ *
+ * The physical sector size, in bytes, of the backend device.
+ *
+ * sectors
+ * Values: <UINT64>
+ *
+ * The size of the backend device, expressed in units of its logical
+ * sector size ("sector-size").
+ *
+ *****************************************************************************
+ * Frontend XenBus Nodes
+ *****************************************************************************
+ *
+ *----------------------- Request Transport Parameters -----------------------
+ *
+ * event-channel
+ * Values: <UINT32>
+ *
+ * The identifier of the Xen event channel used to signal activity
+ * in the ring buffer.
+ *
+ * ring-ref
+ * Values: <UINT32>
+ * Notes: 6
+ *
+ * The Xen grant reference granting permission for the backend to map
+ * the sole page in a single page sized ring buffer.
+ *
+ * ring-ref%u
+ * Values: <UINT32>
+ * Notes: 6
+ *
+ * For a frontend providing a multi-page ring, a "number of ring pages"
+ * sized list of nodes, each containing a Xen grant reference granting
+ * permission for the backend to map the page of the ring located
+ * at page index "%u". Page indexes are zero based.
+ *
+ * protocol
+ * Values: string (XEN_IO_PROTO_ABI_*)
+ * Default Value: XEN_IO_PROTO_ABI_NATIVE
+ *
+ * The machine ABI rules governing the format of all ring request and
+ * response structures.
+ *
+ * ring-page-order
+ * Values: <UINT32>
+ * Default Value: 0
+ * Maximum Value: MAX(ffs(max-ring-pages) - 1, max-ring-page-order)
+ * Notes: 1, 3
+ *
+ * The size of the frontend allocated request ring buffer in units
+ * of lb(machine pages). (e.g. 0 == 1 page, 1 = 2 pages, 2 == 4 pages,
+ * etc.).
+ *
+ * num-ring-pages
+ * Values: <UINT32>
+ * Default Value: 1
+ * Maximum Value: MAX(max-ring-pages,(0x1 << max-ring-page-order))
+ * Notes: DEPRECATED, 2, 3
+ *
+ * The size of the frontend allocated request ring buffer in units of
+ * machine pages. The value must be a power of 2.
+ *
+ * feature-persistent
+ * Values: 0/1 (boolean)
+ * Default Value: 0
+ * Notes: 7, 8, 9
+ *
+ * A value of "1" indicates that the frontend will reuse the same grants
+ * for all transactions, allowing the backend to map them with write
+ * access (even when it should be read-only). If the frontend hits the
+ * maximum number of allowed persistently mapped grants, it can fallback
+ * to non persistent mode. This will cause a performance degradation,
+ * since the the backend driver will still try to map those grants
+ * persistently. Since the persistent grants protocol is compatible with
+ * the previous protocol, a frontend driver can choose to work in
+ * persistent mode even when the backend doesn't support it.
+ *
+ * It is recommended that the frontend driver stores the persistently
+ * mapped grants in a LIFO queue, so a subset of all persistently mapped
+ * grants gets used commonly. This is done in case the backend driver
+ * decides to limit the maximum number of persistently mapped grants
+ * to a value less than RING_SIZE * BLKIF_MAX_SEGMENTS_PER_REQUEST.
+ *
+ *------------------------- Virtual Device Properties -------------------------
+ *
+ * device-type
+ * Values: "disk", "cdrom", "floppy", etc.
+ *
+ * virtual-device
+ * Values: <UINT32>
+ *
+ * A value indicating the physical device to virtualize within the
+ * frontend's domain. (e.g. "The first ATA disk", "The third SCSI
+ * disk", etc.)
+ *
+ * See docs/misc/vbd-interface.txt for details on the format of this
+ * value.
+ *
+ * Notes
+ * -----
+ * (1) Multi-page ring buffer scheme first developed in the Citrix XenServer
+ * PV drivers.
+ * (2) Multi-page ring buffer scheme first used in some RedHat distributions
+ * including a distribution deployed on certain nodes of the Amazon
+ * EC2 cluster.
+ * (3) Support for multi-page ring buffers was implemented independently,
+ * in slightly different forms, by both Citrix and RedHat/Amazon.
+ * For full interoperability, block front and backends should publish
+ * identical ring parameters, adjusted for unit differences, to the
+ * XenStore nodes used in both schemes.
+ * (4) Devices that support discard functionality may internally allocate space
+ * (discardable extents) in units that are larger than the exported logical
+ * block size. If the backing device has such discardable extents the
+ * backend should provide both discard-granularity and discard-alignment.
+ * Providing just one of the two may be considered an error by the frontend.
+ * Backends supporting discard should include discard-granularity and
+ * discard-alignment even if it supports discarding individual sectors.
+ * Frontends should assume discard-alignment == 0 and discard-granularity
+ * == sector size if these keys are missing.
+ * (5) The discard-alignment parameter allows a physical device to be
+ * partitioned into virtual devices that do not necessarily begin or
+ * end on a discardable extent boundary.
+ * (6) When there is only a single page allocated to the request ring,
+ * 'ring-ref' is used to communicate the grant reference for this
+ * page to the backend. When using a multi-page ring, the 'ring-ref'
+ * node is not created. Instead 'ring-ref0' - 'ring-refN' are used.
+ * (7) When using persistent grants data has to be copied from/to the page
+ * where the grant is currently mapped. The overhead of doing this copy
+ * however doesn't suppress the speed improvement of not having to unmap
+ * the grants.
+ * (8) The frontend driver has to allow the backend driver to map all grants
+ * with write access, even when they should be mapped read-only, since
+ * further requests may reuse these grants and require write permissions.
+ * (9) Linux implementation doesn't have a limit on the maximum number of
+ * grants that can be persistently mapped in the frontend driver, but
+ * due to the frontent driver implementation it should never be bigger
+ * than RING_SIZE * BLKIF_MAX_SEGMENTS_PER_REQUEST.
+ *(10) The discard-secure property may be present and will be set to 1 if the
+ * backing device supports secure discard.
+ */
+
+/*
+ * STATE DIAGRAMS
+ *
+ *****************************************************************************
+ * Startup *
+ *****************************************************************************
+ *
+ * Tool stack creates front and back nodes with state XenbusStateInitialising.
+ *
+ * Front Back
+ * ================================= =====================================
+ * XenbusStateInitialising XenbusStateInitialising
+ * o Query virtual device o Query backend device identification
+ * properties. data.
+ * o Setup OS device instance. o Open and validate backend device.
+ * o Publish backend features and
+ * transport parameters.
+ * |
+ * |
+ * V
+ * XenbusStateInitWait
+ *
+ * o Query backend features and
+ * transport parameters.
+ * o Allocate and initialize the
+ * request ring.
+ * o Publish transport parameters
+ * that will be in effect during
+ * this connection.
+ * |
+ * |
+ * V
+ * XenbusStateInitialised
+ *
+ * o Query frontend transport parameters.
+ * o Connect to the request ring and
+ * event channel.
+ * o Publish backend device properties.
+ * |
+ * |
+ * V
+ * XenbusStateConnected
+ *
+ * o Query backend device properties.
+ * o Finalize OS virtual device
+ * instance.
+ * |
+ * |
+ * V
+ * XenbusStateConnected
+ *
+ * Note: Drivers that do not support any optional features, or the negotiation
+ * of transport parameters, can skip certain states in the state machine:
+ *
+ * o A frontend may transition to XenbusStateInitialised without
+ * waiting for the backend to enter XenbusStateInitWait. In this
+ * case, default transport parameters are in effect and any
+ * transport parameters published by the frontend must contain
+ * their default values.
+ *
+ * o A backend may transition to XenbusStateInitialised, bypassing
+ * XenbusStateInitWait, without waiting for the frontend to first
+ * enter the XenbusStateInitialised state. In this case, default
+ * transport parameters are in effect and any transport parameters
+ * published by the backend must contain their default values.
+ *
+ * Drivers that support optional features and/or transport parameter
+ * negotiation must tolerate these additional state transition paths.
+ * In general this means performing the work of any skipped state
+ * transition, if it has not already been performed, in addition to the
+ * work associated with entry into the current state.
+ */
+
+/*
+ * REQUEST CODES.
+ */
+#define BLKIF_OP_READ 0
+#define BLKIF_OP_WRITE 1
+/*
+ * All writes issued prior to a request with the BLKIF_OP_WRITE_BARRIER
+ * operation code ("barrier request") must be completed prior to the
+ * execution of the barrier request. All writes issued after the barrier
+ * request must not execute until after the completion of the barrier request.
+ *
+ * Optional. See "feature-barrier" XenBus node documentation above.
+ */
+#define BLKIF_OP_WRITE_BARRIER 2
+/*
+ * Commit any uncommitted contents of the backing device's volatile cache
+ * to stable storage.
+ *
+ * Optional. See "feature-flush-cache" XenBus node documentation above.
+ */
+#define BLKIF_OP_FLUSH_DISKCACHE 3
+/*
+ * Used in SLES sources for device specific command packet
+ * contained within the request. Reserved for that purpose.
+ */
+#define BLKIF_OP_RESERVED_1 4
+/*
+ * Indicate to the backend device that a region of storage is no longer in
+ * use, and may be discarded at any time without impact to the client. If
+ * the BLKIF_DISCARD_SECURE flag is set on the request, all copies of the
+ * discarded region on the device must be rendered unrecoverable before the
+ * command returns.
+ *
+ * This operation is analogous to performing a trim (ATA) or unamp (SCSI),
+ * command on a native device.
+ *
+ * More information about trim/unmap operations can be found at:
+ * http://t13.org/Documents/UploadedDocuments/docs2008/
+ * e07154r6-Data_Set_Management_Proposal_for_ATA-ACS2.doc
+ * http://www.seagate.com/staticfiles/support/disc/manuals/
+ * Interface%20manuals/100293068c.pdf
+ *
+ * Optional. See "feature-discard", "discard-alignment",
+ * "discard-granularity", and "discard-secure" in the XenBus node
+ * documentation above.
+ */
+#define BLKIF_OP_DISCARD 5
+
+/*
+ * Recognized if "feature-max-indirect-segments" in present in the backend
+ * xenbus info. The "feature-max-indirect-segments" node contains the maximum
+ * number of segments allowed by the backend per request. If the node is
+ * present, the frontend might use blkif_request_indirect structs in order to
+ * issue requests with more than BLKIF_MAX_SEGMENTS_PER_REQUEST (11). The
+ * maximum number of indirect segments is fixed by the backend, but the
+ * frontend can issue requests with any number of indirect segments as long as
+ * it's less than the number provided by the backend. The indirect_grefs field
+ * in blkif_request_indirect should be filled by the frontend with the
+ * grant references of the pages that are holding the indirect segments.
+ * These pages are filled with an array of blkif_request_segment that hold the
+ * information about the segments. The number of indirect pages to use is
+ * determined by the number of segments an indirect request contains. Every
+ * indirect page can contain a maximum of
+ * (PAGE_SIZE / sizeof(struct blkif_request_segment)) segments, so to
+ * calculate the number of indirect pages to use we have to do
+ * ceil(indirect_segments / (PAGE_SIZE / sizeof(struct blkif_request_segment))).
+ *
+ * If a backend does not recognize BLKIF_OP_INDIRECT, it should *not*
+ * create the "feature-max-indirect-segments" node!
+ */
+#define BLKIF_OP_INDIRECT 6
+
+/*
+ * Maximum scatter/gather segments per request.
+ * This is carefully chosen so that sizeof(blkif_ring_t) <= PAGE_SIZE.
+ * NB. This could be 12 if the ring indexes weren't stored in the same page.
+ */
+#define BLKIF_MAX_SEGMENTS_PER_REQUEST 11
+
+/*
+ * Maximum number of indirect pages to use per request.
+ */
+#define BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST 8
+
+/*
+ * NB. first_sect and last_sect in blkif_request_segment, as well as
+ * sector_number in blkif_request, are always expressed in 512-byte units.
+ * However they must be properly aligned to the real sector size of the
+ * physical disk, which is reported in the "physical-sector-size" node in
+ * the backend xenbus info. Also the xenbus "sectors" node is expressed in
+ * 512-byte units.
+ */
+struct blkif_request_segment {
+ grant_ref_t gref; /* reference to I/O buffer frame */
+ /* @first_sect: first sector in frame to transfer (inclusive). */
+ /* @last_sect: last sector in frame to transfer (inclusive). */
+ UINT8 first_sect, last_sect;
+};
+
+/*
+ * Starting ring element for any I/O request.
+ */
+struct blkif_request {
+ UINT8 operation; /* BLKIF_OP_??? */
+ UINT8 nr_segments; /* number of segments */
+ blkif_vdev_t handle; /* only for read/write requests */
+ UINT64 id; /* private guest value, echoed in resp */
+ blkif_sector_t sector_number;/* start sector idx on disk (r/w only) */
+ struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+typedef struct blkif_request blkif_request_t;
+
+/*
+ * Cast to this structure when blkif_request.operation == BLKIF_OP_DISCARD
+ * sizeof(struct blkif_request_discard) <= sizeof(struct blkif_request)
+ */
+struct blkif_request_discard {
+ UINT8 operation; /* BLKIF_OP_DISCARD */
+ UINT8 flag; /* BLKIF_DISCARD_SECURE or zero */
+#define BLKIF_DISCARD_SECURE (1<<0) /* ignored if discard-secure=0 */
+ blkif_vdev_t handle; /* same as for read/write requests */
+ UINT64 id; /* private guest value, echoed in resp */
+ blkif_sector_t sector_number;/* start sector idx on disk */
+ UINT64 nr_sectors; /* number of contiguous sectors to discard*/
+};
+typedef struct blkif_request_discard blkif_request_discard_t;
+
+struct blkif_request_indirect {
+ UINT8 operation; /* BLKIF_OP_INDIRECT */
+ UINT8 indirect_op; /* BLKIF_OP_{READ/WRITE} */
+ UINT16 nr_segments; /* number of segments */
+ UINT64 id; /* private guest value, echoed in resp */
+ blkif_sector_t sector_number;/* start sector idx on disk (r/w only) */
+ blkif_vdev_t handle; /* same as for read/write requests */
+ grant_ref_t indirect_grefs[BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST];
+#ifdef __i386__
+ UINT64 pad; /* Make it 64 byte aligned on i386 */
+#endif
+};
+typedef struct blkif_request_indirect blkif_request_indirect_t;
+
+struct blkif_response {
+ UINT64 id; /* copied from request */
+ UINT8 operation; /* copied from request */
+ INT16 status; /* BLKIF_RSP_??? */
+};
+typedef struct blkif_response blkif_response_t;
+
+/*
+ * STATUS RETURN CODES.
+ */
+ /* Operation not supported (only happens on barrier writes). */
+#define BLKIF_RSP_EOPNOTSUPP -2
+ /* Operation failed for some unspecified reason (-EIO). */
+#define BLKIF_RSP_ERROR -1
+ /* Operation completed successfully. */
+#define BLKIF_RSP_OKAY 0
+
+/*
+ * Generate blkif ring structures and types.
+ */
+DEFINE_RING_TYPES(blkif, struct blkif_request, struct blkif_response);
+
+#define VDISK_CDROM 0x1
+#define VDISK_REMOVABLE 0x2
+#define VDISK_READONLY 0x4
+
+#endif /* __XEN_PUBLIC_IO_BLKIF_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/io/protocols.h b/OvmfPkg/Include/IndustryStandard/Xen/io/protocols.h
new file mode 100644
index 0000000..80b196b
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/io/protocols.h
@@ -0,0 +1,40 @@
+/******************************************************************************
+ * protocols.h
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PROTOCOLS_H__
+#define __XEN_PROTOCOLS_H__
+
+#define XEN_IO_PROTO_ABI_X86_32 "x86_32-abi"
+#define XEN_IO_PROTO_ABI_X86_64 "x86_64-abi"
+#define XEN_IO_PROTO_ABI_ARM "arm-abi"
+
+#if defined(__i386__)
+# define XEN_IO_PROTO_ABI_NATIVE XEN_IO_PROTO_ABI_X86_32
+#elif defined(__x86_64__)
+# define XEN_IO_PROTO_ABI_NATIVE XEN_IO_PROTO_ABI_X86_64
+#elif defined(__arm__) || defined(__aarch64__)
+# define XEN_IO_PROTO_ABI_NATIVE XEN_IO_PROTO_ABI_ARM
+#else
+# error arch fixup needed here
+#endif
+
+#endif
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/io/ring.h b/OvmfPkg/Include/IndustryStandard/Xen/io/ring.h
new file mode 100644
index 0000000..a8e9ea0
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/io/ring.h
@@ -0,0 +1,312 @@
+/******************************************************************************
+ * ring.h
+ *
+ * Shared producer-consumer ring macros.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Tim Deegan and Andrew Warfield November 2004.
+ */
+
+#ifndef __XEN_PUBLIC_IO_RING_H__
+#define __XEN_PUBLIC_IO_RING_H__
+
+#include "../xen-compat.h"
+
+#if __XEN_INTERFACE_VERSION__ < 0x00030208
+#define xen_mb() mb()
+#define xen_rmb() rmb()
+#define xen_wmb() wmb()
+#endif
+
+typedef UINT32 RING_IDX;
+
+/* Round a 32-bit unsigned constant down to the nearest power of two. */
+#define __RD2(_x) (((_x) & 0x00000002) ? 0x2 : ((_x) & 0x1))
+#define __RD4(_x) (((_x) & 0x0000000c) ? __RD2((_x)>>2)<<2 : __RD2(_x))
+#define __RD8(_x) (((_x) & 0x000000f0) ? __RD4((_x)>>4)<<4 : __RD4(_x))
+#define __RD16(_x) (((_x) & 0x0000ff00) ? __RD8((_x)>>8)<<8 : __RD8(_x))
+#define __RD32(_x) (((_x) & 0xffff0000) ? __RD16((_x)>>16)<<16 : __RD16(_x))
+
+/*
+ * Calculate size of a shared ring, given the total available space for the
+ * ring and indexes (_sz), and the name tag of the request/response structure.
+ * A ring contains as many entries as will fit, rounded down to the nearest
+ * power of two (so we can mask with (size-1) to loop around).
+ */
+#define __CONST_RING_SIZE(_s, _sz) \
+ (__RD32(((_sz) - offsetof(struct _s##_sring, ring)) / \
+ sizeof(((struct _s##_sring *)0)->ring[0])))
+/*
+ * The same for passing in an actual pointer instead of a name tag.
+ */
+#define __RING_SIZE(_s, _sz) \
+ (__RD32(((_sz) - (INTN)(_s)->ring + (INTN)(_s)) / sizeof((_s)->ring[0])))
+
+/*
+ * Macros to make the correct C datatypes for a new kind of ring.
+ *
+ * To make a new ring datatype, you need to have two message structures,
+ * let's say request_t, and response_t already defined.
+ *
+ * In a header where you want the ring datatype declared, you then do:
+ *
+ * DEFINE_RING_TYPES(mytag, request_t, response_t);
+ *
+ * These expand out to give you a set of types, as you can see below.
+ * The most important of these are:
+ *
+ * mytag_sring_t - The shared ring.
+ * mytag_front_ring_t - The 'front' half of the ring.
+ * mytag_back_ring_t - The 'back' half of the ring.
+ *
+ * To initialize a ring in your code you need to know the location and size
+ * of the shared memory area (PAGE_SIZE, for instance). To initialise
+ * the front half:
+ *
+ * mytag_front_ring_t front_ring;
+ * SHARED_RING_INIT((mytag_sring_t *)shared_page);
+ * FRONT_RING_INIT(&front_ring, (mytag_sring_t *)shared_page, PAGE_SIZE);
+ *
+ * Initializing the back follows similarly (note that only the front
+ * initializes the shared ring):
+ *
+ * mytag_back_ring_t back_ring;
+ * BACK_RING_INIT(&back_ring, (mytag_sring_t *)shared_page, PAGE_SIZE);
+ */
+
+#define DEFINE_RING_TYPES(__name, __req_t, __rsp_t) \
+ \
+/* Shared ring entry */ \
+union __name##_sring_entry { \
+ __req_t req; \
+ __rsp_t rsp; \
+}; \
+ \
+/* Shared ring page */ \
+struct __name##_sring { \
+ RING_IDX req_prod, req_event; \
+ RING_IDX rsp_prod, rsp_event; \
+ union { \
+ struct { \
+ UINT8 smartpoll_active; \
+ } netif; \
+ struct { \
+ UINT8 msg; \
+ } tapif_user; \
+ UINT8 pvt_pad[4]; \
+ } private; \
+ UINT8 __pad[44]; \
+ union __name##_sring_entry ring[1]; /* variable-length */ \
+}; \
+ \
+/* "Front" end's private variables */ \
+struct __name##_front_ring { \
+ RING_IDX req_prod_pvt; \
+ RING_IDX rsp_cons; \
+ UINT32 nr_ents; \
+ struct __name##_sring *sring; \
+}; \
+ \
+/* "Back" end's private variables */ \
+struct __name##_back_ring { \
+ RING_IDX rsp_prod_pvt; \
+ RING_IDX req_cons; \
+ UINT32 nr_ents; \
+ struct __name##_sring *sring; \
+}; \
+ \
+/* Syntactic sugar */ \
+typedef struct __name##_sring __name##_sring_t; \
+typedef struct __name##_front_ring __name##_front_ring_t; \
+typedef struct __name##_back_ring __name##_back_ring_t
+
+/*
+ * Macros for manipulating rings.
+ *
+ * FRONT_RING_whatever works on the "front end" of a ring: here
+ * requests are pushed on to the ring and responses taken off it.
+ *
+ * BACK_RING_whatever works on the "back end" of a ring: here
+ * requests are taken off the ring and responses put on.
+ *
+ * N.B. these macros do NO INTERLOCKS OR FLOW CONTROL.
+ * This is OK in 1-for-1 request-response situations where the
+ * requestor (front end) never has more than RING_SIZE()-1
+ * outstanding requests.
+ */
+
+/* Initialising empty rings */
+#define SHARED_RING_INIT(_s) do { \
+ (_s)->req_prod = (_s)->rsp_prod = 0; \
+ (_s)->req_event = (_s)->rsp_event = 1; \
+ (VOID)ZeroMem((_s)->private.pvt_pad, sizeof((_s)->private.pvt_pad)); \
+ (VOID)ZeroMem((_s)->__pad, sizeof((_s)->__pad)); \
+} while(0)
+
+#define FRONT_RING_INIT(_r, _s, __size) do { \
+ (_r)->req_prod_pvt = 0; \
+ (_r)->rsp_cons = 0; \
+ (_r)->nr_ents = __RING_SIZE(_s, __size); \
+ (_r)->sring = (_s); \
+} while (0)
+
+#define BACK_RING_INIT(_r, _s, __size) do { \
+ (_r)->rsp_prod_pvt = 0; \
+ (_r)->req_cons = 0; \
+ (_r)->nr_ents = __RING_SIZE(_s, __size); \
+ (_r)->sring = (_s); \
+} while (0)
+
+/* How big is this ring? */
+#define RING_SIZE(_r) \
+ ((_r)->nr_ents)
+
+/* Number of free requests (for use on front side only). */
+#define RING_FREE_REQUESTS(_r) \
+ (RING_SIZE(_r) - ((_r)->req_prod_pvt - (_r)->rsp_cons))
+
+/* Test if there is an empty slot available on the front ring.
+ * (This is only meaningful from the front. )
+ */
+#define RING_FULL(_r) \
+ (RING_FREE_REQUESTS(_r) == 0)
+
+/* Test if there are outstanding messages to be processed on a ring. */
+#define RING_HAS_UNCONSUMED_RESPONSES(_r) \
+ ((_r)->sring->rsp_prod - (_r)->rsp_cons)
+
+#ifdef __GNUC__
+#define RING_HAS_UNCONSUMED_REQUESTS(_r) ({ \
+ UINT32 req = (_r)->sring->req_prod - (_r)->req_cons; \
+ UINT32 rsp = RING_SIZE(_r) - \
+ ((_r)->req_cons - (_r)->rsp_prod_pvt); \
+ req < rsp ? req : rsp; \
+})
+#else
+/* Same as above, but without the nice GCC ({ ... }) syntax. */
+#define RING_HAS_UNCONSUMED_REQUESTS(_r) \
+ ((((_r)->sring->req_prod - (_r)->req_cons) < \
+ (RING_SIZE(_r) - ((_r)->req_cons - (_r)->rsp_prod_pvt))) ? \
+ ((_r)->sring->req_prod - (_r)->req_cons) : \
+ (RING_SIZE(_r) - ((_r)->req_cons - (_r)->rsp_prod_pvt)))
+#endif
+
+/* Direct access to individual ring elements, by index. */
+#define RING_GET_REQUEST(_r, _idx) \
+ (&((_r)->sring->ring[((_idx) & (RING_SIZE(_r) - 1))].req))
+
+#define RING_GET_RESPONSE(_r, _idx) \
+ (&((_r)->sring->ring[((_idx) & (RING_SIZE(_r) - 1))].rsp))
+
+/* Loop termination condition: Would the specified index overflow the ring? */
+#define RING_REQUEST_CONS_OVERFLOW(_r, _cons) \
+ (((_cons) - (_r)->rsp_prod_pvt) >= RING_SIZE(_r))
+
+/* Ill-behaved frontend determination: Can there be this many requests? */
+#define RING_REQUEST_PROD_OVERFLOW(_r, _prod) \
+ (((_prod) - (_r)->rsp_prod_pvt) > RING_SIZE(_r))
+
+#define RING_PUSH_REQUESTS(_r) do { \
+ xen_wmb(); /* back sees requests /before/ updated producer index */ \
+ (_r)->sring->req_prod = (_r)->req_prod_pvt; \
+} while (0)
+
+#define RING_PUSH_RESPONSES(_r) do { \
+ xen_wmb(); /* front sees resps /before/ updated producer index */ \
+ (_r)->sring->rsp_prod = (_r)->rsp_prod_pvt; \
+} while (0)
+
+/*
+ * Notification hold-off (req_event and rsp_event):
+ *
+ * When queueing requests or responses on a shared ring, it may not always be
+ * necessary to notify the remote end. For example, if requests are in flight
+ * in a backend, the front may be able to queue further requests without
+ * notifying the back (if the back checks for new requests when it queues
+ * responses).
+ *
+ * When enqueuing requests or responses:
+ *
+ * Use RING_PUSH_{REQUESTS,RESPONSES}_AND_CHECK_NOTIFY(). The second argument
+ * is a boolean return value. True indicates that the receiver requires an
+ * asynchronous notification.
+ *
+ * After dequeuing requests or responses (before sleeping the connection):
+ *
+ * Use RING_FINAL_CHECK_FOR_REQUESTS() or RING_FINAL_CHECK_FOR_RESPONSES().
+ * The second argument is a boolean return value. True indicates that there
+ * are pending messages on the ring (i.e., the connection should not be put
+ * to sleep).
+ *
+ * These macros will set the req_event/rsp_event field to trigger a
+ * notification on the very next message that is enqueued. If you want to
+ * create batches of work (i.e., only receive a notification after several
+ * messages have been enqueued) then you will need to create a customised
+ * version of the FINAL_CHECK macro in your own code, which sets the event
+ * field appropriately.
+ */
+
+#define RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(_r, _notify) do { \
+ RING_IDX __old = (_r)->sring->req_prod; \
+ RING_IDX __new = (_r)->req_prod_pvt; \
+ xen_wmb(); /* back sees requests /before/ updated producer index */ \
+ (_r)->sring->req_prod = __new; \
+ xen_mb(); /* back sees new requests /before/ we check req_event */ \
+ (_notify) = ((RING_IDX)(__new - (_r)->sring->req_event) < \
+ (RING_IDX)(__new - __old)); \
+} while (0)
+
+#define RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(_r, _notify) do { \
+ RING_IDX __old = (_r)->sring->rsp_prod; \
+ RING_IDX __new = (_r)->rsp_prod_pvt; \
+ xen_wmb(); /* front sees resps /before/ updated producer index */ \
+ (_r)->sring->rsp_prod = __new; \
+ xen_mb(); /* front sees new resps /before/ we check rsp_event */ \
+ (_notify) = ((RING_IDX)(__new - (_r)->sring->rsp_event) < \
+ (RING_IDX)(__new - __old)); \
+} while (0)
+
+#define RING_FINAL_CHECK_FOR_REQUESTS(_r, _work_to_do) do { \
+ (_work_to_do) = RING_HAS_UNCONSUMED_REQUESTS(_r); \
+ if (_work_to_do) break; \
+ (_r)->sring->req_event = (_r)->req_cons + 1; \
+ xen_mb(); \
+ (_work_to_do) = RING_HAS_UNCONSUMED_REQUESTS(_r); \
+} while (0)
+
+#define RING_FINAL_CHECK_FOR_RESPONSES(_r, _work_to_do) do { \
+ (_work_to_do) = RING_HAS_UNCONSUMED_RESPONSES(_r); \
+ if (_work_to_do) break; \
+ (_r)->sring->rsp_event = (_r)->rsp_cons + 1; \
+ xen_mb(); \
+ (_work_to_do) = RING_HAS_UNCONSUMED_RESPONSES(_r); \
+} while (0)
+
+#endif /* __XEN_PUBLIC_IO_RING_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h b/OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h
new file mode 100644
index 0000000..927f9db
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h
@@ -0,0 +1,80 @@
+/*****************************************************************************
+ * xenbus.h
+ *
+ * Xenbus protocol details.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 XenSource Ltd.
+ */
+
+#ifndef _XEN_PUBLIC_IO_XENBUS_H
+#define _XEN_PUBLIC_IO_XENBUS_H
+
+/*
+ * The state of either end of the Xenbus, i.e. the current communication
+ * status of initialisation across the bus. States here imply nothing about
+ * the state of the connection between the driver and the kernel's device
+ * layers.
+ */
+enum xenbus_state {
+ XenbusStateUnknown = 0,
+
+ XenbusStateInitialising = 1,
+
+ /*
+ * InitWait: Finished early initialisation but waiting for information
+ * from the peer or hotplug scripts.
+ */
+ XenbusStateInitWait = 2,
+
+ /*
+ * Initialised: Waiting for a connection from the peer.
+ */
+ XenbusStateInitialised = 3,
+
+ XenbusStateConnected = 4,
+
+ /*
+ * Closing: The device is being closed due to an error or an unplug event.
+ */
+ XenbusStateClosing = 5,
+
+ XenbusStateClosed = 6,
+
+ /*
+ * Reconfiguring: The device is being reconfigured.
+ */
+ XenbusStateReconfiguring = 7,
+
+ XenbusStateReconfigured = 8
+};
+typedef enum xenbus_state XenbusState;
+
+#endif /* _XEN_PUBLIC_IO_XENBUS_H */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h b/OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h
new file mode 100644
index 0000000..4a9e667
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h
@@ -0,0 +1,138 @@
+/*
+ * Details of the "wire" protocol between Xen Store Daemon and client
+ * library or guest kernel.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 Rusty Russell IBM Corporation
+ */
+
+#ifndef _XS_WIRE_H
+#define _XS_WIRE_H
+
+enum xsd_sockmsg_type
+{
+ XS_DEBUG,
+ XS_DIRECTORY,
+ XS_READ,
+ XS_GET_PERMS,
+ XS_WATCH,
+ XS_UNWATCH,
+ XS_TRANSACTION_START,
+ XS_TRANSACTION_END,
+ XS_INTRODUCE,
+ XS_RELEASE,
+ XS_GET_DOMAIN_PATH,
+ XS_WRITE,
+ XS_MKDIR,
+ XS_RM,
+ XS_SET_PERMS,
+ XS_WATCH_EVENT,
+ XS_ERROR,
+ XS_IS_DOMAIN_INTRODUCED,
+ XS_RESUME,
+ XS_SET_TARGET,
+ XS_RESTRICT,
+ XS_RESET_WATCHES
+};
+
+#define XS_WRITE_NONE "NONE"
+#define XS_WRITE_CREATE "CREATE"
+#define XS_WRITE_CREATE_EXCL "CREATE|EXCL"
+
+/* We hand errors as strings, for portability. */
+struct xsd_errors
+{
+ INT32 errnum;
+ const CHAR8 *errstring;
+};
+#ifdef EINVAL
+#define XSD_ERROR(x) { x, #x }
+/* LINTED: static unused */
+static struct xsd_errors xsd_errors[]
+#if defined(__GNUC__)
+__attribute__((unused))
+#endif
+ = {
+ XSD_ERROR(EINVAL),
+ XSD_ERROR(EACCES),
+ XSD_ERROR(EEXIST),
+ XSD_ERROR(EISDIR),
+ XSD_ERROR(ENOENT),
+ XSD_ERROR(ENOMEM),
+ XSD_ERROR(ENOSPC),
+ XSD_ERROR(EIO),
+ XSD_ERROR(ENOTEMPTY),
+ XSD_ERROR(ENOSYS),
+ XSD_ERROR(EROFS),
+ XSD_ERROR(EBUSY),
+ XSD_ERROR(EAGAIN),
+ XSD_ERROR(EISCONN),
+ XSD_ERROR(E2BIG)
+};
+#endif
+
+struct xsd_sockmsg
+{
+ UINT32 type; /* XS_??? */
+ UINT32 req_id;/* Request identifier, echoed in daemon's response. */
+ UINT32 tx_id; /* Transaction id (0 if not related to a transaction). */
+ UINT32 len; /* Length of data following this. */
+
+ /* Generally followed by nul-terminated string(s). */
+};
+
+enum xs_watch_type
+{
+ XS_WATCH_PATH = 0,
+ XS_WATCH_TOKEN
+};
+
+/*
+ * `incontents 150 xenstore_struct XenStore wire protocol.
+ *
+ * Inter-domain shared memory communications. */
+#define XENSTORE_RING_SIZE 1024
+typedef UINT32 XENSTORE_RING_IDX;
+#define MASK_XENSTORE_IDX(idx) ((idx) & (XENSTORE_RING_SIZE-1))
+struct xenstore_domain_interface {
+ CHAR8 req[XENSTORE_RING_SIZE]; /* Requests to xenstore daemon. */
+ CHAR8 rsp[XENSTORE_RING_SIZE]; /* Replies and async watch events. */
+ XENSTORE_RING_IDX req_cons, req_prod;
+ XENSTORE_RING_IDX rsp_cons, rsp_prod;
+};
+
+/* Violating this is very bad. See docs/misc/xenstore.txt. */
+#define XENSTORE_PAYLOAD_MAX 4096
+
+/* Violating these just gets you an error back */
+#define XENSTORE_ABS_PATH_MAX 3072
+#define XENSTORE_REL_PATH_MAX 2048
+
+#endif /* _XS_WIRE_H */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/memory.h b/OvmfPkg/Include/IndustryStandard/Xen/memory.h
new file mode 100644
index 0000000..93b73b5
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/memory.h
@@ -0,0 +1,480 @@
+/******************************************************************************
+ * memory.h
+ *
+ * Memory reservation and information.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_MEMORY_H__
+#define __XEN_PUBLIC_MEMORY_H__
+
+#include "xen.h"
+
+/*
+ * Increase or decrease the specified domain's memory reservation. Returns the
+ * number of extents successfully allocated or freed.
+ * arg == addr of struct xen_memory_reservation.
+ */
+#define XENMEM_increase_reservation 0
+#define XENMEM_decrease_reservation 1
+#define XENMEM_populate_physmap 6
+
+#if __XEN_INTERFACE_VERSION__ >= 0x00030209
+/*
+ * Maximum # bits addressable by the user of the allocated region (e.g., I/O
+ * devices often have a 32-bit limitation even in 64-bit systems). If zero
+ * then the user has no addressing restriction. This field is not used by
+ * XENMEM_decrease_reservation.
+ */
+#define XENMEMF_address_bits(x) (x)
+#define XENMEMF_get_address_bits(x) ((x) & 0xffu)
+/* NUMA node to allocate from. */
+#define XENMEMF_node(x) (((x) + 1) << 8)
+#define XENMEMF_get_node(x) ((((x) >> 8) - 1) & 0xffu)
+/* Flag to populate physmap with populate-on-demand entries */
+#define XENMEMF_populate_on_demand (1<<16)
+/* Flag to request allocation only from the node specified */
+#define XENMEMF_exact_node_request (1<<17)
+#define XENMEMF_exact_node(n) (XENMEMF_node(n) | XENMEMF_exact_node_request)
+#endif
+
+struct xen_memory_reservation {
+
+ /*
+ * XENMEM_increase_reservation:
+ * OUT: MFN (*not* GMFN) bases of extents that were allocated
+ * XENMEM_decrease_reservation:
+ * IN: GMFN bases of extents to free
+ * XENMEM_populate_physmap:
+ * IN: GPFN bases of extents to populate with memory
+ * OUT: GMFN bases of extents that were allocated
+ * (NB. This command also updates the mach_to_phys translation table)
+ * XENMEM_claim_pages:
+ * IN: must be zero
+ */
+ XEN_GUEST_HANDLE(xen_pfn_t) extent_start;
+
+ /* Number of extents, and size/alignment of each (2^extent_order pages). */
+ xen_ulong_t nr_extents;
+ UINT32 extent_order;
+
+#if __XEN_INTERFACE_VERSION__ >= 0x00030209
+ /* XENMEMF flags. */
+ UINT32 mem_flags;
+#else
+ UINT32 address_bits;
+#endif
+
+ /*
+ * Domain whose reservation is being changed.
+ * Unprivileged domains can specify only DOMID_SELF.
+ */
+ domid_t domid;
+};
+typedef struct xen_memory_reservation xen_memory_reservation_t;
+DEFINE_XEN_GUEST_HANDLE(xen_memory_reservation_t);
+
+/*
+ * An atomic exchange of memory pages. If return code is zero then
+ * @out.extent_list provides GMFNs of the newly-allocated memory.
+ * Returns zero on complete success, otherwise a negative error code.
+ * On complete success then always @nr_exchanged == @in.nr_extents.
+ * On partial success @nr_exchanged indicates how much work was done.
+ */
+#define XENMEM_exchange 11
+struct xen_memory_exchange {
+ /*
+ * [IN] Details of memory extents to be exchanged (GMFN bases).
+ * Note that @in.address_bits is ignored and unused.
+ */
+ struct xen_memory_reservation in;
+
+ /*
+ * [IN/OUT] Details of new memory extents.
+ * We require that:
+ * 1. @in.domid == @out.domid
+ * 2. @in.nr_extents << @in.extent_order ==
+ * @out.nr_extents << @out.extent_order
+ * 3. @in.extent_start and @out.extent_start lists must not overlap
+ * 4. @out.extent_start lists GPFN bases to be populated
+ * 5. @out.extent_start is overwritten with allocated GMFN bases
+ */
+ struct xen_memory_reservation out;
+
+ /*
+ * [OUT] Number of input extents that were successfully exchanged:
+ * 1. The first @nr_exchanged input extents were successfully
+ * deallocated.
+ * 2. The corresponding first entries in the output extent list correctly
+ * indicate the GMFNs that were successfully exchanged.
+ * 3. All other input and output extents are untouched.
+ * 4. If not all input exents are exchanged then the return code of this
+ * command will be non-zero.
+ * 5. THIS FIELD MUST BE INITIALISED TO ZERO BY THE CALLER!
+ */
+ xen_ulong_t nr_exchanged;
+};
+typedef struct xen_memory_exchange xen_memory_exchange_t;
+DEFINE_XEN_GUEST_HANDLE(xen_memory_exchange_t);
+
+/*
+ * Returns the maximum machine frame number of mapped RAM in this system.
+ * This command always succeeds (it never returns an error code).
+ * arg == NULL.
+ */
+#define XENMEM_maximum_ram_page 2
+
+/*
+ * Returns the current or maximum memory reservation, in pages, of the
+ * specified domain (may be DOMID_SELF). Returns -ve errcode on failure.
+ * arg == addr of domid_t.
+ */
+#define XENMEM_current_reservation 3
+#define XENMEM_maximum_reservation 4
+
+/*
+ * Returns the maximum GPFN in use by the guest, or -ve errcode on failure.
+ */
+#define XENMEM_maximum_gpfn 14
+
+/*
+ * Returns a list of MFN bases of 2MB extents comprising the machine_to_phys
+ * mapping table. Architectures which do not have a m2p table do not implement
+ * this command.
+ * arg == addr of xen_machphys_mfn_list_t.
+ */
+#define XENMEM_machphys_mfn_list 5
+struct xen_machphys_mfn_list {
+ /*
+ * Size of the 'extent_start' array. Fewer entries will be filled if the
+ * machphys table is smaller than max_extents * 2MB.
+ */
+ UINT32 max_extents;
+
+ /*
+ * Pointer to buffer to fill with list of extent starts. If there are
+ * any large discontiguities in the machine address space, 2MB gaps in
+ * the machphys table will be represented by an MFN base of zero.
+ */
+ XEN_GUEST_HANDLE(xen_pfn_t) extent_start;
+
+ /*
+ * Number of extents written to the above array. This will be smaller
+ * than 'max_extents' if the machphys table is smaller than max_e * 2MB.
+ */
+ UINT32 nr_extents;
+};
+typedef struct xen_machphys_mfn_list xen_machphys_mfn_list_t;
+DEFINE_XEN_GUEST_HANDLE(xen_machphys_mfn_list_t);
+
+/*
+ * Returns the location in virtual address space of the machine_to_phys
+ * mapping table. Architectures which do not have a m2p table, or which do not
+ * map it by default into guest address space, do not implement this command.
+ * arg == addr of xen_machphys_mapping_t.
+ */
+#define XENMEM_machphys_mapping 12
+struct xen_machphys_mapping {
+ xen_ulong_t v_start, v_end; /* Start and end virtual addresses. */
+ xen_ulong_t max_mfn; /* Maximum MFN that can be looked up. */
+};
+typedef struct xen_machphys_mapping xen_machphys_mapping_t;
+DEFINE_XEN_GUEST_HANDLE(xen_machphys_mapping_t);
+
+/* Source mapping space. */
+/* ` enum phys_map_space { */
+#define XENMAPSPACE_shared_info 0 /* shared info page */
+#define XENMAPSPACE_grant_table 1 /* grant table page */
+#define XENMAPSPACE_gmfn 2 /* GMFN */
+#define XENMAPSPACE_gmfn_range 3 /* GMFN range, XENMEM_add_to_physmap only. */
+#define XENMAPSPACE_gmfn_foreign 4 /* GMFN from another dom,
+ * XENMEM_add_to_physmap_batch only. */
+/* ` } */
+
+/*
+ * Sets the GPFN at which a particular page appears in the specified guest's
+ * pseudophysical address space.
+ * arg == addr of xen_add_to_physmap_t.
+ */
+#define XENMEM_add_to_physmap 7
+struct xen_add_to_physmap {
+ /* Which domain to change the mapping for. */
+ domid_t domid;
+
+ /* Number of pages to go through for gmfn_range */
+ UINT16 size;
+
+ UINT32 space; /* => enum phys_map_space */
+
+#define XENMAPIDX_grant_table_status 0x80000000
+
+ /* Index into space being mapped. */
+ xen_ulong_t idx;
+
+ /* GPFN in domid where the source mapping page should appear. */
+ xen_pfn_t gpfn;
+};
+typedef struct xen_add_to_physmap xen_add_to_physmap_t;
+DEFINE_XEN_GUEST_HANDLE(xen_add_to_physmap_t);
+
+/* A batched version of add_to_physmap. */
+#define XENMEM_add_to_physmap_batch 23
+struct xen_add_to_physmap_batch {
+ /* IN */
+ /* Which domain to change the mapping for. */
+ domid_t domid;
+ UINT16 space; /* => enum phys_map_space */
+
+ /* Number of pages to go through */
+ UINT16 size;
+ domid_t foreign_domid; /* IFF gmfn_foreign */
+
+ /* Indexes into space being mapped. */
+ XEN_GUEST_HANDLE(xen_ulong_t) idxs;
+
+ /* GPFN in domid where the source mapping page should appear. */
+ XEN_GUEST_HANDLE(xen_pfn_t) gpfns;
+
+ /* OUT */
+
+ /* Per index error code. */
+ XEN_GUEST_HANDLE(INT32) errs;
+};
+typedef struct xen_add_to_physmap_batch xen_add_to_physmap_batch_t;
+DEFINE_XEN_GUEST_HANDLE(xen_add_to_physmap_batch_t);
+
+#if __XEN_INTERFACE_VERSION__ < 0x00040400
+#define XENMEM_add_to_physmap_range XENMEM_add_to_physmap_batch
+#define xen_add_to_physmap_range xen_add_to_physmap_batch
+typedef struct xen_add_to_physmap_batch xen_add_to_physmap_range_t;
+DEFINE_XEN_GUEST_HANDLE(xen_add_to_physmap_range_t);
+#endif
+
+/*
+ * Unmaps the page appearing at a particular GPFN from the specified guest's
+ * pseudophysical address space.
+ * arg == addr of xen_remove_from_physmap_t.
+ */
+#define XENMEM_remove_from_physmap 15
+struct xen_remove_from_physmap {
+ /* Which domain to change the mapping for. */
+ domid_t domid;
+
+ /* GPFN of the current mapping of the page. */
+ xen_pfn_t gpfn;
+};
+typedef struct xen_remove_from_physmap xen_remove_from_physmap_t;
+DEFINE_XEN_GUEST_HANDLE(xen_remove_from_physmap_t);
+
+/*** REMOVED ***/
+/*#define XENMEM_translate_gpfn_list 8*/
+
+/*
+ * Returns the pseudo-physical memory map as it was when the domain
+ * was started (specified by XENMEM_set_memory_map).
+ * arg == addr of xen_memory_map_t.
+ */
+#define XENMEM_memory_map 9
+struct xen_memory_map {
+ /*
+ * On call the number of entries which can be stored in buffer. On
+ * return the number of entries which have been stored in
+ * buffer.
+ */
+ UINT32 nr_entries;
+
+ /*
+ * Entries in the buffer are in the same format as returned by the
+ * BIOS INT 0x15 EAX=0xE820 call.
+ */
+ XEN_GUEST_HANDLE(VOID) buffer;
+};
+typedef struct xen_memory_map xen_memory_map_t;
+DEFINE_XEN_GUEST_HANDLE(xen_memory_map_t);
+
+/*
+ * Returns the real physical memory map. Passes the same structure as
+ * XENMEM_memory_map.
+ * arg == addr of xen_memory_map_t.
+ */
+#define XENMEM_machine_memory_map 10
+
+/*
+ * Set the pseudo-physical memory map of a domain, as returned by
+ * XENMEM_memory_map.
+ * arg == addr of xen_foreign_memory_map_t.
+ */
+#define XENMEM_set_memory_map 13
+struct xen_foreign_memory_map {
+ domid_t domid;
+ struct xen_memory_map map;
+};
+typedef struct xen_foreign_memory_map xen_foreign_memory_map_t;
+DEFINE_XEN_GUEST_HANDLE(xen_foreign_memory_map_t);
+
+#define XENMEM_set_pod_target 16
+#define XENMEM_get_pod_target 17
+struct xen_pod_target {
+ /* IN */
+ UINT64 target_pages;
+ /* OUT */
+ UINT64 tot_pages;
+ UINT64 pod_cache_pages;
+ UINT64 pod_entries;
+ /* IN */
+ domid_t domid;
+};
+typedef struct xen_pod_target xen_pod_target_t;
+
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+
+#ifndef uint64_aligned_t
+#define uint64_aligned_t UINT64
+#endif
+
+/*
+ * Get the number of MFNs saved through memory sharing.
+ * The call never fails.
+ */
+#define XENMEM_get_sharing_freed_pages 18
+#define XENMEM_get_sharing_shared_pages 19
+
+#define XENMEM_paging_op 20
+#define XENMEM_paging_op_nominate 0
+#define XENMEM_paging_op_evict 1
+#define XENMEM_paging_op_prep 2
+
+#define XENMEM_access_op 21
+#define XENMEM_access_op_resume 0
+
+struct xen_mem_event_op {
+ UINT8 op; /* XENMEM_*_op_* */
+ domid_t domain;
+
+
+ /* PAGING_PREP IN: buffer to immediately fill page in */
+ uint64_aligned_t buffer;
+ /* Other OPs */
+ uint64_aligned_t gfn; /* IN: gfn of page being operated on */
+};
+typedef struct xen_mem_event_op xen_mem_event_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_mem_event_op_t);
+
+#define XENMEM_sharing_op 22
+#define XENMEM_sharing_op_nominate_gfn 0
+#define XENMEM_sharing_op_nominate_gref 1
+#define XENMEM_sharing_op_share 2
+#define XENMEM_sharing_op_resume 3
+#define XENMEM_sharing_op_debug_gfn 4
+#define XENMEM_sharing_op_debug_mfn 5
+#define XENMEM_sharing_op_debug_gref 6
+#define XENMEM_sharing_op_add_physmap 7
+#define XENMEM_sharing_op_audit 8
+
+#define XENMEM_SHARING_OP_S_HANDLE_INVALID (-10)
+#define XENMEM_SHARING_OP_C_HANDLE_INVALID (-9)
+
+/* The following allows sharing of grant refs. This is useful
+ * for sharing utilities sitting as "filters" in IO backends
+ * (e.g. memshr + blktap(2)). The IO backend is only exposed
+ * to grant references, and this allows sharing of the grefs */
+#define XENMEM_SHARING_OP_FIELD_IS_GREF_FLAG (1ULL << 62)
+
+#define XENMEM_SHARING_OP_FIELD_MAKE_GREF(field, val) \
+ (field) = (XENMEM_SHARING_OP_FIELD_IS_GREF_FLAG | val)
+#define XENMEM_SHARING_OP_FIELD_IS_GREF(field) \
+ ((field) & XENMEM_SHARING_OP_FIELD_IS_GREF_FLAG)
+#define XENMEM_SHARING_OP_FIELD_GET_GREF(field) \
+ ((field) & (~XENMEM_SHARING_OP_FIELD_IS_GREF_FLAG))
+
+struct xen_mem_sharing_op {
+ UINT8 op; /* XENMEM_sharing_op_* */
+ domid_t domain;
+
+ union {
+ struct mem_sharing_op_nominate { /* OP_NOMINATE_xxx */
+ union {
+ uint64_aligned_t gfn; /* IN: gfn to nominate */
+ UINT32 grant_ref; /* IN: grant ref to nominate */
+ } u;
+ uint64_aligned_t handle; /* OUT: the handle */
+ } nominate;
+ struct mem_sharing_op_share { /* OP_SHARE/ADD_PHYSMAP */
+ uint64_aligned_t source_gfn; /* IN: the gfn of the source page */
+ uint64_aligned_t source_handle; /* IN: handle to the source page */
+ uint64_aligned_t client_gfn; /* IN: the client gfn */
+ uint64_aligned_t client_handle; /* IN: handle to the client page */
+ domid_t client_domain; /* IN: the client domain id */
+ } share;
+ struct mem_sharing_op_debug { /* OP_DEBUG_xxx */
+ union {
+ uint64_aligned_t gfn; /* IN: gfn to debug */
+ uint64_aligned_t mfn; /* IN: mfn to debug */
+ UINT32 gref; /* IN: gref to debug */
+ } u;
+ } debug;
+ } u;
+};
+typedef struct xen_mem_sharing_op xen_mem_sharing_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_mem_sharing_op_t);
+
+/*
+ * Attempt to stake a claim for a domain on a quantity of pages
+ * of system RAM, but _not_ assign specific pageframes. Only
+ * arithmetic is performed so the hypercall is very fast and need
+ * not be preemptible, thus sidestepping time-of-check-time-of-use
+ * races for memory allocation. Returns 0 if the hypervisor page
+ * allocator has atomically and successfully claimed the requested
+ * number of pages, else non-zero.
+ *
+ * Any domain may have only one active claim. When sufficient memory
+ * has been allocated to resolve the claim, the claim silently expires.
+ * Claiming zero pages effectively resets any outstanding claim and
+ * is always successful.
+ *
+ * Note that a valid claim may be staked even after memory has been
+ * allocated for a domain. In this case, the claim is not incremental,
+ * i.e. if the domain's tot_pages is 3, and a claim is staked for 10,
+ * only 7 additional pages are claimed.
+ *
+ * Caller must be privileged or the hypercall fails.
+ */
+#define XENMEM_claim_pages 24
+
+/*
+ * XENMEM_claim_pages flags - the are no flags at this time.
+ * The zero value is appropiate.
+ */
+
+#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
+
+#endif /* __XEN_PUBLIC_MEMORY_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/sched.h b/OvmfPkg/Include/IndustryStandard/Xen/sched.h
new file mode 100644
index 0000000..b93123f
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/sched.h
@@ -0,0 +1,174 @@
+/******************************************************************************
+ * sched.h
+ *
+ * Scheduler state interactions
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_SCHED_H__
+#define __XEN_PUBLIC_SCHED_H__
+
+#include "event_channel.h"
+
+/*
+ * `incontents 150 sched Guest Scheduler Operations
+ *
+ * The SCHEDOP interface provides mechanisms for a guest to interact
+ * with the scheduler, including yield, blocking and shutting itself
+ * down.
+ */
+
+/*
+ * The prototype for this hypercall is:
+ * ` INTN HYPERVISOR_sched_op(enum sched_op cmd, VOID *arg, ...)
+ *
+ * @cmd == SCHEDOP_??? (scheduler operation).
+ * @arg == Operation-specific extra argument(s), as described below.
+ * ... == Additional Operation-specific extra arguments, described below.
+ *
+ * Versions of Xen prior to 3.0.2 provided only the following legacy version
+ * of this hypercall, supporting only the commands yield, block and shutdown:
+ * INTN sched_op(INT32 cmd, UINTN arg)
+ * @cmd == SCHEDOP_??? (scheduler operation).
+ * @arg == 0 (SCHEDOP_yield and SCHEDOP_block)
+ * == SHUTDOWN_* code (SCHEDOP_shutdown)
+ *
+ * This legacy version is available to new guests as:
+ * ` INTN HYPERVISOR_sched_op_compat(enum sched_op cmd, UINTN arg)
+ */
+
+/* ` enum sched_op { // SCHEDOP_* => struct sched_* */
+/*
+ * Voluntarily yield the CPU.
+ * @arg == NULL.
+ */
+#define SCHEDOP_yield 0
+
+/*
+ * Block execution of this VCPU until an event is received for processing.
+ * If called with event upcalls masked, this operation will atomically
+ * reenable event delivery and check for pending events before blocking the
+ * VCPU. This avoids a "wakeup waiting" race.
+ * @arg == NULL.
+ */
+#define SCHEDOP_block 1
+
+/*
+ * Halt execution of this domain (all VCPUs) and notify the system controller.
+ * @arg == pointer to sched_shutdown_t structure.
+ *
+ * If the sched_shutdown_t reason is SHUTDOWN_suspend then this
+ * hypercall takes an additional extra argument which should be the
+ * MFN of the guest's start_info_t.
+ *
+ * In addition, which reason is SHUTDOWN_suspend this hypercall
+ * returns 1 if suspend was cancelled or the domain was merely
+ * checkpointed, and 0 if it is resuming in a new domain.
+ */
+#define SCHEDOP_shutdown 2
+
+/*
+ * Poll a set of event-channel ports. Return when one or more are pending. An
+ * optional timeout may be specified.
+ * @arg == pointer to sched_poll_t structure.
+ */
+#define SCHEDOP_poll 3
+
+/*
+ * Declare a shutdown for another domain. The main use of this function is
+ * in interpreting shutdown requests and reasons for fully-virtualized
+ * domains. A para-virtualized domain may use SCHEDOP_shutdown directly.
+ * @arg == pointer to sched_remote_shutdown_t structure.
+ */
+#define SCHEDOP_remote_shutdown 4
+
+/*
+ * Latch a shutdown code, so that when the domain later shuts down it
+ * reports this code to the control tools.
+ * @arg == sched_shutdown_t, as for SCHEDOP_shutdown.
+ */
+#define SCHEDOP_shutdown_code 5
+
+/*
+ * Setup, poke and destroy a domain watchdog timer.
+ * @arg == pointer to sched_watchdog_t structure.
+ * With id == 0, setup a domain watchdog timer to cause domain shutdown
+ * after timeout, returns watchdog id.
+ * With id != 0 and timeout == 0, destroy domain watchdog timer.
+ * With id != 0 and timeout != 0, poke watchdog timer and set new timeout.
+ */
+#define SCHEDOP_watchdog 6
+/* ` } */
+
+struct sched_shutdown {
+ UINT32 reason; /* SHUTDOWN_* => enum sched_shutdown_reason */
+};
+typedef struct sched_shutdown sched_shutdown_t;
+DEFINE_XEN_GUEST_HANDLE(sched_shutdown_t);
+
+struct sched_poll {
+ XEN_GUEST_HANDLE(evtchn_port_t) ports;
+ UINT32 nr_ports;
+ UINT64 timeout;
+};
+typedef struct sched_poll sched_poll_t;
+DEFINE_XEN_GUEST_HANDLE(sched_poll_t);
+
+struct sched_remote_shutdown {
+ domid_t domain_id; /* Remote domain ID */
+ UINT32 reason; /* SHUTDOWN_* => enum sched_shutdown_reason */
+};
+typedef struct sched_remote_shutdown sched_remote_shutdown_t;
+DEFINE_XEN_GUEST_HANDLE(sched_remote_shutdown_t);
+
+struct sched_watchdog {
+ UINT32 id; /* watchdog ID */
+ UINT32 timeout; /* timeout */
+};
+typedef struct sched_watchdog sched_watchdog_t;
+DEFINE_XEN_GUEST_HANDLE(sched_watchdog_t);
+
+/*
+ * Reason codes for SCHEDOP_shutdown. These may be interpreted by control
+ * software to determine the appropriate action. For the most part, Xen does
+ * not care about the shutdown code.
+ */
+/* ` enum sched_shutdown_reason { */
+#define SHUTDOWN_poweroff 0 /* Domain exited normally. Clean up and kill. */
+#define SHUTDOWN_reboot 1 /* Clean up, kill, and then restart. */
+#define SHUTDOWN_suspend 2 /* Clean up, save suspend info, kill. */
+#define SHUTDOWN_crash 3 /* Tell controller we've crashed. */
+#define SHUTDOWN_watchdog 4 /* Restart because watchdog time expired. */
+#define SHUTDOWN_MAX 4 /* Maximum valid shutdown reason. */
+/* ` } */
+
+#endif /* __XEN_PUBLIC_SCHED_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/trace.h b/OvmfPkg/Include/IndustryStandard/Xen/trace.h
new file mode 100644
index 0000000..f7334ce
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/trace.h
@@ -0,0 +1,310 @@
+/******************************************************************************
+ * include/public/trace.h
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Mark Williamson, (C) 2004 Intel Research Cambridge
+ * Copyright (C) 2005 Bin Ren
+ */
+
+#ifndef __XEN_PUBLIC_TRACE_H__
+#define __XEN_PUBLIC_TRACE_H__
+
+#define TRACE_EXTRA_MAX 7
+#define TRACE_EXTRA_SHIFT 28
+
+/* Trace classes */
+#define TRC_CLS_SHIFT 16
+#define TRC_GEN 0x0001f000 /* General trace */
+#define TRC_SCHED 0x0002f000 /* Xen Scheduler trace */
+#define TRC_DOM0OP 0x0004f000 /* Xen DOM0 operation trace */
+#define TRC_HVM 0x0008f000 /* Xen HVM trace */
+#define TRC_MEM 0x0010f000 /* Xen memory trace */
+#define TRC_PV 0x0020f000 /* Xen PV traces */
+#define TRC_SHADOW 0x0040f000 /* Xen shadow tracing */
+#define TRC_HW 0x0080f000 /* Xen hardware-related traces */
+#define TRC_GUEST 0x0800f000 /* Guest-generated traces */
+#define TRC_ALL 0x0ffff000
+#define TRC_HD_TO_EVENT(x) ((x)&0x0fffffff)
+#define TRC_HD_CYCLE_FLAG (1UL<<31)
+#define TRC_HD_INCLUDES_CYCLE_COUNT(x) ( !!( (x) & TRC_HD_CYCLE_FLAG ) )
+#define TRC_HD_EXTRA(x) (((x)>>TRACE_EXTRA_SHIFT)&TRACE_EXTRA_MAX)
+
+/* Trace subclasses */
+#define TRC_SUBCLS_SHIFT 12
+
+/* trace subclasses for SVM */
+#define TRC_HVM_ENTRYEXIT 0x00081000 /* VMENTRY and #VMEXIT */
+#define TRC_HVM_HANDLER 0x00082000 /* various HVM handlers */
+
+#define TRC_SCHED_MIN 0x00021000 /* Just runstate changes */
+#define TRC_SCHED_CLASS 0x00022000 /* Scheduler-specific */
+#define TRC_SCHED_VERBOSE 0x00028000 /* More inclusive scheduling */
+
+/*
+ * The highest 3 bits of the last 12 bits of TRC_SCHED_CLASS above are
+ * reserved for encoding what scheduler produced the information. The
+ * actual event is encoded in the last 9 bits.
+ *
+ * This means we have 8 scheduling IDs available (which means at most 8
+ * schedulers generating events) and, in each scheduler, up to 512
+ * different events.
+ */
+#define TRC_SCHED_ID_BITS 3
+#define TRC_SCHED_ID_SHIFT (TRC_SUBCLS_SHIFT - TRC_SCHED_ID_BITS)
+#define TRC_SCHED_ID_MASK (((1UL<<TRC_SCHED_ID_BITS) - 1) << TRC_SCHED_ID_SHIFT)
+#define TRC_SCHED_EVT_MASK (~(TRC_SCHED_ID_MASK))
+
+/* Per-scheduler IDs, to identify scheduler specific events */
+#define TRC_SCHED_CSCHED 0
+#define TRC_SCHED_CSCHED2 1
+#define TRC_SCHED_SEDF 2
+#define TRC_SCHED_ARINC653 3
+
+/* Per-scheduler tracing */
+#define TRC_SCHED_CLASS_EVT(_c, _e) \
+ ( ( TRC_SCHED_CLASS | \
+ ((TRC_SCHED_##_c << TRC_SCHED_ID_SHIFT) & TRC_SCHED_ID_MASK) ) + \
+ (_e & TRC_SCHED_EVT_MASK) )
+
+/* Trace classes for Hardware */
+#define TRC_HW_PM 0x00801000 /* Power management traces */
+#define TRC_HW_IRQ 0x00802000 /* Traces relating to the handling of IRQs */
+
+/* Trace events per class */
+#define TRC_LOST_RECORDS (TRC_GEN + 1)
+#define TRC_TRACE_WRAP_BUFFER (TRC_GEN + 2)
+#define TRC_TRACE_CPU_CHANGE (TRC_GEN + 3)
+
+#define TRC_SCHED_RUNSTATE_CHANGE (TRC_SCHED_MIN + 1)
+#define TRC_SCHED_CONTINUE_RUNNING (TRC_SCHED_MIN + 2)
+#define TRC_SCHED_DOM_ADD (TRC_SCHED_VERBOSE + 1)
+#define TRC_SCHED_DOM_REM (TRC_SCHED_VERBOSE + 2)
+#define TRC_SCHED_SLEEP (TRC_SCHED_VERBOSE + 3)
+#define TRC_SCHED_WAKE (TRC_SCHED_VERBOSE + 4)
+#define TRC_SCHED_YIELD (TRC_SCHED_VERBOSE + 5)
+#define TRC_SCHED_BLOCK (TRC_SCHED_VERBOSE + 6)
+#define TRC_SCHED_SHUTDOWN (TRC_SCHED_VERBOSE + 7)
+#define TRC_SCHED_CTL (TRC_SCHED_VERBOSE + 8)
+#define TRC_SCHED_ADJDOM (TRC_SCHED_VERBOSE + 9)
+#define TRC_SCHED_SWITCH (TRC_SCHED_VERBOSE + 10)
+#define TRC_SCHED_S_TIMER_FN (TRC_SCHED_VERBOSE + 11)
+#define TRC_SCHED_T_TIMER_FN (TRC_SCHED_VERBOSE + 12)
+#define TRC_SCHED_DOM_TIMER_FN (TRC_SCHED_VERBOSE + 13)
+#define TRC_SCHED_SWITCH_INFPREV (TRC_SCHED_VERBOSE + 14)
+#define TRC_SCHED_SWITCH_INFNEXT (TRC_SCHED_VERBOSE + 15)
+#define TRC_SCHED_SHUTDOWN_CODE (TRC_SCHED_VERBOSE + 16)
+
+#define TRC_MEM_PAGE_GRANT_MAP (TRC_MEM + 1)
+#define TRC_MEM_PAGE_GRANT_UNMAP (TRC_MEM + 2)
+#define TRC_MEM_PAGE_GRANT_TRANSFER (TRC_MEM + 3)
+#define TRC_MEM_SET_P2M_ENTRY (TRC_MEM + 4)
+#define TRC_MEM_DECREASE_RESERVATION (TRC_MEM + 5)
+#define TRC_MEM_POD_POPULATE (TRC_MEM + 16)
+#define TRC_MEM_POD_ZERO_RECLAIM (TRC_MEM + 17)
+#define TRC_MEM_POD_SUPERPAGE_SPLINTER (TRC_MEM + 18)
+
+#define TRC_PV_ENTRY 0x00201000 /* Hypervisor entry points for PV guests. */
+#define TRC_PV_SUBCALL 0x00202000 /* Sub-call in a multicall hypercall */
+
+#define TRC_PV_HYPERCALL (TRC_PV_ENTRY + 1)
+#define TRC_PV_TRAP (TRC_PV_ENTRY + 3)
+#define TRC_PV_PAGE_FAULT (TRC_PV_ENTRY + 4)
+#define TRC_PV_FORCED_INVALID_OP (TRC_PV_ENTRY + 5)
+#define TRC_PV_EMULATE_PRIVOP (TRC_PV_ENTRY + 6)
+#define TRC_PV_EMULATE_4GB (TRC_PV_ENTRY + 7)
+#define TRC_PV_MATH_STATE_RESTORE (TRC_PV_ENTRY + 8)
+#define TRC_PV_PAGING_FIXUP (TRC_PV_ENTRY + 9)
+#define TRC_PV_GDT_LDT_MAPPING_FAULT (TRC_PV_ENTRY + 10)
+#define TRC_PV_PTWR_EMULATION (TRC_PV_ENTRY + 11)
+#define TRC_PV_PTWR_EMULATION_PAE (TRC_PV_ENTRY + 12)
+#define TRC_PV_HYPERCALL_V2 (TRC_PV_ENTRY + 13)
+#define TRC_PV_HYPERCALL_SUBCALL (TRC_PV_SUBCALL + 14)
+
+/*
+ * TRC_PV_HYPERCALL_V2 format
+ *
+ * Only some of the hypercall argument are recorded. Bit fields A0 to
+ * A5 in the first extra word are set if the argument is present and
+ * the arguments themselves are packed sequentially in the following
+ * words.
+ *
+ * The TRC_64_FLAG bit is not set for these events (even if there are
+ * 64-bit arguments in the record).
+ *
+ * Word
+ * 0 bit 31 30|29 28|27 26|25 24|23 22|21 20|19 ... 0
+ * A5 |A4 |A3 |A2 |A1 |A0 |Hypercall op
+ * 1 First 32 bit (or low word of first 64 bit) arg in record
+ * 2 Second 32 bit (or high word of first 64 bit) arg in record
+ * ...
+ *
+ * A0-A5 bitfield values:
+ *
+ * 00b Argument not present
+ * 01b 32-bit argument present
+ * 10b 64-bit argument present
+ * 11b Reserved
+ */
+#define TRC_PV_HYPERCALL_V2_ARG_32(i) (0x1 << (20 + 2*(i)))
+#define TRC_PV_HYPERCALL_V2_ARG_64(i) (0x2 << (20 + 2*(i)))
+#define TRC_PV_HYPERCALL_V2_ARG_MASK (0xfff00000)
+
+#define TRC_SHADOW_NOT_SHADOW (TRC_SHADOW + 1)
+#define TRC_SHADOW_FAST_PROPAGATE (TRC_SHADOW + 2)
+#define TRC_SHADOW_FAST_MMIO (TRC_SHADOW + 3)
+#define TRC_SHADOW_FALSE_FAST_PATH (TRC_SHADOW + 4)
+#define TRC_SHADOW_MMIO (TRC_SHADOW + 5)
+#define TRC_SHADOW_FIXUP (TRC_SHADOW + 6)
+#define TRC_SHADOW_DOMF_DYING (TRC_SHADOW + 7)
+#define TRC_SHADOW_EMULATE (TRC_SHADOW + 8)
+#define TRC_SHADOW_EMULATE_UNSHADOW_USER (TRC_SHADOW + 9)
+#define TRC_SHADOW_EMULATE_UNSHADOW_EVTINJ (TRC_SHADOW + 10)
+#define TRC_SHADOW_EMULATE_UNSHADOW_UNHANDLED (TRC_SHADOW + 11)
+#define TRC_SHADOW_WRMAP_BF (TRC_SHADOW + 12)
+#define TRC_SHADOW_PREALLOC_UNPIN (TRC_SHADOW + 13)
+#define TRC_SHADOW_RESYNC_FULL (TRC_SHADOW + 14)
+#define TRC_SHADOW_RESYNC_ONLY (TRC_SHADOW + 15)
+
+/* trace events per subclass */
+#define TRC_HVM_NESTEDFLAG (0x400)
+#define TRC_HVM_VMENTRY (TRC_HVM_ENTRYEXIT + 0x01)
+#define TRC_HVM_VMEXIT (TRC_HVM_ENTRYEXIT + 0x02)
+#define TRC_HVM_VMEXIT64 (TRC_HVM_ENTRYEXIT + TRC_64_FLAG + 0x02)
+#define TRC_HVM_PF_XEN (TRC_HVM_HANDLER + 0x01)
+#define TRC_HVM_PF_XEN64 (TRC_HVM_HANDLER + TRC_64_FLAG + 0x01)
+#define TRC_HVM_PF_INJECT (TRC_HVM_HANDLER + 0x02)
+#define TRC_HVM_PF_INJECT64 (TRC_HVM_HANDLER + TRC_64_FLAG + 0x02)
+#define TRC_HVM_INJ_EXC (TRC_HVM_HANDLER + 0x03)
+#define TRC_HVM_INJ_VIRQ (TRC_HVM_HANDLER + 0x04)
+#define TRC_HVM_REINJ_VIRQ (TRC_HVM_HANDLER + 0x05)
+#define TRC_HVM_IO_READ (TRC_HVM_HANDLER + 0x06)
+#define TRC_HVM_IO_WRITE (TRC_HVM_HANDLER + 0x07)
+#define TRC_HVM_CR_READ (TRC_HVM_HANDLER + 0x08)
+#define TRC_HVM_CR_READ64 (TRC_HVM_HANDLER + TRC_64_FLAG + 0x08)
+#define TRC_HVM_CR_WRITE (TRC_HVM_HANDLER + 0x09)
+#define TRC_HVM_CR_WRITE64 (TRC_HVM_HANDLER + TRC_64_FLAG + 0x09)
+#define TRC_HVM_DR_READ (TRC_HVM_HANDLER + 0x0A)
+#define TRC_HVM_DR_WRITE (TRC_HVM_HANDLER + 0x0B)
+#define TRC_HVM_MSR_READ (TRC_HVM_HANDLER + 0x0C)
+#define TRC_HVM_MSR_WRITE (TRC_HVM_HANDLER + 0x0D)
+#define TRC_HVM_CPUID (TRC_HVM_HANDLER + 0x0E)
+#define TRC_HVM_INTR (TRC_HVM_HANDLER + 0x0F)
+#define TRC_HVM_NMI (TRC_HVM_HANDLER + 0x10)
+#define TRC_HVM_SMI (TRC_HVM_HANDLER + 0x11)
+#define TRC_HVM_VMMCALL (TRC_HVM_HANDLER + 0x12)
+#define TRC_HVM_HLT (TRC_HVM_HANDLER + 0x13)
+#define TRC_HVM_INVLPG (TRC_HVM_HANDLER + 0x14)
+#define TRC_HVM_INVLPG64 (TRC_HVM_HANDLER + TRC_64_FLAG + 0x14)
+#define TRC_HVM_MCE (TRC_HVM_HANDLER + 0x15)
+#define TRC_HVM_IOPORT_READ (TRC_HVM_HANDLER + 0x16)
+#define TRC_HVM_IOMEM_READ (TRC_HVM_HANDLER + 0x17)
+#define TRC_HVM_CLTS (TRC_HVM_HANDLER + 0x18)
+#define TRC_HVM_LMSW (TRC_HVM_HANDLER + 0x19)
+#define TRC_HVM_LMSW64 (TRC_HVM_HANDLER + TRC_64_FLAG + 0x19)
+#define TRC_HVM_RDTSC (TRC_HVM_HANDLER + 0x1a)
+#define TRC_HVM_INTR_WINDOW (TRC_HVM_HANDLER + 0x20)
+#define TRC_HVM_NPF (TRC_HVM_HANDLER + 0x21)
+#define TRC_HVM_REALMODE_EMULATE (TRC_HVM_HANDLER + 0x22)
+#define TRC_HVM_TRAP (TRC_HVM_HANDLER + 0x23)
+#define TRC_HVM_TRAP_DEBUG (TRC_HVM_HANDLER + 0x24)
+#define TRC_HVM_VLAPIC (TRC_HVM_HANDLER + 0x25)
+
+#define TRC_HVM_IOPORT_WRITE (TRC_HVM_HANDLER + 0x216)
+#define TRC_HVM_IOMEM_WRITE (TRC_HVM_HANDLER + 0x217)
+
+/* trace events for per class */
+#define TRC_PM_FREQ_CHANGE (TRC_HW_PM + 0x01)
+#define TRC_PM_IDLE_ENTRY (TRC_HW_PM + 0x02)
+#define TRC_PM_IDLE_EXIT (TRC_HW_PM + 0x03)
+
+/* Trace events for IRQs */
+#define TRC_HW_IRQ_MOVE_CLEANUP_DELAY (TRC_HW_IRQ + 0x1)
+#define TRC_HW_IRQ_MOVE_CLEANUP (TRC_HW_IRQ + 0x2)
+#define TRC_HW_IRQ_BIND_VECTOR (TRC_HW_IRQ + 0x3)
+#define TRC_HW_IRQ_CLEAR_VECTOR (TRC_HW_IRQ + 0x4)
+#define TRC_HW_IRQ_MOVE_FINISH (TRC_HW_IRQ + 0x5)
+#define TRC_HW_IRQ_ASSIGN_VECTOR (TRC_HW_IRQ + 0x6)
+#define TRC_HW_IRQ_UNMAPPED_VECTOR (TRC_HW_IRQ + 0x7)
+#define TRC_HW_IRQ_HANDLED (TRC_HW_IRQ + 0x8)
+
+/*
+ * Event Flags
+ *
+ * Some events (e.g, TRC_PV_TRAP and TRC_HVM_IOMEM_READ) have multiple
+ * record formats. These event flags distinguish between the
+ * different formats.
+ */
+#define TRC_64_FLAG 0x100 /* Addresses are 64 bits (instead of 32 bits) */
+
+/* This structure represents a single trace buffer record. */
+struct t_rec {
+ UINT32 event:28;
+ UINT32 extra_u32:3; /* # entries in trailing extra_u32[] array */
+ UINT32 cycles_included:1; /* u.cycles or u.no_cycles? */
+ union {
+ struct {
+ UINT32 cycles_lo, cycles_hi; /* cycle counter timestamp */
+ UINT32 extra_u32[7]; /* event data items */
+ } cycles;
+ struct {
+ UINT32 extra_u32[7]; /* event data items */
+ } nocycles;
+ } u;
+};
+
+/*
+ * This structure contains the metadata for a single trace buffer. The head
+ * field, indexes into an array of struct t_rec's.
+ */
+struct t_buf {
+ /* Assume the data buffer size is X. X is generally not a power of 2.
+ * CONS and PROD are incremented modulo (2*X):
+ * 0 <= cons < 2*X
+ * 0 <= prod < 2*X
+ * This is done because addition modulo X breaks at 2^32 when X is not a
+ * power of 2:
+ * (((2^32 - 1) % X) + 1) % X != (2^32) % X
+ */
+ UINT32 cons; /* Offset of next item to be consumed by control tools. */
+ UINT32 prod; /* Offset of next item to be produced by Xen. */
+ /* Records follow immediately after the meta-data header. */
+};
+
+/* Structure used to pass MFNs to the trace buffers back to trace consumers.
+ * Offset is an offset into the mapped structure where the mfn list will be held.
+ * MFNs will be at ((UINTN *)(t_info))+(t_info->cpu_offset[cpu]).
+ */
+struct t_info {
+ UINT16 tbuf_size; /* Size in pages of each trace buffer */
+ UINT16 mfn_offset[]; /* Offset within t_info structure of the page list per cpu */
+ /* MFN lists immediately after the header */
+};
+
+#endif /* __XEN_PUBLIC_TRACE_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h b/OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h
new file mode 100644
index 0000000..3eb80a0
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h
@@ -0,0 +1,44 @@
+/******************************************************************************
+ * xen-compat.h
+ *
+ * Guest OS interface to Xen. Compatibility layer.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2006, Christian Limpach
+ */
+
+#ifndef __XEN_PUBLIC_XEN_COMPAT_H__
+#define __XEN_PUBLIC_XEN_COMPAT_H__
+
+#define __XEN_LATEST_INTERFACE_VERSION__ 0x00040400
+
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+/* Xen is built with matching headers and implements the latest interface. */
+#define __XEN_INTERFACE_VERSION__ __XEN_LATEST_INTERFACE_VERSION__
+#elif !defined(__XEN_INTERFACE_VERSION__)
+/* Guests which do not specify a version get the legacy interface. */
+#define __XEN_INTERFACE_VERSION__ 0x00000000
+#endif
+
+#if __XEN_INTERFACE_VERSION__ > __XEN_LATEST_INTERFACE_VERSION__
+#error "These header files do not support the requested interface version."
+#endif
+
+#endif /* __XEN_PUBLIC_XEN_COMPAT_H__ */
diff --git a/OvmfPkg/Include/IndustryStandard/Xen/xen.h b/OvmfPkg/Include/IndustryStandard/Xen/xen.h
new file mode 100644
index 0000000..e19f30c
--- /dev/null
+++ b/OvmfPkg/Include/IndustryStandard/Xen/xen.h
@@ -0,0 +1,897 @@
+/******************************************************************************
+ * xen.h
+ *
+ * Guest OS interface to Xen.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_XEN_H__
+#define __XEN_PUBLIC_XEN_H__
+
+#include "xen-compat.h"
+
+#if defined(__i386__) || defined(__x86_64__)
+#include "arch-x86/xen.h"
+#elif defined(__arm__) || defined (__aarch64__)
+#include "arch-arm.h"
+#else
+#error "Unsupported architecture"
+#endif
+
+#ifndef __ASSEMBLY__
+/* Guest handles for primitive C types. */
+DEFINE_XEN_GUEST_HANDLE(CHAR8);
+/* __DEFINE_XEN_GUEST_HANDLE(uchar, unsigned char); */
+DEFINE_XEN_GUEST_HANDLE(INT32);
+__DEFINE_XEN_GUEST_HANDLE(uint, UINT32);
+#if __XEN_INTERFACE_VERSION__ < 0x00040300
+DEFINE_XEN_GUEST_HANDLE(INTN);
+__DEFINE_XEN_GUEST_HANDLE(ulong, UINTN);
+#endif
+DEFINE_XEN_GUEST_HANDLE(VOID);
+
+DEFINE_XEN_GUEST_HANDLE(UINT64);
+DEFINE_XEN_GUEST_HANDLE(xen_pfn_t);
+DEFINE_XEN_GUEST_HANDLE(xen_ulong_t);
+#endif
+
+/*
+ * HYPERCALLS
+ */
+
+/* `incontents 100 hcalls List of hypercalls
+ * ` enum hypercall_num { // __HYPERVISOR_* => HYPERVISOR_*()
+ */
+
+#define __HYPERVISOR_set_trap_table 0
+#define __HYPERVISOR_mmu_update 1
+#define __HYPERVISOR_set_gdt 2
+#define __HYPERVISOR_stack_switch 3
+#define __HYPERVISOR_set_callbacks 4
+#define __HYPERVISOR_fpu_taskswitch 5
+#define __HYPERVISOR_sched_op_compat 6 /* compat since 0x00030101 */
+#define __HYPERVISOR_platform_op 7
+#define __HYPERVISOR_set_debugreg 8
+#define __HYPERVISOR_get_debugreg 9
+#define __HYPERVISOR_update_descriptor 10
+#define __HYPERVISOR_memory_op 12
+#define __HYPERVISOR_multicall 13
+#define __HYPERVISOR_update_va_mapping 14
+#define __HYPERVISOR_set_timer_op 15
+#define __HYPERVISOR_event_channel_op_compat 16 /* compat since 0x00030202 */
+#define __HYPERVISOR_xen_version 17
+#define __HYPERVISOR_console_io 18
+#define __HYPERVISOR_physdev_op_compat 19 /* compat since 0x00030202 */
+#define __HYPERVISOR_grant_table_op 20
+#define __HYPERVISOR_vm_assist 21
+#define __HYPERVISOR_update_va_mapping_otherdomain 22
+#define __HYPERVISOR_iret 23 /* x86 only */
+#define __HYPERVISOR_vcpu_op 24
+#define __HYPERVISOR_set_segment_base 25 /* x86/64 only */
+#define __HYPERVISOR_mmuext_op 26
+#define __HYPERVISOR_xsm_op 27
+#define __HYPERVISOR_nmi_op 28
+#define __HYPERVISOR_sched_op 29
+#define __HYPERVISOR_callback_op 30
+#define __HYPERVISOR_xenoprof_op 31
+#define __HYPERVISOR_event_channel_op 32
+#define __HYPERVISOR_physdev_op 33
+#define __HYPERVISOR_hvm_op 34
+#define __HYPERVISOR_sysctl 35
+#define __HYPERVISOR_domctl 36
+#define __HYPERVISOR_kexec_op 37
+#define __HYPERVISOR_tmem_op 38
+#define __HYPERVISOR_xc_reserved_op 39 /* reserved for XenClient */
+
+/* Architecture-specific hypercall definitions. */
+#define __HYPERVISOR_arch_0 48
+#define __HYPERVISOR_arch_1 49
+#define __HYPERVISOR_arch_2 50
+#define __HYPERVISOR_arch_3 51
+#define __HYPERVISOR_arch_4 52
+#define __HYPERVISOR_arch_5 53
+#define __HYPERVISOR_arch_6 54
+#define __HYPERVISOR_arch_7 55
+
+/* ` } */
+
+/*
+ * HYPERCALL COMPATIBILITY.
+ */
+
+/* New sched_op hypercall introduced in 0x00030101. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030101
+#undef __HYPERVISOR_sched_op
+#define __HYPERVISOR_sched_op __HYPERVISOR_sched_op_compat
+#endif
+
+/* New event-channel and physdev hypercalls introduced in 0x00030202. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030202
+#undef __HYPERVISOR_event_channel_op
+#define __HYPERVISOR_event_channel_op __HYPERVISOR_event_channel_op_compat
+#undef __HYPERVISOR_physdev_op
+#define __HYPERVISOR_physdev_op __HYPERVISOR_physdev_op_compat
+#endif
+
+/* New platform_op hypercall introduced in 0x00030204. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030204
+#define __HYPERVISOR_dom0_op __HYPERVISOR_platform_op
+#endif
+
+/*
+ * VIRTUAL INTERRUPTS
+ *
+ * Virtual interrupts that a guest OS may receive from Xen.
+ *
+ * In the side comments, 'V.' denotes a per-VCPU VIRQ while 'G.' denotes a
+ * global VIRQ. The former can be bound once per VCPU and cannot be re-bound.
+ * The latter can be allocated only once per guest: they must initially be
+ * allocated to VCPU0 but can subsequently be re-bound.
+ */
+/* ` enum virq { */
+#define VIRQ_TIMER 0 /* V. Timebase update, and/or requested timeout. */
+#define VIRQ_DEBUG 1 /* V. Request guest to dump debug info. */
+#define VIRQ_CONSOLE 2 /* G. (DOM0) Bytes received on emergency console. */
+#define VIRQ_DOM_EXC 3 /* G. (DOM0) Exceptional event for some domain. */
+#define VIRQ_TBUF 4 /* G. (DOM0) Trace buffer has records available. */
+#define VIRQ_DEBUGGER 6 /* G. (DOM0) A domain has paused for debugging. */
+#define VIRQ_XENOPROF 7 /* V. XenOprofile interrupt: new sample available */
+#define VIRQ_CON_RING 8 /* G. (DOM0) Bytes received on console */
+#define VIRQ_PCPU_STATE 9 /* G. (DOM0) PCPU state changed */
+#define VIRQ_MEM_EVENT 10 /* G. (DOM0) A memory event has occured */
+#define VIRQ_XC_RESERVED 11 /* G. Reserved for XenClient */
+#define VIRQ_ENOMEM 12 /* G. (DOM0) Low on heap memory */
+
+/* Architecture-specific VIRQ definitions. */
+#define VIRQ_ARCH_0 16
+#define VIRQ_ARCH_1 17
+#define VIRQ_ARCH_2 18
+#define VIRQ_ARCH_3 19
+#define VIRQ_ARCH_4 20
+#define VIRQ_ARCH_5 21
+#define VIRQ_ARCH_6 22
+#define VIRQ_ARCH_7 23
+/* ` } */
+
+#define NR_VIRQS 24
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_mmu_update(const struct mmu_update reqs[],
+ * ` unsigned count, unsigned *done_out,
+ * ` unsigned foreigndom)
+ * `
+ * @reqs is an array of mmu_update_t structures ((ptr, val) pairs).
+ * @count is the length of the above array.
+ * @pdone is an output parameter indicating number of completed operations
+ * @foreigndom[15:0]: FD, the expected owner of data pages referenced in this
+ * hypercall invocation. Can be DOMID_SELF.
+ * @foreigndom[31:16]: PFD, the expected owner of pagetable pages referenced
+ * in this hypercall invocation. The value of this field
+ * (x) encodes the PFD as follows:
+ * x == 0 => PFD == DOMID_SELF
+ * x != 0 => PFD == x - 1
+ *
+ * Sub-commands: ptr[1:0] specifies the appropriate MMU_* command.
+ * -------------
+ * ptr[1:0] == MMU_NORMAL_PT_UPDATE:
+ * Updates an entry in a page table belonging to PFD. If updating an L1 table,
+ * and the new table entry is valid/present, the mapped frame must belong to
+ * FD. If attempting to map an I/O page then the caller assumes the privilege
+ * of the FD.
+ * FD == DOMID_IO: Permit /only/ I/O mappings, at the priv level of the caller.
+ * FD == DOMID_XEN: Map restricted areas of Xen's heap space.
+ * ptr[:2] -- Machine address of the page-table entry to modify.
+ * val -- Value to write.
+ *
+ * There also certain implicit requirements when using this hypercall. The
+ * pages that make up a pagetable must be mapped read-only in the guest.
+ * This prevents uncontrolled guest updates to the pagetable. Xen strictly
+ * enforces this, and will disallow any pagetable update which will end up
+ * mapping pagetable page RW, and will disallow using any writable page as a
+ * pagetable. In practice it means that when constructing a page table for a
+ * process, thread, etc, we MUST be very dilligient in following these rules:
+ * 1). Start with top-level page (PGD or in Xen language: L4). Fill out
+ * the entries.
+ * 2). Keep on going, filling out the upper (PUD or L3), and middle (PMD
+ * or L2).
+ * 3). Start filling out the PTE table (L1) with the PTE entries. Once
+ * done, make sure to set each of those entries to RO (so writeable bit
+ * is unset). Once that has been completed, set the PMD (L2) for this
+ * PTE table as RO.
+ * 4). When completed with all of the PMD (L2) entries, and all of them have
+ * been set to RO, make sure to set RO the PUD (L3). Do the same
+ * operation on PGD (L4) pagetable entries that have a PUD (L3) entry.
+ * 5). Now before you can use those pages (so setting the cr3), you MUST also
+ * pin them so that the hypervisor can verify the entries. This is done
+ * via the HYPERVISOR_mmuext_op(MMUEXT_PIN_L4_TABLE, guest physical frame
+ * number of the PGD (L4)). And this point the HYPERVISOR_mmuext_op(
+ * MMUEXT_NEW_BASEPTR, guest physical frame number of the PGD (L4)) can be
+ * issued.
+ * For 32-bit guests, the L4 is not used (as there is less pagetables), so
+ * instead use L3.
+ * At this point the pagetables can be modified using the MMU_NORMAL_PT_UPDATE
+ * hypercall. Also if so desired the OS can also try to write to the PTE
+ * and be trapped by the hypervisor (as the PTE entry is RO).
+ *
+ * To deallocate the pages, the operations are the reverse of the steps
+ * mentioned above. The argument is MMUEXT_UNPIN_TABLE for all levels and the
+ * pagetable MUST not be in use (meaning that the cr3 is not set to it).
+ *
+ * ptr[1:0] == MMU_MACHPHYS_UPDATE:
+ * Updates an entry in the machine->pseudo-physical mapping table.
+ * ptr[:2] -- Machine address within the frame whose mapping to modify.
+ * The frame must belong to the FD, if one is specified.
+ * val -- Value to write into the mapping entry.
+ *
+ * ptr[1:0] == MMU_PT_UPDATE_PRESERVE_AD:
+ * As MMU_NORMAL_PT_UPDATE above, but A/D bits currently in the PTE are ORed
+ * with those in @val.
+ *
+ * @val is usually the machine frame number along with some attributes.
+ * The attributes by default follow the architecture defined bits. Meaning that
+ * if this is a X86_64 machine and four page table layout is used, the layout
+ * of val is:
+ * - 63 if set means No execute (NX)
+ * - 46-13 the machine frame number
+ * - 12 available for guest
+ * - 11 available for guest
+ * - 10 available for guest
+ * - 9 available for guest
+ * - 8 global
+ * - 7 PAT (PSE is disabled, must use hypercall to make 4MB or 2MB pages)
+ * - 6 dirty
+ * - 5 accessed
+ * - 4 page cached disabled
+ * - 3 page write through
+ * - 2 userspace accessible
+ * - 1 writeable
+ * - 0 present
+ *
+ * The one bits that does not fit with the default layout is the PAGE_PSE
+ * also called PAGE_PAT). The MMUEXT_[UN]MARK_SUPER arguments to the
+ * HYPERVISOR_mmuext_op serve as mechanism to set a pagetable to be 4MB
+ * (or 2MB) instead of using the PAGE_PSE bit.
+ *
+ * The reason that the PAGE_PSE (bit 7) is not being utilized is due to Xen
+ * using it as the Page Attribute Table (PAT) bit - for details on it please
+ * refer to Intel SDM 10.12. The PAT allows to set the caching attributes of
+ * pages instead of using MTRRs.
+ *
+ * The PAT MSR is as follows (it is a 64-bit value, each entry is 8 bits):
+ * PAT4 PAT0
+ * +-----+-----+----+----+----+-----+----+----+
+ * | UC | UC- | WC | WB | UC | UC- | WC | WB | <= Linux
+ * +-----+-----+----+----+----+-----+----+----+
+ * | UC | UC- | WT | WB | UC | UC- | WT | WB | <= BIOS (default when machine boots)
+ * +-----+-----+----+----+----+-----+----+----+
+ * | rsv | rsv | WP | WC | UC | UC- | WT | WB | <= Xen
+ * +-----+-----+----+----+----+-----+----+----+
+ *
+ * The lookup of this index table translates to looking up
+ * Bit 7, Bit 4, and Bit 3 of val entry:
+ *
+ * PAT/PSE (bit 7) ... PCD (bit 4) .. PWT (bit 3).
+ *
+ * If all bits are off, then we are using PAT0. If bit 3 turned on,
+ * then we are using PAT1, if bit 3 and bit 4, then PAT2..
+ *
+ * As you can see, the Linux PAT1 translates to PAT4 under Xen. Which means
+ * that if a guest that follows Linux's PAT setup and would like to set Write
+ * Combined on pages it MUST use PAT4 entry. Meaning that Bit 7 (PAGE_PAT) is
+ * set. For example, under Linux it only uses PAT0, PAT1, and PAT2 for the
+ * caching as:
+ *
+ * WB = none (so PAT0)
+ * WC = PWT (bit 3 on)
+ * UC = PWT | PCD (bit 3 and 4 are on).
+ *
+ * To make it work with Xen, it needs to translate the WC bit as so:
+ *
+ * PWT (so bit 3 on) --> PAT (so bit 7 is on) and clear bit 3
+ *
+ * And to translate back it would:
+ *
+ * PAT (bit 7 on) --> PWT (bit 3 on) and clear bit 7.
+ */
+#define MMU_NORMAL_PT_UPDATE 0 /* checked '*ptr = val'. ptr is MA. */
+#define MMU_MACHPHYS_UPDATE 1 /* ptr = MA of frame to modify entry for */
+#define MMU_PT_UPDATE_PRESERVE_AD 2 /* atomically: *ptr = val | (*ptr&(A|D)) */
+
+/*
+ * MMU EXTENDED OPERATIONS
+ *
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_mmuext_op(mmuext_op_t uops[],
+ * ` UINT32 count,
+ * ` UINT32 *pdone,
+ * ` UINT32 foreigndom)
+ */
+/* HYPERVISOR_mmuext_op() accepts a list of mmuext_op structures.
+ * A foreigndom (FD) can be specified (or DOMID_SELF for none).
+ * Where the FD has some effect, it is described below.
+ *
+ * cmd: MMUEXT_(UN)PIN_*_TABLE
+ * mfn: Machine frame number to be (un)pinned as a p.t. page.
+ * The frame must belong to the FD, if one is specified.
+ *
+ * cmd: MMUEXT_NEW_BASEPTR
+ * mfn: Machine frame number of new page-table base to install in MMU.
+ *
+ * cmd: MMUEXT_NEW_USER_BASEPTR [x86/64 only]
+ * mfn: Machine frame number of new page-table base to install in MMU
+ * when in user space.
+ *
+ * cmd: MMUEXT_TLB_FLUSH_LOCAL
+ * No additional arguments. Flushes local TLB.
+ *
+ * cmd: MMUEXT_INVLPG_LOCAL
+ * linear_addr: Linear address to be flushed from the local TLB.
+ *
+ * cmd: MMUEXT_TLB_FLUSH_MULTI
+ * vcpumask: Pointer to bitmap of VCPUs to be flushed.
+ *
+ * cmd: MMUEXT_INVLPG_MULTI
+ * linear_addr: Linear address to be flushed.
+ * vcpumask: Pointer to bitmap of VCPUs to be flushed.
+ *
+ * cmd: MMUEXT_TLB_FLUSH_ALL
+ * No additional arguments. Flushes all VCPUs' TLBs.
+ *
+ * cmd: MMUEXT_INVLPG_ALL
+ * linear_addr: Linear address to be flushed from all VCPUs' TLBs.
+ *
+ * cmd: MMUEXT_FLUSH_CACHE
+ * No additional arguments. Writes back and flushes cache contents.
+ *
+ * cmd: MMUEXT_FLUSH_CACHE_GLOBAL
+ * No additional arguments. Writes back and flushes cache contents
+ * on all CPUs in the system.
+ *
+ * cmd: MMUEXT_SET_LDT
+ * linear_addr: Linear address of LDT base (NB. must be page-aligned).
+ * nr_ents: Number of entries in LDT.
+ *
+ * cmd: MMUEXT_CLEAR_PAGE
+ * mfn: Machine frame number to be cleared.
+ *
+ * cmd: MMUEXT_COPY_PAGE
+ * mfn: Machine frame number of the destination page.
+ * src_mfn: Machine frame number of the source page.
+ *
+ * cmd: MMUEXT_[UN]MARK_SUPER
+ * mfn: Machine frame number of head of superpage to be [un]marked.
+ */
+/* ` enum mmuext_cmd { */
+#define MMUEXT_PIN_L1_TABLE 0
+#define MMUEXT_PIN_L2_TABLE 1
+#define MMUEXT_PIN_L3_TABLE 2
+#define MMUEXT_PIN_L4_TABLE 3
+#define MMUEXT_UNPIN_TABLE 4
+#define MMUEXT_NEW_BASEPTR 5
+#define MMUEXT_TLB_FLUSH_LOCAL 6
+#define MMUEXT_INVLPG_LOCAL 7
+#define MMUEXT_TLB_FLUSH_MULTI 8
+#define MMUEXT_INVLPG_MULTI 9
+#define MMUEXT_TLB_FLUSH_ALL 10
+#define MMUEXT_INVLPG_ALL 11
+#define MMUEXT_FLUSH_CACHE 12
+#define MMUEXT_SET_LDT 13
+#define MMUEXT_NEW_USER_BASEPTR 15
+#define MMUEXT_CLEAR_PAGE 16
+#define MMUEXT_COPY_PAGE 17
+#define MMUEXT_FLUSH_CACHE_GLOBAL 18
+#define MMUEXT_MARK_SUPER 19
+#define MMUEXT_UNMARK_SUPER 20
+/* ` } */
+
+#ifndef __ASSEMBLY__
+struct mmuext_op {
+ UINT32 cmd; /* => enum mmuext_cmd */
+ union {
+ /* [UN]PIN_TABLE, NEW_BASEPTR, NEW_USER_BASEPTR
+ * CLEAR_PAGE, COPY_PAGE, [UN]MARK_SUPER */
+ xen_pfn_t mfn;
+ /* INVLPG_LOCAL, INVLPG_ALL, SET_LDT */
+ UINTN linear_addr;
+ } arg1;
+ union {
+ /* SET_LDT */
+ UINT32 nr_ents;
+ /* TLB_FLUSH_MULTI, INVLPG_MULTI */
+#if __XEN_INTERFACE_VERSION__ >= 0x00030205
+ XEN_GUEST_HANDLE(const_void) vcpumask;
+#else
+ const VOID *vcpumask;
+#endif
+ /* COPY_PAGE */
+ xen_pfn_t src_mfn;
+ } arg2;
+};
+typedef struct mmuext_op mmuext_op_t;
+DEFINE_XEN_GUEST_HANDLE(mmuext_op_t);
+#endif
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_update_va_mapping(UINTN va, u64 val,
+ * ` enum uvm_flags flags)
+ * `
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_update_va_mapping_otherdomain(UINTN va, u64 val,
+ * ` enum uvm_flags flags,
+ * ` domid_t domid)
+ * `
+ * ` @va: The virtual address whose mapping we want to change
+ * ` @val: The new page table entry, must contain a machine address
+ * ` @flags: Control TLB flushes
+ */
+/* These are passed as 'flags' to update_va_mapping. They can be ORed. */
+/* When specifying UVMF_MULTI, also OR in a pointer to a CPU bitmap. */
+/* UVMF_LOCAL is merely UVMF_MULTI with a NULL bitmap pointer. */
+/* ` enum uvm_flags { */
+#define UVMF_NONE (0UL<<0) /* No flushing at all. */
+#define UVMF_TLB_FLUSH (1UL<<0) /* Flush entire TLB(s). */
+#define UVMF_INVLPG (2UL<<0) /* Flush only one entry. */
+#define UVMF_FLUSHTYPE_MASK (3UL<<0)
+#define UVMF_MULTI (0UL<<2) /* Flush subset of TLBs. */
+#define UVMF_LOCAL (0UL<<2) /* Flush local TLB. */
+#define UVMF_ALL (1UL<<2) /* Flush all TLBs. */
+/* ` } */
+
+/*
+ * Commands to HYPERVISOR_console_io().
+ */
+#define CONSOLEIO_write 0
+#define CONSOLEIO_read 1
+
+/*
+ * Commands to HYPERVISOR_vm_assist().
+ */
+#define VMASST_CMD_enable 0
+#define VMASST_CMD_disable 1
+
+/* x86/32 guests: simulate full 4GB segment limits. */
+#define VMASST_TYPE_4gb_segments 0
+
+/* x86/32 guests: trap (vector 15) whenever above vmassist is used. */
+#define VMASST_TYPE_4gb_segments_notify 1
+
+/*
+ * x86 guests: support writes to bottom-level PTEs.
+ * NB1. Page-directory entries cannot be written.
+ * NB2. Guest must continue to remove all writable mappings of PTEs.
+ */
+#define VMASST_TYPE_writable_pagetables 2
+
+/* x86/PAE guests: support PDPTs above 4GB. */
+#define VMASST_TYPE_pae_extended_cr3 3
+
+#define MAX_VMASST_TYPE 3
+
+#ifndef __ASSEMBLY__
+
+typedef UINT16 domid_t;
+
+/* Domain ids >= DOMID_FIRST_RESERVED cannot be used for ordinary domains. */
+#define DOMID_FIRST_RESERVED (0x7FF0U)
+
+/* DOMID_SELF is used in certain contexts to refer to oneself. */
+#define DOMID_SELF (0x7FF0U)
+
+/*
+ * DOMID_IO is used to restrict page-table updates to mapping I/O memory.
+ * Although no Foreign Domain need be specified to map I/O pages, DOMID_IO
+ * is useful to ensure that no mappings to the OS's own heap are accidentally
+ * installed. (e.g., in Linux this could cause havoc as reference counts
+ * aren't adjusted on the I/O-mapping code path).
+ * This only makes sense in MMUEXT_SET_FOREIGNDOM, but in that context can
+ * be specified by any calling domain.
+ */
+#define DOMID_IO (0x7FF1U)
+
+/*
+ * DOMID_XEN is used to allow privileged domains to map restricted parts of
+ * Xen's heap space (e.g., the machine_to_phys table).
+ * This only makes sense in MMUEXT_SET_FOREIGNDOM, and is only permitted if
+ * the caller is privileged.
+ */
+#define DOMID_XEN (0x7FF2U)
+
+/*
+ * DOMID_COW is used as the owner of sharable pages */
+#define DOMID_COW (0x7FF3U)
+
+/* DOMID_INVALID is used to identify pages with unknown owner. */
+#define DOMID_INVALID (0x7FF4U)
+
+/* Idle domain. */
+#define DOMID_IDLE (0x7FFFU)
+
+/*
+ * Send an array of these to HYPERVISOR_mmu_update().
+ * NB. The fields are natural pointer/address size for this architecture.
+ */
+struct mmu_update {
+ UINT64 ptr; /* Machine address of PTE. */
+ UINT64 val; /* New contents of PTE. */
+};
+typedef struct mmu_update mmu_update_t;
+DEFINE_XEN_GUEST_HANDLE(mmu_update_t);
+
+/*
+ * ` enum neg_errnoval
+ * ` HYPERVISOR_multicall(multicall_entry_t call_list[],
+ * ` UINT32 nr_calls);
+ *
+ * NB. The fields are natural register size for this architecture.
+ */
+struct multicall_entry {
+ UINTN op, result;
+ UINTN args[6];
+};
+typedef struct multicall_entry multicall_entry_t;
+DEFINE_XEN_GUEST_HANDLE(multicall_entry_t);
+
+#if __XEN_INTERFACE_VERSION__ < 0x00040400
+/*
+ * Event channel endpoints per domain (when using the 2-level ABI):
+ * 1024 if a INTN is 32 bits; 4096 if a INTN is 64 bits.
+ */
+#define NR_EVENT_CHANNELS EVTCHN_2L_NR_CHANNELS
+#endif
+
+struct vcpu_time_info {
+ /*
+ * Updates to the following values are preceded and followed by an
+ * increment of 'version'. The guest can therefore detect updates by
+ * looking for changes to 'version'. If the least-significant bit of
+ * the version number is set then an update is in progress and the guest
+ * must wait to read a consistent set of values.
+ * The correct way to interact with the version number is similar to
+ * Linux's seqlock: see the implementations of read_seqbegin/read_seqretry.
+ */
+ UINT32 version;
+ UINT32 pad0;
+ UINT64 tsc_timestamp; /* TSC at last update of time vals. */
+ UINT64 system_time; /* Time, in nanosecs, since boot. */
+ /*
+ * Current system time:
+ * system_time +
+ * ((((tsc - tsc_timestamp) << tsc_shift) * tsc_to_system_mul) >> 32)
+ * CPU frequency (Hz):
+ * ((10^9 << 32) / tsc_to_system_mul) >> tsc_shift
+ */
+ UINT32 tsc_to_system_mul;
+ INT8 tsc_shift;
+ INT8 pad1[3];
+}; /* 32 bytes */
+typedef struct vcpu_time_info vcpu_time_info_t;
+
+struct vcpu_info {
+ /*
+ * 'evtchn_upcall_pending' is written non-zero by Xen to indicate
+ * a pending notification for a particular VCPU. It is then cleared
+ * by the guest OS /before/ checking for pending work, thus avoiding
+ * a set-and-check race. Note that the mask is only accessed by Xen
+ * on the CPU that is currently hosting the VCPU. This means that the
+ * pending and mask flags can be updated by the guest without special
+ * synchronisation (i.e., no need for the x86 LOCK prefix).
+ * This may seem suboptimal because if the pending flag is set by
+ * a different CPU then an IPI may be scheduled even when the mask
+ * is set. However, note:
+ * 1. The task of 'interrupt holdoff' is covered by the per-event-
+ * channel mask bits. A 'noisy' event that is continually being
+ * triggered can be masked at source at this very precise
+ * granularity.
+ * 2. The main purpose of the per-VCPU mask is therefore to restrict
+ * reentrant execution: whether for concurrency control, or to
+ * prevent unbounded stack usage. Whatever the purpose, we expect
+ * that the mask will be asserted only for short periods at a time,
+ * and so the likelihood of a 'spurious' IPI is suitably small.
+ * The mask is read before making an event upcall to the guest: a
+ * non-zero mask therefore guarantees that the VCPU will not receive
+ * an upcall activation. The mask is cleared when the VCPU requests
+ * to block: this avoids wakeup-waiting races.
+ */
+ UINT8 evtchn_upcall_pending;
+#ifdef XEN_HAVE_PV_UPCALL_MASK
+ UINT8 evtchn_upcall_mask;
+#else /* XEN_HAVE_PV_UPCALL_MASK */
+ UINT8 pad0;
+#endif /* XEN_HAVE_PV_UPCALL_MASK */
+ xen_ulong_t evtchn_pending_sel;
+ struct arch_vcpu_info arch;
+ struct vcpu_time_info time;
+}; /* 64 bytes (x86) */
+#ifndef __XEN__
+typedef struct vcpu_info vcpu_info_t;
+#endif
+
+/*
+ * `incontents 200 startofday_shared Start-of-day shared data structure
+ * Xen/kernel shared data -- pointer provided in start_info.
+ *
+ * This structure is defined to be both smaller than a page, and the
+ * only data on the shared page, but may vary in actual size even within
+ * compatible Xen versions; guests should not rely on the size
+ * of this structure remaining constant.
+ */
+struct shared_info {
+ struct vcpu_info vcpu_info[XEN_LEGACY_MAX_VCPUS];
+
+ /*
+ * A domain can create "event channels" on which it can send and receive
+ * asynchronous event notifications. There are three classes of event that
+ * are delivered by this mechanism:
+ * 1. Bi-directional inter- and intra-domain connections. Domains must
+ * arrange out-of-band to set up a connection (usually by allocating
+ * an unbound 'listener' port and avertising that via a storage service
+ * such as xenstore).
+ * 2. Physical interrupts. A domain with suitable hardware-access
+ * privileges can bind an event-channel port to a physical interrupt
+ * source.
+ * 3. Virtual interrupts ('events'). A domain can bind an event-channel
+ * port to a virtual interrupt source, such as the virtual-timer
+ * device or the emergency console.
+ *
+ * Event channels are addressed by a "port index". Each channel is
+ * associated with two bits of information:
+ * 1. PENDING -- notifies the domain that there is a pending notification
+ * to be processed. This bit is cleared by the guest.
+ * 2. MASK -- if this bit is clear then a 0->1 transition of PENDING
+ * will cause an asynchronous upcall to be scheduled. This bit is only
+ * updated by the guest. It is read-only within Xen. If a channel
+ * becomes pending while the channel is masked then the 'edge' is lost
+ * (i.e., when the channel is unmasked, the guest must manually handle
+ * pending notifications as no upcall will be scheduled by Xen).
+ *
+ * To expedite scanning of pending notifications, any 0->1 pending
+ * transition on an unmasked channel causes a corresponding bit in a
+ * per-vcpu selector word to be set. Each bit in the selector covers a
+ * 'C INTN' in the PENDING bitfield array.
+ */
+ xen_ulong_t evtchn_pending[sizeof(xen_ulong_t) * 8];
+ xen_ulong_t evtchn_mask[sizeof(xen_ulong_t) * 8];
+
+ /*
+ * Wallclock time: updated only by control software. Guests should base
+ * their gettimeofday() syscall on this wallclock-base value.
+ */
+ UINT32 wc_version; /* Version counter: see vcpu_time_info_t. */
+ UINT32 wc_sec; /* Secs 00:00:00 UTC, Jan 1, 1970. */
+ UINT32 wc_nsec; /* Nsecs 00:00:00 UTC, Jan 1, 1970. */
+
+ struct arch_shared_info arch;
+
+};
+#ifndef __XEN__
+typedef struct shared_info shared_info_t;
+#endif
+
+/*
+ * `incontents 200 startofday Start-of-day memory layout
+ *
+ * 1. The domain is started within contiguous virtual-memory region.
+ * 2. The contiguous region ends on an aligned 4MB boundary.
+ * 3. This the order of bootstrap elements in the initial virtual region:
+ * a. relocated kernel image
+ * b. initial ram disk [mod_start, mod_len]
+ * c. list of allocated page frames [mfn_list, nr_pages]
+ * (unless relocated due to XEN_ELFNOTE_INIT_P2M)
+ * d. start_info_t structure [register ESI (x86)]
+ * e. bootstrap page tables [pt_base and CR3 (x86)]
+ * f. bootstrap stack [register ESP (x86)]
+ * 4. Bootstrap elements are packed together, but each is 4kB-aligned.
+ * 5. The initial ram disk may be omitted.
+ * 6. The list of page frames forms a contiguous 'pseudo-physical' memory
+ * layout for the domain. In particular, the bootstrap virtual-memory
+ * region is a 1:1 mapping to the first section of the pseudo-physical map.
+ * 7. All bootstrap elements are mapped read-writable for the guest OS. The
+ * only exception is the bootstrap page table, which is mapped read-only.
+ * 8. There is guaranteed to be at least 512kB padding after the final
+ * bootstrap element. If necessary, the bootstrap virtual region is
+ * extended by an extra 4MB to ensure this.
+ *
+ * Note: Prior to 25833:bb85bbccb1c9. ("x86/32-on-64 adjust Dom0 initial page
+ * table layout") a bug caused the pt_base (3.e above) and cr3 to not point
+ * to the start of the guest page tables (it was offset by two pages).
+ * This only manifested itself on 32-on-64 dom0 kernels and not 32-on-64 domU
+ * or 64-bit kernels of any colour. The page tables for a 32-on-64 dom0 got
+ * allocated in the order: 'first L1','first L2', 'first L3', so the offset
+ * to the page table base is by two pages back. The initial domain if it is
+ * 32-bit and runs under a 64-bit hypervisor should _NOT_ use two of the
+ * pages preceding pt_base and mark them as reserved/unused.
+ */
+#ifdef XEN_HAVE_PV_GUEST_ENTRY
+struct start_info {
+ /* THE FOLLOWING ARE FILLED IN BOTH ON INITIAL BOOT AND ON RESUME. */
+ CHAR8 magic[32]; /* "xen-<version>-<platform>". */
+ UINTN nr_pages; /* Total pages allocated to this domain. */
+ UINTN shared_info; /* MACHINE address of shared info struct. */
+ UINT32 flags; /* SIF_xxx flags. */
+ xen_pfn_t store_mfn; /* MACHINE page number of shared page. */
+ UINT32 store_evtchn; /* Event channel for store communication. */
+ union {
+ struct {
+ xen_pfn_t mfn; /* MACHINE page number of console page. */
+ UINT32 evtchn; /* Event channel for console page. */
+ } domU;
+ struct {
+ UINT32 info_off; /* Offset of console_info struct. */
+ UINT32 info_size; /* Size of console_info struct from start.*/
+ } dom0;
+ } console;
+ /* THE FOLLOWING ARE ONLY FILLED IN ON INITIAL BOOT (NOT RESUME). */
+ UINTN pt_base; /* VIRTUAL address of page directory. */
+ UINTN nr_pt_frames; /* Number of bootstrap p.t. frames. */
+ UINTN mfn_list; /* VIRTUAL address of page-frame list. */
+ UINTN mod_start; /* VIRTUAL address of pre-loaded module */
+ /* (PFN of pre-loaded module if */
+ /* SIF_MOD_START_PFN set in flags). */
+ UINTN mod_len; /* Size (bytes) of pre-loaded module. */
+#define MAX_GUEST_CMDLINE 1024
+ INT8 cmd_line[MAX_GUEST_CMDLINE];
+ /* The pfn range here covers both page table and p->m table frames. */
+ UINTN first_p2m_pfn;/* 1st pfn forming initial P->M table. */
+ UINTN nr_p2m_frames;/* # of pfns forming initial P->M table. */
+};
+typedef struct start_info start_info_t;
+
+/* New console union for dom0 introduced in 0x00030203. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030203
+#define console_mfn console.domU.mfn
+#define console_evtchn console.domU.evtchn
+#endif
+#endif /* XEN_HAVE_PV_GUEST_ENTRY */
+
+/* These flags are passed in the 'flags' field of start_info_t. */
+#define SIF_PRIVILEGED (1<<0) /* Is the domain privileged? */
+#define SIF_INITDOMAIN (1<<1) /* Is this the initial control domain? */
+#define SIF_MULTIBOOT_MOD (1<<2) /* Is mod_start a multiboot module? */
+#define SIF_MOD_START_PFN (1<<3) /* Is mod_start a PFN? */
+#define SIF_PM_MASK (0xFF<<8) /* reserve 1 byte for xen-pm options */
+
+/*
+ * A multiboot module is a package containing modules very similar to a
+ * multiboot module array. The only differences are:
+ * - the array of module descriptors is by convention simply at the beginning
+ * of the multiboot module,
+ * - addresses in the module descriptors are based on the beginning of the
+ * multiboot module,
+ * - the number of modules is determined by a termination descriptor that has
+ * mod_start == 0.
+ *
+ * This permits to both build it statically and reference it in a configuration
+ * file, and let the PV guest easily rebase the addresses to virtual addresses
+ * and at the same time count the number of modules.
+ */
+struct xen_multiboot_mod_list
+{
+ /* Address of first byte of the module */
+ UINT32 mod_start;
+ /* Address of last byte of the module (inclusive) */
+ UINT32 mod_end;
+ /* Address of zero-terminated command line */
+ UINT32 cmdline;
+ /* Unused, must be zero */
+ UINT32 pad;
+};
+/*
+ * `incontents 200 startofday_dom0_console Dom0_console
+ *
+ * The console structure in start_info.console.dom0
+ *
+ * This structure includes a variety of information required to
+ * have a working VGA/VESA console.
+ */
+typedef struct dom0_vga_console_info {
+ UINT8 video_type; /* DOM0_VGA_CONSOLE_??? */
+#define XEN_VGATYPE_TEXT_MODE_3 0x03
+#define XEN_VGATYPE_VESA_LFB 0x23
+#define XEN_VGATYPE_EFI_LFB 0x70
+
+ union {
+ struct {
+ /* Font height, in pixels. */
+ UINT16 font_height;
+ /* Cursor location (column, row). */
+ UINT16 cursor_x, cursor_y;
+ /* Number of rows and columns (dimensions in characters). */
+ UINT16 rows, columns;
+ } text_mode_3;
+
+ struct {
+ /* Width and height, in pixels. */
+ UINT16 width, height;
+ /* Bytes per scan line. */
+ UINT16 bytes_per_line;
+ /* Bits per pixel. */
+ UINT16 bits_per_pixel;
+ /* LFB physical address, and size (in units of 64kB). */
+ UINT32 lfb_base;
+ UINT32 lfb_size;
+ /* RGB mask offsets and sizes, as defined by VBE 1.2+ */
+ UINT8 red_pos, red_size;
+ UINT8 green_pos, green_size;
+ UINT8 blue_pos, blue_size;
+ UINT8 rsvd_pos, rsvd_size;
+#if __XEN_INTERFACE_VERSION__ >= 0x00030206
+ /* VESA capabilities (offset 0xa, VESA command 0x4f00). */
+ UINT32 gbl_caps;
+ /* Mode attributes (offset 0x0, VESA command 0x4f01). */
+ UINT16 mode_attrs;
+#endif
+ } vesa_lfb;
+ } u;
+} dom0_vga_console_info_t;
+#define xen_vga_console_info dom0_vga_console_info
+#define xen_vga_console_info_t dom0_vga_console_info_t
+
+typedef UINT8 xen_domain_handle_t[16];
+
+/* Turn a plain number into a C UINTN constant. */
+#define __mk_unsigned_long(x) x ## UL
+#define mk_unsigned_long(x) __mk_unsigned_long(x)
+
+__DEFINE_XEN_GUEST_HANDLE(uint8, UINT8);
+__DEFINE_XEN_GUEST_HANDLE(uint16, UINT16);
+__DEFINE_XEN_GUEST_HANDLE(uint32, UINT32);
+__DEFINE_XEN_GUEST_HANDLE(uint64, UINT64);
+
+#else /* __ASSEMBLY__ */
+
+/* In assembly code we cannot use C numeric constant suffixes. */
+#define mk_unsigned_long(x) x
+
+#endif /* !__ASSEMBLY__ */
+
+/* Default definitions for macros used by domctl/sysctl. */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+
+#ifndef uint64_aligned_t
+#define uint64_aligned_t UINT64
+#endif
+#ifndef XEN_GUEST_HANDLE_64
+#define XEN_GUEST_HANDLE_64(name) XEN_GUEST_HANDLE(name)
+#endif
+
+#ifndef __ASSEMBLY__
+struct xenctl_bitmap {
+ XEN_GUEST_HANDLE_64(uint8) bitmap;
+ UINT32 nr_bits;
+};
+#endif
+
+#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
+
+#endif /* __XEN_PUBLIC_XEN_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
Anthony PERARD
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
2014-07-16 15:15 ` [PATCH RFC 01/18] OvmfPkg: Add public headers from Xen Project Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event Anthony PERARD
` (25 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This includes Component Name and Driver Binding.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/OvmfPkgX64.dsc | 1 +
OvmfPkg/OvmfPkgX64.fdf | 1 +
OvmfPkg/XenbusDxe/ComponentName.c | 148 ++++++++++++++++++
OvmfPkg/XenbusDxe/ComponentName.h | 91 +++++++++++
OvmfPkg/XenbusDxe/DriverBinding.h | 139 +++++++++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.c | 314 ++++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.h | 78 ++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 56 +++++++
8 files changed, 828 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/ComponentName.c
create mode 100644 OvmfPkg/XenbusDxe/ComponentName.h
create mode 100644 OvmfPkg/XenbusDxe/DriverBinding.h
create mode 100644 OvmfPkg/XenbusDxe/XenbusDxe.c
create mode 100644 OvmfPkg/XenbusDxe/XenbusDxe.h
create mode 100644 OvmfPkg/XenbusDxe/XenbusDxe.inf
diff --git a/OvmfPkg/OvmfPkgX64.dsc b/OvmfPkg/OvmfPkgX64.dsc
index 66459c2..4ea1a83 100644
--- a/OvmfPkg/OvmfPkgX64.dsc
+++ b/OvmfPkg/OvmfPkgX64.dsc
@@ -430,6 +430,7 @@
OvmfPkg/VirtioBlkDxe/VirtioBlk.inf
OvmfPkg/VirtioScsiDxe/VirtioScsi.inf
OvmfPkg/QemuFlashFvbServicesRuntimeDxe/FvbServicesRuntimeDxe.inf
+ OvmfPkg/XenbusDxe/XenbusDxe.inf
OvmfPkg/EmuVariableFvbRuntimeDxe/Fvb.inf {
<LibraryClasses>
PlatformFvbLib|OvmfPkg/Library/EmuVariableFvbLib/EmuVariableFvbLib.inf
diff --git a/OvmfPkg/OvmfPkgX64.fdf b/OvmfPkg/OvmfPkgX64.fdf
index b9e9c59..7f1aab4 100644
--- a/OvmfPkg/OvmfPkgX64.fdf
+++ b/OvmfPkg/OvmfPkgX64.fdf
@@ -278,6 +278,7 @@ INF OvmfPkg/VirtioScsiDxe/VirtioScsi.inf
INF OvmfPkg/QemuFlashFvbServicesRuntimeDxe/FvbServicesRuntimeDxe.inf
INF OvmfPkg/EmuVariableFvbRuntimeDxe/Fvb.inf
INF MdeModulePkg/Universal/FaultTolerantWriteDxe/FaultTolerantWriteDxe.inf
+INF OvmfPkg/XenbusDxe/XenbusDxe.inf
!if $(SECURE_BOOT_ENABLE) == TRUE
INF SecurityPkg/VariableAuthenticated/RuntimeDxe/VariableRuntimeDxe.inf
diff --git a/OvmfPkg/XenbusDxe/ComponentName.c b/OvmfPkg/XenbusDxe/ComponentName.c
new file mode 100644
index 0000000..2cf11b5
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/ComponentName.c
@@ -0,0 +1,148 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenbusDxe
+
+ TODO: Detailed Description of UEFI Driver XenbusDxe
+
+ TODO: Copyright for UEFI Driver XenbusDxe
+
+ TODO: License for UEFI Driver XenbusDxe
+
+**/
+
+#include "XenbusDxe.h"
+
+///
+/// Component Name Protocol instance
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName = {
+ (EFI_COMPONENT_NAME_GET_DRIVER_NAME) XenbusDxeComponentNameGetDriverName,
+ (EFI_COMPONENT_NAME_GET_CONTROLLER_NAME)XenbusDxeComponentNameGetControllerName,
+ "eng"
+};
+
+///
+/// Component Name 2 Protocol instance
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_COMPONENT_NAME2_PROTOCOL gXenbusDxeComponentName2 = {
+ XenbusDxeComponentNameGetDriverName,
+ XenbusDxeComponentNameGetControllerName,
+ "en"
+};
+
+///
+/// Table of driver names
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_UNICODE_STRING_TABLE mXenbusDxeDriverNameTable[] = {
+ { "eng;en", (CHAR16 *)L"Xenbus Driver" },
+ { NULL, NULL }
+};
+
+/**
+ Retrieves a Unicode string that is the user-readable name of the EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param Language A pointer to a three-character ISO 639-2 language identifier.
+ This is the language of the driver name that that the caller
+ is requesting, and it must match one of the languages specified
+ in SupportedLanguages. The number of languages supported by a
+ driver is up to the driver writer.
+ @param DriverName A pointer to the Unicode string to return. This Unicode string
+ is the name of the driver specified by This in the language
+ specified by Language.
+
+ @retval EFI_SUCCESS The Unicode string for the Driver specified by This
+ and the language specified by Language was returned
+ in DriverName.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER DriverName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeComponentNameGetDriverName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN CHAR8 *Language,
+ OUT CHAR16 **DriverName
+ )
+{
+ return LookupUnicodeString2 (
+ Language,
+ This->SupportedLanguages,
+ mXenbusDxeDriverNameTable,
+ DriverName,
+ (BOOLEAN)(This != &gXenbusDxeComponentName2)
+ );
+}
+
+/**
+ Retrieves a Unicode string that is the user readable name of the controller
+ that is being managed by an EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param ControllerHandle The handle of a controller that the driver specified by
+ This is managing. This handle specifies the controller
+ whose name is to be returned.
+ @param ChildHandle The handle of the child controller to retrieve the name
+ of. This is an optional parameter that may be NULL. It
+ will be NULL for device drivers. It will also be NULL
+ for a bus drivers that wish to retrieve the name of the
+ bus controller. It will not be NULL for a bus driver
+ that wishes to retrieve the name of a child controller.
+ @param Language A pointer to a three character ISO 639-2 language
+ identifier. This is the language of the controller name
+ that the caller is requesting, and it must match one
+ of the languages specified in SupportedLanguages. The
+ number of languages supported by a driver is up to the
+ driver writer.
+ @param ControllerName A pointer to the Unicode string to return. This Unicode
+ string is the name of the controller specified by
+ ControllerHandle and ChildHandle in the language specified
+ by Language, from the point of view of the driver specified
+ by This.
+
+ @retval EFI_SUCCESS The Unicode string for the user-readable name in the
+ language specified by Language for the driver
+ specified by This was returned in DriverName.
+ @retval EFI_INVALID_PARAMETER ControllerHandle is NULL.
+ @retval EFI_INVALID_PARAMETER ChildHandle is not NULL and it is not a valid EFI_HANDLE.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER ControllerName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This is not currently managing
+ the controller specified by ControllerHandle and
+ ChildHandle.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeComponentNameGetControllerName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_HANDLE ChildHandle OPTIONAL,
+ IN CHAR8 *Language,
+ OUT CHAR16 **ControllerName
+ )
+{
+ EFI_STATUS Status;
+
+ //
+ // ChildHandle must be NULL for a Device Driver
+ //
+ if (ChildHandle != NULL) {
+ return EFI_UNSUPPORTED;
+ }
+
+ //
+ // Lookup name of controller specified by ControllerHandle
+ //
+ Status = EFI_UNSUPPORTED;
+
+ return Status;
+}
diff --git a/OvmfPkg/XenbusDxe/ComponentName.h b/OvmfPkg/XenbusDxe/ComponentName.h
new file mode 100644
index 0000000..aa2e808
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/ComponentName.h
@@ -0,0 +1,91 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenbusDxe
+
+ TODO: Detailed Description of UEFI Driver XenbusDxe
+
+ TODO: Copyright for UEFI Driver XenbusDxe
+
+ TODO: License for UEFI Driver XenbusDxe
+
+**/
+
+/**
+ Retrieves a Unicode string that is the user-readable name of the EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param Language A pointer to a three-character ISO 639-2 language identifier.
+ This is the language of the driver name that that the caller
+ is requesting, and it must match one of the languages specified
+ in SupportedLanguages. The number of languages supported by a
+ driver is up to the driver writer.
+ @param DriverName A pointer to the Unicode string to return. This Unicode string
+ is the name of the driver specified by This in the language
+ specified by Language.
+
+ @retval EFI_SUCCESS The Unicode string for the Driver specified by This
+ and the language specified by Language was returned
+ in DriverName.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER DriverName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeComponentNameGetDriverName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN CHAR8 *Language,
+ OUT CHAR16 **DriverName
+ );
+
+/**
+ Retrieves a Unicode string that is the user readable name of the controller
+ that is being managed by an EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param ControllerHandle The handle of a controller that the driver specified by
+ This is managing. This handle specifies the controller
+ whose name is to be returned.
+ @param ChildHandle The handle of the child controller to retrieve the name
+ of. This is an optional parameter that may be NULL. It
+ will be NULL for device drivers. It will also be NULL
+ for a bus drivers that wish to retrieve the name of the
+ bus controller. It will not be NULL for a bus driver
+ that wishes to retrieve the name of a child controller.
+ @param Language A pointer to a three character ISO 639-2 language
+ identifier. This is the language of the controller name
+ that the caller is requesting, and it must match one
+ of the languages specified in SupportedLanguages. The
+ number of languages supported by a driver is up to the
+ driver writer.
+ @param ControllerName A pointer to the Unicode string to return. This Unicode
+ string is the name of the controller specified by
+ ControllerHandle and ChildHandle in the language specified
+ by Language, from the point of view of the driver specified
+ by This.
+
+ @retval EFI_SUCCESS The Unicode string for the user-readable name in the
+ language specified by Language for the driver
+ specified by This was returned in DriverName.
+ @retval EFI_INVALID_PARAMETER ControllerHandle is NULL.
+ @retval EFI_INVALID_PARAMETER ChildHandle is not NULL and it is not a valid EFI_HANDLE.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER ControllerName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This is not currently managing
+ the controller specified by ControllerHandle and
+ ChildHandle.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeComponentNameGetControllerName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_HANDLE ChildHandle OPTIONAL,
+ IN CHAR8 *Language,
+ OUT CHAR16 **ControllerName
+ );
diff --git a/OvmfPkg/XenbusDxe/DriverBinding.h b/OvmfPkg/XenbusDxe/DriverBinding.h
new file mode 100644
index 0000000..d54bd1a
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/DriverBinding.h
@@ -0,0 +1,139 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenbusDxe
+
+ TODO: Detailed Description of UEFI Driver XenbusDxe
+
+ TODO: Copyright for UEFI Driver XenbusDxe
+
+ TODO: License for UEFI Driver XenbusDxe
+
+**/
+
+/**
+ Tests to see if this driver supports a given controller. If a child device is provided,
+ it further tests to see if this driver supports creating a handle for the specified child device.
+
+ This function checks to see if the driver specified by This supports the device specified by
+ ControllerHandle. Drivers will typically use the device path attached to
+ ControllerHandle and/or the services from the bus I/O abstraction attached to
+ ControllerHandle to determine if the driver supports ControllerHandle. This function
+ may be called many times during platform initialization. In order to reduce boot times, the tests
+ performed by this function must be very small, and take as little time as possible to execute. This
+ function must not change the state of any hardware devices, and this function must be aware that the
+ device specified by ControllerHandle may already be managed by the same driver or a
+ different driver. This function must match its calls to AllocatePages() with FreePages(),
+ AllocatePool() with FreePool(), and OpenProtocol() with CloseProtocol().
+ Because ControllerHandle may have been previously started by the same driver, if a protocol is
+ already in the opened state, then it must not be closed with CloseProtocol(). This is required
+ to guarantee the state of ControllerHandle is not modified by this function.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to test. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For bus drivers, if this parameter is not NULL, then
+ the bus driver must determine if the bus controller specified
+ by ControllerHandle and the child controller specified
+ by RemainingDevicePath are both supported by this
+ bus driver.
+
+ @retval EFI_SUCCESS The device specified by ControllerHandle and
+ RemainingDevicePath is supported by the driver specified by This.
+ @retval EFI_ALREADY_STARTED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by the driver
+ specified by This.
+ @retval EFI_ACCESS_DENIED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by a different
+ driver or an application that requires exclusive access.
+ Currently not implemented.
+ @retval EFI_UNSUPPORTED The device specified by ControllerHandle and
+ RemainingDevicePath is not supported by the driver specified by This.
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverBindingSupported (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ );
+
+/**
+ Starts a device controller or a bus controller.
+
+ The Start() function is designed to be invoked from the EFI boot service ConnectController().
+ As a result, much of the error checking on the parameters to Start() has been moved into this
+ common boot service. It is legal to call Start() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE.
+ 2. If RemainingDevicePath is not NULL, then it must be a pointer to a naturally aligned
+ EFI_DEVICE_PATH_PROTOCOL.
+ 3. Prior to calling Start(), the Supported() function for the driver specified by This must
+ have been called with the same calling parameters, and Supported() must have returned EFI_SUCCESS.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to start. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For a bus driver, if this parameter is NULL, then handles
+ for all the children of Controller are created by this driver.
+ If this parameter is not NULL and the first Device Path Node is
+ not the End of Device Path Node, then only the handle for the
+ child device specified by the first Device Path Node of
+ RemainingDevicePath is created by this driver.
+ If the first Device Path Node of RemainingDevicePath is
+ the End of Device Path Node, no child handle is created by this
+ driver.
+
+ @retval EFI_SUCCESS The device was started.
+ @retval EFI_DEVICE_ERROR The device could not be started due to a device error.Currently not implemented.
+ @retval EFI_OUT_OF_RESOURCES The request could not be completed due to a lack of resources.
+ @retval Others The driver failded to start the device.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverBindingStart (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ );
+
+/**
+ Stops a device controller or a bus controller.
+
+ The Stop() function is designed to be invoked from the EFI boot service DisconnectController().
+ As a result, much of the error checking on the parameters to Stop() has been moved
+ into this common boot service. It is legal to call Stop() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE that was used on a previous call to this
+ same driver's Start() function.
+ 2. The first NumberOfChildren handles of ChildHandleBuffer must all be a valid
+ EFI_HANDLE. In addition, all of these handles must have been created in this driver's
+ Start() function, and the Start() function must have called OpenProtocol() on
+ ControllerHandle with an Attribute of EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle A handle to the device being stopped. The handle must
+ support a bus specific I/O protocol for the driver
+ to use to stop the device.
+ @param[in] NumberOfChildren The number of child device handles in ChildHandleBuffer.
+ @param[in] ChildHandleBuffer An array of child handles to be freed. May be NULL
+ if NumberOfChildren is 0.
+
+ @retval EFI_SUCCESS The device was stopped.
+ @retval EFI_DEVICE_ERROR The device could not be stopped due to a device error.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverBindingStop (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN UINTN NumberOfChildren,
+ IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
+ );
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
new file mode 100644
index 0000000..14113ad
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -0,0 +1,314 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenbusDxe
+
+ TODO: Detailed Description of UEFI Driver XenbusDxe
+
+ TODO: Copyright for UEFI Driver XenbusDxe
+
+ TODO: License for UEFI Driver XenbusDxe
+
+**/
+
+#include <IndustryStandard/Pci.h>
+#include <IndustryStandard/Acpi.h>
+#include <Library/DebugLib.h>
+
+#include "XenbusDxe.h"
+
+
+///
+/// Driver Binding Protocol instance
+///
+EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding = {
+ XenbusDxeDriverBindingSupported,
+ XenbusDxeDriverBindingStart,
+ XenbusDxeDriverBindingStop,
+ XENBUS_DXE_VERSION,
+ NULL,
+ NULL
+};
+
+
+/**
+ Unloads an image.
+
+ @param ImageHandle Handle that identifies the image to be unloaded.
+
+ @retval EFI_SUCCESS The image has been unloaded.
+ @retval EFI_INVALID_PARAMETER ImageHandle is not a valid image handle.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeUnload (
+ IN EFI_HANDLE ImageHandle
+ )
+{
+ EFI_STATUS Status;
+
+ EFI_HANDLE *HandleBuffer;
+ UINTN HandleCount;
+ UINTN Index;
+
+
+ Status = EFI_SUCCESS;
+
+ //
+ // Retrieve array of all handles in the handle database
+ //
+ Status = gBS->LocateHandleBuffer (
+ AllHandles,
+ NULL,
+ NULL,
+ &HandleCount,
+ &HandleBuffer
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ //
+ // Disconnect the current driver from handles in the handle database
+ //
+ for (Index = 0; Index < HandleCount; Index++) {
+ Status = gBS->DisconnectController (HandleBuffer[Index], gImageHandle, NULL);
+ }
+
+ //
+ // Free the array of handles
+ //
+ FreePool (HandleBuffer);
+
+
+ //
+ // Uninstall protocols installed in the driver entry point
+ //
+ Status = gBS->UninstallMultipleProtocolInterfaces (
+ ImageHandle,
+ &gEfiDriverBindingProtocolGuid, &gXenbusDxeDriverBinding,
+ &gEfiComponentNameProtocolGuid, &gXenbusDxeComponentName,
+ &gEfiComponentName2ProtocolGuid, &gXenbusDxeComponentName2,
+ NULL
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+
+ //
+ // Do any additional cleanup that is required for this driver
+ //
+
+ return EFI_SUCCESS;
+}
+
+/**
+ This is the declaration of an EFI image entry point. This entry point is
+ the same for UEFI Applications, UEFI OS Loaders, and UEFI Drivers including
+ both device drivers and bus drivers.
+
+ @param ImageHandle The firmware allocated handle for the UEFI image.
+ @param SystemTable A pointer to the EFI System Table.
+
+ @retval EFI_SUCCESS The operation completed successfully.
+ @retval Others An unexpected error occurred.
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverEntryPoint (
+ IN EFI_HANDLE ImageHandle,
+ IN EFI_SYSTEM_TABLE *SystemTable
+ )
+{
+ EFI_STATUS Status;
+
+
+ Status = EFI_SUCCESS;
+
+
+ //
+ // Install UEFI Driver Model protocol(s).
+ //
+ Status = EfiLibInstallDriverBindingComponentName2 (
+ ImageHandle,
+ SystemTable,
+ &gXenbusDxeDriverBinding,
+ ImageHandle,
+ &gXenbusDxeComponentName,
+ &gXenbusDxeComponentName2
+ );
+ ASSERT_EFI_ERROR (Status);
+
+
+ return Status;
+}
+
+
+/**
+ Tests to see if this driver supports a given controller. If a child device is provided,
+ it further tests to see if this driver supports creating a handle for the specified child device.
+
+ This function checks to see if the driver specified by This supports the device specified by
+ ControllerHandle. Drivers will typically use the device path attached to
+ ControllerHandle and/or the services from the bus I/O abstraction attached to
+ ControllerHandle to determine if the driver supports ControllerHandle. This function
+ may be called many times during platform initialization. In order to reduce boot times, the tests
+ performed by this function must be very small, and take as little time as possible to execute. This
+ function must not change the state of any hardware devices, and this function must be aware that the
+ device specified by ControllerHandle may already be managed by the same driver or a
+ different driver. This function must match its calls to AllocatePages() with FreePages(),
+ AllocatePool() with FreePool(), and OpenProtocol() with CloseProtocol().
+ Because ControllerHandle may have been previously started by the same driver, if a protocol is
+ already in the opened state, then it must not be closed with CloseProtocol(). This is required
+ to guarantee the state of ControllerHandle is not modified by this function.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to test. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For bus drivers, if this parameter is not NULL, then
+ the bus driver must determine if the bus controller specified
+ by ControllerHandle and the child controller specified
+ by RemainingDevicePath are both supported by this
+ bus driver.
+
+ @retval EFI_SUCCESS The device specified by ControllerHandle and
+ RemainingDevicePath is supported by the driver specified by This.
+ @retval EFI_ALREADY_STARTED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by the driver
+ specified by This.
+ @retval EFI_ACCESS_DENIED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by a different
+ driver or an application that requires exclusive access.
+ Currently not implemented.
+ @retval EFI_UNSUPPORTED The device specified by ControllerHandle and
+ RemainingDevicePath is not supported by the driver specified by This.
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverBindingSupported (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ )
+{
+ EFI_STATUS Status;
+ EFI_PCI_IO_PROTOCOL *PciIo;
+ PCI_TYPE00 Pci;
+
+ Status = gBS->OpenProtocol (
+ ControllerHandle,
+ &gEfiPciIoProtocolGuid,
+ (VOID **)&PciIo,
+ This->DriverBindingHandle,
+ ControllerHandle,
+ EFI_OPEN_PROTOCOL_BY_DRIVER
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ Status = PciIo->Pci.Read (PciIo, EfiPciIoWidthUint32, 0,
+ sizeof Pci / sizeof (UINT32), &Pci);
+
+ if (Status == EFI_SUCCESS) {
+ if (Pci.Hdr.VendorId == PCI_VENDOR_ID_XEN &&
+ Pci.Hdr.DeviceId == PCI_DEVICE_ID_XEN_PLATFORM) {
+ Status = EFI_SUCCESS;
+ } else {
+ Status = EFI_UNSUPPORTED;
+ }
+ }
+
+ gBS->CloseProtocol (ControllerHandle, &gEfiPciIoProtocolGuid,
+ This->DriverBindingHandle, ControllerHandle);
+
+ return Status;
+}
+
+/**
+ Starts a device controller or a bus controller.
+
+ The Start() function is designed to be invoked from the EFI boot service ConnectController().
+ As a result, much of the error checking on the parameters to Start() has been moved into this
+ common boot service. It is legal to call Start() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE.
+ 2. If RemainingDevicePath is not NULL, then it must be a pointer to a naturally aligned
+ EFI_DEVICE_PATH_PROTOCOL.
+ 3. Prior to calling Start(), the Supported() function for the driver specified by This must
+ have been called with the same calling parameters, and Supported() must have returned EFI_SUCCESS.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to start. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For a bus driver, if this parameter is NULL, then handles
+ for all the children of Controller are created by this driver.
+ If this parameter is not NULL and the first Device Path Node is
+ not the End of Device Path Node, then only the handle for the
+ child device specified by the first Device Path Node of
+ RemainingDevicePath is created by this driver.
+ If the first Device Path Node of RemainingDevicePath is
+ the End of Device Path Node, no child handle is created by this
+ driver.
+
+ @retval EFI_SUCCESS The device was started.
+ @retval EFI_DEVICE_ERROR The device could not be started due to a device error.Currently not implemented.
+ @retval EFI_OUT_OF_RESOURCES The request could not be completed due to a lack of resources.
+ @retval Others The driver failded to start the device.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverBindingStart (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ )
+{
+ return EFI_UNSUPPORTED;
+}
+
+/**
+ Stops a device controller or a bus controller.
+
+ The Stop() function is designed to be invoked from the EFI boot service DisconnectController().
+ As a result, much of the error checking on the parameters to Stop() has been moved
+ into this common boot service. It is legal to call Stop() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE that was used on a previous call to this
+ same driver's Start() function.
+ 2. The first NumberOfChildren handles of ChildHandleBuffer must all be a valid
+ EFI_HANDLE. In addition, all of these handles must have been created in this driver's
+ Start() function, and the Start() function must have called OpenProtocol() on
+ ControllerHandle with an Attribute of EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle A handle to the device being stopped. The handle must
+ support a bus specific I/O protocol for the driver
+ to use to stop the device.
+ @param[in] NumberOfChildren The number of child device handles in ChildHandleBuffer.
+ @param[in] ChildHandleBuffer An array of child handles to be freed. May be NULL
+ if NumberOfChildren is 0.
+
+ @retval EFI_SUCCESS The device was stopped.
+ @retval EFI_DEVICE_ERROR The device could not be stopped due to a device error.
+
+**/
+EFI_STATUS
+EFIAPI
+XenbusDxeDriverBindingStop (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN UINTN NumberOfChildren,
+ IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
+ )
+{
+ return EFI_UNSUPPORTED;
+}
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
new file mode 100644
index 0000000..6f22732
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -0,0 +1,78 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenbusDxe
+
+ TODO: Detailed Description of UEFI Driver XenbusDxe
+
+ TODO: Copyright for UEFI Driver XenbusDxe
+
+ TODO: License for UEFI Driver XenbusDxe
+
+**/
+
+#ifndef __EFI_XENBUS_DXE_H__
+#define __EFI_XENBUS_DXE_H__
+
+#include <Uefi.h>
+
+//
+// Libraries
+//
+#include <Library/UefiBootServicesTableLib.h>
+#include <Library/MemoryAllocationLib.h>
+#include <Library/BaseMemoryLib.h>
+#include <Library/BaseLib.h>
+#include <Library/UefiLib.h>
+#include <Library/DevicePathLib.h>
+#include <Library/DebugLib.h>
+
+
+//
+// UEFI Driver Model Protocols
+//
+#include <Protocol/DriverBinding.h>
+
+
+//
+// Consumed Protocols
+//
+#include <Protocol/PciIo.h>
+
+
+//
+// Produced Protocols
+//
+
+
+//
+// Guids
+//
+
+//
+// Driver Version
+//
+#define XENBUS_DXE_VERSION 0x00000000
+
+
+//
+// Protocol instances
+//
+extern EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding;
+extern EFI_COMPONENT_NAME2_PROTOCOL gXenbusDxeComponentName2;
+extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
+
+
+//
+// Include files with function prototypes
+//
+#include "DriverBinding.h"
+#include "ComponentName.h"
+
+//
+// Other stuff
+//
+#define PCI_VENDOR_ID_XEN 0x5853
+#define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
+
+
+#endif
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
new file mode 100644
index 0000000..d6685b3
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -0,0 +1,56 @@
+
+## @file
+# TODO: Brief Description of UEFI Driver XenbusDxe
+#
+# TODO: Detailed Description of UEFI Driver XenbusDxe
+#
+# TODO: Copyright for UEFI Driver XenbusDxe
+#
+# TODO: License for UEFI Driver XenbusDxe
+#
+##
+
+[Defines]
+ INF_VERSION = 0x00010005
+ BASE_NAME = XenbusDxe
+ FILE_GUID = 565ec8ba-a484-11e3-802b-b8ac6f7d65e6
+ MODULE_TYPE = UEFI_DRIVER
+
+ VERSION_STRING = 0.1
+ ENTRY_POINT = XenbusDxeDriverEntryPoint
+ UNLOAD_IMAGE = XenbusDxeUnload
+
+
+[Packages]
+ MdePkg/MdePkg.dec
+ OvmfPkg/OvmfPkg.dec
+
+[Sources]
+ XenbusDxe.h
+ XenbusDxe.c
+ DriverBinding.h
+ ComponentName.c
+ ComponentName.h
+
+[LibraryClasses]
+ UefiDriverEntryPoint
+ UefiBootServicesTableLib
+ MemoryAllocationLib
+ BaseMemoryLib
+ BaseLib
+ UefiLib
+ DevicePathLib
+ DebugLib
+ HobLib
+
+
+[Protocols]
+ gEfiDriverBindingProtocolGuid
+ gEfiPciIoProtocolGuid
+ gEfiComponentName2ProtocolGuid
+ gEfiComponentNameProtocolGuid
+
+
+[Guids]
+ gEfiXenInfoGuid
+
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
2014-07-16 15:15 ` [PATCH RFC 01/18] OvmfPkg: Add public headers from Xen Project Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls Anthony PERARD
` (24 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
The ExitBoot event is used to disconnect from the device before the
next operating system start using them.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/XenbusDxe.c | 39 ++++++++++++++++++++++++++++++++++++---
OvmfPkg/XenbusDxe/XenbusDxe.h | 10 ++++++++++
2 files changed, 46 insertions(+), 3 deletions(-)
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
index 14113ad..2c3d9b8 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.c
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -16,7 +16,6 @@
#include "XenbusDxe.h"
-
///
/// Driver Binding Protocol instance
///
@@ -30,6 +29,8 @@ EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding = {
};
+XENBUS_DEVICE *mMyDevice;
+
/**
Unloads an image.
@@ -229,6 +230,19 @@ XenbusDxeDriverBindingSupported (
return Status;
}
+VOID
+EFIAPI
+NotifyExitBoot (
+ IN EFI_EVENT Event,
+ IN VOID *Context
+ )
+{
+ XENBUS_DEVICE *Dev = Context;
+
+ gBS->DisconnectController(Dev->ControllerHandle,
+ Dev->This->DriverBindingHandle, NULL);
+}
+
/**
Starts a device controller or a bus controller.
@@ -272,7 +286,22 @@ XenbusDxeDriverBindingStart (
IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
)
{
- return EFI_UNSUPPORTED;
+ EFI_STATUS Status;
+ XENBUS_DEVICE *Dev;
+
+ Dev = AllocateZeroPool (sizeof (*Dev));
+ Dev->Signature = XENBUS_DEVICE_SIGNATURE;
+ Dev->This = This;
+ Dev->ControllerHandle = ControllerHandle;
+
+ Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
+ NotifyExitBoot,
+ (VOID*) Dev,
+ &Dev->ExitBootEvent);
+ ASSERT_EFI_ERROR (Status);
+
+ mMyDevice = Dev;
+ return EFI_SUCCESS;
}
/**
@@ -310,5 +339,9 @@ XenbusDxeDriverBindingStop (
IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
)
{
- return EFI_UNSUPPORTED;
+ XENBUS_DEVICE *Dev = mMyDevice;
+
+ gBS->CloseEvent (Dev->ExitBootEvent);
+
+ return EFI_SUCCESS;
}
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index 6f22732..c1ca87a 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -75,4 +75,14 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
#define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
+typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
+
+#define XENBUS_DEVICE_SIGNATURE SIGNATURE_32 ('X','B','b','d')
+struct _XENBUS_DEVICE {
+ UINT32 Signature;
+ EFI_DRIVER_BINDING_PROTOCOL *This;
+ EFI_HANDLE ControllerHandle;
+ EFI_EVENT ExitBootEvent;
+};
+
#endif
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (2 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol Anthony PERARD
` (23 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/X64/hypercall.S | 16 ++++++
OvmfPkg/XenbusDxe/X64/hypercall.asm | 19 +++++++
OvmfPkg/XenbusDxe/XenHypercall.c | 101 ++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/XenHypercall.h | 44 ++++++++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.c | 8 +++
OvmfPkg/XenbusDxe/XenbusDxe.h | 5 ++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 6 +++
7 files changed, 199 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/X64/hypercall.S
create mode 100644 OvmfPkg/XenbusDxe/X64/hypercall.asm
create mode 100644 OvmfPkg/XenbusDxe/XenHypercall.c
create mode 100644 OvmfPkg/XenbusDxe/XenHypercall.h
diff --git a/OvmfPkg/XenbusDxe/X64/hypercall.S b/OvmfPkg/XenbusDxe/X64/hypercall.S
new file mode 100644
index 0000000..80982e3
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/X64/hypercall.S
@@ -0,0 +1,16 @@
+# Hypercall with 2 arguments
+# INTN
+# EFIAPI
+# XenHypercall2 (
+# IN VOID *hypercall_addr,
+# IN OUT INTN arg1,
+# IN OUT INTN arg2
+# );
+# XXX push and pop ?
+ASM_GLOBAL ASM_PFX(XenHypercall2)
+ASM_PFX(XenHypercall2):
+ movq %rcx, %rax
+ movq %rdx, %rdi
+ movq %r8, %rsi
+ call *%rax
+ ret
diff --git a/OvmfPkg/XenbusDxe/X64/hypercall.asm b/OvmfPkg/XenbusDxe/X64/hypercall.asm
new file mode 100644
index 0000000..8bc41a2
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/X64/hypercall.asm
@@ -0,0 +1,19 @@
+.code
+
+; Hypercall with 2 arguments
+; INTN
+; EFIAPI
+; XenHypercall2 (
+; IN VOID *hypercall_addr,
+; IN OUT INTN arg1,
+; IN OUT INTN arg2
+; );
+XenHypercall2 PROC
+ mov rax, rcx
+ mov rdi, rdx
+ mov rsi, r8
+ call rax
+ ret
+XenHypercall2 ENDP
+
+END
diff --git a/OvmfPkg/XenbusDxe/XenHypercall.c b/OvmfPkg/XenbusDxe/XenHypercall.c
new file mode 100644
index 0000000..5e98dd5
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenHypercall.c
@@ -0,0 +1,101 @@
+#include <PiDxe.h>
+#include <Library/HobLib.h>
+#include <Guid/XenInfo.h>
+
+#include "XenbusDxe.h"
+#include "XenHypercall.h"
+
+#include <IndustryStandard/Xen/hvm/params.h>
+#include <IndustryStandard/Xen/memory.h>
+
+EFI_STATUS
+XenHyperpageInit (
+ XENBUS_DEVICE *Dev
+ )
+{
+ EFI_HOB_GUID_TYPE *GuidHob;
+ EFI_XEN_INFO *XenInfo;
+
+ GuidHob = GetFirstGuidHob (&gEfiXenInfoGuid);
+ if (GuidHob == NULL) {
+ DEBUG ((EFI_D_INFO, "XenbusInit: No xeninfo ?\n"));
+ return EFI_NOT_FOUND;
+ }
+ XenInfo = (EFI_XEN_INFO*)GET_GUID_HOB_DATA (GuidHob);
+ Dev->Hyperpage = XenInfo->HyperPages;
+ return EFI_SUCCESS;
+}
+
+UINTN
+XenHypercallHvmGetParam (
+ XENBUS_DEVICE *Dev,
+ INTN Index
+ )
+{
+ xen_hvm_param_t xhv;
+ INTN error;
+
+ ASSERT (Dev->Hyperpage != NULL);
+
+ xhv.domid = DOMID_SELF;
+ xhv.index = Index;
+ error = XenHypercall2(Dev->Hyperpage + __HYPERVISOR_hvm_op * 32,
+ HVMOP_get_param, (INTN)&xhv);
+ if (error) {
+ DEBUG ((EFI_D_ERROR, "%a: error %d trying to get %d\n", __func__,
+ error, Index));
+ return (0);
+ }
+ return (xhv.value);
+}
+
+INTN
+XenHypercallMemoryOp (
+ IN XENBUS_DEVICE *Dev,
+ IN UINTN Operation,
+ IN OUT VOID *Arguments
+ )
+{
+ ASSERT (Dev->Hyperpage != NULL);
+ return XenHypercall2(Dev->Hyperpage + __HYPERVISOR_memory_op * 32,
+ Operation, (INTN)Arguments);
+}
+
+INTN
+XenHypercallEventChannelOp (
+ IN XENBUS_DEVICE *Dev,
+ IN INTN Operation,
+ IN OUT VOID *Arguments
+ )
+{
+ ASSERT (Dev->Hyperpage != NULL);
+ return XenHypercall2(Dev->Hyperpage + __HYPERVISOR_event_channel_op * 32,
+ Operation, (INTN)Arguments);
+}
+
+EFI_STATUS
+XenGetSharedInfoPage (
+ IN OUT XENBUS_DEVICE *Dev
+ )
+{
+ struct xen_add_to_physmap xatp;
+
+ ASSERT (Dev->SharedInfo == NULL);
+
+ xatp.domid = DOMID_SELF;
+ xatp.space = XENMAPSPACE_shared_info;
+ xatp.idx = 0;
+ /* using reserved page because the page is not released when linux is
+ * starting because of the add_to_physmap. QEMU might try to access the
+ * page, and fail because it have no right to do so (segv).
+ */
+ Dev->SharedInfo = AllocateReservedPages (1);
+ xatp.gpfn = (UINTN)Dev->SharedInfo >> EFI_PAGE_SHIFT;
+ if (XenHypercallMemoryOp (Dev, XENMEM_add_to_physmap, &xatp) != 0) {
+ FreePages (Dev->SharedInfo, 1);
+ Dev->SharedInfo = NULL;
+ return EFI_LOAD_ERROR;
+ }
+
+ return EFI_SUCCESS;
+}
diff --git a/OvmfPkg/XenbusDxe/XenHypercall.h b/OvmfPkg/XenbusDxe/XenHypercall.h
new file mode 100644
index 0000000..1afaf73
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenHypercall.h
@@ -0,0 +1,44 @@
+#ifndef __XENBUS_DXE_HYPERCALL_H__
+#define __XENBUS_DXE_HYPERCALL_H__
+
+typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
+
+INTN
+EFIAPI
+XenHypercall2 (
+ IN VOID *hypercall_addr,
+ IN OUT INTN arg1,
+ IN OUT INTN arg2
+ );
+
+EFI_STATUS
+XenHyperpageInit (
+ XENBUS_DEVICE *Dev
+ );
+
+UINTN
+XenHypercallHvmGetParam (
+ XENBUS_DEVICE *Dev,
+ INTN Index
+ );
+
+INTN
+XenHypercallMemoryOp (
+ IN XENBUS_DEVICE *Dev,
+ IN UINTN Operation,
+ IN OUT VOID *Arguments
+ );
+
+INTN
+XenHypercallEventChannelOp (
+ IN XENBUS_DEVICE *Dev,
+ IN INTN Operation,
+ IN OUT VOID *Arguments
+ );
+
+EFI_STATUS
+XenGetSharedInfoPage (
+ IN OUT XENBUS_DEVICE *Dev
+ );
+
+#endif
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
index 2c3d9b8..63ea31b 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.c
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -16,6 +16,8 @@
#include "XenbusDxe.h"
+#include "XenHypercall.h"
+
///
/// Driver Binding Protocol instance
///
@@ -294,6 +296,12 @@ XenbusDxeDriverBindingStart (
Dev->This = This;
Dev->ControllerHandle = ControllerHandle;
+ Status = XenHyperpageInit (Dev);
+ ASSERT_EFI_ERROR (Status);
+
+ Status = XenGetSharedInfoPage (Dev);
+ ASSERT_EFI_ERROR (Status);
+
Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
NotifyExitBoot,
(VOID*) Dev,
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index c1ca87a..6140f94 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -71,6 +71,8 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
//
// Other stuff
//
+#include <IndustryStandard/Xen/xen.h>
+
#define PCI_VENDOR_ID_XEN 0x5853
#define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
@@ -83,6 +85,9 @@ struct _XENBUS_DEVICE {
EFI_DRIVER_BINDING_PROTOCOL *This;
EFI_HANDLE ControllerHandle;
EFI_EVENT ExitBootEvent;
+
+ VOID *Hyperpage;
+ shared_info_t *SharedInfo;
};
#endif
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index d6685b3..8b69f93 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -31,6 +31,12 @@
DriverBinding.h
ComponentName.c
ComponentName.h
+ XenHypercall.c
+ XenHypercall.h
+
+[Sources.X64]
+ X64/hypercall.S
+ X64/hypercall.asm
[LibraryClasses]
UefiDriverEntryPoint
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (3 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol Anthony PERARD
` (22 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/XenbusDxe.c | 16 ++++++++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.h | 1 +
2 files changed, 17 insertions(+)
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
index 63ea31b..ba5e8f4 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.c
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -290,11 +290,25 @@ XenbusDxeDriverBindingStart (
{
EFI_STATUS Status;
XENBUS_DEVICE *Dev;
+ EFI_PCI_IO_PROTOCOL *PciIo;
+
+ Status = gBS->OpenProtocol (
+ ControllerHandle,
+ &gEfiPciIoProtocolGuid,
+ (VOID **)&PciIo,
+ This->DriverBindingHandle,
+ ControllerHandle,
+ EFI_OPEN_PROTOCOL_BY_DRIVER
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
Dev = AllocateZeroPool (sizeof (*Dev));
Dev->Signature = XENBUS_DEVICE_SIGNATURE;
Dev->This = This;
Dev->ControllerHandle = ControllerHandle;
+ Dev->PciIo = PciIo;
Status = XenHyperpageInit (Dev);
ASSERT_EFI_ERROR (Status);
@@ -351,5 +365,7 @@ XenbusDxeDriverBindingStop (
gBS->CloseEvent (Dev->ExitBootEvent);
+ gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
+ This->DriverBindingHandle, ControllerHandle);
return EFI_SUCCESS;
}
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index 6140f94..d57e3c8 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -84,6 +84,7 @@ struct _XENBUS_DEVICE {
UINT32 Signature;
EFI_DRIVER_BINDING_PROTOCOL *This;
EFI_HANDLE ControllerHandle;
+ EFI_PCI_IO_PROTOCOL *PciIo;
EFI_EVENT ExitBootEvent;
VOID *Hyperpage;
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (4 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 07/18] OvmfPkg/XenbusDxe: Add InterlockedCompareExchange16 Anthony PERARD
` (21 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This protocol will be used for communication between a PV driver (like a
PV block driver) and the Xenbus/Xenstore.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/Include/Protocol/Xenbus.h | 62 +++++++++++++++++++++++++++++++++++++++
OvmfPkg/OvmfPkg.dec | 2 ++
OvmfPkg/XenbusDxe/ComponentName.c | 12 ++++++++
OvmfPkg/XenbusDxe/XenbusDxe.h | 3 ++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 1 +
5 files changed, 80 insertions(+)
create mode 100644 OvmfPkg/Include/Protocol/Xenbus.h
diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
new file mode 100644
index 0000000..191cee1
--- /dev/null
+++ b/OvmfPkg/Include/Protocol/Xenbus.h
@@ -0,0 +1,62 @@
+
+/** @file
+ TODO: Brief Description of Protocol Xenbus
+
+ TODO: Detailed Description of Protocol Xenbus
+
+ TODO: Copyright for Protocol Xenbus
+
+ TODO: License for Protocol Xenbus
+
+**/
+
+#ifndef __PROTOCOL_XENBUS_H__
+#define __PROTOCOL_XENBUS_H__
+
+#define XENBUS_PROTOCOL_GUID \
+ {0x3d3ca290, 0xb9a5, 0x11e3, {0xb7, 0x5d, 0xb8, 0xac, 0x6f, 0x7d, 0x65, 0xe6}}
+
+///
+/// Forward declaration
+///
+typedef struct _XENBUS_PROTOCOL XENBUS_PROTOCOL;
+
+
+#include <IndustryStandard/Xen/grant_table.h>
+
+///
+/// Function prototypes
+///
+
+typedef
+EFI_STATUS
+(EFIAPI *XENBUS_GRANT_ACCESS)(
+ IN XENBUS_PROTOCOL *This,
+ IN domid_t DomainId,
+ IN UINTN Frame,
+ IN BOOLEAN ReadOnly,
+ OUT grant_ref_t *refp
+ );
+
+typedef
+EFI_STATUS
+(EFIAPI *XENBUS_GRANT_END_ACCESS)(
+ IN XENBUS_PROTOCOL *This,
+ IN grant_ref_t Ref
+ );
+
+
+///
+/// Protocol structure
+///
+struct _XENBUS_PROTOCOL {
+ XENBUS_GRANT_ACCESS GrantAccess;
+ XENBUS_GRANT_END_ACCESS GrantEndAccess;
+ //
+ // Place protocol data fields here
+ //
+};
+
+extern EFI_GUID gXenbusProtocolGuid;
+
+#endif
diff --git a/OvmfPkg/OvmfPkg.dec b/OvmfPkg/OvmfPkg.dec
index f829247..095d40a 100644
--- a/OvmfPkg/OvmfPkg.dec
+++ b/OvmfPkg/OvmfPkg.dec
@@ -47,6 +47,8 @@
[Protocols]
gVirtioDeviceProtocolGuid = {0xfa920010, 0x6785, 0x4941, {0xb6, 0xec, 0x49, 0x8c, 0x57, 0x9f, 0x16, 0x0a}}
gBlockMmioProtocolGuid = {0x6b558ce3, 0x69e5, 0x4c67, {0xa6, 0x34, 0xf7, 0xfe, 0x72, 0xad, 0xbe, 0x84}}
+ ## Include/Protocol/Xenbus.h
+ gXenbusProtocolGuid = {0x3d3ca290, 0xb9a5, 0x11e3, {0xb7, 0x5d, 0xb8, 0xac, 0x6f, 0x7d, 0x65, 0xe6}}
[PcdsFixedAtBuild]
gUefiOvmfPkgTokenSpaceGuid.PcdOvmfPeiMemFvBase|0x0|UINT32|0
diff --git a/OvmfPkg/XenbusDxe/ComponentName.c b/OvmfPkg/XenbusDxe/ComponentName.c
index 2cf11b5..fd46cc8 100644
--- a/OvmfPkg/XenbusDxe/ComponentName.c
+++ b/OvmfPkg/XenbusDxe/ComponentName.c
@@ -140,6 +140,18 @@ XenbusDxeComponentNameGetControllerName (
}
//
+ // Make sure this driver is currently managing ControllerHandle
+ //
+ Status = EfiTestManagedDevice (
+ ControllerHandle,
+ gXenbusDxeDriverBinding.DriverBindingHandle,
+ &gXenbusProtocolGuid
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ //
// Lookup name of controller specified by ControllerHandle
//
Status = EFI_UNSUPPORTED;
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index d57e3c8..975fb6b 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -42,6 +42,9 @@
//
// Produced Protocols
//
+// Xen interface version used:
+#define __XEN_INTERFACE_VERSION__ 0x00040400
+#include <Protocol/Xenbus.h>
//
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index 8b69f93..1a6b131 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -55,6 +55,7 @@
gEfiPciIoProtocolGuid
gEfiComponentName2ProtocolGuid
gEfiComponentNameProtocolGuid
+ gXenbusProtocolGuid
[Guids]
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 07/18] OvmfPkg/XenbusDxe: Add InterlockedCompareExchange16.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (5 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions Anthony PERARD
` (20 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This is a copy of InterlockedCompareExchange32 from the
BaseSynchronizationLib.
The function will be used in the next patch.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h | 7 ++++
.../XenbusDxe/X64/InterlockedCompareExchange16.c | 41 ++++++++++++++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 2 ++
3 files changed, 50 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h
create mode 100644 OvmfPkg/XenbusDxe/X64/InterlockedCompareExchange16.c
diff --git a/OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h b/OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h
new file mode 100644
index 0000000..55f61a8
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h
@@ -0,0 +1,7 @@
+UINT16
+EFIAPI
+InterlockedCompareExchange16 (
+ IN OUT volatile UINT16 *Value,
+ IN UINT16 CompareValue,
+ IN UINT16 ExchangeValue
+ );
diff --git a/OvmfPkg/XenbusDxe/X64/InterlockedCompareExchange16.c b/OvmfPkg/XenbusDxe/X64/InterlockedCompareExchange16.c
new file mode 100644
index 0000000..64962ba
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/X64/InterlockedCompareExchange16.c
@@ -0,0 +1,41 @@
+// Took from BaseSynchronizationLib, and replaced 32 by 16
+/**
+ Performs an atomic compare exchange operation on a 16-bit unsigned integer.
+
+ Performs an atomic compare exchange operation on the 16-bit unsigned integer
+ specified by Value. If Value is equal to CompareValue, then Value is set to
+ ExchangeValue and CompareValue is returned. If Value is not equal to CompareValue,
+ then Value is returned. The compare exchange operation must be performed using
+ MP safe mechanisms.
+
+
+ @param Value A pointer to the 16-bit value for the compare exchange
+ operation.
+ @param CompareValue 16-bit value used in compare operation.
+ @param ExchangeValue 16-bit value used in exchange operation.
+
+ @return The original *Value before exchange.
+
+**/
+UINT16
+EFIAPI
+InterlockedCompareExchange16 (
+ IN OUT volatile UINT16 *Value,
+ IN UINT16 CompareValue,
+ IN UINT16 ExchangeValue
+ )
+{
+ __asm__ __volatile__ (
+ "lock \n\t"
+ "cmpxchgw %3, %1 "
+ : "=a" (CompareValue), // %0
+ "=m" (*Value) // %1
+ : "a" (CompareValue), // %2
+ "r" (ExchangeValue), // %3
+ "m" (*Value)
+ : "memory",
+ "cc"
+ );
+
+ return CompareValue;
+}
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index 1a6b131..8f58821 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -33,10 +33,12 @@
ComponentName.h
XenHypercall.c
XenHypercall.h
+ InterlockedCompareExchange16.h
[Sources.X64]
X64/hypercall.S
X64/hypercall.asm
+ X64/InterlockedCompareExchange16.c | GCC
[LibraryClasses]
UefiDriverEntryPoint
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (6 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 07/18] OvmfPkg/XenbusDxe: Add InterlockedCompareExchange16 Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 09/18] OvmfPkg/XenbusDxe: Add Event Channel Notify Anthony PERARD
` (19 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Steven Smith, Grzegorz Milos, Xen Devel
There are used to grant access of pages to other Xen domains.
This code originaly comes from the Xen Project, and more precisely from
MiniOS.
Signed-off-by: Steven Smith <sos22@cam.ac.uk>
Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/GrantTable.c | 204 ++++++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/GrantTable.h | 34 +++++++
OvmfPkg/XenbusDxe/XenbusDxe.c | 15 +++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 2 +
4 files changed, 255 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/GrantTable.c
create mode 100644 OvmfPkg/XenbusDxe/GrantTable.h
diff --git a/OvmfPkg/XenbusDxe/GrantTable.c b/OvmfPkg/XenbusDxe/GrantTable.c
new file mode 100644
index 0000000..97bd15a
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/GrantTable.c
@@ -0,0 +1,204 @@
+/*
+ ****************************************************************************
+ * (C) 2006 - Cambridge University
+ ****************************************************************************
+ *
+ * File: GrantTable.c
+ * Author: Steven Smith (sos22@cam.ac.uk)
+ * Changes: Grzegorz Milos (gm281@cam.ac.uk)
+ *
+ * Date: July 2006
+ *
+ * Environment: Xen Minimal OS
+ * Description: Simple grant tables implementation. About as stupid as it's
+ * possible to be and still work.
+ *
+ ****************************************************************************
+ */
+#include "XenbusDxe.h"
+
+#include <IndustryStandard/Xen/memory.h>
+
+#include "XenHypercall.h"
+
+#include "GrantTable.h"
+#include "InterlockedCompareExchange16.h"
+
+#define NR_RESERVED_ENTRIES 8
+
+/* NR_GRANT_FRAMES must be less than or equal to that configured in Xen */
+#define NR_GRANT_FRAMES 4
+#define NR_GRANT_ENTRIES (NR_GRANT_FRAMES * EFI_PAGE_SIZE / sizeof(grant_entry_v1_t))
+
+STATIC grant_entry_v1_t *GrantTable = NULL;
+STATIC grant_ref_t GrantList[NR_GRANT_ENTRIES];
+#ifdef GNT_DEBUG
+STATIC BOOLEAN GrantInUseList[NR_GRANT_ENTRIES];
+#endif
+
+STATIC
+VOID
+XenGrantTablePutFreeEntry (
+ grant_ref_t Ref
+ )
+{
+ //local_irq_save(flags);
+ /* MemoryFence (); */
+ // lock ?
+#ifdef GNT_DEBUG
+ ASSERT (GrantInUseList[Ref]);
+ GrantInUseList[Ref] = FALSE;
+#endif
+ GrantList[Ref] = GrantList[0];
+ GrantList[0] = Ref;
+ //local_irq_restore(flags);
+ /* MemoryFence (); */
+}
+
+STATIC
+grant_ref_t
+XenGrantTableGetFreeEntry (
+ VOID
+ )
+{
+ UINTN Ref;
+ /* local_irq_save(flags); */
+ // lock ??
+ Ref = GrantList[0];
+ ASSERT (Ref >= NR_RESERVED_ENTRIES && Ref < NR_GRANT_ENTRIES);
+ GrantList[0] = GrantList[Ref];
+#ifdef GNT_DEBUG
+ ASSERT (!GrantInUseList[Ref]);
+ GrantInUseList[Ref] = TRUE;
+#endif
+ /* local_irq_restore(flags); */
+ return Ref;
+}
+
+STATIC
+grant_ref_t
+XenGrantTableGrantAccess (
+ IN domid_t DomainId,
+ IN UINTN Frame,
+ IN BOOLEAN ReadOnly
+ )
+{
+ grant_ref_t Ref;
+ UINT32 Flags;
+
+ ASSERT (GrantTable != NULL);
+ Ref = XenGrantTableGetFreeEntry ();
+ GrantTable[Ref].frame = Frame;
+ GrantTable[Ref].domid = DomainId;
+ MemoryFence ();
+ Flags = GTF_permit_access;
+ if (ReadOnly) {
+ Flags |= GTF_readonly;
+ }
+ GrantTable[Ref].flags = Flags;
+
+ return Ref;
+}
+
+STATIC
+EFI_STATUS
+XenGrantTableEndAccess (
+ grant_ref_t Ref
+ )
+{
+ UINT16 Flags, OldFlags;
+
+ ASSERT (GrantTable != NULL);
+ ASSERT (Ref >= NR_RESERVED_ENTRIES && Ref < NR_GRANT_ENTRIES);
+
+ OldFlags = GrantTable[Ref].flags;
+ do {
+ if ((Flags = OldFlags) & (GTF_reading | GTF_writing)) {
+ DEBUG ((EFI_D_WARN, "WARNING: g.e. still in use! (%x)\n", Flags));
+ return EFI_NOT_READY;
+ }
+ OldFlags = InterlockedCompareExchange16 (&GrantTable[Ref].flags, Flags, 0);
+ } while (OldFlags != Flags);
+
+ XenGrantTablePutFreeEntry (Ref);
+ return EFI_SUCCESS;
+}
+
+VOID
+XenGrantTableInit (
+ IN XENBUS_DEVICE *Dev,
+ IN UINT64 MmioAddr
+ )
+{
+ xen_add_to_physmap_t Args;
+ INTN Index;
+ INTN ReturnCode;
+
+#ifdef GNT_DEBUG
+ SetMem(GrantInUseList, sizeof (GrantInUseList), 1);
+#endif
+ for (Index = NR_RESERVED_ENTRIES; Index < NR_GRANT_ENTRIES; Index++) {
+ XenGrantTablePutFreeEntry (Index);
+ }
+
+ GrantTable = (VOID*)(UINTN) MmioAddr;
+ for (Index = 0; Index < NR_GRANT_FRAMES; Index++) {
+ Args.domid = DOMID_SELF;
+ Args.idx = Index;
+ Args.space = XENMAPSPACE_grant_table;
+ Args.gpfn = (((xen_pfn_t) GrantTable) >> EFI_PAGE_SHIFT) + Index;
+ ReturnCode = XenHypercallMemoryOp (Dev, XENMEM_add_to_physmap, &Args);
+ if (ReturnCode != 0) {
+ DEBUG ((EFI_D_ERROR, "Xen GrantTable, add_to_physmap hypercall error: %d\n", ReturnCode));
+ }
+ }
+}
+
+VOID
+XenGrantTableDeinit (
+ XENBUS_DEVICE *Dev
+ )
+{
+ INTN ReturnCode, Index;
+ xen_remove_from_physmap_t Args;
+
+ if (GrantTable == NULL) {
+ return;
+ }
+
+ for (Index = NR_GRANT_FRAMES - 1; Index >= 0; Index--) {
+ Args.domid = DOMID_SELF;
+ Args.gpfn = (((xen_pfn_t) GrantTable) >> EFI_PAGE_SHIFT) + Index;
+ DEBUG ((EFI_D_INFO, "%a %d, removing %X\n", __func__, __LINE__, Args.gpfn));
+ ReturnCode = XenHypercallMemoryOp (Dev, XENMEM_remove_from_physmap, &Args);
+ if (ReturnCode != 0) {
+ DEBUG ((EFI_D_ERROR, "Xen GrantTable, remove_from_physmap hypercall error: %d\n", ReturnCode));
+ }
+ }
+ // XXX remove it from the physmap?
+ GrantTable = NULL;
+}
+
+EFI_STATUS
+EFIAPI
+XenbusGrantAccess (
+ IN XENBUS_PROTOCOL *This,
+ IN domid_t DomainId,
+ IN UINTN Frame, // MFN
+ IN BOOLEAN ReadOnly,
+ OUT grant_ref_t *RefPtr
+ )
+{
+ *RefPtr = XenGrantTableGrantAccess (DomainId, Frame, ReadOnly);
+ return EFI_SUCCESS;
+}
+
+EFI_STATUS
+EFIAPI
+XenbusGrantEndAccess (
+ IN XENBUS_PROTOCOL *This,
+ IN grant_ref_t Ref
+ )
+{
+ return XenGrantTableEndAccess (Ref);
+}
diff --git a/OvmfPkg/XenbusDxe/GrantTable.h b/OvmfPkg/XenbusDxe/GrantTable.h
new file mode 100644
index 0000000..05ab4be
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/GrantTable.h
@@ -0,0 +1,34 @@
+#ifndef __GNTTAB_H__
+#define __GNTTAB_H__
+
+#include <IndustryStandard/Xen/grant_table.h>
+
+VOID
+XenGrantTableInit (
+ IN XENBUS_DEVICE *Dev,
+ IN UINT64 MmioAddr
+ );
+
+VOID
+XenGrantTableDeinit (
+ IN XENBUS_DEVICE *Dev
+ );
+
+EFI_STATUS
+EFIAPI
+XenbusGrantAccess (
+ IN XENBUS_PROTOCOL *This,
+ IN domid_t DomainId,
+ IN UINTN Frame, // MFN
+ IN BOOLEAN ReadOnly,
+ OUT grant_ref_t *RefPtr
+ );
+
+EFI_STATUS
+EFIAPI
+XenbusGrantEndAccess (
+ IN XENBUS_PROTOCOL *This,
+ IN grant_ref_t Ref
+ );
+
+#endif /* !__GNTTAB_H__ */
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
index ba5e8f4..b19055d 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.c
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -17,6 +17,7 @@
#include "XenbusDxe.h"
#include "XenHypercall.h"
+#include "GrantTable.h"
///
/// Driver Binding Protocol instance
@@ -291,6 +292,8 @@ XenbusDxeDriverBindingStart (
EFI_STATUS Status;
XENBUS_DEVICE *Dev;
EFI_PCI_IO_PROTOCOL *PciIo;
+ EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR *BarDesc;
+ UINT64 MmioAddr;
Status = gBS->OpenProtocol (
ControllerHandle,
@@ -310,12 +313,23 @@ XenbusDxeDriverBindingStart (
Dev->ControllerHandle = ControllerHandle;
Dev->PciIo = PciIo;
+ Status = PciIo->GetBarAttributes (PciIo, PCI_BAR_IDX1, NULL, (VOID**) &BarDesc);
+ ASSERT_EFI_ERROR (Status);
+ ASSERT (BarDesc->ResType == ACPI_ADDRESS_SPACE_TYPE_MEM);
+
+ /* Get a Memory address for mapping the Grant Table. */
+ DEBUG((EFI_D_INFO, "%a, bar at %LX\n", __func__, BarDesc->AddrRangeMin));
+ MmioAddr = BarDesc->AddrRangeMin;
+ FreePool (BarDesc);
+
Status = XenHyperpageInit (Dev);
ASSERT_EFI_ERROR (Status);
Status = XenGetSharedInfoPage (Dev);
ASSERT_EFI_ERROR (Status);
+ XenGrantTableInit (Dev, MmioAddr);
+
Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
NotifyExitBoot,
(VOID*) Dev,
@@ -364,6 +378,7 @@ XenbusDxeDriverBindingStop (
XENBUS_DEVICE *Dev = mMyDevice;
gBS->CloseEvent (Dev->ExitBootEvent);
+ XenGrantTableDeinit (Dev);
gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
This->DriverBindingHandle, ControllerHandle);
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index 8f58821..093932b 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -34,6 +34,8 @@
XenHypercall.c
XenHypercall.h
InterlockedCompareExchange16.h
+ GrantTable.c
+ GrantTable.h
[Sources.X64]
X64/hypercall.S
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 09/18] OvmfPkg/XenbusDxe: Add Event Channel Notify.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (7 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 10/18] OvmfPkg/XenbusDxe: Add TestAndClearBit Anthony PERARD
` (18 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This first function is used to notify the other side that there is
something to do. The other side is another Xen domain.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/EventChannel.c | 16 ++++++++++++++++
OvmfPkg/XenbusDxe/EventChannel.h | 14 ++++++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 2 ++
3 files changed, 32 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/EventChannel.c
create mode 100644 OvmfPkg/XenbusDxe/EventChannel.h
diff --git a/OvmfPkg/XenbusDxe/EventChannel.c b/OvmfPkg/XenbusDxe/EventChannel.c
new file mode 100644
index 0000000..4ab8a40
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/EventChannel.c
@@ -0,0 +1,16 @@
+#include "EventChannel.h"
+#include "XenHypercall.h"
+
+VOID
+XenEventChannelNotify (
+ IN XENBUS_DEVICE *Dev,
+ IN evtchn_port_t Port
+ )
+{
+ INTN ReturnCode;
+ evtchn_send_t send;
+
+ send.port = Port;
+ ReturnCode = XenHypercallEventChannelOp (Dev, EVTCHNOP_send, &send);
+ ASSERT (ReturnCode == 0);
+}
diff --git a/OvmfPkg/XenbusDxe/EventChannel.h b/OvmfPkg/XenbusDxe/EventChannel.h
new file mode 100644
index 0000000..8a3f6d4
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/EventChannel.h
@@ -0,0 +1,14 @@
+#ifndef __XENBUS_EVENT_CHANNEL_H
+#define __XENBUS_EVENT_CHANNEL_H
+
+#include "XenbusDxe.h"
+
+#include <IndustryStandard/Xen/event_channel.h>
+
+VOID
+XenEventChannelNotify (
+ IN XENBUS_DEVICE *Dev,
+ IN evtchn_port_t Port
+ );
+
+#endif
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index 093932b..d4007db 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -36,6 +36,8 @@
InterlockedCompareExchange16.h
GrantTable.c
GrantTable.h
+ EventChannel.c
+ EventChannel.h
[Sources.X64]
X64/hypercall.S
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 10/18] OvmfPkg/XenbusDxe: Add TestAndClearBit.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (8 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 09/18] OvmfPkg/XenbusDxe: Add Event Channel Notify Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 11/18] OvmfPkg/XenbusDxe: Add XenStore client implementation Anthony PERARD
` (17 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This atomically test and clear a bit.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/X64/TestAndClearBit.S | 10 ++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.h | 11 +++++++++++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 1 +
3 files changed, 22 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/X64/TestAndClearBit.S
diff --git a/OvmfPkg/XenbusDxe/X64/TestAndClearBit.S b/OvmfPkg/XenbusDxe/X64/TestAndClearBit.S
new file mode 100644
index 0000000..e266f68
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/X64/TestAndClearBit.S
@@ -0,0 +1,10 @@
+# TestAndClearBit (
+# IN INT32 Bit, // rcx
+# IN volatile VOID* Address // rdx
+# );
+ASM_GLOBAL ASM_PFX(TestAndClearBit)
+ASM_PFX(TestAndClearBit):
+ lock
+ btrl %ecx, (%rdx)
+ sbbl %eax, %eax
+ ret
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index 975fb6b..50d0925 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -94,4 +94,15 @@ struct _XENBUS_DEVICE {
shared_info_t *SharedInfo;
};
+/*
+ * Helpers
+ */
+
+INT32
+EFIAPI
+TestAndClearBit (
+ IN INT32 Bit,
+ IN volatile VOID *Address
+ );
+
#endif
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index d4007db..dc27c13 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -43,6 +43,7 @@
X64/hypercall.S
X64/hypercall.asm
X64/InterlockedCompareExchange16.c | GCC
+ X64/TestAndClearBit.S
[LibraryClasses]
UefiDriverEntryPoint
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 11/18] OvmfPkg/XenbusDxe: Add XenStore client implementation
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (9 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 10/18] OvmfPkg/XenbusDxe: Add TestAndClearBit Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 12/18] OvmfPkg/XenbusDxe: Add an helper AsciiStrDup Anthony PERARD
` (16 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
XenStore is a key/value database, which is running on another virtual
machine. It can be accessed through shared memory. This is a client
implementation.
Origin: FreeBSD 10.0
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/Include/Protocol/Xenbus.h | 28 +
OvmfPkg/XenbusDxe/XenbusDxe.c | 5 +
OvmfPkg/XenbusDxe/XenbusDxe.inf | 2 +
OvmfPkg/XenbusDxe/Xenstore.c | 1313 +++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/Xenstore.h | 281 ++++++++
5 files changed, 1629 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/Xenstore.c
create mode 100644 OvmfPkg/XenbusDxe/Xenstore.h
diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
index 191cee1..1e0d01d 100644
--- a/OvmfPkg/Include/Protocol/Xenbus.h
+++ b/OvmfPkg/Include/Protocol/Xenbus.h
@@ -21,6 +21,34 @@
///
typedef struct _XENBUS_PROTOCOL XENBUS_PROTOCOL;
+typedef enum xenbus_state XenbusState;
+
+typedef struct
+{
+ UINT32 Id;
+} XENSTORE_TRANSACTION;
+
+#define XST_NIL ((XENSTORE_TRANSACTION) { 0 })
+
+typedef enum {
+ XENSTORE_STATUS_SUCCESS = 0,
+ XENSTORE_STATUS_EINVAL,
+ XENSTORE_STATUS_EACCES,
+ XENSTORE_STATUS_EEXIST,
+ XENSTORE_STATUS_EISDIR,
+ XENSTORE_STATUS_ENOENT,
+ XENSTORE_STATUS_ENOMEM,
+ XENSTORE_STATUS_ENOSPC,
+ XENSTORE_STATUS_EIO,
+ XENSTORE_STATUS_ENOTEMPTY,
+ XENSTORE_STATUS_ENOSYS,
+ XENSTORE_STATUS_EROFS,
+ XENSTORE_STATUS_EBUSY,
+ XENSTORE_STATUS_EAGAIN,
+ XENSTORE_STATUS_EISCONN,
+ XENSTORE_STATUS_E2BIG
+} XENSTORE_STATUS;
+
#include <IndustryStandard/Xen/grant_table.h>
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
index b19055d..2c85d5e 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.c
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -18,6 +18,7 @@
#include "XenHypercall.h"
#include "GrantTable.h"
+#include "Xenstore.h"
///
/// Driver Binding Protocol instance
@@ -330,6 +331,9 @@ XenbusDxeDriverBindingStart (
XenGrantTableInit (Dev, MmioAddr);
+ Status = XenstoreInit (Dev);
+ ASSERT_EFI_ERROR (Status);
+
Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
NotifyExitBoot,
(VOID*) Dev,
@@ -378,6 +382,7 @@ XenbusDxeDriverBindingStop (
XENBUS_DEVICE *Dev = mMyDevice;
gBS->CloseEvent (Dev->ExitBootEvent);
+ // XXX xenstore cleanup
XenGrantTableDeinit (Dev);
gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index dc27c13..6c8260f 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -38,6 +38,8 @@
GrantTable.h
EventChannel.c
EventChannel.h
+ Xenstore.c
+ Xenstore.h
[Sources.X64]
X64/hypercall.S
diff --git a/OvmfPkg/XenbusDxe/Xenstore.c b/OvmfPkg/XenbusDxe/Xenstore.c
new file mode 100644
index 0000000..2252c4f
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/Xenstore.c
@@ -0,0 +1,1313 @@
+/******************************************************************************
+ * xenstore.c
+ *
+ * Low-level kernel interface to the XenStore.
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2009,2010 Spectra Logic Corporation
+ *
+ * This file may be distributed separately from the Linux kernel, or
+ * incorporated into other software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "Xenstore.h"
+
+#include <Library/PrintLib.h>
+
+#include <IndustryStandard/Xen/hvm/params.h>
+
+#include "XenHypercall.h"
+#include "EventChannel.h"
+
+/**
+ * \file xenstore.c
+ * \brief XenStore interface
+ *
+ * The XenStore interface is a simple storage system that is a means of
+ * communicating state and configuration data between the Xen Domain 0
+ * and the various guest domains. All configuration data other than
+ * a small amount of essential information required during the early
+ * boot process of launching a Xen aware guest, is managed using the
+ * XenStore.
+ *
+ * The XenStore is ASCII string based, and has a structure and semantics
+ * similar to a filesystem. There are files and directories, the directories
+ * able to contain files or other directories. The depth of the hierachy
+ * is only limited by the XenStore's maximum path length.
+ *
+ * The communication channel between the XenStore service and other
+ * domains is via two, guest specific, ring buffers in a shared memory
+ * area. One ring buffer is used for communicating in each direction.
+ * The grant table references for this shared memory are given to the
+ * guest either via the xen_start_info structure for a fully para-
+ * virtualized guest, or via HVM hypercalls for a hardware virtualized
+ * guest.
+ *
+ * The XenStore communication relies on an event channel and thus
+ * interrupts. For this reason, the attachment of the XenStore
+ * relies on an interrupt driven configuration hook to hold off
+ * boot processing until communication with the XenStore service
+ * can be established.
+ *
+ * Several Xen services depend on the XenStore, most notably the
+ * XenBus used to discover and manage Xen devices. These services
+ * are implemented as NewBus child attachments to a bus exported
+ * by this XenStore driver.
+ */
+
+/*-------------------------- Private Data Structures ------------------------*/
+
+typedef struct {
+ CONST VOID *Data;
+ UINTN Len;
+} WRITE_REQUEST;
+
+/* Register callback to watch subtree (node) in the XenStore. */
+#define XENSTORE_WATCH_SIGNATURE SIGNATURE_32 ('X','S','w','a')
+struct _XENSTORE_WATCH
+{
+ UINT32 Signature;
+ LIST_ENTRY Link;
+
+ /* Path being watched. */
+ CHAR8 *Node;
+};
+
+#define XENSTORE_WATCH_FROM_LINK(l) \
+ CR (l, XENSTORE_WATCH, Link, XENSTORE_WATCH_SIGNATURE)
+
+
+/**
+ * Structure capturing messages received from the XenStore service.
+ */
+#define XENSTORE_MESSAGE_SIGNATURE SIGNATURE_32 ('X', 'S', 's', 'm')
+typedef struct {
+ UINT32 Signature;
+ LIST_ENTRY Link;
+
+ struct xsd_sockmsg Header;
+
+ union {
+ /* Queued replies. */
+ struct {
+ CHAR8 *Body;
+ } Reply;
+
+ /* Queued watch events. */
+ struct {
+ XENSTORE_WATCH *Handle;
+ CONST CHAR8 **Vector;
+ UINT32 VectorSize;
+ } Watch;
+ } u;
+} XENSTORE_MESSAGE;
+#define XENSTORE_MESSAGE_FROM_LINK(r) \
+ CR (r, XENSTORE_MESSAGE, Link, XENSTORE_MESSAGE_SIGNATURE)
+
+/**
+ * Container for all XenStore related state.
+ */
+typedef struct {
+ /**
+ * Pointer to shared memory communication structures allowing us
+ * to communicate with the XenStore service.
+ */
+ struct xenstore_domain_interface *Xenstore;
+
+ XENBUS_DEVICE *Dev;
+
+ /**
+ * Lock serializing access to ring producer/consumer
+ * indexes. Use of this lock guarantees that wakeups
+ * of blocking readers/writers are not missed due to
+ * races with the XenStore service.
+ */
+ // might need an other way of locking, more ovmf
+ //SPIN_LOCK ring_lock;
+
+ /*
+ * Mutex used to insure exclusive access to the outgoing
+ * communication ring. We use a lock type that can be
+ * held while sleeping so that xs_write() can block waiting
+ * for space in the ring to free up, without allowing another
+ * writer to come in and corrupt a partial message write.
+ */
+ //struct sx request_mutex;
+
+ /**
+ * A list of replies to our requests.
+ *
+ * The reply list is filled by xs_rcv_thread(). It
+ * is consumed by the context that issued the request
+ * to which a reply is made. The requester blocks in
+ * XenstoreReadReply ().
+ *
+ * /note Only one requesting context can be active at a time.
+ * This is guaranteed by the request_mutex and insures
+ * that the requester sees replies matching the order
+ * of its requests.
+ */
+ LIST_ENTRY ReplyList;
+
+ /** Lock protecting the reply list. */
+ EFI_LOCK ReplyLock;
+
+ /**
+ * List of registered watches.
+ */
+ LIST_ENTRY RegisteredWatches;
+
+ /** Lock protecting the registered watches list. */
+ EFI_LOCK RegisteredWatchesLock;
+
+ /**
+ * List of pending watch callback events.
+ */
+ LIST_ENTRY WatchEvents;
+
+ /** Lock protecting the watch calback list. */
+ EFI_LOCK WatchEventsLock;
+
+ /**
+ * The event channel for communicating with the
+ * XenStore service.
+ */
+ UINT32 EventChannel;
+
+ /** Handle for XenStore events. */
+ EFI_EVENT EventChannelEvent;
+} XENSTORE_PRIVATE;
+
+/*-------------------------------- Global Data ------------------------------*/
+static XENSTORE_PRIVATE xs;
+
+/*------------------------- Private Utility Functions -----------------------*/
+/**
+ * Count and optionally record pointers to a number of NUL terminated
+ * strings in a buffer.
+ *
+ * \param strings A pointer to a contiguous buffer of NUL terminated strings.
+ * \param len The length of the buffer pointed to by strings.
+ * \param dest An array to store pointers to each string found in strings.
+ *
+ * \return A count of the number of strings found.
+ */
+STATIC
+UINT32
+ExtractStrings (
+ IN CONST CHAR8 *Strings,
+ IN UINTN Len,
+ OUT CONST CHAR8 **Dst OPTIONAL
+ )
+{
+ UINT32 Num = 0;
+ CONST CHAR8 *Ptr;
+
+ for (Ptr = Strings; Ptr < Strings + Len; Ptr += AsciiStrSize (Ptr)) {
+ if (Dst != NULL) {
+ *Dst++ = Ptr;
+ }
+ Num++;
+ }
+
+ return Num;
+}
+
+/**
+ * Convert a contiguous buffer containing a series of NUL terminated
+ * strings into an array of pointers to strings.
+ *
+ * The returned pointer references the array of string pointers which
+ * is followed by the storage for the string data. It is the client's
+ * responsibility to free this storage.
+ *
+ * The storage addressed by strings is free'd prior to Split returning.
+ *
+ * \param strings A pointer to a contiguous buffer of NUL terminated strings.
+ * \param len The length of the buffer pointed to by strings.
+ * \param num The number of strings found and returned in the strings
+ * array.
+ *
+ * \return An array of pointers to the strings found in the input buffer.
+ */
+STATIC
+CONST CHAR8 **
+Split (
+ IN CHAR8 *Strings,
+ IN UINTN Len,
+ OUT UINT32 *NumPtr
+ )
+{
+ // XXX Why const and later a cast to not const?
+ CONST CHAR8 **Dst;
+
+ ASSERT(NumPtr != NULL);
+ ASSERT(Strings != NULL);
+
+ /* Protect against unterminated buffers. */
+ if (Len > 0) {
+ Strings[Len - 1] = '\0';
+ }
+
+ /* Count the Strings. */
+ *NumPtr = ExtractStrings (Strings, Len, NULL);
+
+ /* Transfer to one big alloc for easy freeing by the caller. */
+ Dst = AllocatePool (*NumPtr * sizeof (CHAR8 *) + Len);
+ // XXX is Dst + *NumPtr better ?
+ CopyMem (&Dst[*NumPtr], Strings, Len);
+ FreePool (Strings);
+
+ /* Extract pointers to newly allocated array. */
+ Strings = (CHAR8 *) &Dst[*NumPtr];
+ ExtractStrings (Strings, Len, Dst);
+
+ return (Dst);
+}
+
+/**
+ * Convert from watch token (unique identifier) to the associated
+ * internal tracking structure for this watch.
+ *
+ * \param tocken The unique identifier for the watch to find.
+ *
+ * \return A pointer to the found watch structure or NULL.
+ */
+STATIC
+XENSTORE_WATCH *
+XenstoreFindWatch (
+ IN CONST CHAR8 *Token
+ )
+{
+ XENSTORE_WATCH *Watch, *WantedWatch;
+ LIST_ENTRY *Entry;
+
+ WantedWatch = (VOID *) AsciiStrHexToUintn (Token);
+
+ if (IsListEmpty (&xs.RegisteredWatches)) {
+ return NULL;
+ }
+ for (Entry = GetFirstNode (&xs.RegisteredWatches);
+ !IsNull (&xs.RegisteredWatches, Entry);
+ Entry = GetNextNode (&xs.RegisteredWatches, Entry)) {
+ Watch = XENSTORE_WATCH_FROM_LINK (Entry);
+ if (Watch == WantedWatch)
+ return Watch;
+ }
+
+ return NULL;
+}
+
+/*------------------------- Public Utility Functions ----------------------*/
+/*------- API comments for these methods can be found in Xenstore.h -------*/
+CHAR8 *
+XenstoreJoin (
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node
+ )
+{
+ CHAR8 *Buf;
+
+ /* +1 for '/' and +1 for '\0' */
+ Buf = AllocateZeroPool (
+ AsciiStrLen (DirectoryPath) + AsciiStrLen (Node) + 2);
+ AsciiStrCat (Buf, DirectoryPath);
+ if (Node[0] != '\0') {
+ AsciiStrCat (Buf, "/");
+ AsciiStrCat (Buf, Node);
+ }
+
+ return Buf;
+}
+
+/*-------------------- Low Level Communication Management --------------------*/
+/**
+ * Verify that the indexes for a ring are valid.
+ *
+ * The difference between the producer and consumer cannot
+ * exceed the size of the ring.
+ *
+ * \param cons The consumer index for the ring to test.
+ * \param prod The producer index for the ring to test.
+ *
+ * \retval 1 If indexes are in range.
+ * \retval 0 If the indexes are out of range.
+ */
+STATIC
+BOOLEAN
+XenstoreCheckIndexes (
+ XENSTORE_RING_IDX Cons,
+ XENSTORE_RING_IDX Prod
+ )
+{
+ return ((Prod - Cons) <= XENSTORE_RING_SIZE);
+}
+
+/**
+ * Return a pointer to, and the length of, the contiguous
+ * free region available for output in a ring buffer.
+ *
+ * \param cons The consumer index for the ring.
+ * \param prod The producer index for the ring.
+ * \param buf The base address of the ring's storage.
+ * \param len The amount of contiguous storage available.
+ *
+ * \return A pointer to the start location of the free region.
+ */
+STATIC
+VOID *
+XenstoreGetOutputChunk (
+ XENSTORE_RING_IDX Cons,
+ XENSTORE_RING_IDX Prod,
+ CHAR8 *Buffer,
+ UINT32 *LenPtr
+ )
+{
+ UINT32 Len;
+ Len = XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(Prod);
+ if ((XENSTORE_RING_SIZE - (Prod - Cons)) < Len) {
+ Len = XENSTORE_RING_SIZE - (Prod - Cons);
+ }
+ *LenPtr = Len;
+ return (Buffer + MASK_XENSTORE_IDX(Prod));
+}
+
+/**
+ * Return a pointer to, and the length of, the contiguous
+ * data available to read from a ring buffer.
+ *
+ * \param cons The consumer index for the ring.
+ * \param prod The producer index for the ring.
+ * \param buf The base address of the ring's storage.
+ * \param len The amount of contiguous data available to read.
+ *
+ * \return A pointer to the start location of the available data.
+ */
+STATIC
+CONST VOID *
+XenstoreGetInputChunk (
+ XENSTORE_RING_IDX Cons,
+ XENSTORE_RING_IDX Prod,
+ CONST CHAR8 *Buffer,
+ UINT32 *LenPtr
+ )
+{
+ UINT32 Len;
+
+ Len = XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(Cons);
+ if ((Prod - Cons) < Len) {
+ Len = Prod - Cons;
+ }
+ *LenPtr = Len;
+ return (Buffer + MASK_XENSTORE_IDX(Cons));
+}
+
+/**
+ * Transmit data to the XenStore service.
+ *
+ * \param tdata A pointer to the contiguous data to send.
+ * \param len The amount of data to send.
+ *
+ * \return On success 0, otherwise an errno value indicating the
+ * cause of failure.
+ *
+ * \invariant Called from thread context.
+ * \invariant The buffer pointed to by tdata is at least len bytes
+ * in length.
+ * \invariant xs.request_mutex exclusively locked.
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreWriteStore (
+ CONST VOID *tdata,
+ UINTN Len
+ )
+{
+ XENSTORE_RING_IDX Cons, Prod;
+ CONST CHAR8 *Data = (CONST CHAR8 *)tdata;
+
+ // an assert that check if it's still a locket mutex
+ // sx_assert(&xs.request_mutex, SX_XLOCKED);
+ while (Len != 0) {
+ void *Dest;
+ UINT32 Available;
+
+ /* Hold lock so we can't miss wakeups should we block. */
+ //AcquireSpinLock(&xs.ring_lock);
+ Cons = xs.Xenstore->req_cons;
+ Prod = xs.Xenstore->req_prod;
+ if ((Prod - Cons) == XENSTORE_RING_SIZE) {
+ /*
+ * Output ring is full. Wait for a ring event.
+ *
+ * Note that the events from both queues are combined, so being woken
+ * does not guarantee that data exist in the read ring.
+ */
+ UINTN Index;
+ EFI_STATUS Status;
+
+ Status = gBS->WaitForEvent (1, &xs.EventChannelEvent, &Index);
+ ASSERT (Status != EFI_INVALID_PARAMETER);
+ if (Status == EFI_UNSUPPORTED) {
+ // see XenstoreReadStore
+ }
+
+ /* Try again. */
+ continue;
+ }
+ //mtx_unlock(&xs.ring_lock);
+
+ /* Verify queue sanity. */
+ if (!XenstoreCheckIndexes (Cons, Prod)) {
+ xs.Xenstore->req_cons = xs.Xenstore->req_prod = 0;
+ return XENSTORE_STATUS_EIO;
+ }
+
+ Dest = XenstoreGetOutputChunk (Cons, Prod, xs.Xenstore->req, &Available);
+ if (Available > Len) {
+ Available = Len;
+ }
+
+ CopyMem (Dest, Data, Available);
+ Data += Available;
+ Len -= Available;
+
+ /*
+ * The store to the producer index, which indicates
+ * to the other side that new data has arrived, must
+ * be visible only after our copy of the data into the
+ * ring has completed.
+ */
+ MemoryFence ();
+ xs.Xenstore->req_prod += Available;
+
+ /*
+ * The other side will see the change to req_prod at the time of the
+ * interrupt.
+ */
+ MemoryFence ();
+ XenEventChannelNotify (xs.Dev, xs.EventChannel);
+ }
+
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+/**
+ * Receive data from the XenStore service.
+ *
+ * \param tdata A pointer to the contiguous buffer to receive the data.
+ * \param len The amount of data to receive.
+ *
+ * \return On success 0, otherwise an errno value indicating the
+ * cause of failure.
+ *
+ * \invariant Called from thread context.
+ * \invariant The buffer pointed to by tdata is at least len bytes
+ * in length.
+ *
+ * \note xs_read does not perform any internal locking to guarantee
+ * serial access to the incoming ring buffer. However, there
+ * is only one context processing reads: xs_rcv_thread().
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreReadStore (
+ OUT VOID *tdata,
+ IN UINTN Len
+ )
+{
+ XENSTORE_RING_IDX Cons, Prod;
+ CHAR8 *Data = (CHAR8 *)tdata;
+
+ while (Len != 0) {
+ UINT32 Available;
+ CONST CHAR8 *Src;
+
+ /* Hold lock so we can't miss wakeups should we block. */
+ //mtx_lock(&xs.ring_lock); //SpinLock ?
+ Cons = xs.Xenstore->rsp_cons;
+ Prod = xs.Xenstore->rsp_prod;
+ if (Cons == Prod) {
+ /*
+ * Nothing to read. Wait for a ring event.
+ *
+ * Note that the events from both queues are combined, so being woken
+ * does not guarantee that data exist in the read ring.
+ */
+ UINTN Index;
+ EFI_STATUS Status;
+
+ Status = gBS->WaitForEvent (1, &xs.EventChannelEvent, &Index);
+ ASSERT (Status != EFI_INVALID_PARAMETER);
+ if (Status == EFI_UNSUPPORTED) {
+ // The current TPL is not TPL_APPLICATION.
+ // could:
+ // this could happen when the device is deinitialized
+ // from the ExitBootServices notification
+ /* OldTpl = EfiGetCurrentTpl(); */
+ /* gBS->RestoreTPL (TPL_APPLICATION); */
+ /* gBS->WaitForEvent (XXX); */
+ /* gBS->RaiseTPL (OldTpl); */
+ }
+ continue;
+ }
+ //mtx_unlock(&xs.ring_lock);
+
+ /* Verify queue sanity. */
+ if (!XenstoreCheckIndexes (Cons, Prod)) {
+ xs.Xenstore->rsp_cons = xs.Xenstore->rsp_prod = 0;
+ return XENSTORE_STATUS_EIO;
+ }
+
+ Src = XenstoreGetInputChunk (Cons, Prod, xs.Xenstore->rsp, &Available);
+ if (Available > Len) {
+ Available = Len;
+ }
+
+ /*
+ * Insure the data we read is related to the indexes
+ * we read above.
+ */
+ MemoryFence ();
+
+ CopyMem (Data, Src, Available);
+ Data += Available;
+ Len -= Available;
+
+ /*
+ * Insure that the producer of this ring does not see
+ * the ring space as free until after we have copied it
+ * out.
+ */
+ MemoryFence ();
+ xs.Xenstore->rsp_cons += Available;
+
+ /*
+ * The producer will see the updated consumer index when the event is
+ * delivered.
+ */
+ MemoryFence ();
+ XenEventChannelNotify (xs.Dev, xs.EventChannel);
+ }
+
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+/*----------------------- Received Message Processing ------------------------*/
+/**
+ * Block reading the next message from the XenStore service and
+ * process the result.
+ *
+ * \param type The returned type of the XenStore message received.
+ *
+ * \return 0 on success. Otherwise an errno value indicating the
+ * type of failure encountered.
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreProcessMessage (
+ VOID
+ )
+{
+ XENSTORE_MESSAGE *Message;
+ CHAR8 *Body;
+ XENSTORE_STATUS Status;
+
+ Message = AllocateZeroPool (sizeof (XENSTORE_MESSAGE));
+ Message->Signature = XENSTORE_MESSAGE_SIGNATURE;
+ Status = XenstoreReadStore (&Message->Header, sizeof (Message->Header));
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ FreePool (Message);
+ DEBUG ((EFI_D_ERROR, "%a %d, error read store %d\n",
+ __func__, __LINE__, Status));
+ return Status;
+ }
+
+ Body = AllocatePool (Message->Header.len + 1);
+ Status = XenstoreReadStore (Body, Message->Header.len);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ FreePool (Body);
+ FreePool (Message);
+ DEBUG ((EFI_D_ERROR, "%a %d, error read store %d\n",
+ __func__, __LINE__, Status));
+ return Status;
+ }
+ Body[Message->Header.len] = '\0';
+
+ if (Message->Header.type == XS_WATCH_EVENT) {
+ Message->u.Watch.Vector = Split(Body, Message->Header.len,
+ &Message->u.Watch.VectorSize);
+
+ EfiAcquireLock (&xs.RegisteredWatchesLock);
+ Message->u.Watch.Handle =
+ XenstoreFindWatch (Message->u.Watch.Vector[XS_WATCH_TOKEN]);
+ DEBUG ((EFI_D_INFO, "%a %d, watch event %a\n", __func__, __LINE__,
+ Message->u.Watch.Vector[XS_WATCH_TOKEN]));
+ if (Message->u.Watch.Handle != NULL) {
+ EfiAcquireLock (&xs.WatchEventsLock);
+ InsertHeadList (&xs.WatchEvents, &Message->Link);
+ EfiReleaseLock (&xs.WatchEventsLock);
+ } else {
+ DEBUG ((EFI_D_WARN, "%a %d, watch handle %a not found\n",
+ __func__, __LINE__, Message->u.Watch.Vector[XS_WATCH_TOKEN]));
+ FreePool(Message->u.Watch.Vector);
+ FreePool(Message);
+ }
+ EfiReleaseLock (&xs.RegisteredWatchesLock);
+ } else {
+ Message->u.Reply.Body = Body;
+ EfiAcquireLock (&xs.ReplyLock);
+ InsertTailList (&xs.ReplyList, &Message->Link);
+ EfiReleaseLock (&xs.ReplyLock);
+ }
+
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+/*---------------- XenStore Message Request/Reply Processing -----------------*/
+
+
+/**
+ * Convert a XenStore error string into an errno number.
+ *
+ * \param errorstring The error string to convert.
+ *
+ * \return The errno best matching the input string.
+ *
+ * \note Unknown error strings are converted to EINVAL.
+ */
+
+typedef struct {
+ XENSTORE_STATUS Status;
+ CONST CHAR8 *ErrorStr;
+} XenstoreErrors;
+
+static XenstoreErrors gXenstoreErrors[] = {
+ { XENSTORE_STATUS_EINVAL, "EINVAL" },
+ { XENSTORE_STATUS_EACCES, "EACCES" },
+ { XENSTORE_STATUS_EEXIST, "EEXIST" },
+ { XENSTORE_STATUS_EISDIR, "EISDIR" },
+ { XENSTORE_STATUS_ENOENT, "ENOENT" },
+ { XENSTORE_STATUS_ENOMEM, "ENOMEM" },
+ { XENSTORE_STATUS_ENOSPC, "ENOSPC" },
+ { XENSTORE_STATUS_EIO, "EIO" },
+ { XENSTORE_STATUS_ENOTEMPTY, "ENOTEMPTY" },
+ { XENSTORE_STATUS_ENOSYS, "ENOSYS" },
+ { XENSTORE_STATUS_EROFS, "EROFS" },
+ { XENSTORE_STATUS_EBUSY, "EBUSY" },
+ { XENSTORE_STATUS_EAGAIN, "EAGAIN" },
+ { XENSTORE_STATUS_EISCONN, "EISCONN" },
+ { XENSTORE_STATUS_E2BIG, "E2BIG" }
+};
+#define ARRAY_SIZE(x) (sizeof(x) / sizeof(x[0]))
+
+STATIC
+XENSTORE_STATUS
+XenstoreGetError (
+ CONST CHAR8 *ErrorStr
+ )
+{
+ UINT32 Index;
+
+ for (Index = 0; Index < ARRAY_SIZE(gXenstoreErrors); Index++) {
+ if (!AsciiStrCmp (ErrorStr, gXenstoreErrors[Index].ErrorStr)) {
+ return gXenstoreErrors[Index].Status;
+ }
+ }
+ DEBUG ((EFI_D_WARN, "Xenstore gave unknown error %a\n", ErrorStr));
+ return XENSTORE_STATUS_EINVAL;
+}
+
+/**
+ * Block waiting for a reply to a message request.
+ *
+ * \param type The returned type of the reply.
+ * \param len The returned body length of the reply.
+ * \param result The returned body of the reply.
+ *
+ * \return 0 on success. Otherwise an errno indicating the
+ * cause of failure.
+ */
+STATIC
+VOID
+XenstoreReadReply (
+ OUT enum xsd_sockmsg_type *type,
+ OUT UINT32 *LenPtr OPTIONAL,
+ OUT VOID **Result
+ )
+{
+ XENSTORE_MESSAGE *Message;
+ LIST_ENTRY *Entry;
+ CHAR8 *Body;
+
+ while (IsListEmpty (&xs.ReplyList)) {
+ XenstoreProcessMessage ();
+ }
+ EfiAcquireLock (&xs.ReplyLock);
+ Entry = GetFirstNode (&xs.ReplyList);
+ Message = XENSTORE_MESSAGE_FROM_LINK (Entry);
+ RemoveEntryList (Entry);
+ EfiReleaseLock (&xs.ReplyLock);
+
+ *type = Message->Header.type;
+ if (LenPtr != NULL) {
+ *LenPtr = Message->Header.len;
+ }
+ Body = Message->u.Reply.Body;
+
+ FreePool (Message);
+ *Result = Body;
+}
+
+/**
+ * Send a message with an optionally muti-part body to the XenStore service.
+ *
+ * \param Transaction The transaction to use for this request.
+ * \param request_type The type of message to send.
+ * \param iovec Pointers to the body sections of the request.
+ * \param NumRequests The number of body sections in the request.
+ * \param LenPtr The returned length of the reply.
+ * \param ResultPtr The returned body of the reply.
+ *
+ * \return 0 on success. Otherwise an errno indicating
+ * the cause of failure.
+ *
+ * \note The returned result is provided in malloced storage and thus
+ * must be free'd by the caller with 'free(*result, M_XENSTORE);
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreTalkv (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN enum xsd_sockmsg_type RequestType,
+ IN CONST WRITE_REQUEST *WriteRequest,
+ IN UINT32 NumRequests,
+ OUT UINT32 *LenPtr OPTIONAL,
+ OUT VOID **ResultPtr OPTIONAL
+ )
+{
+ struct xsd_sockmsg Message;
+ void *Return = NULL;
+ UINT32 Index;
+ XENSTORE_STATUS Status;
+
+ Message.tx_id = Transaction.Id;
+ Message.req_id = 0;
+ Message.type = RequestType;
+ Message.len = 0;
+ for (Index = 0; Index < NumRequests; Index++) {
+ Message.len += WriteRequest[Index].Len;
+ }
+
+ //sx_xlock(&xs.request_mutex);
+ Status = XenstoreWriteStore (&Message, sizeof (Message));
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "XenstoreTalkv failed %d\n", Status));
+ goto ErrorLockHeld;
+ }
+
+ for (Index = 0; Index < NumRequests; Index++) {
+ Status = XenstoreWriteStore (WriteRequest[Index].Data, WriteRequest[Index].Len);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "XenstoreTalkv failed %d\n", Status));
+ goto ErrorLockHeld;
+ }
+ }
+
+ XenstoreReadReply (&Message.type, LenPtr, &Return);
+
+ErrorLockHeld:
+ //sx_xunlock(&xs.request_mutex);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return Status;
+ }
+
+ if (Message.type == XS_ERROR) {
+ Status = XenstoreGetError (Return);
+ FreePool (Return);
+ return Status;
+ }
+
+ /* Reply is either error or an echo of our request message type. */
+ ASSERT (Message.type == RequestType);
+
+ if (ResultPtr) {
+ *ResultPtr = Return;
+ } else {
+ FreePool (Return);
+ }
+
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+/**
+ * Wrapper for XenstoreTalkv allowing easy transmission of a message with
+ * a single, contiguous, message body.
+ *
+ * \param Transaction The transaction to use for this request.
+ * \param request_type The type of message to send.
+ * \param Body The body of the request.
+ * \param len The returned length of the reply.
+ * \param result The returned body of the reply.
+ *
+ * \return 0 on success. Otherwise an errno indicating
+ * the cause of failure.
+ *
+ * \note The returned result is provided in malloced storage and thus
+ * must be free'd by the caller with 'free(*result, M_XENSTORE);
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreSingle (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN enum xsd_sockmsg_type RequestType,
+ IN CONST CHAR8 *Body,
+ OUT UINT32 *LenPtr OPTIONAL,
+ OUT VOID **Result OPTIONAL
+ )
+{
+ WRITE_REQUEST WriteRequest;
+
+ WriteRequest.Data = (VOID *) Body;
+ WriteRequest.Len = AsciiStrSize (Body);
+
+ return XenstoreTalkv (Transaction, RequestType, &WriteRequest, 1,
+ LenPtr, Result);
+}
+
+/*------------------------- XenStore Watch Support ---------------------------*/
+/**
+ * Transmit a watch request to the XenStore service.
+ *
+ * \param path The path in the XenStore to watch.
+ * \param tocken A unique identifier for this watch.
+ *
+ * \return 0 on success. Otherwise an errno indicating the
+ * cause of failure.
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreWatch (
+ CONST CHAR8 *Path,
+ CONST CHAR8 *Token
+ )
+{
+ WRITE_REQUEST WriteRequest[2];
+
+ WriteRequest[0].Data = (VOID *) Path;
+ WriteRequest[0].Len = AsciiStrSize (Path);
+ WriteRequest[1].Data = (VOID *) Token;
+ WriteRequest[1].Len = AsciiStrSize (Token);
+
+ return XenstoreTalkv (XST_NIL, XS_WATCH, WriteRequest, 2, NULL, NULL);
+}
+
+/**
+ * Transmit an uwatch request to the XenStore service.
+ *
+ * \param Path The path in the XenStore to watch.
+ * \param Tocken A unique identifier for this watch.
+ *
+ * \return 0 on success. Otherwise an errno indicating the
+ * cause of failure.
+ */
+STATIC
+XENSTORE_STATUS
+XenstoreUnwatch (
+ CONST CHAR8 *Path,
+ CONST CHAR8 *Token
+ )
+{
+ WRITE_REQUEST WriteRequest[2];
+
+ WriteRequest[0].Data = (VOID *) Path;
+ WriteRequest[0].Len = AsciiStrSize (Path);
+ WriteRequest[1].Data = (VOID *) Token;
+ WriteRequest[1].Len = AsciiStrSize (Token);
+
+ return XenstoreTalkv (XST_NIL, XS_UNWATCH, WriteRequest, 2, NULL, NULL);
+}
+
+VOID
+EFIAPI
+NotifyEventChannelCheckForEvent (
+ IN EFI_EVENT Event,
+ IN VOID *Context
+ )
+{
+ XENSTORE_PRIVATE *xs;
+ xs = (XENSTORE_PRIVATE *)Context;
+ if (TestAndClearBit (xs->EventChannel, xs->Dev->SharedInfo->evtchn_pending)) {
+ gBS->SignalEvent (Event);
+ }
+}
+
+/*----------- XenStore Configuration, Initialization, and Control ------------*/
+/**
+ * Setup communication channels with the XenStore service.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ */
+STATIC
+EFI_STATUS
+XenstoreInitComms (
+ XENSTORE_PRIVATE *xs
+ )
+{
+ EFI_STATUS Status;
+ struct xenstore_domain_interface *Xenstore = xs->Xenstore;
+
+ if (Xenstore->rsp_prod != Xenstore->rsp_cons) {
+ DEBUG ((EFI_D_WARN, "XENSTORE response ring is not quiescent "
+ "(%08x:%08x): fixing up\n",
+ Xenstore->rsp_cons, Xenstore->rsp_prod));
+ Xenstore->rsp_cons = Xenstore->rsp_prod;
+ }
+
+ Status = gBS->CreateEvent (EVT_NOTIFY_WAIT, TPL_NOTIFY,
+ NotifyEventChannelCheckForEvent, xs,
+ &xs->EventChannelEvent);
+ ASSERT_EFI_ERROR (Status);
+
+ return Status;
+}
+
+/*------------------ Private Device Attachment Functions --------------------*/
+/**
+ * Attach to the XenStore.
+ *
+ * This routine also prepares for the probe/attach of drivers that rely
+ * on the XenStore.
+ */
+EFI_STATUS
+XenstoreInit (
+ XENBUS_DEVICE *Dev
+ )
+{
+ EFI_STATUS Status;
+ /**
+ * The HVM guest pseudo-physical frame number. This is Xen's mapping
+ * of the true machine frame number into our "physical address space".
+ */
+ UINTN XenstoreGpfn;
+
+ xs.Dev = Dev;
+
+ xs.EventChannel = XenHypercallHvmGetParam (Dev, HVM_PARAM_STORE_EVTCHN);
+ XenstoreGpfn = XenHypercallHvmGetParam (Dev, HVM_PARAM_STORE_PFN);
+ xs.Xenstore = (VOID *) (XenstoreGpfn << EFI_PAGE_SHIFT);
+ DEBUG ((EFI_D_INFO, "XenbusInit: Xenbus rings @0x%lx, event channel %ld\n",
+ (UINTN) xs.Xenstore, (UINTN) xs.EventChannel));
+
+ InitializeListHead (&xs.ReplyList);
+ InitializeListHead (&xs.WatchEvents);
+ InitializeListHead (&xs.RegisteredWatches);
+
+ //mtx_init(&xs.ring_lock, "ring lock", NULL, MTX_DEF);
+ EfiInitializeLock (&xs.ReplyLock, TPL_NOTIFY);
+ //sx_init(&xs.request_mutex, "xenstore request");
+ EfiInitializeLock (&xs.RegisteredWatchesLock, TPL_NOTIFY);
+ EfiInitializeLock (&xs.WatchEventsLock, TPL_NOTIFY);
+
+ /* Initialize the shared memory rings to talk to xenstored */
+ Status = XenstoreInitComms (&xs);
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ return Status;
+}
+
+/*-------------------------------- Public API -----------------------------*/
+/*------- API comments for these methods can be found in Xenstore.h -------*/
+XENSTORE_STATUS
+XenstoreListDirectory (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ OUT UINT32 *DirectoryCountPtr,
+ OUT CONST CHAR8 ***DirectoryListPtr
+ )
+{
+ CHAR8 *Path;
+ CHAR8 *TempStr;
+ UINT32 Len = 0;
+ XENSTORE_STATUS Status;
+
+ Path = XenstoreJoin (DirectoryPath, Node);
+ Status = XenstoreSingle (Transaction, XS_DIRECTORY, Path, &Len,
+ (VOID **) &TempStr);
+ FreePool (Path);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return Status;
+ }
+
+ *DirectoryListPtr = Split (TempStr, Len, DirectoryCountPtr);
+
+ return (0);
+}
+
+BOOLEAN
+XenstorePathExists (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Directory,
+ IN CONST CHAR8 *Node
+ )
+{
+ CONST CHAR8 **TempStr;
+ XENSTORE_STATUS Status;
+ UINT32 TempNum;
+
+ Status = XenstoreListDirectory (Transaction, Directory, Node,
+ &TempNum, &TempStr);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return FALSE;
+ }
+ FreePool (TempStr);
+ return TRUE;
+}
+
+XENSTORE_STATUS
+XenstoreRead (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ OUT UINT32 *LenPtr OPTIONAL,
+ OUT VOID **Result
+ )
+{
+ CHAR8 *Path;
+ VOID *Value;
+ XENSTORE_STATUS Status;
+
+ Path = XenstoreJoin (DirectoryPath, Node);
+ Status = XenstoreSingle (Transaction, XS_READ, Path, LenPtr, &Value);
+ FreePool (Path);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return Status;
+ }
+
+ *Result = Value;
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+XENSTORE_STATUS
+XenstoreWrite (
+ XENSTORE_TRANSACTION Transaction,
+ CONST CHAR8 *DirectoryPath,
+ CONST CHAR8 *Node,
+ CONST CHAR8 *Str
+ )
+{
+ CHAR8 *Path;
+ WRITE_REQUEST WriteRequest[2];
+ XENSTORE_STATUS Status;
+
+ Path = XenstoreJoin (DirectoryPath, Node);
+
+ WriteRequest[0].Data = (VOID *) Path;
+ WriteRequest[0].Len = AsciiStrSize (Path);
+ WriteRequest[1].Data = (VOID *) Str;
+ WriteRequest[1].Len = AsciiStrLen (Str);
+
+ Status = XenstoreTalkv (Transaction, XS_WRITE, WriteRequest, 2, NULL, NULL);
+ FreePool (Path);
+
+ return Status;
+}
+
+XENSTORE_STATUS
+XenstoreRemove (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node
+ )
+{
+ CHAR8 *Path;
+ XENSTORE_STATUS Status;
+
+ Path = XenstoreJoin (DirectoryPath, Node);
+ Status = XenstoreSingle (Transaction, XS_RM, Path, NULL, NULL);
+ FreePool (Path);
+
+ return Status;
+}
+
+XENSTORE_STATUS
+XenstoreTransactionStart (
+ OUT XENSTORE_TRANSACTION *Transaction
+ )
+{
+ CHAR8 *IdStr;
+ XENSTORE_STATUS Status;
+
+ Status = XenstoreSingle (XST_NIL, XS_TRANSACTION_START, "", NULL,
+ (VOID **) &IdStr);
+ if (Status == XENSTORE_STATUS_SUCCESS) {
+ Transaction->Id = AsciiStrDecimalToUintn (IdStr);
+ FreePool (IdStr);
+ }
+
+ return Status;
+}
+
+XENSTORE_STATUS
+XenstoreTransactionEnd (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN BOOLEAN Abort
+ )
+{
+ CHAR8 AbortStr[2];
+
+ if (Abort) {
+ AsciiStrCpy (AbortStr, "F");
+ } else {
+ AsciiStrCpy (AbortStr, "T");
+ }
+
+ return XenstoreSingle (Transaction, XS_TRANSACTION_END, AbortStr, NULL, NULL);
+}
+
+XENSTORE_STATUS
+XenstoreVSPrint (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ IN CONST CHAR8 *FormatString,
+ IN VA_LIST Marker
+ )
+{
+ CHAR8 *Buf;
+ XENSTORE_STATUS Status;
+ UINTN BufSize;
+
+ BufSize = SPrintLengthAsciiFormat (FormatString, Marker) + 1;
+ Buf = AllocateZeroPool (BufSize);
+ AsciiVSPrint (Buf, BufSize, FormatString, Marker);
+ Status = XenstoreWrite (Transaction, DirectoryPath, Node, Buf);
+ FreePool (Buf);
+
+ return Status;
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenstoreSPrint (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ IN CONST CHAR8 *FormatString,
+ ...
+ )
+{
+ VA_LIST Marker;
+ XENSTORE_STATUS Status;
+
+ VA_START (Marker, FormatString);
+ Status = XenstoreVSPrint (Transaction, DirectoryPath, Node, FormatString, Marker);
+ VA_END (Marker);
+
+ return Status;
+}
+
+XENSTORE_STATUS
+XenstoreRegisterWatch (
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ OUT XENSTORE_WATCH **WatchPtr
+ )
+{
+ /* Pointer in ascii is the token. */
+ CHAR8 Token[sizeof (XENSTORE_WATCH) * 2 + 1];
+ XENSTORE_STATUS Status;
+ XENSTORE_WATCH *Watch;
+
+ Watch = AllocateZeroPool (sizeof (XENSTORE_WATCH));
+ Watch->Signature = XENSTORE_WATCH_SIGNATURE;
+ Watch->Node = XenstoreJoin (DirectoryPath, Node);
+
+ EfiAcquireLock (&xs.RegisteredWatchesLock);
+ InsertTailList (&xs.RegisteredWatches, &Watch->Link);
+ EfiReleaseLock (&xs.RegisteredWatchesLock);
+
+ AsciiSPrint (Token, sizeof (Token), "%p", (VOID*) Watch);
+ Status = XenstoreWatch (Watch->Node, Token);
+
+ /* Ignore errors due to multiple registration. */
+ if (Status == XENSTORE_STATUS_EEXIST) {
+ Status = XENSTORE_STATUS_SUCCESS;
+ }
+
+ if (Status == XENSTORE_STATUS_SUCCESS) {
+ *WatchPtr = Watch;
+ } else {
+ EfiAcquireLock (&xs.RegisteredWatchesLock);
+ RemoveEntryList (&Watch->Link);
+ EfiReleaseLock (&xs.RegisteredWatchesLock);
+ FreePool (Watch->Node);
+ FreePool (Watch);
+ }
+
+ return Status;
+}
+
+VOID
+XenstoreUnregisterWatch (
+ IN XENSTORE_WATCH *Watch
+ )
+{
+ CHAR8 Token[sizeof (Watch) * 2 + 1];
+ XENSTORE_STATUS Status;
+ LIST_ENTRY *Entry;
+
+ ASSERT (Watch->Signature == XENSTORE_WATCH_SIGNATURE);
+
+ AsciiSPrint (Token, sizeof (Token), "%p", (VOID *) Watch);
+ // XXX do I need a lock ?
+ if (XenstoreFindWatch (Token) == NULL) {
+ return;
+ }
+ // XXX need cleaning (remove from two list, free everything)
+
+ EfiAcquireLock (&xs.RegisteredWatchesLock);
+ RemoveEntryList (&Watch->Link);
+ EfiReleaseLock (&xs.RegisteredWatchesLock);
+
+ Status = XenstoreUnwatch (Watch->Node, Token);
+
+ /* Cancel pending watch events. */
+ EfiAcquireLock (&xs.WatchEventsLock);
+ Entry = GetFirstNode (&xs.WatchEvents);
+ while (!IsNull (&xs.WatchEvents, Entry)) {
+ XENSTORE_MESSAGE *Message = XENSTORE_MESSAGE_FROM_LINK (Entry);
+ Entry = GetNextNode (&xs.WatchEvents, Entry);
+ if (Message->u.Watch.Handle == Watch) {
+ RemoveEntryList (&Message->Link);
+ FreePool (Message->u.Watch.Vector);
+ FreePool (Message);
+ }
+ }
+ EfiReleaseLock (&xs.WatchEventsLock);
+
+ FreePool (Watch->Node);
+ FreePool (Watch);
+}
diff --git a/OvmfPkg/XenbusDxe/Xenstore.h b/OvmfPkg/XenbusDxe/Xenstore.h
new file mode 100644
index 0000000..3a056c0
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/Xenstore.h
@@ -0,0 +1,281 @@
+/******************************************************************************
+ * Xenstore.h
+ *
+ * Method declarations and structures for accessing the Xenstore
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2005 XenSource Ltd.
+ * Copyright (C) 2009,2010 Spectra Logic Corporation
+ *
+ * This file may be distributed separately from the Linux kernel, or
+ * incorporated into other software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * $FreeBSD: release/10.0.0/sys/xen/xenstore/xenstorevar.h 255040 2013-08-29 19:52:18Z gibbs $
+ */
+
+#ifndef _XEN_XENSTORE_XENSTOREVAR_H
+#define _XEN_XENSTORE_XENSTOREVAR_H
+
+#include "XenbusDxe.h"
+
+#include <IndustryStandard/Xen/io/xs_wire.h>
+
+typedef struct _XENSTORE_WATCH XENSTORE_WATCH;
+
+/**
+ * Fetch the contents of a directory in the XenStore.
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the path to read.
+ * \param node The basename of the path to read.
+ * \param num The returned number of directory entries.
+ * \param result An array of directory entry strings.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ *
+ * \note The results buffer is malloced and should be free'd by the
+ * caller with 'free(*result, M_XENSTORE)'.
+ */
+XENSTORE_STATUS
+XenstoreListDirectory (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ OUT UINT32 *DirectoryCountPtr,
+ OUT CONST CHAR8 ***DirectoryListPtr
+ );
+
+/**
+ * Determine if a path exists in the XenStore.
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the path to read.
+ * \param node The basename of the path to read.
+ *
+ * \retval 1 The path exists.
+ * \retval 0 The path does not exist or an error occurred attempting
+ * to make that determination.
+ */
+BOOLEAN
+XenstorePathExists (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Directory,
+ IN CONST CHAR8 *Node
+ );
+
+/**
+ * Get the contents of a single "file". Returns the contents in
+ * *result which should be freed with free(*result, M_XENSTORE) after
+ * use. The length of the value in bytes is returned in *len.
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the file to read.
+ * \param node The basename of the file to read.
+ * \param len The amount of data read.
+ * \param result The returned contents from this file.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ *
+ * \note The results buffer is malloced and should be free'd by the
+ * caller with 'free(*result, M_XENSTORE)'.
+ */
+XENSTORE_STATUS
+XenstoreRead (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ OUT UINT32 *LenPtr OPTIONAL,
+ OUT VOID **Result
+ );
+
+/**
+ * Write to a single file.
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the file to write.
+ * \param node The basename of the file to write.
+ * \param string The NUL terminated string of data to write.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ */
+XENSTORE_STATUS
+XenstoreWrite (
+ XENSTORE_TRANSACTION Transaction,
+ CONST CHAR8 *DirectoryPath,
+ CONST CHAR8 *Node,
+ CONST CHAR8 *Str
+ );
+
+/**
+ * Remove a file or directory (directories must be empty).
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the directory to remove.
+ * \param node The basename of the directory to remove.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ */
+XENSTORE_STATUS
+XenstoreRemove (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node
+ );
+
+/**
+ * Start a transaction.
+ *
+ * Changes by others will not be seen during the lifetime of this
+ * transaction, and changes will not be visible to others until it
+ * is committed (xs_transaction_end).
+ *
+ * \param Transaction The returned transaction.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ */
+XENSTORE_STATUS
+XenstoreTransactionStart (
+ OUT XENSTORE_TRANSACTION *Transaction
+ );
+
+/**
+ * End a transaction.
+ *
+ * \param Transaction The transaction to end/commit.
+ * \param abort If non-zero, the transaction is discarded
+ * instead of committed.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ */
+XENSTORE_STATUS
+XenstoreTransactionEnd (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN BOOLEAN Abort
+ );
+
+/**
+ * Printf formatted write to a XenStore file.
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the path to read.
+ * \param node The basename of the path to read.
+ * \param fmt Printf format string followed by a variable number of
+ * printf arguments.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of write failure.
+ */
+XENSTORE_STATUS
+EFIAPI
+XenstoreSPrint (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *dir,
+ IN CONST CHAR8 *node,
+ IN CONST CHAR8 *FormatString,
+ ...
+ );
+
+/**
+ * va_list version of xenbus_printf().
+ *
+ * \param Transaction The XenStore transaction covering this request.
+ * \param dir The dirname of the path to read.
+ * \param node The basename of the path to read.
+ * \param fmt Printf format string.
+ * \param ap Va_list of printf arguments.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of write failure.
+ */
+XENSTORE_STATUS
+XenstoreVSPrint (
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ IN CONST CHAR8 *FormatString,
+ IN VA_LIST Marker
+ );
+
+/**
+ * Register a XenStore watch.
+ *
+ * XenStore watches allow a client to be notified via a callback (embedded
+ * within the watch object) of changes to an object in the XenStore.
+ *
+ * \param watch An XENSTORE_WATCH struct with it's node and callback fields
+ * properly initialized.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of write failure. EEXIST errors from the XenStore
+ * are supressed, allowing multiple, physically different,
+ * xenbus_watch objects, to watch the same path in the XenStore.
+ */
+XENSTORE_STATUS
+XenstoreRegisterWatch (
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ OUT XENSTORE_WATCH **WatchPtr
+ );
+
+/**
+ * Unregister a XenStore watch.
+ *
+ * \param watch An XENSTORE_WATCH object previously used in a successful call
+ * to XenstoreRegisterWatch ().
+ *
+ * The XENSTORE_WATCH object's node field is not altered by this call.
+ * It is the caller's responsibility to properly dispose of both the
+ * watch object and the data pointed to by watch->node.
+ */
+VOID
+XenstoreUnregisterWatch (
+ IN XENSTORE_WATCH *Watch
+ );
+
+/**
+ * Allocate and return an sbuf containing the XenStore path string
+ * <dir>/<name>. If name is the NUL string, the returned sbuf contains
+ * the path string <dir>.
+ *
+ * \param dir The NUL terminated directory prefix for new path.
+ * \param name The NUL terminated basename for the new path.
+ *
+ * \return A buffer containing the joined path.
+ */
+CHAR8 *
+XenstoreJoin (
+ CONST CHAR8 *DirectoryPath,
+ CONST CHAR8 *Node
+ );
+
+
+EFI_STATUS
+XenstoreInit (
+ XENBUS_DEVICE *Dev
+ );
+
+#endif /* _XEN_XENSTORE_XENSTOREVAR_H */
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 12/18] OvmfPkg/XenbusDxe: Add an helper AsciiStrDup.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (10 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 11/18] OvmfPkg/XenbusDxe: Add XenStore client implementation Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 13/18] OvmfPkg/XenbusDxe: Add Xenstore function into the Xenbus protocol Anthony PERARD
` (15 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenbusDxe/Helpers.c | 9 +++++++++
OvmfPkg/XenbusDxe/XenbusDxe.h | 5 +++++
2 files changed, 14 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/Helpers.c
diff --git a/OvmfPkg/XenbusDxe/Helpers.c b/OvmfPkg/XenbusDxe/Helpers.c
new file mode 100644
index 0000000..4ea63db
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/Helpers.c
@@ -0,0 +1,9 @@
+#include "XenbusDxe.h"
+
+CHAR8*
+AsciiStrDup (
+ IN CONST CHAR8* Str
+ )
+{
+ return AllocateCopyPool (AsciiStrSize (Str), Str);
+}
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index 50d0925..4d9db72 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -105,4 +105,9 @@ TestAndClearBit (
IN volatile VOID *Address
);
+CHAR8*
+AsciiStrDup (
+ IN CONST CHAR8* Str
+ );
+
#endif
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 13/18] OvmfPkg/XenbusDxe: Add Xenstore function into the Xenbus protocol
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (11 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 12/18] OvmfPkg/XenbusDxe: Add an helper AsciiStrDup Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself Anthony PERARD
` (14 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/Include/Protocol/Xenbus.h | 105 +++++++++++++++++++++++++
OvmfPkg/XenbusDxe/Xenstore.c | 159 ++++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/Xenstore.h | 86 +++++++++++++++++++++
3 files changed, 350 insertions(+)
diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
index 1e0d01d..ef4d0c2 100644
--- a/OvmfPkg/Include/Protocol/Xenbus.h
+++ b/OvmfPkg/Include/Protocol/Xenbus.h
@@ -56,6 +56,62 @@ typedef enum {
/// Function prototypes
///
+/*
+ * Return error from xs_read
+ * Return value of the node ${This->Node}/${node}
+ */
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_XS_READ)(
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Result
+ );
+
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_XS_BACKEND_READ)(
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Result
+ );
+
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_XS_PRINTF) (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Directory,
+ IN CONST CHAR8 *Node,
+ IN CONST CHAR8 *Format,
+ ...
+ );
+
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_XS_REMOVE) (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node
+ );
+
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_XS_TRANSACTION_START)(
+ IN XENBUS_PROTOCOL *This,
+ OUT XENSTORE_TRANSACTION *Transaction
+ );
+
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_XS_TRANSACTION_END) (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN BOOLEAN Abort
+ );
+
typedef
EFI_STATUS
(EFIAPI *XENBUS_GRANT_ACCESS)(
@@ -73,16 +129,65 @@ EFI_STATUS
IN grant_ref_t Ref
);
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_REGISTER_WATCH) (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ // XENSTORE_WATCH ?
+ OUT VOID **Token
+ );
+
+typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_REGISTER_WATCH_BACKEND) (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Token
+ );
+
+typedef
+VOID
+(EFIAPI *XENBUS_UNREGISTER_WATCH) (
+ IN XENBUS_PROTOCOL *This,
+ IN VOID *Token
+ );
+
+typedef
+VOID
+(EFIAPI *XENBUS_WAIT_FOR_WATCH) (
+ IN XENBUS_PROTOCOL *This,
+ IN VOID *Token
+ );
+
///
/// Protocol structure
///
struct _XENBUS_PROTOCOL {
+ XENBUS_XS_READ XsRead;
+ XENBUS_XS_BACKEND_READ XsBackendRead;
+ XENBUS_XS_PRINTF XsPrintf;
+ XENBUS_XS_REMOVE XsRemove;
+ XENBUS_XS_TRANSACTION_START XsTransactionStart;
+ XENBUS_XS_TRANSACTION_END XsTransactionEnd;
+
XENBUS_GRANT_ACCESS GrantAccess;
XENBUS_GRANT_END_ACCESS GrantEndAccess;
+
+ XENBUS_REGISTER_WATCH RegisterWatch;
+ XENBUS_REGISTER_WATCH_BACKEND RegisterWatchBackend;
+ XENBUS_UNREGISTER_WATCH UnregisterWatch;
+ XENBUS_WAIT_FOR_WATCH WaitForWatch;
//
// Place protocol data fields here
//
+ CONST CHAR8 *Type;
+ // helper, since the value is in Node, avoid calculating twice,
+ // in XenbusDxe and in XenPvBlkDxe
+ UINT16 DeviceId;
+ CONST CHAR8 *Node;
+ CONST CHAR8 *Backend;
};
extern EFI_GUID gXenbusProtocolGuid;
diff --git a/OvmfPkg/XenbusDxe/Xenstore.c b/OvmfPkg/XenbusDxe/Xenstore.c
index 2252c4f..6d92fc4 100644
--- a/OvmfPkg/XenbusDxe/Xenstore.c
+++ b/OvmfPkg/XenbusDxe/Xenstore.c
@@ -946,6 +946,42 @@ XenstoreUnwatch (
return XenstoreTalkv (XST_NIL, XS_UNWATCH, WriteRequest, 2, NULL, NULL);
}
+STATIC
+VOID
+XenstoreWaitWatch (
+ VOID *Token
+ )
+{
+ XENSTORE_MESSAGE *Message;
+ LIST_ENTRY *Entry = NULL;
+ LIST_ENTRY *Last = NULL;
+
+ while (TRUE) {
+ EfiAcquireLock (&xs.WatchEventsLock);
+ if (IsListEmpty (&xs.WatchEvents) ||
+ Last == GetFirstNode (&xs.WatchEvents)) {
+ EfiReleaseLock (&xs.WatchEventsLock);
+ XenstoreProcessMessage ();
+ continue;
+ }
+
+ for (Entry = GetFirstNode (&xs.WatchEvents);
+ Entry != Last && !IsNull (&xs.WatchEvents, Entry);
+ Entry = GetNextNode (&xs.WatchEvents, Entry)) {
+ Message = XENSTORE_MESSAGE_FROM_LINK (Entry);
+ if (Message->u.Watch.Handle == Token) {
+ RemoveEntryList (Entry);
+ EfiReleaseLock (&xs.WatchEventsLock);
+ FreePool(Message->u.Watch.Vector);
+ FreePool(Message);
+ return;
+ }
+ }
+ Last = GetFirstNode (&xs.WatchEvents);
+ EfiReleaseLock (&xs.WatchEventsLock);
+ }
+}
+
VOID
EFIAPI
NotifyEventChannelCheckForEvent (
@@ -1311,3 +1347,126 @@ XenstoreUnregisterWatch (
FreePool (Watch->Node);
FreePool (Watch);
}
+
+
+/*
+ * XENBUS
+ */
+VOID
+EFIAPI
+XenbusWaitForWatch (
+ IN XENBUS_PROTOCOL *This,
+ IN VOID *Token
+ )
+{
+ XenstoreWaitWatch (Token);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreRead (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Value
+ )
+{
+ return XenstoreRead (Transaction, This->Node, Node, NULL, Value);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreBackendRead (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Value
+ )
+{
+ return XenstoreRead (Transaction, This->Backend, Node, NULL, Value);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreRemove (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN const char *Node
+ )
+{
+ return XenstoreRemove (Transaction, This->Node, Node);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreTransactionStart (
+ IN XENBUS_PROTOCOL *This,
+ OUT XENSTORE_TRANSACTION *Transaction
+ )
+{
+ return XenstoreTransactionStart (Transaction);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreTransactionEnd (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN BOOLEAN Abort
+ )
+{
+ return XenstoreTransactionEnd (Transaction, Abort);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreSPrint (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ IN CONST CHAR8 *FormatString,
+ ...
+ )
+{
+ VA_LIST Marker;
+ XENSTORE_STATUS Status;
+
+ VA_START (Marker, FormatString);
+ Status = XenstoreVSPrint (Transaction, DirectoryPath, Node, FormatString, Marker);
+ VA_END (Marker);
+
+ return Status;
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusRegisterWatch (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Token
+ )
+{
+ return XenstoreRegisterWatch (This->Node, Node, (XENSTORE_WATCH **) Token);
+}
+
+XENSTORE_STATUS
+EFIAPI
+XenbusRegisterWatchBackend (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Token
+ )
+{
+ return XenstoreRegisterWatch (This->Backend, Node, (XENSTORE_WATCH **) Token);
+}
+
+VOID
+EFIAPI
+XenbusUnregisterWatch (
+ IN XENBUS_PROTOCOL *This,
+ IN VOID *Token
+ )
+{
+ XenstoreUnregisterWatch ((XENSTORE_WATCH *) Token);
+}
diff --git a/OvmfPkg/XenbusDxe/Xenstore.h b/OvmfPkg/XenbusDxe/Xenstore.h
index 3a056c0..276ad7a 100644
--- a/OvmfPkg/XenbusDxe/Xenstore.h
+++ b/OvmfPkg/XenbusDxe/Xenstore.h
@@ -278,4 +278,90 @@ XenstoreInit (
XENBUS_DEVICE *Dev
);
+/*
+ * XENBUS
+ */
+
+VOID
+EFIAPI
+XenbusWaitForWatch (
+ IN XENBUS_PROTOCOL *This,
+ IN VOID *Token
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreRead (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Value
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreBackendRead (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Value
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreRemove (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *Node
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreTransactionStart (
+ IN XENBUS_PROTOCOL *This,
+ OUT XENSTORE_TRANSACTION *Transaction
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreTransactionEnd (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN BOOLEAN Abort
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusXenstoreSPrint (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN CONST CHAR8 *DirectoryPath,
+ IN CONST CHAR8 *Node,
+ IN CONST CHAR8 *FormatString,
+ ...
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusRegisterWatch (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Token
+ );
+
+XENSTORE_STATUS
+EFIAPI
+XenbusRegisterWatchBackend (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ OUT VOID **Token
+ );
+
+VOID
+EFIAPI
+XenbusUnregisterWatch (
+ IN XENBUS_PROTOCOL *This,
+ IN VOID *Token
+ );
+
#endif /* _XEN_XENSTORE_XENSTOREVAR_H */
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (12 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 13/18] OvmfPkg/XenbusDxe: Add Xenstore function into the Xenbus protocol Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 15/18] OvmfPkg/XenbusDxe: Add Event Channel into XenBus protocol Anthony PERARD
` (13 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
This is a bus-like on top of Xenstore. It will look for advertised
ParaVirtualized devices and initialize them by producing XenBus
protocol.
Origin: FreeBSD 10.0
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/Include/Protocol/Xenbus.h | 9 +
OvmfPkg/XenbusDxe/XenBus.c | 360 ++++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/XenBus.h | 76 ++++++++
OvmfPkg/XenbusDxe/XenbusDxe.c | 61 +++++++
OvmfPkg/XenbusDxe/XenbusDxe.h | 18 ++
OvmfPkg/XenbusDxe/XenbusDxe.inf | 3 +
6 files changed, 527 insertions(+)
create mode 100644 OvmfPkg/XenbusDxe/XenBus.c
create mode 100644 OvmfPkg/XenbusDxe/XenBus.h
diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
index ef4d0c2..25408c1 100644
--- a/OvmfPkg/Include/Protocol/Xenbus.h
+++ b/OvmfPkg/Include/Protocol/Xenbus.h
@@ -113,6 +113,14 @@ XENSTORE_STATUS
);
typedef
+XENSTORE_STATUS
+(EFIAPI *XENBUS_SET_STATE)(
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN XenbusState State
+ );
+
+typedef
EFI_STATUS
(EFIAPI *XENBUS_GRANT_ACCESS)(
IN XENBUS_PROTOCOL *This,
@@ -171,6 +179,7 @@ struct _XENBUS_PROTOCOL {
XENBUS_XS_REMOVE XsRemove;
XENBUS_XS_TRANSACTION_START XsTransactionStart;
XENBUS_XS_TRANSACTION_END XsTransactionEnd;
+ XENBUS_SET_STATE SetState;
XENBUS_GRANT_ACCESS GrantAccess;
XENBUS_GRANT_END_ACCESS GrantEndAccess;
diff --git a/OvmfPkg/XenbusDxe/XenBus.c b/OvmfPkg/XenbusDxe/XenBus.c
new file mode 100644
index 0000000..b0cf1ba
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenBus.c
@@ -0,0 +1,360 @@
+/******************************************************************************
+ * Copyright (C) 2010 Spectra Logic Corporation
+ * Copyright (C) 2008 Doug Rabson
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2005 Mike Wray, Hewlett-Packard
+ * Copyright (C) 2005 XenSource Ltd
+ *
+ * This file may be distributed separately from the Linux kernel, or
+ * incorporated into other software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "XenbusDxe.h"
+
+#include <Library/PrintLib.h>
+
+#include "XenBus.h"
+#include "GrantTable.h"
+#include "Xenstore.h"
+#include "EventChannel.h"
+
+#include <IndustryStandard/Xen/io/xenbus.h>
+
+STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData;
+
+STATIC XENBUS_DEVICE_PATH gXenbusDevicePathTemplate = {
+ .Vendor.Header.Type = HARDWARE_DEVICE_PATH,
+ .Vendor.Header.SubType = HW_VENDOR_DP,
+ .Vendor.Header.Length[0] = (UINT8) sizeof (XENBUS_DEVICE_PATH),
+ .Vendor.Header.Length[1] = (UINT8) (sizeof (XENBUS_DEVICE_PATH) >> 8),
+ .Vendor.Guid = XENBUS_PROTOCOL_GUID,
+ .Type = 0,
+ .DeviceId = 0
+};
+
+
+/**
+ * Search our internal record of configured devices (not the XenStore)
+ * to determine if the XenBus device indicated by \a node is known to
+ * the system.
+ *
+ * \param Dev The XenBus bus instance to search for device children.
+ * \param node The XenStore node path for the device to find.
+ *
+ * \return The device_t of the found device if any, or NULL.
+ *
+ * \note device_t is a pointer type, so it can be compared against
+ * NULL for validity.
+ */
+STATIC
+XENBUS_PRIVATE_DATA *
+XenbusDeviceInitialized (
+ IN XENBUS_DEVICE *Dev,
+ IN CONST CHAR8 *Node
+ )
+{
+ LIST_ENTRY *Entry;
+ XENBUS_PRIVATE_DATA *Child;
+ XENBUS_PRIVATE_DATA *Result;
+
+ if (IsListEmpty (&Dev->ChildList)) {
+ return NULL;
+ }
+
+ Result = NULL;
+ for (Entry = GetFirstNode (&Dev->ChildList);
+ !IsNodeAtEnd (&Dev->ChildList, Entry);
+ Entry = GetNextNode (&Dev->ChildList, Entry)) {
+ Child = XENBUS_PRIVATE_DATA_FROM_LINK (Entry);
+ if (!AsciiStrCmp (Child->XenbusIo.Node, Node)) {
+ Result = Child;
+ break;
+ }
+ }
+
+ return (Result);
+}
+
+STATIC
+XenbusState
+XenbusReadDriverState (
+ IN CONST CHAR8 *Path
+ )
+{
+ XenbusState State;
+ CHAR8 *Ptr;
+ XENSTORE_STATUS Status;
+
+ Status = XenstoreRead (XST_NIL, Path, "state", NULL, (VOID **)&Ptr);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ State = XenbusStateClosed;
+ } else {
+ State = AsciiStrDecimalToUintn (Ptr);
+ FreePool (Ptr);
+ }
+
+ return State;
+}
+
+STATIC
+EFI_STATUS
+XenbusAddDevice (
+ XENBUS_DEVICE *Dev,
+ CONST CHAR8 *Type,
+ CONST CHAR8 *Id)
+{
+ CHAR8 DevicePath[XENSTORE_ABS_PATH_MAX];
+ XENSTORE_STATUS StatusXenstore;
+ XENBUS_PRIVATE_DATA *Private;
+ EFI_STATUS Status;
+ XENBUS_DEVICE_PATH *TempXenbusPath;
+ VOID *ChildPciIo;
+
+ AsciiSPrint (DevicePath, sizeof (DevicePath),
+ XENBUS_XENSTORE_NODE "/%a/%a", Type, Id);
+
+ Status = EFI_NOT_STARTED; //ENXIO; // Device not configured
+
+ if (XenstorePathExists (XST_NIL, DevicePath, "")) {
+ XENBUS_PRIVATE_DATA *Child;
+ enum xenbus_state State;
+ CHAR8 *BackendPath;
+
+ Child = XenbusDeviceInitialized (Dev, DevicePath);
+ if (Child != NULL) {
+ /*
+ * We are already tracking this node
+ */
+ Status = EFI_SUCCESS;
+ goto out;
+ }
+
+ State = XenbusReadDriverState (DevicePath);
+ if (State != XenbusStateInitialising) {
+ /*
+ * Device is not new, so ignore it. This can
+ * happen if a device is going away after
+ * switching to Closed.
+ */
+ DEBUG ((EFI_D_INFO, "Xenbus: Device %a ignored. "
+ "State %d\n", DevicePath, State));
+ Status = EFI_SUCCESS;
+ goto out;
+ }
+
+ StatusXenstore = XenstoreRead (XST_NIL, DevicePath, "backend",
+ NULL, (VOID **) &BackendPath);
+ if (StatusXenstore != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "xenbus: %a no backend path.\n", DevicePath));
+ Status = EFI_NOT_FOUND;
+ goto out;
+ }
+
+ Private = AllocateCopyPool (sizeof *Private, &gXenbusPrivateData);
+ InsertTailList (&Dev->ChildList, &Private->Link);
+
+ Private->XenbusIo.Type = AsciiStrDup (Type);
+ Private->XenbusIo.Node = AsciiStrDup (DevicePath);
+ Private->XenbusIo.Backend = BackendPath;
+ Private->XenbusIo.DeviceId = AsciiStrDecimalToUintn (Id);
+ Private->Dev = Dev;
+
+ TempXenbusPath = AllocateCopyPool (sizeof (XENBUS_DEVICE_PATH),
+ &gXenbusDevicePathTemplate);
+ if (!AsciiStrCmp (Private->XenbusIo.Type, "vbd")) {
+ TempXenbusPath->Type = XENBUS_DEVICE_PATH_TYPE_VBD;
+ }
+ TempXenbusPath->DeviceId = Private->XenbusIo.DeviceId;
+ Private->DevicePath = (XENBUS_DEVICE_PATH *)AppendDevicePathNode (
+ Dev->DevicePath,
+ &TempXenbusPath->Vendor.Header);
+ FreePool (TempXenbusPath);
+
+ Status = gBS->InstallMultipleProtocolInterfaces (
+ &Private->Handle,
+ &gEfiDevicePathProtocolGuid, Private->DevicePath,
+ &gXenbusProtocolGuid, &Private->XenbusIo,
+ NULL);
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ Status = gBS->OpenProtocol (Dev->ControllerHandle,
+ &gEfiPciIoProtocolGuid,
+ &ChildPciIo, Dev->This->DriverBindingHandle,
+ Private->Handle,
+ EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER);
+ if (EFI_ERROR (Status)) {
+ DEBUG ((EFI_D_ERROR, "open by child controller fail (%r)\n",
+ Status));
+ }
+ } else {
+ DEBUG ((EFI_D_INFO, "Xenbus: does not exist: %a\n", DevicePath));
+ }
+
+out:
+ return Status;
+}
+
+/**
+ * \brief Enumerate all devices of the given type on this bus.
+ *
+ * \param Dev NewBus device_t for this XenBus front bus instance.
+ * \param type String indicating the device sub-tree (e.g. "vfb", "vif")
+ * to enumerate.
+ *
+ * Devices that are found are entered into the NewBus hierarchy via
+ * xenbusb_add_device(). xenbusb_add_device() ignores duplicate detects
+ * and ignores duplicate devices, so it can be called unconditionally
+ * for any device found in the XenStore.
+ */
+STATIC
+VOID
+XenbusEnumerateDeviceType (
+ XENBUS_DEVICE *Dev,
+ CONST CHAR8 *Type
+ )
+{
+ CONST CHAR8 **Directory;
+ UINTN Index;
+ UINT32 Count;
+ XENSTORE_STATUS Status;
+
+ Status = XenstoreListDirectory (XST_NIL,
+ XENBUS_XENSTORE_NODE, Type,
+ &Count, &Directory);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return;
+ }
+ for (Index = 0; Index < Count; Index++) {
+ XenbusAddDevice (Dev, Type, Directory[Index]);
+ }
+
+ FreePool (Directory);
+}
+
+
+/**
+ * \brief Enumerate the devices on a XenBus bus and register them with
+ * the NewBus device tree.
+ *
+ * xenbusb_enumerate_bus() will create entries (in state DS_NOTPRESENT)
+ * for nodes that appear in the XenStore, but will not invoke probe/attach
+ * operations on drivers. Probe/Attach processing must be separately
+ * performed via an invocation of xenbusb_probe_children(). This is usually
+ * done via the xbs_probe_children task.
+ *
+ * \param Dev XenBus Bus device softc of the owner of the bus to enumerate.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ */
+XENSTORE_STATUS
+XenbusEnumerateBus (
+ XENBUS_DEVICE *Dev
+ )
+{
+ CONST CHAR8 **Types;
+ UINTN Index;
+ UINT32 Count;
+ XENSTORE_STATUS Status;
+
+ Status = XenstoreListDirectory (XST_NIL,
+ XENBUS_XENSTORE_NODE, "",
+ &Count, &Types);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return Status;
+ }
+
+ for (Index = 0; Index < Count; Index++) {
+ XenbusEnumerateDeviceType (Dev, Types[Index]);
+ }
+
+ FreePool (Types);
+
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+STATIC
+XENSTORE_STATUS
+EFIAPI
+XenbusSetState (
+ IN XENBUS_PROTOCOL *This,
+ IN XENSTORE_TRANSACTION Transaction,
+ IN enum xenbus_state NewState
+ )
+{
+ XENBUS_PRIVATE_DATA *Private;
+ enum xenbus_state CurrentState;
+ XENSTORE_STATUS Status;
+ CHAR8 *Temp;
+
+ DEBUG ((EFI_D_INFO, "Xenbus: Set state to %d\n", NewState));
+
+ Private = XENBUS_PRIVATE_DATA_FROM_THIS (This);
+
+ Status = XenstoreRead (Transaction, This->Node, "state", NULL, (VOID **)&Temp);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ goto Error;
+ }
+ CurrentState = AsciiStrDecimalToUintn (Temp);
+ FreePool (Temp);
+ if (CurrentState == NewState) {
+ DEBUG ((EFI_D_INFO, "%a %d, same state\n", __func__, __LINE__));
+ goto Error;
+ }
+
+ do {
+ Status = XenstoreSPrint (Transaction, This->Node, "state", "%d", NewState);
+ } while (Status == XENSTORE_STATUS_EAGAIN);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "Xenbus: fail to writing new state\n"));
+ goto Error;
+ }
+ DEBUG ((EFI_D_INFO, "Xenbus: Set state to %d, done\n", NewState));
+
+Error:
+ return Status;
+}
+
+STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData = {
+ .Signature = XENBUS_PRIVATE_DATA_SIGNATURE,
+
+ .XenbusIo.XsRead = XenbusXenstoreRead,
+ .XenbusIo.XsBackendRead = XenbusXenstoreBackendRead,
+ .XenbusIo.XsPrintf = XenbusXenstoreSPrint,
+ .XenbusIo.XsRemove = XenbusXenstoreRemove,
+ .XenbusIo.XsTransactionStart = XenbusXenstoreTransactionStart,
+ .XenbusIo.XsTransactionEnd = XenbusXenstoreTransactionEnd,
+ .XenbusIo.SetState = XenbusSetState,
+ .XenbusIo.GrantAccess = XenbusGrantAccess,
+ .XenbusIo.GrantEndAccess = XenbusGrantEndAccess,
+ .XenbusIo.RegisterWatch = XenbusRegisterWatch,
+ .XenbusIo.RegisterWatchBackend = XenbusRegisterWatchBackend,
+ .XenbusIo.UnregisterWatch = XenbusUnregisterWatch,
+ .XenbusIo.WaitForWatch = XenbusWaitForWatch,
+
+ .XenbusIo.Type = NULL,
+ .XenbusIo.Node = NULL,
+ .XenbusIo.Backend = NULL,
+
+ .Dev = NULL
+};
diff --git a/OvmfPkg/XenbusDxe/XenBus.h b/OvmfPkg/XenbusDxe/XenBus.h
new file mode 100644
index 0000000..63d0815
--- /dev/null
+++ b/OvmfPkg/XenbusDxe/XenBus.h
@@ -0,0 +1,76 @@
+/*-
+ * Core definitions and data structures shareable across OS platforms.
+ *
+ * Copyright (c) 2010 Spectra Logic Corporation
+ * Copyright (C) 2008 Doug Rabson
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions, and the following disclaimer,
+ * without modification.
+ * 2. Redistributions in binary form must reproduce at minimum a disclaimer
+ * substantially similar to the "NO WARRANTY" disclaimer below
+ * ("Disclaimer") and any redistribution must be conditioned upon
+ * including a substantially similar Disclaimer requirement for further
+ * binary redistribution.
+ *
+ * NO WARRANTY
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
+ * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGES.
+ *
+ * $FreeBSD: release/10.0.0/sys/xen/xenbus/xenbusb.h 222975 2011-06-11 04:59:01Z gibbs $
+ */
+#ifndef _XEN_XENBUS_XENBUSB_H
+#define _XEN_XENBUS_XENBUSB_H
+
+#include "XenbusDxe.h"
+
+#define XENBUS_DEVICE_PATH_TYPE_VBD 0x1
+struct _XENBUS_DEVICE_PATH {
+ VENDOR_DEVICE_PATH Vendor;
+ UINT8 Type;
+ UINT16 DeviceId;
+};
+
+/**
+ * The VM relative path to the XenStore subtree this
+ * bus attachment manages.
+ *
+ * OR
+ *
+ * The XenStore path to the XenStore subtree for this XenBus bus.
+ */
+#define XENBUS_XENSTORE_NODE "device"
+
+
+/**
+ * \brief Perform common XenBus bus attach processing.
+ *
+ * \param Dev The NewBus device representing this XenBus bus.
+ *
+ * \return On success, 0. Otherwise an errno value indicating the
+ * type of failure.
+ *
+ * Intiailizes the softc for this bus, installs an interrupt driven
+ * configuration hook to block boot processing until XenBus devices fully
+ * configure, performs an initial probe/attach of the bus, and registers
+ * a XenStore watch so we are notified when the bus topology changes.
+ */
+XENSTORE_STATUS
+XenbusEnumerateBus (
+ XENBUS_DEVICE *Dev
+ );
+
+#endif /* _XEN_XENBUS_XENBUSB_H */
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
index 2c85d5e..56225c4 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.c
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
@@ -19,6 +19,7 @@
#include "XenHypercall.h"
#include "GrantTable.h"
#include "Xenstore.h"
+#include "XenBus.h"
///
/// Driver Binding Protocol instance
@@ -295,6 +296,7 @@ XenbusDxeDriverBindingStart (
EFI_PCI_IO_PROTOCOL *PciIo;
EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR *BarDesc;
UINT64 MmioAddr;
+ EFI_DEVICE_PATH_PROTOCOL *DevicePath;
Status = gBS->OpenProtocol (
ControllerHandle,
@@ -308,11 +310,26 @@ XenbusDxeDriverBindingStart (
return Status;
}
+ Status = gBS->OpenProtocol (
+ ControllerHandle,
+ &gEfiDevicePathProtocolGuid,
+ (VOID **) &DevicePath,
+ This->DriverBindingHandle,
+ ControllerHandle,
+ EFI_OPEN_PROTOCOL_BY_DRIVER
+ );
+
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
Dev = AllocateZeroPool (sizeof (*Dev));
Dev->Signature = XENBUS_DEVICE_SIGNATURE;
Dev->This = This;
Dev->ControllerHandle = ControllerHandle;
Dev->PciIo = PciIo;
+ Dev->DevicePath = DevicePath;
+ InitializeListHead (&Dev->ChildList);
Status = PciIo->GetBarAttributes (PciIo, PCI_BAR_IDX1, NULL, (VOID**) &BarDesc);
ASSERT_EFI_ERROR (Status);
@@ -334,6 +351,8 @@ XenbusDxeDriverBindingStart (
Status = XenstoreInit (Dev);
ASSERT_EFI_ERROR (Status);
+ XenbusEnumerateBus (Dev);
+
Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
NotifyExitBoot,
(VOID*) Dev,
@@ -379,12 +398,54 @@ XenbusDxeDriverBindingStop (
IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
)
{
+ UINTN Index;
+ XENBUS_PROTOCOL *XenbusIo;
+ XENBUS_PRIVATE_DATA *ChildData;
+ EFI_STATUS Status;
XENBUS_DEVICE *Dev = mMyDevice;
+ for (Index = 0; Index < NumberOfChildren; Index++) {
+ Status = gBS->OpenProtocol (
+ ChildHandleBuffer[Index],
+ &gXenbusProtocolGuid,
+ (VOID **) &XenbusIo,
+ This->DriverBindingHandle,
+ ControllerHandle,
+ EFI_OPEN_PROTOCOL_GET_PROTOCOL);
+ if (EFI_ERROR (Status)) {
+ DEBUG ((EFI_D_ERROR, "XenbusDxe: get childre proto: %r\n", Status));
+ continue;
+ }
+ ChildData = XENBUS_PRIVATE_DATA_FROM_THIS(XenbusIo);
+ Status = gBS->DisconnectController(ChildData->Handle, NULL, NULL);
+ if (EFI_ERROR (Status)) {
+ DEBUG ((EFI_D_ERROR, "XenbusDxe: error disconnecting child: %r\n",
+ Status));
+ continue;
+ }
+
+ Status = gBS->UninstallMultipleProtocolInterfaces (
+ ChildData->Handle,
+ &gEfiDevicePathProtocolGuid, ChildData->DevicePath,
+ &gXenbusProtocolGuid, &ChildData->XenbusIo,
+ NULL);
+ ASSERT_EFI_ERROR (Status);
+
+ FreePool ((VOID*)ChildData->XenbusIo.Type);
+ FreePool ((VOID*)ChildData->XenbusIo.Node);
+ FreePool ((VOID*)ChildData->XenbusIo.Backend);
+ FreePool (ChildData->DevicePath);
+ RemoveEntryList (&ChildData->Link);
+ FreePool (ChildData);
+ }
+
gBS->CloseEvent (Dev->ExitBootEvent);
+ // XXX xenbus cleanup
// XXX xenstore cleanup
XenGrantTableDeinit (Dev);
+ gBS->CloseProtocol(ControllerHandle, &gEfiDevicePathProtocolGuid,
+ This->DriverBindingHandle, ControllerHandle);
gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
This->DriverBindingHandle, ControllerHandle);
return EFI_SUCCESS;
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
index 4d9db72..3ff08dd 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.h
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
@@ -80,6 +80,7 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
#define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
+typedef struct _XENBUS_DEVICE_PATH XENBUS_DEVICE_PATH;
typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
#define XENBUS_DEVICE_SIGNATURE SIGNATURE_32 ('X','B','b','d')
@@ -89,11 +90,28 @@ struct _XENBUS_DEVICE {
EFI_HANDLE ControllerHandle;
EFI_PCI_IO_PROTOCOL *PciIo;
EFI_EVENT ExitBootEvent;
+ EFI_DEVICE_PATH_PROTOCOL *DevicePath;
+ LIST_ENTRY ChildList;
VOID *Hyperpage;
shared_info_t *SharedInfo;
};
+#define XENBUS_PRIVATE_DATA_SIGNATURE SIGNATURE_32 ('X', 'B', 'u', 's')
+typedef struct {
+ UINTN Signature;
+ LIST_ENTRY Link;
+ EFI_HANDLE Handle;
+ XENBUS_PROTOCOL XenbusIo;
+ XENBUS_DEVICE *Dev;
+ XENBUS_DEVICE_PATH *DevicePath;
+} XENBUS_PRIVATE_DATA;
+
+#define XENBUS_PRIVATE_DATA_FROM_THIS(a) \
+ CR (a, XENBUS_PRIVATE_DATA, XenbusIo, XENBUS_PRIVATE_DATA_SIGNATURE)
+#define XENBUS_PRIVATE_DATA_FROM_LINK(a) \
+ CR (a, XENBUS_PRIVATE_DATA, Link, XENBUS_PRIVATE_DATA_SIGNATURE)
+
/*
* Helpers
*/
diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
index 6c8260f..708fae6 100644
--- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
+++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
@@ -40,6 +40,9 @@
EventChannel.h
Xenstore.c
Xenstore.h
+ XenBus.c
+ XenBus.h
+ Helpers.c
[Sources.X64]
X64/hypercall.S
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 15/18] OvmfPkg/XenbusDxe: Add Event Channel into XenBus protocol.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (13 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 16/18] OvmfPkg/XenPvBlkDxe: Xen PV Block device, initial skeleton Anthony PERARD
` (12 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/Include/Protocol/Xenbus.h | 27 +++++++++++++++++++
OvmfPkg/XenbusDxe/EventChannel.c | 55 +++++++++++++++++++++++++++++++++++++++
OvmfPkg/XenbusDxe/EventChannel.h | 26 ++++++++++++++++++
OvmfPkg/XenbusDxe/XenBus.c | 3 +++
4 files changed, 111 insertions(+)
diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
index 25408c1..aa37369 100644
--- a/OvmfPkg/Include/Protocol/Xenbus.h
+++ b/OvmfPkg/Include/Protocol/Xenbus.h
@@ -51,6 +51,7 @@ typedef enum {
#include <IndustryStandard/Xen/grant_table.h>
+#include <IndustryStandard/Xen/event_channel.h>
///
/// Function prototypes
@@ -138,6 +139,28 @@ EFI_STATUS
);
typedef
+int
+(EFIAPI *XENBUS_EVENT_CHANNEL_ALLOCATE) (
+ IN XENBUS_PROTOCOL *This,
+ IN domid_t DomainId,
+ OUT evtchn_port_t *Port
+ );
+
+typedef
+VOID
+(EFIAPI *XENBUS_EVENT_CHANNEL_NOTIFY) (
+ IN XENBUS_PROTOCOL *This,
+ IN evtchn_port_t Port
+ );
+
+typedef
+int
+(EFIAPI *XENBUS_EVENT_CHANNEL_CLOSE) (
+ IN XENBUS_PROTOCOL *This,
+ IN evtchn_port_t Port
+ );
+
+typedef
XENSTORE_STATUS
(EFIAPI *XENBUS_REGISTER_WATCH) (
IN XENBUS_PROTOCOL *This,
@@ -184,6 +207,10 @@ struct _XENBUS_PROTOCOL {
XENBUS_GRANT_ACCESS GrantAccess;
XENBUS_GRANT_END_ACCESS GrantEndAccess;
+ XENBUS_EVENT_CHANNEL_ALLOCATE EventChannelAllocate;
+ XENBUS_EVENT_CHANNEL_NOTIFY EventChannelNotify;
+ XENBUS_EVENT_CHANNEL_CLOSE EventChannelClose;
+
XENBUS_REGISTER_WATCH RegisterWatch;
XENBUS_REGISTER_WATCH_BACKEND RegisterWatchBackend;
XENBUS_UNREGISTER_WATCH UnregisterWatch;
diff --git a/OvmfPkg/XenbusDxe/EventChannel.c b/OvmfPkg/XenbusDxe/EventChannel.c
index 4ab8a40..101a1e4 100644
--- a/OvmfPkg/XenbusDxe/EventChannel.c
+++ b/OvmfPkg/XenbusDxe/EventChannel.c
@@ -14,3 +14,58 @@ XenEventChannelNotify (
ReturnCode = XenHypercallEventChannelOp (Dev, EVTCHNOP_send, &send);
ASSERT (ReturnCode == 0);
}
+
+int
+EFIAPI
+XenbusEventChannelAllocate (
+ IN XENBUS_PROTOCOL *This,
+ IN domid_t DomainId,
+ OUT evtchn_port_t *Port
+ )
+{
+ XENBUS_PRIVATE_DATA *Private;
+ evtchn_alloc_unbound_t op;
+ int rc;
+
+ Private = XENBUS_PRIVATE_DATA_FROM_THIS (This);
+
+ op.dom = DOMID_SELF;
+ op.remote_dom = DomainId;
+ rc = XenHypercallEventChannelOp (Private->Dev,
+ EVTCHNOP_alloc_unbound,
+ &op);
+ if (rc != 0) {
+ DEBUG ((EFI_D_ERROR, "ERROR: alloc_unbound failed with rc=%d", rc));
+ return rc;
+ }
+ *Port = op.port;
+ return rc;
+}
+
+VOID
+EFIAPI
+XenbusEventChannelNotify (
+ IN XENBUS_PROTOCOL *This,
+ IN evtchn_port_t Port
+ )
+{
+ XENBUS_PRIVATE_DATA *Private;
+
+ Private = XENBUS_PRIVATE_DATA_FROM_THIS(This);
+ XenEventChannelNotify (Private->Dev, Port);
+}
+
+int
+EFIAPI
+XenbusEventChannelClose (
+ IN XENBUS_PROTOCOL *This,
+ IN evtchn_port_t Port
+ )
+{
+ XENBUS_PRIVATE_DATA *Private;
+ evtchn_close_t close;
+
+ Private = XENBUS_PRIVATE_DATA_FROM_THIS(This);
+ close.port = Port;
+ return XenHypercallEventChannelOp (Private->Dev, EVTCHNOP_close, &close);
+}
diff --git a/OvmfPkg/XenbusDxe/EventChannel.h b/OvmfPkg/XenbusDxe/EventChannel.h
index 8a3f6d4..b1e16f0 100644
--- a/OvmfPkg/XenbusDxe/EventChannel.h
+++ b/OvmfPkg/XenbusDxe/EventChannel.h
@@ -11,4 +11,30 @@ XenEventChannelNotify (
IN evtchn_port_t Port
);
+/*
+ * Xenbus protocol
+ */
+
+int
+EFIAPI
+XenbusEventChannelAllocate (
+ IN XENBUS_PROTOCOL *This,
+ IN domid_t DomainId,
+ OUT evtchn_port_t *Port
+ );
+
+VOID
+EFIAPI
+XenbusEventChannelNotify (
+ IN XENBUS_PROTOCOL *This,
+ IN evtchn_port_t Port
+ );
+
+int
+EFIAPI
+XenbusEventChannelClose (
+ IN XENBUS_PROTOCOL *This,
+ IN evtchn_port_t Port
+ );
+
#endif
diff --git a/OvmfPkg/XenbusDxe/XenBus.c b/OvmfPkg/XenbusDxe/XenBus.c
index b0cf1ba..f93423f 100644
--- a/OvmfPkg/XenbusDxe/XenBus.c
+++ b/OvmfPkg/XenbusDxe/XenBus.c
@@ -347,6 +347,9 @@ STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData = {
.XenbusIo.SetState = XenbusSetState,
.XenbusIo.GrantAccess = XenbusGrantAccess,
.XenbusIo.GrantEndAccess = XenbusGrantEndAccess,
+ .XenbusIo.EventChannelAllocate = XenbusEventChannelAllocate,
+ .XenbusIo.EventChannelNotify = XenbusEventChannelNotify,
+ .XenbusIo.EventChannelClose = XenbusEventChannelClose,
.XenbusIo.RegisterWatch = XenbusRegisterWatch,
.XenbusIo.RegisterWatchBackend = XenbusRegisterWatchBackend,
.XenbusIo.UnregisterWatch = XenbusUnregisterWatch,
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 16/18] OvmfPkg/XenPvBlkDxe: Xen PV Block device, initial skeleton
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (14 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 15/18] OvmfPkg/XenbusDxe: Add Event Channel into XenBus protocol Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client Anthony PERARD
` (11 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
A ParaVirtualize block driver.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/OvmfPkgX64.dsc | 1 +
OvmfPkg/OvmfPkgX64.fdf | 1 +
OvmfPkg/XenPvBlkDxe/ComponentName.c | 160 ++++++++++++++++++
OvmfPkg/XenPvBlkDxe/ComponentName.h | 91 ++++++++++
OvmfPkg/XenPvBlkDxe/DriverBinding.h | 139 ++++++++++++++++
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 321 ++++++++++++++++++++++++++++++++++++
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h | 82 +++++++++
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 57 +++++++
8 files changed, 852 insertions(+)
create mode 100644 OvmfPkg/XenPvBlkDxe/ComponentName.c
create mode 100644 OvmfPkg/XenPvBlkDxe/ComponentName.h
create mode 100644 OvmfPkg/XenPvBlkDxe/DriverBinding.h
create mode 100644 OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
create mode 100644 OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
create mode 100644 OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
diff --git a/OvmfPkg/OvmfPkgX64.dsc b/OvmfPkg/OvmfPkgX64.dsc
index 4ea1a83..1564c4c 100644
--- a/OvmfPkg/OvmfPkgX64.dsc
+++ b/OvmfPkg/OvmfPkgX64.dsc
@@ -431,6 +431,7 @@
OvmfPkg/VirtioScsiDxe/VirtioScsi.inf
OvmfPkg/QemuFlashFvbServicesRuntimeDxe/FvbServicesRuntimeDxe.inf
OvmfPkg/XenbusDxe/XenbusDxe.inf
+ OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
OvmfPkg/EmuVariableFvbRuntimeDxe/Fvb.inf {
<LibraryClasses>
PlatformFvbLib|OvmfPkg/Library/EmuVariableFvbLib/EmuVariableFvbLib.inf
diff --git a/OvmfPkg/OvmfPkgX64.fdf b/OvmfPkg/OvmfPkgX64.fdf
index 7f1aab4..ef2aee7 100644
--- a/OvmfPkg/OvmfPkgX64.fdf
+++ b/OvmfPkg/OvmfPkgX64.fdf
@@ -279,6 +279,7 @@ INF OvmfPkg/QemuFlashFvbServicesRuntimeDxe/FvbServicesRuntimeDxe.inf
INF OvmfPkg/EmuVariableFvbRuntimeDxe/Fvb.inf
INF MdeModulePkg/Universal/FaultTolerantWriteDxe/FaultTolerantWriteDxe.inf
INF OvmfPkg/XenbusDxe/XenbusDxe.inf
+INF OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
!if $(SECURE_BOOT_ENABLE) == TRUE
INF SecurityPkg/VariableAuthenticated/RuntimeDxe/VariableRuntimeDxe.inf
diff --git a/OvmfPkg/XenPvBlkDxe/ComponentName.c b/OvmfPkg/XenPvBlkDxe/ComponentName.c
new file mode 100644
index 0000000..aa9ae6d
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/ComponentName.c
@@ -0,0 +1,160 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+#include "XenPvBlkDxe.h"
+
+///
+/// Component Name Protocol instance
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_COMPONENT_NAME_PROTOCOL gXenPvBlkDxeComponentName = {
+ (EFI_COMPONENT_NAME_GET_DRIVER_NAME) XenPvBlkDxeComponentNameGetDriverName,
+ (EFI_COMPONENT_NAME_GET_CONTROLLER_NAME)XenPvBlkDxeComponentNameGetControllerName,
+ "eng"
+};
+
+///
+/// Component Name 2 Protocol instance
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_COMPONENT_NAME2_PROTOCOL gXenPvBlkDxeComponentName2 = {
+ XenPvBlkDxeComponentNameGetDriverName,
+ XenPvBlkDxeComponentNameGetControllerName,
+ "en"
+};
+
+///
+/// Table of driver names
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_UNICODE_STRING_TABLE mXenPvBlkDxeDriverNameTable[] = {
+ { "eng;en", (CHAR16 *)L"Xen PV Block Driver" },
+ { NULL, NULL }
+};
+
+/**
+ Retrieves a Unicode string that is the user-readable name of the EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param Language A pointer to a three-character ISO 639-2 language identifier.
+ This is the language of the driver name that that the caller
+ is requesting, and it must match one of the languages specified
+ in SupportedLanguages. The number of languages supported by a
+ driver is up to the driver writer.
+ @param DriverName A pointer to the Unicode string to return. This Unicode string
+ is the name of the driver specified by This in the language
+ specified by Language.
+
+ @retval EFI_SUCCESS The Unicode string for the Driver specified by This
+ and the language specified by Language was returned
+ in DriverName.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER DriverName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeComponentNameGetDriverName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN CHAR8 *Language,
+ OUT CHAR16 **DriverName
+ )
+{
+ return LookupUnicodeString2 (
+ Language,
+ This->SupportedLanguages,
+ mXenPvBlkDxeDriverNameTable,
+ DriverName,
+ (BOOLEAN)(This != &gXenPvBlkDxeComponentName2)
+ );
+}
+
+/**
+ Retrieves a Unicode string that is the user readable name of the controller
+ that is being managed by an EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param ControllerHandle The handle of a controller that the driver specified by
+ This is managing. This handle specifies the controller
+ whose name is to be returned.
+ @param ChildHandle The handle of the child controller to retrieve the name
+ of. This is an optional parameter that may be NULL. It
+ will be NULL for device drivers. It will also be NULL
+ for a bus drivers that wish to retrieve the name of the
+ bus controller. It will not be NULL for a bus driver
+ that wishes to retrieve the name of a child controller.
+ @param Language A pointer to a three character ISO 639-2 language
+ identifier. This is the language of the controller name
+ that the caller is requesting, and it must match one
+ of the languages specified in SupportedLanguages. The
+ number of languages supported by a driver is up to the
+ driver writer.
+ @param ControllerName A pointer to the Unicode string to return. This Unicode
+ string is the name of the controller specified by
+ ControllerHandle and ChildHandle in the language specified
+ by Language, from the point of view of the driver specified
+ by This.
+
+ @retval EFI_SUCCESS The Unicode string for the user-readable name in the
+ language specified by Language for the driver
+ specified by This was returned in DriverName.
+ @retval EFI_INVALID_PARAMETER ControllerHandle is NULL.
+ @retval EFI_INVALID_PARAMETER ChildHandle is not NULL and it is not a valid EFI_HANDLE.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER ControllerName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This is not currently managing
+ the controller specified by ControllerHandle and
+ ChildHandle.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeComponentNameGetControllerName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_HANDLE ChildHandle OPTIONAL,
+ IN CHAR8 *Language,
+ OUT CHAR16 **ControllerName
+ )
+{
+ EFI_STATUS Status;
+
+ //
+ // ChildHandle must be NULL for a Device Driver
+ //
+ if (ChildHandle != NULL) {
+ return EFI_UNSUPPORTED;
+ }
+
+ //
+ // Make sure this driver is currently managing ControllerHandle
+ //
+ Status = EfiTestManagedDevice (
+ ControllerHandle,
+ gXenPvBlkDxeDriverBinding.DriverBindingHandle,
+ &gXenbusProtocolGuid
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ //
+ // Lookup name of controller specified by ControllerHandle
+ //
+ Status = EFI_UNSUPPORTED;
+
+ return Status;
+}
diff --git a/OvmfPkg/XenPvBlkDxe/ComponentName.h b/OvmfPkg/XenPvBlkDxe/ComponentName.h
new file mode 100644
index 0000000..11a9255
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/ComponentName.h
@@ -0,0 +1,91 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+/**
+ Retrieves a Unicode string that is the user-readable name of the EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param Language A pointer to a three-character ISO 639-2 language identifier.
+ This is the language of the driver name that that the caller
+ is requesting, and it must match one of the languages specified
+ in SupportedLanguages. The number of languages supported by a
+ driver is up to the driver writer.
+ @param DriverName A pointer to the Unicode string to return. This Unicode string
+ is the name of the driver specified by This in the language
+ specified by Language.
+
+ @retval EFI_SUCCESS The Unicode string for the Driver specified by This
+ and the language specified by Language was returned
+ in DriverName.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER DriverName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeComponentNameGetDriverName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN CHAR8 *Language,
+ OUT CHAR16 **DriverName
+ );
+
+/**
+ Retrieves a Unicode string that is the user readable name of the controller
+ that is being managed by an EFI Driver.
+
+ @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
+ @param ControllerHandle The handle of a controller that the driver specified by
+ This is managing. This handle specifies the controller
+ whose name is to be returned.
+ @param ChildHandle The handle of the child controller to retrieve the name
+ of. This is an optional parameter that may be NULL. It
+ will be NULL for device drivers. It will also be NULL
+ for a bus drivers that wish to retrieve the name of the
+ bus controller. It will not be NULL for a bus driver
+ that wishes to retrieve the name of a child controller.
+ @param Language A pointer to a three character ISO 639-2 language
+ identifier. This is the language of the controller name
+ that the caller is requesting, and it must match one
+ of the languages specified in SupportedLanguages. The
+ number of languages supported by a driver is up to the
+ driver writer.
+ @param ControllerName A pointer to the Unicode string to return. This Unicode
+ string is the name of the controller specified by
+ ControllerHandle and ChildHandle in the language specified
+ by Language, from the point of view of the driver specified
+ by This.
+
+ @retval EFI_SUCCESS The Unicode string for the user-readable name in the
+ language specified by Language for the driver
+ specified by This was returned in DriverName.
+ @retval EFI_INVALID_PARAMETER ControllerHandle is NULL.
+ @retval EFI_INVALID_PARAMETER ChildHandle is not NULL and it is not a valid EFI_HANDLE.
+ @retval EFI_INVALID_PARAMETER Language is NULL.
+ @retval EFI_INVALID_PARAMETER ControllerName is NULL.
+ @retval EFI_UNSUPPORTED The driver specified by This is not currently managing
+ the controller specified by ControllerHandle and
+ ChildHandle.
+ @retval EFI_UNSUPPORTED The driver specified by This does not support the
+ language specified by Language.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeComponentNameGetControllerName (
+ IN EFI_COMPONENT_NAME2_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_HANDLE ChildHandle OPTIONAL,
+ IN CHAR8 *Language,
+ OUT CHAR16 **ControllerName
+ );
diff --git a/OvmfPkg/XenPvBlkDxe/DriverBinding.h b/OvmfPkg/XenPvBlkDxe/DriverBinding.h
new file mode 100644
index 0000000..96a0155
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/DriverBinding.h
@@ -0,0 +1,139 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+/**
+ Tests to see if this driver supports a given controller. If a child device is provided,
+ it further tests to see if this driver supports creating a handle for the specified child device.
+
+ This function checks to see if the driver specified by This supports the device specified by
+ ControllerHandle. Drivers will typically use the device path attached to
+ ControllerHandle and/or the services from the bus I/O abstraction attached to
+ ControllerHandle to determine if the driver supports ControllerHandle. This function
+ may be called many times during platform initialization. In order to reduce boot times, the tests
+ performed by this function must be very small, and take as little time as possible to execute. This
+ function must not change the state of any hardware devices, and this function must be aware that the
+ device specified by ControllerHandle may already be managed by the same driver or a
+ different driver. This function must match its calls to AllocatePages() with FreePages(),
+ AllocatePool() with FreePool(), and OpenProtocol() with CloseProtocol().
+ Because ControllerHandle may have been previously started by the same driver, if a protocol is
+ already in the opened state, then it must not be closed with CloseProtocol(). This is required
+ to guarantee the state of ControllerHandle is not modified by this function.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to test. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For bus drivers, if this parameter is not NULL, then
+ the bus driver must determine if the bus controller specified
+ by ControllerHandle and the child controller specified
+ by RemainingDevicePath are both supported by this
+ bus driver.
+
+ @retval EFI_SUCCESS The device specified by ControllerHandle and
+ RemainingDevicePath is supported by the driver specified by This.
+ @retval EFI_ALREADY_STARTED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by the driver
+ specified by This.
+ @retval EFI_ACCESS_DENIED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by a different
+ driver or an application that requires exclusive access.
+ Currently not implemented.
+ @retval EFI_UNSUPPORTED The device specified by ControllerHandle and
+ RemainingDevicePath is not supported by the driver specified by This.
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverBindingSupported (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ );
+
+/**
+ Starts a device controller or a bus controller.
+
+ The Start() function is designed to be invoked from the EFI boot service ConnectController().
+ As a result, much of the error checking on the parameters to Start() has been moved into this
+ common boot service. It is legal to call Start() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE.
+ 2. If RemainingDevicePath is not NULL, then it must be a pointer to a naturally aligned
+ EFI_DEVICE_PATH_PROTOCOL.
+ 3. Prior to calling Start(), the Supported() function for the driver specified by This must
+ have been called with the same calling parameters, and Supported() must have returned EFI_SUCCESS.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to start. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For a bus driver, if this parameter is NULL, then handles
+ for all the children of Controller are created by this driver.
+ If this parameter is not NULL and the first Device Path Node is
+ not the End of Device Path Node, then only the handle for the
+ child device specified by the first Device Path Node of
+ RemainingDevicePath is created by this driver.
+ If the first Device Path Node of RemainingDevicePath is
+ the End of Device Path Node, no child handle is created by this
+ driver.
+
+ @retval EFI_SUCCESS The device was started.
+ @retval EFI_DEVICE_ERROR The device could not be started due to a device error.Currently not implemented.
+ @retval EFI_OUT_OF_RESOURCES The request could not be completed due to a lack of resources.
+ @retval Others The driver failded to start the device.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverBindingStart (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ );
+
+/**
+ Stops a device controller or a bus controller.
+
+ The Stop() function is designed to be invoked from the EFI boot service DisconnectController().
+ As a result, much of the error checking on the parameters to Stop() has been moved
+ into this common boot service. It is legal to call Stop() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE that was used on a previous call to this
+ same driver's Start() function.
+ 2. The first NumberOfChildren handles of ChildHandleBuffer must all be a valid
+ EFI_HANDLE. In addition, all of these handles must have been created in this driver's
+ Start() function, and the Start() function must have called OpenProtocol() on
+ ControllerHandle with an Attribute of EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle A handle to the device being stopped. The handle must
+ support a bus specific I/O protocol for the driver
+ to use to stop the device.
+ @param[in] NumberOfChildren The number of child device handles in ChildHandleBuffer.
+ @param[in] ChildHandleBuffer An array of child handles to be freed. May be NULL
+ if NumberOfChildren is 0.
+
+ @retval EFI_SUCCESS The device was stopped.
+ @retval EFI_DEVICE_ERROR The device could not be stopped due to a device error.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverBindingStop (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN UINTN NumberOfChildren,
+ IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
+ );
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
new file mode 100644
index 0000000..0ce1e86
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
@@ -0,0 +1,321 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+#include "XenPvBlkDxe.h"
+
+
+///
+/// Driver Binding Protocol instance
+///
+EFI_DRIVER_BINDING_PROTOCOL gXenPvBlkDxeDriverBinding = {
+ XenPvBlkDxeDriverBindingSupported,
+ XenPvBlkDxeDriverBindingStart,
+ XenPvBlkDxeDriverBindingStop,
+ XEN_PV_BLK_DXE_VERSION,
+ NULL,
+ NULL
+};
+
+
+/**
+ Unloads an image.
+
+ @param ImageHandle Handle that identifies the image to be unloaded.
+
+ @retval EFI_SUCCESS The image has been unloaded.
+ @retval EFI_INVALID_PARAMETER ImageHandle is not a valid image handle.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeUnload (
+ IN EFI_HANDLE ImageHandle
+ )
+{
+ EFI_STATUS Status;
+
+ EFI_HANDLE *HandleBuffer;
+ UINTN HandleCount;
+ UINTN Index;
+
+
+ DEBUG ((EFI_D_INFO, "%a %d\n", __func__, __LINE__));
+ Status = EFI_SUCCESS;
+
+ //
+ // Retrieve array of all handles in the handle database
+ //
+ Status = gBS->LocateHandleBuffer (
+ AllHandles,
+ NULL,
+ NULL,
+ &HandleCount,
+ &HandleBuffer
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ //
+ // Disconnect the current driver from handles in the handle database
+ //
+ for (Index = 0; Index < HandleCount; Index++) {
+ Status = gBS->DisconnectController (HandleBuffer[Index], gImageHandle, NULL);
+ }
+
+ //
+ // Free the array of handles
+ //
+ FreePool (HandleBuffer);
+
+
+ //
+ // Uninstall protocols installed in the driver entry point
+ //
+ Status = gBS->UninstallMultipleProtocolInterfaces (
+ ImageHandle,
+ &gEfiDriverBindingProtocolGuid, &gXenPvBlkDxeDriverBinding,
+ &gEfiComponentNameProtocolGuid, &gXenPvBlkDxeComponentName,
+ &gEfiComponentName2ProtocolGuid, &gXenPvBlkDxeComponentName2,
+ NULL
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+
+ //
+ // Do any additional cleanup that is required for this driver
+ //
+
+ return EFI_SUCCESS;
+}
+
+/**
+ This is the declaration of an EFI image entry point. This entry point is
+ the same for UEFI Applications, UEFI OS Loaders, and UEFI Drivers including
+ both device drivers and bus drivers.
+
+ @param ImageHandle The firmware allocated handle for the UEFI image.
+ @param SystemTable A pointer to the EFI System Table.
+
+ @retval EFI_SUCCESS The operation completed successfully.
+ @retval Others An unexpected error occurred.
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverEntryPoint (
+ IN EFI_HANDLE ImageHandle,
+ IN EFI_SYSTEM_TABLE *SystemTable
+ )
+{
+ EFI_STATUS Status;
+
+
+ Status = EFI_SUCCESS;
+
+
+ //
+ // Install UEFI Driver Model protocol(s).
+ //
+ Status = EfiLibInstallDriverBindingComponentName2 (
+ ImageHandle,
+ SystemTable,
+ &gXenPvBlkDxeDriverBinding,
+ ImageHandle,
+ &gXenPvBlkDxeComponentName,
+ &gXenPvBlkDxeComponentName2
+ );
+ ASSERT_EFI_ERROR (Status);
+
+
+ return Status;
+}
+
+
+/**
+ Tests to see if this driver supports a given controller. If a child device is provided,
+ it further tests to see if this driver supports creating a handle for the specified child device.
+
+ This function checks to see if the driver specified by This supports the device specified by
+ ControllerHandle. Drivers will typically use the device path attached to
+ ControllerHandle and/or the services from the bus I/O abstraction attached to
+ ControllerHandle to determine if the driver supports ControllerHandle. This function
+ may be called many times during platform initialization. In order to reduce boot times, the tests
+ performed by this function must be very small, and take as little time as possible to execute. This
+ function must not change the state of any hardware devices, and this function must be aware that the
+ device specified by ControllerHandle may already be managed by the same driver or a
+ different driver. This function must match its calls to AllocatePages() with FreePages(),
+ AllocatePool() with FreePool(), and OpenProtocol() with CloseProtocol().
+ Because ControllerHandle may have been previously started by the same driver, if a protocol is
+ already in the opened state, then it must not be closed with CloseProtocol(). This is required
+ to guarantee the state of ControllerHandle is not modified by this function.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to test. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For bus drivers, if this parameter is not NULL, then
+ the bus driver must determine if the bus controller specified
+ by ControllerHandle and the child controller specified
+ by RemainingDevicePath are both supported by this
+ bus driver.
+
+ @retval EFI_SUCCESS The device specified by ControllerHandle and
+ RemainingDevicePath is supported by the driver specified by This.
+ @retval EFI_ALREADY_STARTED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by the driver
+ specified by This.
+ @retval EFI_ACCESS_DENIED The device specified by ControllerHandle and
+ RemainingDevicePath is already being managed by a different
+ driver or an application that requires exclusive access.
+ Currently not implemented.
+ @retval EFI_UNSUPPORTED The device specified by ControllerHandle and
+ RemainingDevicePath is not supported by the driver specified by This.
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverBindingSupported (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ )
+{
+ EFI_STATUS Status;
+ XENBUS_PROTOCOL *XenbusIo;
+
+ Status = gBS->OpenProtocol (
+ ControllerHandle,
+ &gXenbusProtocolGuid,
+ (VOID **)&XenbusIo,
+ This->DriverBindingHandle,
+ ControllerHandle,
+ EFI_OPEN_PROTOCOL_BY_DRIVER
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+ if (AsciiStrCmp (XenbusIo->Type, "vbd") == 0) {
+ Status = EFI_SUCCESS;
+ } else {
+ Status = EFI_UNSUPPORTED;
+ }
+
+ gBS->CloseProtocol (ControllerHandle, &gXenbusProtocolGuid,
+ This->DriverBindingHandle, ControllerHandle);
+
+ return Status;
+}
+
+/**
+ Starts a device controller or a bus controller.
+
+ The Start() function is designed to be invoked from the EFI boot service ConnectController().
+ As a result, much of the error checking on the parameters to Start() has been moved into this
+ common boot service. It is legal to call Start() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE.
+ 2. If RemainingDevicePath is not NULL, then it must be a pointer to a naturally aligned
+ EFI_DEVICE_PATH_PROTOCOL.
+ 3. Prior to calling Start(), the Supported() function for the driver specified by This must
+ have been called with the same calling parameters, and Supported() must have returned EFI_SUCCESS.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle The handle of the controller to start. This handle
+ must support a protocol interface that supplies
+ an I/O abstraction to the driver.
+ @param[in] RemainingDevicePath A pointer to the remaining portion of a device path. This
+ parameter is ignored by device drivers, and is optional for bus
+ drivers. For a bus driver, if this parameter is NULL, then handles
+ for all the children of Controller are created by this driver.
+ If this parameter is not NULL and the first Device Path Node is
+ not the End of Device Path Node, then only the handle for the
+ child device specified by the first Device Path Node of
+ RemainingDevicePath is created by this driver.
+ If the first Device Path Node of RemainingDevicePath is
+ the End of Device Path Node, no child handle is created by this
+ driver.
+
+ @retval EFI_SUCCESS The device was started.
+ @retval EFI_DEVICE_ERROR The device could not be started due to a device error.Currently not implemented.
+ @retval EFI_OUT_OF_RESOURCES The request could not be completed due to a lack of resources.
+ @retval Others The driver failded to start the device.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverBindingStart (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
+ )
+{
+ EFI_STATUS Status;
+ XENBUS_PROTOCOL *XenbusIo;
+
+ Status = gBS->OpenProtocol (
+ ControllerHandle,
+ &gXenbusProtocolGuid,
+ (VOID **)&XenbusIo,
+ This->DriverBindingHandle,
+ ControllerHandle,
+ EFI_OPEN_PROTOCOL_BY_DRIVER
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ return EFI_SUCCESS;
+}
+
+/**
+ Stops a device controller or a bus controller.
+
+ The Stop() function is designed to be invoked from the EFI boot service DisconnectController().
+ As a result, much of the error checking on the parameters to Stop() has been moved
+ into this common boot service. It is legal to call Stop() from other locations,
+ but the following calling restrictions must be followed, or the system behavior will not be deterministic.
+ 1. ControllerHandle must be a valid EFI_HANDLE that was used on a previous call to this
+ same driver's Start() function.
+ 2. The first NumberOfChildren handles of ChildHandleBuffer must all be a valid
+ EFI_HANDLE. In addition, all of these handles must have been created in this driver's
+ Start() function, and the Start() function must have called OpenProtocol() on
+ ControllerHandle with an Attribute of EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER.
+
+ @param[in] This A pointer to the EFI_DRIVER_BINDING_PROTOCOL instance.
+ @param[in] ControllerHandle A handle to the device being stopped. The handle must
+ support a bus specific I/O protocol for the driver
+ to use to stop the device.
+ @param[in] NumberOfChildren The number of child device handles in ChildHandleBuffer.
+ @param[in] ChildHandleBuffer An array of child handles to be freed. May be NULL
+ if NumberOfChildren is 0.
+
+ @retval EFI_SUCCESS The device was stopped.
+ @retval EFI_DEVICE_ERROR The device could not be stopped due to a device error.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeDriverBindingStop (
+ IN EFI_DRIVER_BINDING_PROTOCOL *This,
+ IN EFI_HANDLE ControllerHandle,
+ IN UINTN NumberOfChildren,
+ IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
+ )
+{
+ gBS->CloseProtocol (ControllerHandle, &gXenbusProtocolGuid,
+ This->DriverBindingHandle, ControllerHandle);
+
+ return EFI_SUCCESS;
+}
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
new file mode 100644
index 0000000..186d0ee
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
@@ -0,0 +1,82 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+#ifndef __EFI_XEN_PV_BLK_DXE_H__
+#define __EFI_XEN_PV_BLK_DXE_H__
+
+#include <Uefi.h>
+
+//
+// Libraries
+//
+#include <Library/UefiBootServicesTableLib.h>
+#include <Library/MemoryAllocationLib.h>
+#include <Library/BaseMemoryLib.h>
+#include <Library/BaseLib.h>
+#include <Library/UefiLib.h>
+#include <Library/DevicePathLib.h>
+#include <Library/DebugLib.h>
+
+
+//
+// UEFI Driver Model Protocols
+//
+#include <Protocol/DriverBinding.h>
+#include <Protocol/ComponentName2.h>
+#include <Protocol/ComponentName.h>
+
+
+//
+// Consumed Protocols
+//
+// Xen interface version used:
+#define __XEN_INTERFACE_VERSION__ 0x00040400
+#define xen_mb() MemoryFence()
+#define xen_rmb() MemoryFence()
+#define xen_wmb() MemoryFence()
+#include <Protocol/Xenbus.h>
+
+
+//
+// Produced Protocols
+//
+#include <Protocol/BlockIo.h>
+#include <Protocol/BlockIo2.h>
+
+
+//
+// Guids
+//
+
+//
+// Driver Version
+//
+#define XEN_PV_BLK_DXE_VERSION 0x00000000
+
+
+//
+// Protocol instances
+//
+extern EFI_DRIVER_BINDING_PROTOCOL gXenPvBlkDxeDriverBinding;
+extern EFI_COMPONENT_NAME2_PROTOCOL gXenPvBlkDxeComponentName2;
+extern EFI_COMPONENT_NAME_PROTOCOL gXenPvBlkDxeComponentName;
+
+
+//
+// Include files with function prototypes
+//
+#include "DriverBinding.h"
+#include "ComponentName.h"
+
+
+
+#endif
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
new file mode 100644
index 0000000..4444cda
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
@@ -0,0 +1,57 @@
+
+## @file
+# TODO: Brief Description of UEFI Driver XenPvBlkDxe
+#
+# TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+#
+# TODO: Copyright for UEFI Driver XenPvBlkDxe
+#
+# TODO: License for UEFI Driver XenPvBlkDxe
+#
+##
+
+[Defines]
+ INF_VERSION = 0x00010005
+ BASE_NAME = XenPvBlkDxe
+ FILE_GUID = 8c2487ea-9af3-11e3-b966-b8ac6f7d65e6
+ MODULE_TYPE = UEFI_DRIVER
+
+ VERSION_STRING = 0.1
+ ENTRY_POINT = XenPvBlkDxeDriverEntryPoint
+ UNLOAD_IMAGE = XenPvBlkDxeUnload
+
+
+[Packages]
+ MdePkg/MdePkg.dec
+ OvmfPkg/OvmfPkg.dec
+
+[Sources]
+ XenPvBlkDxe.h
+ XenPvBlkDxe.c
+ ComponentName.c
+ ComponentName.h
+
+
+[LibraryClasses]
+ UefiDriverEntryPoint
+ UefiBootServicesTableLib
+ MemoryAllocationLib
+ BaseMemoryLib
+ BaseLib
+ UefiLib
+ DevicePathLib
+ DebugLib
+
+
+[Protocols]
+ gEfiDriverBindingProtocolGuid
+ gEfiBlockIoProtocolGuid
+ gEfiBlockIo2ProtocolGuid
+ gEfiComponentName2ProtocolGuid
+ gEfiComponentNameProtocolGuid
+ gXenbusProtocolGuid
+ gEfiDevicePathProtocolGuid ## TO_START
+
+
+[Guids]
+
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (15 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 16/18] OvmfPkg/XenPvBlkDxe: Xen PV Block device, initial skeleton Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo Anthony PERARD
` (10 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Samuel Thibault, Xen Devel
This is the code that will do the actual communication between OVMF and
a PV block backend, where the block device lives.
This implementation originally comes from Mini-OS, a part of the Xen
Project.
Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenPvBlkDxe/BlockFront.c | 583 ++++++++++++++++++++++++++++++++++++
OvmfPkg/XenPvBlkDxe/BlockFront.h | 88 ++++++
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 8 +
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 2 +
4 files changed, 681 insertions(+)
create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.c
create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.h
diff --git a/OvmfPkg/XenPvBlkDxe/BlockFront.c b/OvmfPkg/XenPvBlkDxe/BlockFront.c
new file mode 100644
index 0000000..ff5641c
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/BlockFront.c
@@ -0,0 +1,583 @@
+/* Minimal block driver for Mini-OS.
+ * Copyright (c) 2007-2008 Samuel Thibault.
+ * Based on netfront.c.
+ */
+
+#include <Library/PrintLib.h>
+#include <Library/DebugLib.h>
+
+#include "BlockFront.h"
+
+#include <IndustryStandard/Xen/io/protocols.h>
+#include <IndustryStandard/Xen/io/xenbus.h>
+
+#include "inttypes.h"
+
+STATIC
+XENSTORE_STATUS
+XenbusReadInteger (
+ IN XENBUS_PROTOCOL *This,
+ IN CONST CHAR8 *Node,
+ IN BOOLEAN FromBackend,
+ OUT UINT64 *ValuePtr
+ )
+{
+ XENSTORE_STATUS Status;
+ CHAR8 *p;
+
+ if (!FromBackend) {
+ Status = This->XsRead (This, XST_NIL, Node, (VOID**)&p);
+ } else {
+ Status = This->XsBackendRead (This, XST_NIL, Node, (VOID**)&p);
+ }
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return Status;
+ }
+ // XXX Will assert if p overflow UINT64 ...
+ *ValuePtr = AsciiStrDecimalToUint64 (p);
+ FreePool (p);
+ return Status;
+}
+
+STATIC
+VOID
+XenPvBlockFree (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ )
+{
+ XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
+
+ if (Dev->RingRef != 0) {
+ XenbusIo->GrantEndAccess (XenbusIo, Dev->RingRef);
+ }
+ if (Dev->Ring.sring != NULL) {
+ FreePages (Dev->Ring.sring, 1);
+ }
+ if (Dev->EventChannel != 0) {
+ XenbusIo->EventChannelClose (XenbusIo, Dev->EventChannel);
+ }
+ FreePool (Dev);
+}
+
+XENSTORE_STATUS
+XenPvBlkWaitForBackendState (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev,
+ IN XenbusState WaitedState,
+ OUT XenbusState *LastStatePtr OPTIONAL
+ )
+{
+ XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
+ XenbusState State;
+ UINT64 Value;
+ XENSTORE_STATUS Status;
+
+ while (TRUE) {
+ Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ return Status;
+ }
+ if (Value > XenbusStateReconfigured) {
+ return XENSTORE_STATUS_EIO;
+ }
+ State = Value;
+ if (State >= WaitedState) {
+ break;
+ }
+ DEBUG ((EFI_D_INFO,
+ "XenPvBlk: waiting backend state %d, current: %d\n",
+ WaitedState, State));
+ XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
+ }
+
+ if (LastStatePtr != NULL) {
+ *LastStatePtr = State;
+ }
+
+ return XENSTORE_STATUS_SUCCESS;
+}
+
+EFI_STATUS
+XenPvBlockFrontInitialization (
+ IN XENBUS_PROTOCOL *XenbusIo,
+ IN CONST CHAR8 *NodeName,
+ OUT XEN_BLOCK_FRONT_DEVICE **DevPtr
+ )
+{
+ XENSTORE_TRANSACTION xbt;
+ CHAR8 *DeviceType;
+ blkif_sring_t *SharedRing;
+ XENSTORE_STATUS Status;
+ XEN_BLOCK_FRONT_DEVICE *Dev;
+ XenbusState State;
+ UINT64 Value;
+
+ ASSERT (NodeName != NULL);
+
+ Dev = AllocateZeroPool (sizeof (XEN_BLOCK_FRONT_DEVICE));
+ Dev->Signature = XEN_BLOCK_FRONT_SIGNATURE;
+ Dev->NodeName = NodeName;
+ Dev->XenbusIo = XenbusIo;
+ Dev->DeviceId = XenbusIo->DeviceId;
+
+ XenbusIo->XsRead (XenbusIo, XST_NIL, "device-type", (VOID**)&DeviceType);
+ if (AsciiStrCmp (DeviceType, "cdrom") == 0) {
+ Dev->MediaInfo.CdRom = TRUE;
+ } else {
+ Dev->MediaInfo.CdRom = FALSE;
+ }
+ FreePool (DeviceType);
+
+ Status = XenbusReadInteger (XenbusIo, "backend-id", FALSE, &Value);
+ if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT16_MAX) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to get backend-id (%d)\n",
+ Status));
+ goto Error;
+ }
+ Dev->DomainId = Value;
+ XenbusIo->EventChannelAllocate (XenbusIo, Dev->DomainId, &Dev->EventChannel);
+
+ SharedRing = (blkif_sring_t*) AllocatePages (1);
+ SHARED_RING_INIT (SharedRing);
+ FRONT_RING_INIT (&Dev->Ring, SharedRing, EFI_PAGE_SIZE);
+ XenbusIo->GrantAccess (XenbusIo,
+ Dev->DomainId,
+ (INTN) SharedRing >> EFI_PAGE_SHIFT,
+ FALSE,
+ &Dev->RingRef);
+
+Again:
+ Status = XenbusIo->XsTransactionStart (XenbusIo, &xbt);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_WARN, "XenPvBlk: Failed to start transaction, %d\n", Status));
+ }
+
+ Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName, "ring-ref", "%d",
+ Dev->RingRef);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write ring-ref.\n"));
+ goto AbortTransaction;
+ }
+ Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
+ "event-channel", "%d", Dev->EventChannel);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write event-channel.\n"));
+ goto AbortTransaction;
+ }
+ Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
+ "protocol", "%a", XEN_IO_PROTO_ABI_NATIVE);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write protocol.\n"));
+ goto AbortTransaction;
+ }
+
+ Status = XenbusIo->SetState (XenbusIo, xbt, XenbusStateConnected);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed switch state.\n"));
+ goto AbortTransaction;
+ }
+
+ Status = XenbusIo->XsTransactionEnd (XenbusIo, xbt, FALSE);
+ if (Status == XENSTORE_STATUS_EAGAIN) {
+ goto Again;
+ }
+
+ XenbusIo->RegisterWatchBackend (XenbusIo, "state", &Dev->StateWatchToken);
+
+ // Waiting backend
+ Status = XenPvBlkWaitForBackendState (Dev, XenbusStateConnected, &State);
+ if (Status != XENSTORE_STATUS_SUCCESS || State != XenbusStateConnected) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: backend for %a/%d not available, rc=%d state=%d\n",
+ XenbusIo->Type, XenbusIo->DeviceId, Status, State));
+ goto Error2;
+ }
+
+ Status = XenbusReadInteger (XenbusIo, "info", TRUE, &Value);
+ if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
+ goto Error2;
+ }
+ Dev->MediaInfo.VDiskInfo = Value;
+ if (Dev->MediaInfo.VDiskInfo & VDISK_READONLY) {
+ Dev->MediaInfo.ReadWrite = FALSE;
+ } else {
+ Dev->MediaInfo.ReadWrite = TRUE;
+ }
+
+ Status = XenbusReadInteger (XenbusIo, "sectors", TRUE, &Dev->MediaInfo.Sectors);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ goto Error2;
+ }
+
+ Status = XenbusReadInteger (XenbusIo, "sector-size", TRUE, &Value);
+ if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
+ goto Error2;
+ }
+ Dev->MediaInfo.SectorSize = Value;
+
+ // Default value
+ Value = 0;
+ Status = XenbusReadInteger (XenbusIo, "feature-barrier", TRUE, &Value);
+ if (Value == 1) {
+ Dev->MediaInfo.FeatureBarrier = TRUE;
+ } else {
+ Dev->MediaInfo.FeatureBarrier = FALSE;
+ }
+
+ // Default value
+ Value = 0;
+ XenbusReadInteger (XenbusIo, "feature-flush-cache", TRUE, &Value);
+ if (Value == 1) {
+ Dev->MediaInfo.FeatureFlushCache = TRUE;
+ } else {
+ Dev->MediaInfo.FeatureFlushCache = FALSE;
+ }
+
+ DEBUG ((EFI_D_INFO, "XenPvBlk: New disk with %d sectors of %d bytes\n",
+ Dev->MediaInfo.Sectors, Dev->MediaInfo.SectorSize));
+
+ *DevPtr = Dev;
+ return EFI_SUCCESS;
+
+Error2:
+ XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
+ XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
+ XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
+ goto Error;
+AbortTransaction:
+ XenbusIo->XsTransactionEnd (XenbusIo, xbt, TRUE);
+Error:
+ XenPvBlockFree (Dev);
+ return EFI_DEVICE_ERROR;
+}
+
+VOID
+XenPvBlockFrontShutdown (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ )
+{
+ XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
+ XENSTORE_STATUS Status;
+ UINT64 Value;
+
+ XenPvBlockSync (Dev);
+
+ Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosing);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: error while changing state to Closing: %d\n",
+ Status));
+ goto Close;
+ }
+
+ Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosing, NULL);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: error while waiting for closing backend state: %d\n",
+ Status));
+ goto Close;
+ }
+
+ Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosed);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: error while changing state to Closed: %d\n",
+ Status));
+ goto Close;
+ }
+
+ Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosed, NULL);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: error while waiting for closed backend state: %d\n",
+ Status));
+ goto Close;
+ }
+
+ Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateInitialising);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: error while changing state to initialising: %d\n",
+ Status));
+ goto Close;
+ }
+
+ while (TRUE) {
+ Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
+ if (Status != XENSTORE_STATUS_SUCCESS) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: error while waiting for new backend state: %d\n",
+ Status));
+ goto Close;
+ }
+ if (Value < XenbusStateInitWait || Value >= XenbusStateClosed) {
+ break;
+ }
+ DEBUG ((EFI_D_INFO,
+ "XenPvBlk: waiting backend state %d, current: %d\n",
+ XenbusStateInitWait, Value));
+ XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
+ }
+
+Close:
+ XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
+ XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
+ XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
+
+ XenPvBlockFree (Dev);
+}
+
+STATIC
+VOID
+XenPvBlockWaitSlot (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ )
+{
+ /* Wait for a slot */
+ if (RING_FULL (&Dev->Ring)) {
+ while (TRUE) {
+ XenPvBlockAsyncIoPoll (Dev);
+ if (!RING_FULL(&Dev->Ring)) {
+ break;
+ }
+ /* Really no slot, go to sleep. */
+ // XXX wait for an event on Dev->EventChannel
+ DEBUG ((EFI_D_INFO, "%a %d, waiting\n", __func__, __LINE__));
+ }
+ }
+}
+
+/* Issue an aio */
+VOID
+XenPvBlockAsyncIo (
+ IN OUT XEN_BLOCK_FRONT_IO *IoData,
+ IN BOOLEAN IsWrite
+ )
+{
+ XEN_BLOCK_FRONT_DEVICE *Dev = IoData->Dev;
+ XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
+ blkif_request_t *Request;
+ RING_IDX RingIndex;
+ BOOLEAN Notify;
+ INT32 NumSegments, Index;
+ UINTN Start, End;
+
+ // Can't io at non-sector-aligned location
+ ASSERT(!(IoData->Offset & (Dev->MediaInfo.SectorSize - 1)));
+ // Can't io non-sector-sized amounts
+ ASSERT(!(IoData->Size & (Dev->MediaInfo.SectorSize - 1)));
+ // Can't io non-sector-aligned buffer
+ ASSERT(!((UINTN) IoData->Buffer & (Dev->MediaInfo.SectorSize - 1)));
+
+ Start = (UINTN) IoData->Buffer & ~EFI_PAGE_MASK;
+ End = ((UINTN) IoData->Buffer + IoData->Size + EFI_PAGE_SIZE - 1) & ~EFI_PAGE_MASK;
+ IoData->NumRef = NumSegments = (End - Start) / EFI_PAGE_SIZE;
+
+ // XXX is this comment still true ?
+ /* qemu's IDE max multsect is 16 (8KB) and SCSI max DMA was set to 32KB,
+ * so max 44KB can't happen */
+ ASSERT (NumSegments <= BLKIF_MAX_SEGMENTS_PER_REQUEST);
+
+ XenPvBlockWaitSlot (Dev);
+ RingIndex = Dev->Ring.req_prod_pvt;
+ Request = RING_GET_REQUEST (&Dev->Ring, RingIndex);
+
+ Request->operation = IsWrite ? BLKIF_OP_WRITE : BLKIF_OP_READ;
+ Request->nr_segments = NumSegments;
+ Request->handle = Dev->DeviceId;
+ Request->id = (UINTN) IoData;
+ Request->sector_number = IoData->Offset / 512;
+
+ for (Index = 0; Index < NumSegments; Index++) {
+ Request->seg[Index].first_sect = 0;
+ Request->seg[Index].last_sect = EFI_PAGE_SIZE / 512 - 1;
+ }
+ Request->seg[0].first_sect = ((UINTN) IoData->Buffer & EFI_PAGE_MASK) / 512;
+ Request->seg[NumSegments - 1].last_sect =
+ (((UINTN) IoData->Buffer + IoData->Size - 1) & EFI_PAGE_MASK) / 512;
+ for (Index = 0; Index < NumSegments; Index++) {
+ UINTN Data = Start + Index * EFI_PAGE_SIZE;
+ if (!IsWrite) {
+ /* Trigger CoW if needed */
+ *(CHAR8 *)(Data + (Request->seg[Index].first_sect << 9)) = 0;
+ MemoryFence ();
+ }
+ XenbusIo->GrantAccess (XenbusIo, Dev->DomainId,
+ Data >> EFI_PAGE_SHIFT, IsWrite,
+ &Request->seg[Index].gref);
+ IoData->GrantRef[Index] = Request->seg[Index].gref;
+ }
+
+ Dev->Ring.req_prod_pvt = RingIndex + 1;
+
+ MemoryFence ();
+ RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
+
+ if (Notify) {
+ XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
+ }
+}
+
+STATIC
+VOID
+XenPvBlockAsyncIoCallback (
+ IN OUT XEN_BLOCK_FRONT_IO *IoData,
+ IN EFI_STATUS Status
+ )
+{
+ IoData->Data = (VOID *) 1;
+ IoData->Callback = NULL;
+}
+
+VOID
+XenPvBlockIo (
+ IN OUT XEN_BLOCK_FRONT_IO *IoData,
+ IN BOOLEAN IsWrite
+ )
+{
+ ASSERT (IoData->Callback == NULL);
+ // XXX Signal/Event instead ?
+ IoData->Callback = XenPvBlockAsyncIoCallback;
+ XenPvBlockAsyncIo (IoData, IsWrite);
+ IoData->Data = NULL;
+
+ while (TRUE) {
+ XenPvBlockAsyncIoPoll (IoData->Dev);
+ if (IoData->Data) {
+ break;
+ }
+ }
+}
+
+STATIC
+VOID
+XenPvBlockPushOperation (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev,
+ IN UINT8 Operation,
+ IN UINT64 Id
+ )
+{
+ INT32 Index;
+ blkif_request_t *Request;
+ BOOLEAN Notify;
+
+ XenPvBlockWaitSlot (Dev);
+ Index = Dev->Ring.req_prod_pvt;
+ Request = RING_GET_REQUEST(&Dev->Ring, Index);
+ Request->operation = Operation;
+ Request->nr_segments = 0;
+ Request->handle = Dev->DeviceId;
+ Request->id = Id;
+ /* Not needed anyway, but the backend will check it */
+ Request->sector_number = 0;
+ Dev->Ring.req_prod_pvt = Index + 1;
+ MemoryFence ();
+ RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
+ if (Notify) {
+ XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
+ XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
+ }
+}
+
+VOID
+XenPvBlockSync (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ )
+{
+ if (Dev->MediaInfo.ReadWrite) {
+ if (Dev->MediaInfo.FeatureBarrier) {
+ XenPvBlockPushOperation (Dev, BLKIF_OP_WRITE_BARRIER, 0);
+ }
+
+ if (Dev->MediaInfo.FeatureFlushCache) {
+ XenPvBlockPushOperation (Dev, BLKIF_OP_FLUSH_DISKCACHE, 0);
+ }
+ }
+
+ /* Note: This won't finish if another thread enqueues requests. */
+ while (TRUE) {
+ XenPvBlockAsyncIoPoll (Dev);
+ if (RING_FREE_REQUESTS (&Dev->Ring) == RING_SIZE (&Dev->Ring)) {
+ break;
+ }
+ }
+}
+
+VOID
+XenPvBlockAsyncIoPoll (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ )
+{
+ RING_IDX ProducerIndex, ConsumerIndex;
+ blkif_response_t *Response;
+ INT32 More;
+
+ do {
+ ProducerIndex = Dev->Ring.sring->rsp_prod;
+ /* Ensure we see queued responses up to 'ProducerIndex'. */
+ MemoryFence ();
+ ConsumerIndex = Dev->Ring.rsp_cons;
+
+ while (ConsumerIndex != ProducerIndex) {
+ XEN_BLOCK_FRONT_IO *IoData;
+ INT16 Status;
+
+ Response = RING_GET_RESPONSE (&Dev->Ring, ConsumerIndex);
+
+ IoData = (VOID *) (UINTN) Response->id;
+ Status = Response->status;
+
+ switch (Response->operation) {
+ case BLKIF_OP_READ:
+ case BLKIF_OP_WRITE:
+ {
+ INT32 Index;
+
+ if (Status != BLKIF_RSP_OKAY) {
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: "
+ "%a error %d on %a at offset %Ld, num bytes %Ld\n",
+ Response->operation == BLKIF_OP_READ ? "read" : "write",
+ Status, IoData->Dev->NodeName,
+ IoData->Offset,
+ IoData->Size));
+ }
+
+ for (Index = 0; Index < IoData->NumRef; Index++) {
+ Dev->XenbusIo->GrantEndAccess (Dev->XenbusIo, IoData->GrantRef[Index]);
+ }
+
+ break;
+ }
+
+ case BLKIF_OP_WRITE_BARRIER:
+ if (Status != BLKIF_RSP_OKAY) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: write barrier error %d\n", Status));
+ }
+ break;
+ case BLKIF_OP_FLUSH_DISKCACHE:
+ if (Status != BLKIF_RSP_OKAY) {
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: flush error %d\n", Status));
+ }
+ break;
+
+ default:
+ DEBUG ((EFI_D_ERROR,
+ "XenPvBlk: unrecognized block operation %d response (status %d)\n",
+ Response->operation, Status));
+ break;
+ }
+
+ Dev->Ring.rsp_cons = ++ConsumerIndex;
+ /* Nota: callback frees IoData itself */
+ if (IoData && IoData->Callback) {
+ IoData->Callback (IoData, Status ? EFI_DEVICE_ERROR : EFI_SUCCESS);
+ }
+ if (Dev->Ring.rsp_cons != ConsumerIndex) {
+ /* We reentered, we must not continue here */
+ break;
+ }
+ }
+
+ RING_FINAL_CHECK_FOR_RESPONSES (&Dev->Ring, More);
+ } while (More != 0);
+}
diff --git a/OvmfPkg/XenPvBlkDxe/BlockFront.h b/OvmfPkg/XenPvBlkDxe/BlockFront.h
new file mode 100644
index 0000000..d319e0d
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/BlockFront.h
@@ -0,0 +1,88 @@
+#include "XenPvBlkDxe.h"
+
+#include <IndustryStandard/Xen/event_channel.h>
+#include <IndustryStandard/Xen/io/blkif.h>
+
+typedef struct _XEN_BLOCK_FRONT_DEVICE XEN_BLOCK_FRONT_DEVICE;
+typedef struct _XEN_BLOCK_FRONT_IO XEN_BLOCK_FRONT_IO;
+
+struct _XEN_BLOCK_FRONT_IO
+{
+ XEN_BLOCK_FRONT_DEVICE *Dev;
+ UINT8 *Buffer;
+ UINTN Size;
+ UINTN Offset;
+
+ grant_ref_t GrantRef[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+ INT32 NumRef;
+
+ VOID (*Callback)(XEN_BLOCK_FRONT_IO *aiocb, EFI_STATUS Status);
+ VOID *Data;
+};
+
+typedef struct
+{
+ UINT64 Sectors;
+ UINT32 SectorSize;
+ UINT32 VDiskInfo;
+ BOOLEAN ReadWrite;
+ BOOLEAN CdRom;
+ BOOLEAN FeatureBarrier;
+ BOOLEAN FeatureFlushCache;
+} XEN_BLOCK_FRONT_MEDIA_INFO;
+
+#define XEN_BLOCK_FRONT_SIGNATURE SIGNATURE_32 ('X', 'p', 'v', 'B')
+struct _XEN_BLOCK_FRONT_DEVICE {
+ UINT32 Signature;
+ EFI_BLOCK_IO_PROTOCOL BlockIo;
+ domid_t DomainId;
+
+ blkif_front_ring_t Ring;
+ grant_ref_t RingRef;
+ evtchn_port_t EventChannel;
+ blkif_vdev_t DeviceId;
+
+ CONST CHAR8 *NodeName;
+ XEN_BLOCK_FRONT_MEDIA_INFO MediaInfo;
+
+ VOID *StateWatchToken;
+
+ XENBUS_PROTOCOL *XenbusIo;
+};
+
+#define XEN_BLOCK_FRONT_FROM_BLOCK_IO(b) \
+ CR (b, XEN_BLOCK_FRONT_DEVICE, BlockIo, XEN_BLOCK_FRONT_SIGNATURE)
+
+EFI_STATUS
+XenPvBlockFrontInitialization (
+ IN XENBUS_PROTOCOL *XenbusIo,
+ IN CONST CHAR8 *NodeName,
+ OUT XEN_BLOCK_FRONT_DEVICE **DevPtr
+ );
+
+VOID
+XenPvBlockFrontShutdown (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ );
+
+VOID
+XenPvBlockAsyncIo (
+ IN OUT XEN_BLOCK_FRONT_IO *IoData,
+ IN BOOLEAN IsWrite
+ );
+
+VOID
+XenPvBlockIo (
+ IN OUT XEN_BLOCK_FRONT_IO *IoData,
+ IN BOOLEAN IsWrite
+ );
+
+VOID
+XenPvBlockAsyncIoPoll (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ );
+
+VOID
+XenPvBlockSync (
+ IN XEN_BLOCK_FRONT_DEVICE *Dev
+ );
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
index 0ce1e86..7ec7d60 100644
--- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
@@ -12,6 +12,8 @@
#include "XenPvBlkDxe.h"
+#include "BlockFront.h"
+
///
/// Driver Binding Protocol instance
@@ -263,6 +265,7 @@ XenPvBlkDxeDriverBindingStart (
{
EFI_STATUS Status;
XENBUS_PROTOCOL *XenbusIo;
+ XEN_BLOCK_FRONT_DEVICE *Dev;
Status = gBS->OpenProtocol (
ControllerHandle,
@@ -276,6 +279,11 @@ XenPvBlkDxeDriverBindingStart (
return Status;
}
+ Status = XenPvBlockFrontInitialization (XenbusIo, XenbusIo->Node, &Dev);
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
return EFI_SUCCESS;
}
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
index 4444cda..a43d006 100644
--- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
@@ -30,6 +30,8 @@
XenPvBlkDxe.c
ComponentName.c
ComponentName.h
+ BlockFront.c
+ BlockFront.h
[LibraryClasses]
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo.
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (16 preceding siblings ...)
2014-07-16 15:15 ` [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client Anthony PERARD
@ 2014-07-16 15:15 ` Anthony PERARD
[not found] ` <1405523747-5024-3-git-send-email-anthony.perard@citrix.com>
` (9 subsequent siblings)
27 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-16 15:15 UTC (permalink / raw)
To: EDK2 devel; +Cc: Anthony PERARD, Xen Devel
Install the BlockIo protocol.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
---
OvmfPkg/XenPvBlkDxe/BlockIo.c | 256 ++++++++++++++++++++++++++++++++++++
OvmfPkg/XenPvBlkDxe/BlockIo.h | 109 +++++++++++++++
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 59 +++++++++
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h | 1 +
OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 2 +
5 files changed, 427 insertions(+)
create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.c
create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.h
diff --git a/OvmfPkg/XenPvBlkDxe/BlockIo.c b/OvmfPkg/XenPvBlkDxe/BlockIo.c
new file mode 100644
index 0000000..4ac2411
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/BlockIo.c
@@ -0,0 +1,256 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+#include "XenPvBlkDxe.h"
+
+#include "BlockFront.h"
+
+///
+/// Block I/O Media structure
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_BLOCK_IO_MEDIA gXenPvBlkDxeBlockIoMedia = {
+ 0, // MediaId
+ FALSE, // RemovableMedia
+ FALSE, // MediaPresent
+ FALSE, // LogicalPartition
+ TRUE, // ReadOnly
+ FALSE, // WriteCaching
+ 512, // BlockSize
+ 512, // IoAlign
+ 0, // LastBlock
+ 0, // LowestAlignedLba
+ 0, // LogicalBlocksPerPhysicalBlock
+ 0 // OptimalTransferLengthGranularity
+};
+
+///
+/// Block I/O Protocol instance
+///
+GLOBAL_REMOVE_IF_UNREFERENCED
+EFI_BLOCK_IO_PROTOCOL gXenPvBlkDxeBlockIo = {
+ EFI_BLOCK_IO_PROTOCOL_REVISION3, // Revision
+ &gXenPvBlkDxeBlockIoMedia, // Media
+ (EFI_BLOCK_RESET)XenPvBlkDxeBlockIoReset, // Reset
+ XenPvBlkDxeBlockIoReadBlocks, // ReadBlocks
+ XenPvBlkDxeBlockIoWriteBlocks, // WriteBlocks
+ XenPvBlkDxeBlockIoFlushBlocks // FlushBlocks
+};
+
+
+
+
+STATIC
+EFI_STATUS
+XenPvBlkDxeBlockIoReadWriteBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This,
+ IN UINT32 MediaId,
+ IN EFI_LBA Lba,
+ IN UINTN BufferSize,
+ IN OUT VOID *Buffer,
+ IN BOOLEAN IsWrite
+ )
+{
+ XEN_BLOCK_FRONT_IO IoData;
+ EFI_BLOCK_IO_MEDIA *Media;
+ UINTN NextOffset;
+
+ if (Buffer == NULL) {
+ DEBUG ((EFI_D_ERROR, "%a %d, buffer null\n", __func__, __LINE__));
+ return EFI_INVALID_PARAMETER;
+ }
+ if (BufferSize == 0) {
+ return EFI_SUCCESS;
+ }
+
+ Media = This->Media;
+
+ if (BufferSize % Media->BlockSize != 0) {
+ DEBUG ((EFI_D_ERROR, "%a %d, bad buffer size\n", __func__, __LINE__));
+ return EFI_BAD_BUFFER_SIZE;
+ }
+
+ if (Lba > Media->LastBlock ||
+ (BufferSize / Media->BlockSize) - 1 > Media->LastBlock - Lba) {
+ DEBUG ((EFI_D_ERROR, "%a %d, read too large\n", __func__, __LINE__));
+ return EFI_INVALID_PARAMETER;
+ }
+
+ if (IsWrite && Media->ReadOnly) {
+ DEBUG ((EFI_D_ERROR, "%a %d, read only\n", __func__, __LINE__));
+ return EFI_WRITE_PROTECTED;
+ }
+
+ if ((Media->IoAlign > 1) && (UINTN)Buffer & (Media->IoAlign - 1)) {
+ // This happen often when using grub2
+ VOID *NewBuffer;
+ EFI_STATUS Status;
+
+ /* Try again with a properly aligned buffer. */
+ NewBuffer = AllocateAlignedPages((BufferSize + EFI_PAGE_SIZE) / EFI_PAGE_SIZE,
+ Media->IoAlign);
+ if (!IsWrite) {
+ Status = XenPvBlkDxeBlockIoReadBlocks (This, MediaId,
+ Lba, BufferSize, NewBuffer);
+ CopyMem (Buffer, NewBuffer, BufferSize);
+ } else {
+ CopyMem (NewBuffer, Buffer, BufferSize);
+ Status = XenPvBlkDxeBlockIoWriteBlocks (This, MediaId,
+ Lba, BufferSize, NewBuffer);
+ }
+ FreeAlignedPages (NewBuffer, (BufferSize + EFI_PAGE_SIZE) / EFI_PAGE_SIZE);
+ return Status;
+ }
+
+
+ // XXX Check MediaId
+
+ IoData.Dev = XEN_BLOCK_FRONT_FROM_BLOCK_IO (This);
+ IoData.Offset = Lba * Media->BlockSize;
+ IoData.Callback = NULL;
+ IoData.Data = NULL;
+ IoData.Size = BufferSize;
+ IoData.Buffer = Buffer;
+
+ NextOffset = Lba * Media->BlockSize;
+ while (BufferSize > 0) {
+ // This should work if every Buffer is page aligned
+ // but block accept sector aligned, so FIXME
+ IoData.Size = MIN(EFI_PAGE_SIZE * 8,//BLKIF_MAX_SEGMENTS_PER_REQUEST,
+ BufferSize);
+ IoData.Buffer = Buffer;
+ IoData.Offset = NextOffset;
+ BufferSize -= IoData.Size;
+ Buffer = (VOID*) ((UINTN) Buffer + IoData.Size);
+ NextOffset += IoData.Size;
+ XenPvBlockIo (&IoData, IsWrite);
+ if (!IoData.Data) {
+ DEBUG ((EFI_D_ERROR, "%a %d, XenPvBlockIo no data...\n", __func__, __LINE__));
+ return EFI_DEVICE_ERROR;
+ }
+ }
+ return EFI_SUCCESS;
+}
+
+
+/**
+ Read BufferSize bytes from Lba into Buffer.
+
+ @param This Indicates a pointer to the calling context.
+ @param MediaId Id of the media, changes every time the media is replaced.
+ @param Lba The starting Logical Block Address to read from
+ @param BufferSize Size of Buffer, must be a multiple of device block size.
+ @param Buffer A pointer to the destination buffer for the data. The caller is
+ responsible for either having implicit or explicit ownership of the buffer.
+
+ @retval EFI_SUCCESS The data was read correctly from the device.
+ @retval EFI_DEVICE_ERROR The device reported an error while performing the read.
+ @retval EFI_NO_MEDIA There is no media in the device.
+ @retval EFI_MEDIA_CHANGED The MediaId does not matched the current device.
+ @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
+ @retval EFI_INVALID_PARAMETER The read request contains LBAs that are not valid,
+ or the buffer is not on proper alignment.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoReadBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This,
+ IN UINT32 MediaId,
+ IN EFI_LBA Lba,
+ IN UINTN BufferSize,
+ OUT VOID *Buffer
+ )
+{
+ return XenPvBlkDxeBlockIoReadWriteBlocks (This,
+ MediaId, Lba, BufferSize, Buffer, FALSE);
+}
+
+/**
+ Write BufferSize bytes from Lba into Buffer.
+
+ @param This Indicates a pointer to the calling context.
+ @param MediaId The media ID that the write request is for.
+ @param Lba The starting logical block address to be written. The caller is
+ responsible for writing to only legitimate locations.
+ @param BufferSize Size of Buffer, must be a multiple of device block size.
+ @param Buffer A pointer to the source buffer for the data.
+
+ @retval EFI_SUCCESS The data was written correctly to the device.
+ @retval EFI_WRITE_PROTECTED The device can not be written to.
+ @retval EFI_DEVICE_ERROR The device reported an error while performing the write.
+ @retval EFI_NO_MEDIA There is no media in the device.
+ @retval EFI_MEDIA_CHNAGED The MediaId does not matched the current device.
+ @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
+ @retval EFI_INVALID_PARAMETER The write request contains LBAs that are not valid,
+ or the buffer is not on proper alignment.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoWriteBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This,
+ IN UINT32 MediaId,
+ IN EFI_LBA Lba,
+ IN UINTN BufferSize,
+ IN VOID *Buffer
+ )
+{
+ return XenPvBlkDxeBlockIoReadWriteBlocks (This,
+ MediaId, Lba, BufferSize, Buffer, TRUE);
+}
+
+/**
+ Flush the Block Device.
+
+ @param This Indicates a pointer to the calling context.
+
+ @retval EFI_SUCCESS All outstanding data was written to the device
+ @retval EFI_DEVICE_ERROR The device reported an error while writting back the data
+ @retval EFI_NO_MEDIA There is no media in the device.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoFlushBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This
+ )
+{
+ XenPvBlockSync (XEN_BLOCK_FRONT_FROM_BLOCK_IO (This));
+ return EFI_SUCCESS;
+}
+
+/**
+ Reset the block device hardware.
+
+ @param[in] This Indicates a pointer to the calling context.
+ @param[in] ExtendedVerification Indicates that the driver may perform a more
+ exhausive verfication operation of the device
+ during reset.
+
+ @retval EFI_SUCCESS The device was reset.
+ @retval EFI_DEVICE_ERROR The device is not functioning properly and could
+ not be reset.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoReset (
+ IN EFI_BLOCK_IO2_PROTOCOL *This,
+ IN BOOLEAN ExtendedVerification
+ )
+{
+ // Is there anything to do ?
+ // since initialisation is done in Start (driver binding)
+ DEBUG ((EFI_D_INFO, "%a %d\n", __func__, __LINE__));
+ return EFI_SUCCESS;
+}
diff --git a/OvmfPkg/XenPvBlkDxe/BlockIo.h b/OvmfPkg/XenPvBlkDxe/BlockIo.h
new file mode 100644
index 0000000..ec0fe2b
--- /dev/null
+++ b/OvmfPkg/XenPvBlkDxe/BlockIo.h
@@ -0,0 +1,109 @@
+
+/** @file
+ TODO: Brief Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Detailed Description of UEFI Driver XenPvBlkDxe
+
+ TODO: Copyright for UEFI Driver XenPvBlkDxe
+
+ TODO: License for UEFI Driver XenPvBlkDxe
+
+**/
+
+/**
+ Read BufferSize bytes from Lba into Buffer.
+
+ @param This Indicates a pointer to the calling context.
+ @param MediaId Id of the media, changes every time the media is replaced.
+ @param Lba The starting Logical Block Address to read from
+ @param BufferSize Size of Buffer, must be a multiple of device block size.
+ @param Buffer A pointer to the destination buffer for the data. The caller is
+ responsible for either having implicit or explicit ownership of the buffer.
+
+ @retval EFI_SUCCESS The data was read correctly from the device.
+ @retval EFI_DEVICE_ERROR The device reported an error while performing the read.
+ @retval EFI_NO_MEDIA There is no media in the device.
+ @retval EFI_MEDIA_CHANGED The MediaId does not matched the current device.
+ @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
+ @retval EFI_INVALID_PARAMETER The read request contains LBAs that are not valid,
+ or the buffer is not on proper alignment.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoReadBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This,
+ IN UINT32 MediaId,
+ IN EFI_LBA Lba,
+ IN UINTN BufferSize,
+ OUT VOID *Buffer
+ );
+
+/**
+ Write BufferSize bytes from Lba into Buffer.
+
+ @param This Indicates a pointer to the calling context.
+ @param MediaId The media ID that the write request is for.
+ @param Lba The starting logical block address to be written. The caller is
+ responsible for writing to only legitimate locations.
+ @param BufferSize Size of Buffer, must be a multiple of device block size.
+ @param Buffer A pointer to the source buffer for the data.
+
+ @retval EFI_SUCCESS The data was written correctly to the device.
+ @retval EFI_WRITE_PROTECTED The device can not be written to.
+ @retval EFI_DEVICE_ERROR The device reported an error while performing the write.
+ @retval EFI_NO_MEDIA There is no media in the device.
+ @retval EFI_MEDIA_CHNAGED The MediaId does not matched the current device.
+ @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
+ @retval EFI_INVALID_PARAMETER The write request contains LBAs that are not valid,
+ or the buffer is not on proper alignment.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoWriteBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This,
+ IN UINT32 MediaId,
+ IN EFI_LBA Lba,
+ IN UINTN BufferSize,
+ IN VOID *Buffer
+ );
+
+/**
+ Flush the Block Device.
+
+ @param This Indicates a pointer to the calling context.
+
+ @retval EFI_SUCCESS All outstanding data was written to the device
+ @retval EFI_DEVICE_ERROR The device reported an error while writting back the data
+ @retval EFI_NO_MEDIA There is no media in the device.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoFlushBlocks (
+ IN EFI_BLOCK_IO_PROTOCOL *This
+ );
+
+/**
+ Reset the block device hardware.
+
+ @param[in] This Indicates a pointer to the calling context.
+ @param[in] ExtendedVerification Indicates that the driver may perform a more
+ exhausive verfication operation of the device
+ during reset.
+
+ @retval EFI_SUCCESS The device was reset.
+ @retval EFI_DEVICE_ERROR The device is not functioning properly and could
+ not be reset.
+
+**/
+EFI_STATUS
+EFIAPI
+XenPvBlkDxeBlockIoReset (
+ IN EFI_BLOCK_IO2_PROTOCOL *This,
+ IN BOOLEAN ExtendedVerification
+ );
+
+extern EFI_BLOCK_IO_MEDIA gXenPvBlkDxeBlockIoMedia;
+extern EFI_BLOCK_IO_PROTOCOL gXenPvBlkDxeBlockIo;
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
index 7ec7d60..3991e83 100644
--- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
@@ -266,6 +266,7 @@ XenPvBlkDxeDriverBindingStart (
EFI_STATUS Status;
XENBUS_PROTOCOL *XenbusIo;
XEN_BLOCK_FRONT_DEVICE *Dev;
+ EFI_BLOCK_IO_MEDIA *Media;
Status = gBS->OpenProtocol (
ControllerHandle,
@@ -284,6 +285,37 @@ XenPvBlkDxeDriverBindingStart (
return Status;
}
+ CopyMem (&Dev->BlockIo, &gXenPvBlkDxeBlockIo, sizeof (EFI_BLOCK_IO_PROTOCOL));
+ Media = AllocateCopyPool (sizeof (EFI_BLOCK_IO_MEDIA),
+ &gXenPvBlkDxeBlockIoMedia);
+ if (Dev->MediaInfo.VDiskInfo & VDISK_REMOVABLE) {
+ Media->RemovableMedia = TRUE;
+ }
+ Media->MediaPresent = TRUE;
+ Media->ReadOnly = !Dev->MediaInfo.ReadWrite;
+ if (Dev->MediaInfo.CdRom) {
+ // If it's a cdrom, the blocksize value need to be 2048 for OVMF to
+ // recognize it as a cdrom:
+ // MdeModulePkg/Universal/Disk/PartitionDxe/ElTorito.c
+ Media->BlockSize = 2048;
+ } else {
+ Media->BlockSize = Dev->MediaInfo.SectorSize;
+ }
+ Media->LastBlock = Dev->MediaInfo.Sectors - 1;
+ // XXX What about LowestAlignedLba, LogicalBlocksPerPhysicalBlock and OptimalTransferLengthGranularity?
+ Dev->BlockIo.Media = Media;
+
+ Status = gBS->InstallMultipleProtocolInterfaces (
+ &ControllerHandle,
+ &gEfiBlockIoProtocolGuid, &Dev->BlockIo,
+ NULL
+ );
+ if (EFI_ERROR (Status)) {
+ // XXX: uninit everythings
+ DEBUG ((EFI_D_ERROR, "XenPvBlk: install protocol fail: %r\n", Status));
+ return Status;
+ }
+
return EFI_SUCCESS;
}
@@ -322,6 +354,33 @@ XenPvBlkDxeDriverBindingStop (
IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
)
{
+ EFI_BLOCK_IO_PROTOCOL *BlockIo;
+ XEN_BLOCK_FRONT_DEVICE *Dev;
+ EFI_BLOCK_IO_MEDIA *Media;
+ EFI_STATUS Status;
+
+ Status = gBS->OpenProtocol (
+ ControllerHandle, &gEfiBlockIoProtocolGuid,
+ (VOID **)&BlockIo,
+ This->DriverBindingHandle, ControllerHandle,
+ EFI_OPEN_PROTOCOL_GET_PROTOCOL
+ );
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ Status = gBS->UninstallProtocolInterface (ControllerHandle,
+ &gEfiBlockIoProtocolGuid, BlockIo);
+ if (EFI_ERROR (Status)) {
+ return Status;
+ }
+
+ Media = BlockIo->Media;
+ Dev = XEN_BLOCK_FRONT_FROM_BLOCK_IO (BlockIo);
+ XenPvBlockFrontShutdown (Dev);
+
+ FreePool (Media);
+
gBS->CloseProtocol (ControllerHandle, &gXenbusProtocolGuid,
This->DriverBindingHandle, ControllerHandle);
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
index 186d0ee..3120137 100644
--- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
@@ -76,6 +76,7 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenPvBlkDxeComponentName;
//
#include "DriverBinding.h"
#include "ComponentName.h"
+#include "BlockIo.h"
diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
index a43d006..b379020 100644
--- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
+++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
@@ -32,6 +32,8 @@
ComponentName.h
BlockFront.c
BlockFront.h
+ BlockIo.c
+ BlockIo.h
[LibraryClasses]
--
Anthony PERARD
^ permalink raw reply related [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver.
[not found] ` <1405523747-5024-3-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 17:25 ` Konrad Rzeszutek Wilk
2014-07-18 10:30 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 17:25 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:31PM +0100, Anthony PERARD wrote:
> This includes Component Name and Driver Binding.
The title says driver, but the cover letter said 'bus driver'.
Should it have bus driver in it?
.. snip..
> +/**
> + Retrieves a Unicode string that is the user readable name of the controller
> + that is being managed by an EFI Driver.
> +
> + @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
> + @param ControllerHandle The handle of a controller that the driver specified by
> + This is managing. This handle specifies the controller
> + whose name is to be returned.
> + @param ChildHandle The handle of the child controller to retrieve the name
> + of. This is an optional parameter that may be NULL. It
> + will be NULL for device drivers. It will also be NULL
> + for a bus drivers that wish to retrieve the name of the
> + bus controller. It will not be NULL for a bus driver
> + that wishes to retrieve the name of a child controller.
.. snip..
> +EFI_STATUS
> +EFIAPI
> +XenbusDxeComponentNameGetControllerName (
> + IN EFI_COMPONENT_NAME2_PROTOCOL *This,
> + IN EFI_HANDLE ControllerHandle,
> + IN EFI_HANDLE ChildHandle OPTIONAL,
> + IN CHAR8 *Language,
> + OUT CHAR16 **ControllerName
> + )
> +{
> + EFI_STATUS Status;
> +
> + //
> + // ChildHandle must be NULL for a Device Driver
OK, but this is a bus driver in which case ChildHandle can be NULL or not (at
least that is what the comment says) ?
> + //
> + if (ChildHandle != NULL) {
> + return EFI_UNSUPPORTED;
> + }
> +
> + //
> + // Lookup name of controller specified by ControllerHandle
> + //
> + Status = EFI_UNSUPPORTED;
> +
> + return Status;
> +}
.. snip..
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> new file mode 100644
> index 0000000..14113ad
> --- /dev/null
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> @@ -0,0 +1,314 @@
> +
> +/** @file
> + TODO: Brief Description of UEFI Driver XenbusDxe
> +
> + TODO: Detailed Description of UEFI Driver XenbusDxe
> +
> + TODO: Copyright for UEFI Driver XenbusDxe
> +
> + TODO: License for UEFI Driver XenbusDxe
> +
> +**/
> +
> +#include <IndustryStandard/Pci.h>
> +#include <IndustryStandard/Acpi.h>
> +#include <Library/DebugLib.h>
> +
> +#include "XenbusDxe.h"
> +
> +
> +///
> +/// Driver Binding Protocol instance
> +///
> +EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding = {
> + XenbusDxeDriverBindingSupported,
> + XenbusDxeDriverBindingStart,
> + XenbusDxeDriverBindingStop,
> + XENBUS_DXE_VERSION,
> + NULL,
> + NULL
> +};
> +
> +
> +/**
> + Unloads an image.
> +
> + @param ImageHandle Handle that identifies the image to be unloaded.
> +
> + @retval EFI_SUCCESS The image has been unloaded.
> + @retval EFI_INVALID_PARAMETER ImageHandle is not a valid image handle.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenbusDxeUnload (
> + IN EFI_HANDLE ImageHandle
> + )
> +{
> + EFI_STATUS Status;
> +
> + EFI_HANDLE *HandleBuffer;
> + UINTN HandleCount;
> + UINTN Index;
> +
> +
> + Status = EFI_SUCCESS;
Is that neccessary considering the next line you overwrite it?
> +
> + //
> + // Retrieve array of all handles in the handle database
> + //
> + Status = gBS->LocateHandleBuffer (
> + AllHandles,
> + NULL,
> + NULL,
> + &HandleCount,
> + &HandleBuffer
> + );
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
> +
> + //
> + // Disconnect the current driver from handles in the handle database
> + //
> + for (Index = 0; Index < HandleCount; Index++) {
> + Status = gBS->DisconnectController (HandleBuffer[Index], gImageHandle, NULL);
> + }
Should you check the Status?
I guess it doesn't matter at all. In which case why don't you just
(void)gBS->DisconnectController ..
> +
> + //
> + // Free the array of handles
> + //
> + FreePool (HandleBuffer);
> +
> +
> + //
> + // Uninstall protocols installed in the driver entry point
> + //
> + Status = gBS->UninstallMultipleProtocolInterfaces (
> + ImageHandle,
> + &gEfiDriverBindingProtocolGuid, &gXenbusDxeDriverBinding,
> + &gEfiComponentNameProtocolGuid, &gXenbusDxeComponentName,
> + &gEfiComponentName2ProtocolGuid, &gXenbusDxeComponentName2,
> + NULL
> + );
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
> +
> +
> + //
> + // Do any additional cleanup that is required for this driver
> + //
> +
> + return EFI_SUCCESS;
> +}
> +
> +/**
> + This is the declaration of an EFI image entry point. This entry point is
> + the same for UEFI Applications, UEFI OS Loaders, and UEFI Drivers including
> + both device drivers and bus drivers.
> +
> + @param ImageHandle The firmware allocated handle for the UEFI image.
> + @param SystemTable A pointer to the EFI System Table.
> +
> + @retval EFI_SUCCESS The operation completed successfully.
> + @retval Others An unexpected error occurred.
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenbusDxeDriverEntryPoint (
> + IN EFI_HANDLE ImageHandle,
> + IN EFI_SYSTEM_TABLE *SystemTable
> + )
> +{
> + EFI_STATUS Status;
> +
> +
> + Status = EFI_SUCCESS;
Not needed.
> +
> +
> + //
> + // Install UEFI Driver Model protocol(s).
> + //
> + Status = EfiLibInstallDriverBindingComponentName2 (
> + ImageHandle,
> + SystemTable,
> + &gXenbusDxeDriverBinding,
> + ImageHandle,
> + &gXenbusDxeComponentName,
> + &gXenbusDxeComponentName2
> + );
> + ASSERT_EFI_ERROR (Status);
> +
> +
> + return Status;
> +}
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event.
[not found] ` <1405523747-5024-4-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 17:32 ` Konrad Rzeszutek Wilk
2014-07-18 10:32 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 17:32 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:32PM +0100, Anthony PERARD wrote:
> The ExitBoot event is used to disconnect from the device before the
> next operating system start using them.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> ---
> OvmfPkg/XenbusDxe/XenbusDxe.c | 39 ++++++++++++++++++++++++++++++++++++---
> OvmfPkg/XenbusDxe/XenbusDxe.h | 10 ++++++++++
> 2 files changed, 46 insertions(+), 3 deletions(-)
>
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> index 14113ad..2c3d9b8 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> @@ -16,7 +16,6 @@
>
> #include "XenbusDxe.h"
>
> -
> ///
> /// Driver Binding Protocol instance
> ///
> @@ -30,6 +29,8 @@ EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding = {
> };
>
>
> +XENBUS_DEVICE *mMyDevice;
> +
> /**
> Unloads an image.
>
> @@ -229,6 +230,19 @@ XenbusDxeDriverBindingSupported (
> return Status;
> }
>
> +VOID
> +EFIAPI
> +NotifyExitBoot (
> + IN EFI_EVENT Event,
> + IN VOID *Context
> + )
> +{
> + XENBUS_DEVICE *Dev = Context;
> +
> + gBS->DisconnectController(Dev->ControllerHandle,
> + Dev->This->DriverBindingHandle, NULL);
> +}
> +
> /**
> Starts a device controller or a bus controller.
>
> @@ -272,7 +286,22 @@ XenbusDxeDriverBindingStart (
> IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
> )
> {
> - return EFI_UNSUPPORTED;
> + EFI_STATUS Status;
> + XENBUS_DEVICE *Dev;
> +
> + Dev = AllocateZeroPool (sizeof (*Dev));
> + Dev->Signature = XENBUS_DEVICE_SIGNATURE;
> + Dev->This = This;
> + Dev->ControllerHandle = ControllerHandle;
> +
> + Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
> + NotifyExitBoot,
> + (VOID*) Dev,
> + &Dev->ExitBootEvent);
> + ASSERT_EFI_ERROR (Status);
> +
> + mMyDevice = Dev;
> + return EFI_SUCCESS;
> }
>
> /**
> @@ -310,5 +339,9 @@ XenbusDxeDriverBindingStop (
> IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
> )
> {
> - return EFI_UNSUPPORTED;
> + XENBUS_DEVICE *Dev = mMyDevice;
> +
> + gBS->CloseEvent (Dev->ExitBootEvent);
> +
> + return EFI_SUCCESS;
> }
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> index 6f22732..c1ca87a 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> @@ -75,4 +75,14 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
> #define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
>
>
> +typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
> +
> +#define XENBUS_DEVICE_SIGNATURE SIGNATURE_32 ('X','B','b','d')
I get the 'XB', but 'bd' ? Should it be 'XenB' ? Or perhaps
'XBus' ?
> +struct _XENBUS_DEVICE {
> + UINT32 Signature;
> + EFI_DRIVER_BINDING_PROTOCOL *This;
> + EFI_HANDLE ControllerHandle;
> + EFI_EVENT ExitBootEvent;
> +};
> +
> #endif
> --
> Anthony PERARD
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls.
[not found] ` <1405523747-5024-5-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 17:37 ` Konrad Rzeszutek Wilk
2014-07-17 9:43 ` Wei Liu
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 17:37 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
> diff --git a/OvmfPkg/XenbusDxe/XenHypercall.c b/OvmfPkg/XenbusDxe/XenHypercall.c
> new file mode 100644
> index 0000000..5e98dd5
> --- /dev/null
> +++ b/OvmfPkg/XenbusDxe/XenHypercall.c
> @@ -0,0 +1,101 @@
> +#include <PiDxe.h>
> +#include <Library/HobLib.h>
> +#include <Guid/XenInfo.h>
> +
> +#include "XenbusDxe.h"
> +#include "XenHypercall.h"
> +
> +#include <IndustryStandard/Xen/hvm/params.h>
> +#include <IndustryStandard/Xen/memory.h>
> +
> +EFI_STATUS
> +XenHyperpageInit (
> + XENBUS_DEVICE *Dev
> + )
> +{
> + EFI_HOB_GUID_TYPE *GuidHob;
> + EFI_XEN_INFO *XenInfo;
> +
> + GuidHob = GetFirstGuidHob (&gEfiXenInfoGuid);
> + if (GuidHob == NULL) {
> + DEBUG ((EFI_D_INFO, "XenbusInit: No xeninfo ?\n"));
> + return EFI_NOT_FOUND;
> + }
> + XenInfo = (EFI_XEN_INFO*)GET_GUID_HOB_DATA (GuidHob);
> + Dev->Hyperpage = XenInfo->HyperPages;
> + return EFI_SUCCESS;
> +}
Could you describe in the git commit a bit on how/where in hvmloader
this EfiXenInfoGuid is set and how the OVMF gets a handle on it?
Thank you.
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol.
[not found] ` <1405523747-5024-6-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 17:39 ` Konrad Rzeszutek Wilk
2014-07-18 10:36 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 17:39 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:34PM +0100, Anthony PERARD wrote:
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Could you describe a bit why you need to use it?
I think I know why - that is you need to poke at its IO ports
but it would be good to mention that in here.
> ---
> OvmfPkg/XenbusDxe/XenbusDxe.c | 16 ++++++++++++++++
> OvmfPkg/XenbusDxe/XenbusDxe.h | 1 +
> 2 files changed, 17 insertions(+)
>
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> index 63ea31b..ba5e8f4 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> @@ -290,11 +290,25 @@ XenbusDxeDriverBindingStart (
> {
> EFI_STATUS Status;
> XENBUS_DEVICE *Dev;
> + EFI_PCI_IO_PROTOCOL *PciIo;
> +
> + Status = gBS->OpenProtocol (
> + ControllerHandle,
> + &gEfiPciIoProtocolGuid,
> + (VOID **)&PciIo,
> + This->DriverBindingHandle,
> + ControllerHandle,
> + EFI_OPEN_PROTOCOL_BY_DRIVER
> + );
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
>
> Dev = AllocateZeroPool (sizeof (*Dev));
> Dev->Signature = XENBUS_DEVICE_SIGNATURE;
> Dev->This = This;
> Dev->ControllerHandle = ControllerHandle;
> + Dev->PciIo = PciIo;
>
> Status = XenHyperpageInit (Dev);
> ASSERT_EFI_ERROR (Status);
> @@ -351,5 +365,7 @@ XenbusDxeDriverBindingStop (
>
> gBS->CloseEvent (Dev->ExitBootEvent);
>
> + gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
> + This->DriverBindingHandle, ControllerHandle);
> return EFI_SUCCESS;
> }
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> index 6140f94..d57e3c8 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> @@ -84,6 +84,7 @@ struct _XENBUS_DEVICE {
> UINT32 Signature;
> EFI_DRIVER_BINDING_PROTOCOL *This;
> EFI_HANDLE ControllerHandle;
> + EFI_PCI_IO_PROTOCOL *PciIo;
> EFI_EVENT ExitBootEvent;
>
> VOID *Hyperpage;
> --
> Anthony PERARD
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol.
[not found] ` <1405523747-5024-7-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 17:42 ` Konrad Rzeszutek Wilk
2014-07-18 10:40 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 17:42 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:35PM +0100, Anthony PERARD wrote:
> This protocol will be used for communication between a PV driver (like a
> PV block driver) and the Xenbus/Xenstore.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> ---
> OvmfPkg/Include/Protocol/Xenbus.h | 62 +++++++++++++++++++++++++++++++++++++++
> OvmfPkg/OvmfPkg.dec | 2 ++
> OvmfPkg/XenbusDxe/ComponentName.c | 12 ++++++++
> OvmfPkg/XenbusDxe/XenbusDxe.h | 3 ++
> OvmfPkg/XenbusDxe/XenbusDxe.inf | 1 +
> 5 files changed, 80 insertions(+)
> create mode 100644 OvmfPkg/Include/Protocol/Xenbus.h
>
> diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
> new file mode 100644
> index 0000000..191cee1
> --- /dev/null
> +++ b/OvmfPkg/Include/Protocol/Xenbus.h
> @@ -0,0 +1,62 @@
> +
> +/** @file
> + TODO: Brief Description of Protocol Xenbus
> +
> + TODO: Detailed Description of Protocol Xenbus
> +
> + TODO: Copyright for Protocol Xenbus
> +
> + TODO: License for Protocol Xenbus
> +
> +**/
> +
> +#ifndef __PROTOCOL_XENBUS_H__
> +#define __PROTOCOL_XENBUS_H__
> +
> +#define XENBUS_PROTOCOL_GUID \
> + {0x3d3ca290, 0xb9a5, 0x11e3, {0xb7, 0x5d, 0xb8, 0xac, 0x6f, 0x7d, 0x65, 0xe6}}
For those who are not familiar with these GUIDs and live in the Linux world - how
are they constructed? Is there a chance for collision ?
> +
> +///
> +/// Forward declaration
> +///
> +typedef struct _XENBUS_PROTOCOL XENBUS_PROTOCOL;
> +
> +
> +#include <IndustryStandard/Xen/grant_table.h>
> +
> +///
> +/// Function prototypes
> +///
> +
> +typedef
> +EFI_STATUS
> +(EFIAPI *XENBUS_GRANT_ACCESS)(
> + IN XENBUS_PROTOCOL *This,
> + IN domid_t DomainId,
> + IN UINTN Frame,
> + IN BOOLEAN ReadOnly,
> + OUT grant_ref_t *refp
> + );
> +
> +typedef
> +EFI_STATUS
> +(EFIAPI *XENBUS_GRANT_END_ACCESS)(
> + IN XENBUS_PROTOCOL *This,
> + IN grant_ref_t Ref
> + );
> +
> +
> +///
> +/// Protocol structure
> +///
> +struct _XENBUS_PROTOCOL {
> + XENBUS_GRANT_ACCESS GrantAccess;
> + XENBUS_GRANT_END_ACCESS GrantEndAccess;
> + //
> + // Place protocol data fields here
> + //
> +};
> +
> +extern EFI_GUID gXenbusProtocolGuid;
> +
> +#endif
> diff --git a/OvmfPkg/OvmfPkg.dec b/OvmfPkg/OvmfPkg.dec
> index f829247..095d40a 100644
> --- a/OvmfPkg/OvmfPkg.dec
> +++ b/OvmfPkg/OvmfPkg.dec
> @@ -47,6 +47,8 @@
> [Protocols]
> gVirtioDeviceProtocolGuid = {0xfa920010, 0x6785, 0x4941, {0xb6, 0xec, 0x49, 0x8c, 0x57, 0x9f, 0x16, 0x0a}}
> gBlockMmioProtocolGuid = {0x6b558ce3, 0x69e5, 0x4c67, {0xa6, 0x34, 0xf7, 0xfe, 0x72, 0xad, 0xbe, 0x84}}
> + ## Include/Protocol/Xenbus.h
> + gXenbusProtocolGuid = {0x3d3ca290, 0xb9a5, 0x11e3, {0xb7, 0x5d, 0xb8, 0xac, 0x6f, 0x7d, 0x65, 0xe6}}
Formatting? Looks like the others use tabs?
>
> [PcdsFixedAtBuild]
> gUefiOvmfPkgTokenSpaceGuid.PcdOvmfPeiMemFvBase|0x0|UINT32|0
> diff --git a/OvmfPkg/XenbusDxe/ComponentName.c b/OvmfPkg/XenbusDxe/ComponentName.c
> index 2cf11b5..fd46cc8 100644
> --- a/OvmfPkg/XenbusDxe/ComponentName.c
> +++ b/OvmfPkg/XenbusDxe/ComponentName.c
> @@ -140,6 +140,18 @@ XenbusDxeComponentNameGetControllerName (
> }
>
> //
> + // Make sure this driver is currently managing ControllerHandle
> + //
> + Status = EfiTestManagedDevice (
> + ControllerHandle,
> + gXenbusDxeDriverBinding.DriverBindingHandle,
> + &gXenbusProtocolGuid
> + );
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
> +
> + //
> // Lookup name of controller specified by ControllerHandle
> //
> Status = EFI_UNSUPPORTED;
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> index d57e3c8..975fb6b 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> @@ -42,6 +42,9 @@
> //
> // Produced Protocols
> //
> +// Xen interface version used:
> +#define __XEN_INTERFACE_VERSION__ 0x00040400
> +#include <Protocol/Xenbus.h>
>
>
> //
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> index 8b69f93..1a6b131 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> @@ -55,6 +55,7 @@
> gEfiPciIoProtocolGuid
> gEfiComponentName2ProtocolGuid
> gEfiComponentNameProtocolGuid
> + gXenbusProtocolGuid
>
>
> [Guids]
> --
> Anthony PERARD
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions.
[not found] ` <1405523747-5024-9-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 17:48 ` Konrad Rzeszutek Wilk
2014-07-18 10:53 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 17:48 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Steven Smith, Grzegorz Milos, Xen Devel
On Wed, Jul 16, 2014 at 04:15:37PM +0100, Anthony PERARD wrote:
> There are used to grant access of pages to other Xen domains.
>
> This code originaly comes from the Xen Project, and more precisely from
> MiniOS.
Should you mention something about its license?
>
> Signed-off-by: Steven Smith <sos22@cam.ac.uk>
> Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> ---
> OvmfPkg/XenbusDxe/GrantTable.c | 204 ++++++++++++++++++++++++++++++++++++++++
> OvmfPkg/XenbusDxe/GrantTable.h | 34 +++++++
> OvmfPkg/XenbusDxe/XenbusDxe.c | 15 +++
> OvmfPkg/XenbusDxe/XenbusDxe.inf | 2 +
> 4 files changed, 255 insertions(+)
> create mode 100644 OvmfPkg/XenbusDxe/GrantTable.c
> create mode 100644 OvmfPkg/XenbusDxe/GrantTable.h
>
> diff --git a/OvmfPkg/XenbusDxe/GrantTable.c b/OvmfPkg/XenbusDxe/GrantTable.c
> new file mode 100644
> index 0000000..97bd15a
> --- /dev/null
> +++ b/OvmfPkg/XenbusDxe/GrantTable.c
> @@ -0,0 +1,204 @@
> +/*
> + ****************************************************************************
> + * (C) 2006 - Cambridge University
> + ****************************************************************************
> + *
> + * File: GrantTable.c
> + * Author: Steven Smith (sos22@cam.ac.uk)
> + * Changes: Grzegorz Milos (gm281@cam.ac.uk)
> + *
> + * Date: July 2006
> + *
> + * Environment: Xen Minimal OS
> + * Description: Simple grant tables implementation. About as stupid as it's
> + * possible to be and still work.
> + *
> + ****************************************************************************
> + */
> +#include "XenbusDxe.h"
> +
> +#include <IndustryStandard/Xen/memory.h>
> +
> +#include "XenHypercall.h"
> +
> +#include "GrantTable.h"
> +#include "InterlockedCompareExchange16.h"
> +
> +#define NR_RESERVED_ENTRIES 8
> +
> +/* NR_GRANT_FRAMES must be less than or equal to that configured in Xen */
> +#define NR_GRANT_FRAMES 4
> +#define NR_GRANT_ENTRIES (NR_GRANT_FRAMES * EFI_PAGE_SIZE / sizeof(grant_entry_v1_t))
> +
> +STATIC grant_entry_v1_t *GrantTable = NULL;
> +STATIC grant_ref_t GrantList[NR_GRANT_ENTRIES];
> +#ifdef GNT_DEBUG
> +STATIC BOOLEAN GrantInUseList[NR_GRANT_ENTRIES];
> +#endif
> +
> +STATIC
> +VOID
> +XenGrantTablePutFreeEntry (
> + grant_ref_t Ref
> + )
> +{
> + //local_irq_save(flags);
> + /* MemoryFence (); */
> + // lock ?
You just need some form of locking. If EFI has some simple Mutex() you can
use that.
> +#ifdef GNT_DEBUG
> + ASSERT (GrantInUseList[Ref]);
> + GrantInUseList[Ref] = FALSE;
> +#endif
> + GrantList[Ref] = GrantList[0];
> + GrantList[0] = Ref;
> + //local_irq_restore(flags);
> + /* MemoryFence (); */
> +}
> +
> +STATIC
> +grant_ref_t
> +XenGrantTableGetFreeEntry (
> + VOID
> + )
> +{
> + UINTN Ref;
> + /* local_irq_save(flags); */
> + // lock ??
> + Ref = GrantList[0];
> + ASSERT (Ref >= NR_RESERVED_ENTRIES && Ref < NR_GRANT_ENTRIES);
> + GrantList[0] = GrantList[Ref];
> +#ifdef GNT_DEBUG
> + ASSERT (!GrantInUseList[Ref]);
> + GrantInUseList[Ref] = TRUE;
> +#endif
> + /* local_irq_restore(flags); */
> + return Ref;
> +}
> +
> +STATIC
> +grant_ref_t
> +XenGrantTableGrantAccess (
> + IN domid_t DomainId,
> + IN UINTN Frame,
> + IN BOOLEAN ReadOnly
> + )
> +{
> + grant_ref_t Ref;
> + UINT32 Flags;
> +
> + ASSERT (GrantTable != NULL);
> + Ref = XenGrantTableGetFreeEntry ();
> + GrantTable[Ref].frame = Frame;
> + GrantTable[Ref].domid = DomainId;
> + MemoryFence ();
> + Flags = GTF_permit_access;
> + if (ReadOnly) {
> + Flags |= GTF_readonly;
> + }
> + GrantTable[Ref].flags = Flags;
> +
> + return Ref;
> +}
> +
> +STATIC
> +EFI_STATUS
> +XenGrantTableEndAccess (
> + grant_ref_t Ref
> + )
> +{
> + UINT16 Flags, OldFlags;
> +
> + ASSERT (GrantTable != NULL);
> + ASSERT (Ref >= NR_RESERVED_ENTRIES && Ref < NR_GRANT_ENTRIES);
> +
> + OldFlags = GrantTable[Ref].flags;
> + do {
> + if ((Flags = OldFlags) & (GTF_reading | GTF_writing)) {
> + DEBUG ((EFI_D_WARN, "WARNING: g.e. still in use! (%x)\n", Flags));
> + return EFI_NOT_READY;
> + }
> + OldFlags = InterlockedCompareExchange16 (&GrantTable[Ref].flags, Flags, 0);
> + } while (OldFlags != Flags);
> +
> + XenGrantTablePutFreeEntry (Ref);
> + return EFI_SUCCESS;
> +}
> +
> +VOID
> +XenGrantTableInit (
> + IN XENBUS_DEVICE *Dev,
> + IN UINT64 MmioAddr
> + )
> +{
> + xen_add_to_physmap_t Args;
> + INTN Index;
> + INTN ReturnCode;
> +
> +#ifdef GNT_DEBUG
> + SetMem(GrantInUseList, sizeof (GrantInUseList), 1);
> +#endif
> + for (Index = NR_RESERVED_ENTRIES; Index < NR_GRANT_ENTRIES; Index++) {
> + XenGrantTablePutFreeEntry (Index);
> + }
> +
> + GrantTable = (VOID*)(UINTN) MmioAddr;
> + for (Index = 0; Index < NR_GRANT_FRAMES; Index++) {
> + Args.domid = DOMID_SELF;
> + Args.idx = Index;
> + Args.space = XENMAPSPACE_grant_table;
> + Args.gpfn = (((xen_pfn_t) GrantTable) >> EFI_PAGE_SHIFT) + Index;
> + ReturnCode = XenHypercallMemoryOp (Dev, XENMEM_add_to_physmap, &Args);
> + if (ReturnCode != 0) {
> + DEBUG ((EFI_D_ERROR, "Xen GrantTable, add_to_physmap hypercall error: %d\n", ReturnCode));
> + }
> + }
> +}
> +
> +VOID
> +XenGrantTableDeinit (
> + XENBUS_DEVICE *Dev
> + )
> +{
> + INTN ReturnCode, Index;
> + xen_remove_from_physmap_t Args;
> +
> + if (GrantTable == NULL) {
> + return;
> + }
> +
> + for (Index = NR_GRANT_FRAMES - 1; Index >= 0; Index--) {
> + Args.domid = DOMID_SELF;
> + Args.gpfn = (((xen_pfn_t) GrantTable) >> EFI_PAGE_SHIFT) + Index;
> + DEBUG ((EFI_D_INFO, "%a %d, removing %X\n", __func__, __LINE__, Args.gpfn));
> + ReturnCode = XenHypercallMemoryOp (Dev, XENMEM_remove_from_physmap, &Args);
> + if (ReturnCode != 0) {
> + DEBUG ((EFI_D_ERROR, "Xen GrantTable, remove_from_physmap hypercall error: %d\n", ReturnCode));
> + }
> + }
> + // XXX remove it from the physmap?
Didn't you just do that?
> + GrantTable = NULL;
> +}
> +
> +EFI_STATUS
> +EFIAPI
> +XenbusGrantAccess (
> + IN XENBUS_PROTOCOL *This,
> + IN domid_t DomainId,
> + IN UINTN Frame, // MFN
> + IN BOOLEAN ReadOnly,
> + OUT grant_ref_t *RefPtr
> + )
> +{
> + *RefPtr = XenGrantTableGrantAccess (DomainId, Frame, ReadOnly);
> + return EFI_SUCCESS;
> +}
> +
> +EFI_STATUS
> +EFIAPI
> +XenbusGrantEndAccess (
> + IN XENBUS_PROTOCOL *This,
> + IN grant_ref_t Ref
> + )
> +{
> + return XenGrantTableEndAccess (Ref);
> +}
> diff --git a/OvmfPkg/XenbusDxe/GrantTable.h b/OvmfPkg/XenbusDxe/GrantTable.h
> new file mode 100644
> index 0000000..05ab4be
> --- /dev/null
> +++ b/OvmfPkg/XenbusDxe/GrantTable.h
> @@ -0,0 +1,34 @@
> +#ifndef __GNTTAB_H__
> +#define __GNTTAB_H__
> +
> +#include <IndustryStandard/Xen/grant_table.h>
> +
> +VOID
> +XenGrantTableInit (
> + IN XENBUS_DEVICE *Dev,
> + IN UINT64 MmioAddr
> + );
> +
> +VOID
> +XenGrantTableDeinit (
> + IN XENBUS_DEVICE *Dev
> + );
> +
> +EFI_STATUS
> +EFIAPI
> +XenbusGrantAccess (
> + IN XENBUS_PROTOCOL *This,
> + IN domid_t DomainId,
> + IN UINTN Frame, // MFN
> + IN BOOLEAN ReadOnly,
> + OUT grant_ref_t *RefPtr
> + );
> +
> +EFI_STATUS
> +EFIAPI
> +XenbusGrantEndAccess (
> + IN XENBUS_PROTOCOL *This,
> + IN grant_ref_t Ref
> + );
> +
> +#endif /* !__GNTTAB_H__ */
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> index ba5e8f4..b19055d 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> @@ -17,6 +17,7 @@
> #include "XenbusDxe.h"
>
> #include "XenHypercall.h"
> +#include "GrantTable.h"
>
> ///
> /// Driver Binding Protocol instance
> @@ -291,6 +292,8 @@ XenbusDxeDriverBindingStart (
> EFI_STATUS Status;
> XENBUS_DEVICE *Dev;
> EFI_PCI_IO_PROTOCOL *PciIo;
> + EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR *BarDesc;
> + UINT64 MmioAddr;
>
> Status = gBS->OpenProtocol (
> ControllerHandle,
> @@ -310,12 +313,23 @@ XenbusDxeDriverBindingStart (
> Dev->ControllerHandle = ControllerHandle;
> Dev->PciIo = PciIo;
>
> + Status = PciIo->GetBarAttributes (PciIo, PCI_BAR_IDX1, NULL, (VOID**) &BarDesc);
I think we have a spec somewhere for this device. Do you think
it might make sense to reference it here?
> + ASSERT_EFI_ERROR (Status);
> + ASSERT (BarDesc->ResType == ACPI_ADDRESS_SPACE_TYPE_MEM);
I sooo hope the spec mandates that it is always this type of BAR.
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself.
[not found] ` <1405523747-5024-15-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 19:04 ` Konrad Rzeszutek Wilk
2014-07-18 11:04 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 19:04 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:43PM +0100, Anthony PERARD wrote:
> This is a bus-like on top of Xenstore. It will look for advertised
> ParaVirtualized devices and initialize them by producing XenBus
> protocol.
>
> Origin: FreeBSD 10.0
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> ---
> OvmfPkg/Include/Protocol/Xenbus.h | 9 +
> OvmfPkg/XenbusDxe/XenBus.c | 360 ++++++++++++++++++++++++++++++++++++++
> OvmfPkg/XenbusDxe/XenBus.h | 76 ++++++++
> OvmfPkg/XenbusDxe/XenbusDxe.c | 61 +++++++
> OvmfPkg/XenbusDxe/XenbusDxe.h | 18 ++
> OvmfPkg/XenbusDxe/XenbusDxe.inf | 3 +
> 6 files changed, 527 insertions(+)
> create mode 100644 OvmfPkg/XenbusDxe/XenBus.c
> create mode 100644 OvmfPkg/XenbusDxe/XenBus.h
>
> diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
> index ef4d0c2..25408c1 100644
> --- a/OvmfPkg/Include/Protocol/Xenbus.h
> +++ b/OvmfPkg/Include/Protocol/Xenbus.h
> @@ -113,6 +113,14 @@ XENSTORE_STATUS
> );
>
> typedef
> +XENSTORE_STATUS
> +(EFIAPI *XENBUS_SET_STATE)(
> + IN XENBUS_PROTOCOL *This,
> + IN XENSTORE_TRANSACTION Transaction,
> + IN XenbusState State
> + );
> +
> +typedef
> EFI_STATUS
> (EFIAPI *XENBUS_GRANT_ACCESS)(
> IN XENBUS_PROTOCOL *This,
> @@ -171,6 +179,7 @@ struct _XENBUS_PROTOCOL {
> XENBUS_XS_REMOVE XsRemove;
> XENBUS_XS_TRANSACTION_START XsTransactionStart;
> XENBUS_XS_TRANSACTION_END XsTransactionEnd;
> + XENBUS_SET_STATE SetState;
>
> XENBUS_GRANT_ACCESS GrantAccess;
> XENBUS_GRANT_END_ACCESS GrantEndAccess;
> diff --git a/OvmfPkg/XenbusDxe/XenBus.c b/OvmfPkg/XenbusDxe/XenBus.c
> new file mode 100644
> index 0000000..b0cf1ba
> --- /dev/null
> +++ b/OvmfPkg/XenbusDxe/XenBus.c
> @@ -0,0 +1,360 @@
> +/******************************************************************************
> + * Copyright (C) 2010 Spectra Logic Corporation
> + * Copyright (C) 2008 Doug Rabson
> + * Copyright (C) 2005 Rusty Russell, IBM Corporation
> + * Copyright (C) 2005 Mike Wray, Hewlett-Packard
> + * Copyright (C) 2005 XenSource Ltd
> + *
> + * This file may be distributed separately from the Linux kernel, or
> + * incorporated into other software packages, subject to the following license:
> + *
> + * Permission is hereby granted, free of charge, to any person obtaining a copy
> + * of this source file (the "Software"), to deal in the Software without
> + * restriction, including without limitation the rights to use, copy, modify,
> + * merge, publish, distribute, sublicense, and/or sell copies of the Software,
> + * and to permit persons to whom the Software is furnished to do so, subject to
> + * the following conditions:
> + *
> + * The above copyright notice and this permission notice shall be included in
> + * all copies or substantial portions of the Software.
> + *
> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
> + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
> + * IN THE SOFTWARE.
> + */
> +
> +#include "XenbusDxe.h"
> +
> +#include <Library/PrintLib.h>
> +
> +#include "XenBus.h"
> +#include "GrantTable.h"
> +#include "Xenstore.h"
> +#include "EventChannel.h"
> +
> +#include <IndustryStandard/Xen/io/xenbus.h>
> +
> +STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData;
> +
> +STATIC XENBUS_DEVICE_PATH gXenbusDevicePathTemplate = {
> + .Vendor.Header.Type = HARDWARE_DEVICE_PATH,
> + .Vendor.Header.SubType = HW_VENDOR_DP,
> + .Vendor.Header.Length[0] = (UINT8) sizeof (XENBUS_DEVICE_PATH),
> + .Vendor.Header.Length[1] = (UINT8) (sizeof (XENBUS_DEVICE_PATH) >> 8),
> + .Vendor.Guid = XENBUS_PROTOCOL_GUID,
> + .Type = 0,
> + .DeviceId = 0
> +};
> +
> +
> +/**
> + * Search our internal record of configured devices (not the XenStore)
> + * to determine if the XenBus device indicated by \a node is known to
> + * the system.
> + *
> + * \param Dev The XenBus bus instance to search for device children.
> + * \param node The XenStore node path for the device to find.
> + *
> + * \return The device_t of the found device if any, or NULL.
> + *
> + * \note device_t is a pointer type, so it can be compared against
> + * NULL for validity.
> + */
> +STATIC
> +XENBUS_PRIVATE_DATA *
> +XenbusDeviceInitialized (
> + IN XENBUS_DEVICE *Dev,
> + IN CONST CHAR8 *Node
> + )
> +{
> + LIST_ENTRY *Entry;
> + XENBUS_PRIVATE_DATA *Child;
> + XENBUS_PRIVATE_DATA *Result;
> +
> + if (IsListEmpty (&Dev->ChildList)) {
> + return NULL;
> + }
> +
> + Result = NULL;
> + for (Entry = GetFirstNode (&Dev->ChildList);
> + !IsNodeAtEnd (&Dev->ChildList, Entry);
> + Entry = GetNextNode (&Dev->ChildList, Entry)) {
> + Child = XENBUS_PRIVATE_DATA_FROM_LINK (Entry);
> + if (!AsciiStrCmp (Child->XenbusIo.Node, Node)) {
> + Result = Child;
> + break;
> + }
> + }
> +
> + return (Result);
> +}
> +
> +STATIC
> +XenbusState
> +XenbusReadDriverState (
> + IN CONST CHAR8 *Path
> + )
> +{
> + XenbusState State;
> + CHAR8 *Ptr;
> + XENSTORE_STATUS Status;
> +
> + Status = XenstoreRead (XST_NIL, Path, "state", NULL, (VOID **)&Ptr);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + State = XenbusStateClosed;
> + } else {
> + State = AsciiStrDecimalToUintn (Ptr);
> + FreePool (Ptr);
> + }
I think the FreePool(Ptr) should be here. You want to it
irregardless of what the error might be (thought the implemenation
in XenstoreRead et.all is pretty clear that we won't initialize
*Ptr to anything if !Status. But just in case this changes in
the future..
> +
> + return State;
> +}
> +
> +STATIC
> +EFI_STATUS
> +XenbusAddDevice (
> + XENBUS_DEVICE *Dev,
> + CONST CHAR8 *Type,
> + CONST CHAR8 *Id)
> +{
> + CHAR8 DevicePath[XENSTORE_ABS_PATH_MAX];
> + XENSTORE_STATUS StatusXenstore;
> + XENBUS_PRIVATE_DATA *Private;
> + EFI_STATUS Status;
> + XENBUS_DEVICE_PATH *TempXenbusPath;
> + VOID *ChildPciIo;
> +
> + AsciiSPrint (DevicePath, sizeof (DevicePath),
> + XENBUS_XENSTORE_NODE "/%a/%a", Type, Id);
> +
> + Status = EFI_NOT_STARTED; //ENXIO; // Device not configured
> +
> + if (XenstorePathExists (XST_NIL, DevicePath, "")) {
> + XENBUS_PRIVATE_DATA *Child;
> + enum xenbus_state State;
> + CHAR8 *BackendPath;
> +
> + Child = XenbusDeviceInitialized (Dev, DevicePath);
> + if (Child != NULL) {
> + /*
> + * We are already tracking this node
> + */
> + Status = EFI_SUCCESS;
> + goto out;
> + }
> +
> + State = XenbusReadDriverState (DevicePath);
> + if (State != XenbusStateInitialising) {
> + /*
> + * Device is not new, so ignore it. This can
> + * happen if a device is going away after
> + * switching to Closed.
> + */
> + DEBUG ((EFI_D_INFO, "Xenbus: Device %a ignored. "
> + "State %d\n", DevicePath, State));
> + Status = EFI_SUCCESS;
> + goto out;
> + }
> +
> + StatusXenstore = XenstoreRead (XST_NIL, DevicePath, "backend",
> + NULL, (VOID **) &BackendPath);
> + if (StatusXenstore != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR, "xenbus: %a no backend path.\n", DevicePath));
> + Status = EFI_NOT_FOUND;
> + goto out;
> + }
> +
> + Private = AllocateCopyPool (sizeof *Private, &gXenbusPrivateData);
> + InsertTailList (&Dev->ChildList, &Private->Link);
> +
> + Private->XenbusIo.Type = AsciiStrDup (Type);
> + Private->XenbusIo.Node = AsciiStrDup (DevicePath);
> + Private->XenbusIo.Backend = BackendPath;
> + Private->XenbusIo.DeviceId = AsciiStrDecimalToUintn (Id);
> + Private->Dev = Dev;
> +
> + TempXenbusPath = AllocateCopyPool (sizeof (XENBUS_DEVICE_PATH),
> + &gXenbusDevicePathTemplate);
> + if (!AsciiStrCmp (Private->XenbusIo.Type, "vbd")) {
> + TempXenbusPath->Type = XENBUS_DEVICE_PATH_TYPE_VBD;
> + }
> + TempXenbusPath->DeviceId = Private->XenbusIo.DeviceId;
> + Private->DevicePath = (XENBUS_DEVICE_PATH *)AppendDevicePathNode (
> + Dev->DevicePath,
> + &TempXenbusPath->Vendor.Header);
Could you provide a comment explaining why you do that please?
> + FreePool (TempXenbusPath);
> +
> + Status = gBS->InstallMultipleProtocolInterfaces (
> + &Private->Handle,
> + &gEfiDevicePathProtocolGuid, Private->DevicePath,
> + &gXenbusProtocolGuid, &Private->XenbusIo,
> + NULL);
> + if (EFI_ERROR (Status)) {
Shouldn't you free 'Private' and 'BackendPath' ? And
also de-link yourself from 'Dev->ChildList' ?
> + return Status;
> + }
> +
> + Status = gBS->OpenProtocol (Dev->ControllerHandle,
> + &gEfiPciIoProtocolGuid,
> + &ChildPciIo, Dev->This->DriverBindingHandle,
> + Private->Handle,
> + EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER);
> + if (EFI_ERROR (Status)) {
> + DEBUG ((EFI_D_ERROR, "open by child controller fail (%r)\n",
> + Status));
> + }
> + } else {
> + DEBUG ((EFI_D_INFO, "Xenbus: does not exist: %a\n", DevicePath));
> + }
> +
> +out:
> + return Status;
> +}
> +
> +/**
> + * \brief Enumerate all devices of the given type on this bus.
> + *
> + * \param Dev NewBus device_t for this XenBus front bus instance.
> + * \param type String indicating the device sub-tree (e.g. "vfb", "vif")
> + * to enumerate.
> + *
> + * Devices that are found are entered into the NewBus hierarchy via
> + * xenbusb_add_device(). xenbusb_add_device() ignores duplicate detects
> + * and ignores duplicate devices, so it can be called unconditionally
> + * for any device found in the XenStore.
> + */
> +STATIC
> +VOID
> +XenbusEnumerateDeviceType (
> + XENBUS_DEVICE *Dev,
> + CONST CHAR8 *Type
> + )
> +{
> + CONST CHAR8 **Directory;
> + UINTN Index;
> + UINT32 Count;
> + XENSTORE_STATUS Status;
> +
> + Status = XenstoreListDirectory (XST_NIL,
> + XENBUS_XENSTORE_NODE, Type,
> + &Count, &Directory);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + return;
> + }
> + for (Index = 0; Index < Count; Index++) {
> + XenbusAddDevice (Dev, Type, Directory[Index]);
> + }
> +
> + FreePool (Directory);
> +}
> +
> +
> +/**
> + * \brief Enumerate the devices on a XenBus bus and register them with
> + * the NewBus device tree.
> + *
> + * xenbusb_enumerate_bus() will create entries (in state DS_NOTPRESENT)
> + * for nodes that appear in the XenStore, but will not invoke probe/attach
> + * operations on drivers. Probe/Attach processing must be separately
> + * performed via an invocation of xenbusb_probe_children(). This is usually
> + * done via the xbs_probe_children task.
> + *
> + * \param Dev XenBus Bus device softc of the owner of the bus to enumerate.
> + *
> + * \return On success, 0. Otherwise an errno value indicating the
> + * type of failure.
> + */
> +XENSTORE_STATUS
> +XenbusEnumerateBus (
> + XENBUS_DEVICE *Dev
> + )
> +{
> + CONST CHAR8 **Types;
> + UINTN Index;
> + UINT32 Count;
> + XENSTORE_STATUS Status;
> +
> + Status = XenstoreListDirectory (XST_NIL,
> + XENBUS_XENSTORE_NODE, "",
> + &Count, &Types);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + return Status;
> + }
> +
> + for (Index = 0; Index < Count; Index++) {
> + XenbusEnumerateDeviceType (Dev, Types[Index]);
> + }
> +
> + FreePool (Types);
> +
> + return XENSTORE_STATUS_SUCCESS;
> +}
> +
> +STATIC
> +XENSTORE_STATUS
> +EFIAPI
> +XenbusSetState (
> + IN XENBUS_PROTOCOL *This,
> + IN XENSTORE_TRANSACTION Transaction,
> + IN enum xenbus_state NewState
> + )
> +{
> + XENBUS_PRIVATE_DATA *Private;
> + enum xenbus_state CurrentState;
> + XENSTORE_STATUS Status;
> + CHAR8 *Temp;
> +
> + DEBUG ((EFI_D_INFO, "Xenbus: Set state to %d\n", NewState));
> +
> + Private = XENBUS_PRIVATE_DATA_FROM_THIS (This);
> +
> + Status = XenstoreRead (Transaction, This->Node, "state", NULL, (VOID **)&Temp);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + goto Error;
> + }
> + CurrentState = AsciiStrDecimalToUintn (Temp);
> + FreePool (Temp);
> + if (CurrentState == NewState) {
> + DEBUG ((EFI_D_INFO, "%a %d, same state\n", __func__, __LINE__));
> + goto Error;
> + }
> +
> + do {
> + Status = XenstoreSPrint (Transaction, This->Node, "state", "%d", NewState);
> + } while (Status == XENSTORE_STATUS_EAGAIN);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR, "Xenbus: fail to writing new state\n"));
failed to write new state
> + goto Error;
> + }
> + DEBUG ((EFI_D_INFO, "Xenbus: Set state to %d, done\n", NewState));
> +
> +Error:
s/Error/Out/
as it is OK for to set to the same state twice.
> + return Status;
> +}
> +
> +STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData = {
> + .Signature = XENBUS_PRIVATE_DATA_SIGNATURE,
> +
> + .XenbusIo.XsRead = XenbusXenstoreRead,
> + .XenbusIo.XsBackendRead = XenbusXenstoreBackendRead,
> + .XenbusIo.XsPrintf = XenbusXenstoreSPrint,
> + .XenbusIo.XsRemove = XenbusXenstoreRemove,
> + .XenbusIo.XsTransactionStart = XenbusXenstoreTransactionStart,
> + .XenbusIo.XsTransactionEnd = XenbusXenstoreTransactionEnd,
> + .XenbusIo.SetState = XenbusSetState,
> + .XenbusIo.GrantAccess = XenbusGrantAccess,
> + .XenbusIo.GrantEndAccess = XenbusGrantEndAccess,
> + .XenbusIo.RegisterWatch = XenbusRegisterWatch,
> + .XenbusIo.RegisterWatchBackend = XenbusRegisterWatchBackend,
> + .XenbusIo.UnregisterWatch = XenbusUnregisterWatch,
> + .XenbusIo.WaitForWatch = XenbusWaitForWatch,
> +
> + .XenbusIo.Type = NULL,
> + .XenbusIo.Node = NULL,
> + .XenbusIo.Backend = NULL,
> +
> + .Dev = NULL
> +};
> diff --git a/OvmfPkg/XenbusDxe/XenBus.h b/OvmfPkg/XenbusDxe/XenBus.h
> new file mode 100644
> index 0000000..63d0815
> --- /dev/null
> +++ b/OvmfPkg/XenbusDxe/XenBus.h
> @@ -0,0 +1,76 @@
> +/*-
> + * Core definitions and data structures shareable across OS platforms.
> + *
> + * Copyright (c) 2010 Spectra Logic Corporation
> + * Copyright (C) 2008 Doug Rabson
> + * All rights reserved.
> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions
> + * are met:
> + * 1. Redistributions of source code must retain the above copyright
> + * notice, this list of conditions, and the following disclaimer,
> + * without modification.
> + * 2. Redistributions in binary form must reproduce at minimum a disclaimer
> + * substantially similar to the "NO WARRANTY" disclaimer below
> + * ("Disclaimer") and any redistribution must be conditioned upon
> + * including a substantially similar Disclaimer requirement for further
> + * binary redistribution.
> + *
> + * NO WARRANTY
> + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
> + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
> + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
> + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
> + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
> + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
> + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
> + * POSSIBILITY OF SUCH DAMAGES.
> + *
> + * $FreeBSD: release/10.0.0/sys/xen/xenbus/xenbusb.h 222975 2011-06-11 04:59:01Z gibbs $
> + */
> +#ifndef _XEN_XENBUS_XENBUSB_H
> +#define _XEN_XENBUS_XENBUSB_H
> +
> +#include "XenbusDxe.h"
> +
> +#define XENBUS_DEVICE_PATH_TYPE_VBD 0x1
> +struct _XENBUS_DEVICE_PATH {
> + VENDOR_DEVICE_PATH Vendor;
> + UINT8 Type;
> + UINT16 DeviceId;
> +};
> +
> +/**
> + * The VM relative path to the XenStore subtree this
> + * bus attachment manages.
> + *
> + * OR
> + *
> + * The XenStore path to the XenStore subtree for this XenBus bus.
> + */
> +#define XENBUS_XENSTORE_NODE "device"
> +
> +
> +/**
> + * \brief Perform common XenBus bus attach processing.
> + *
> + * \param Dev The NewBus device representing this XenBus bus.
> + *
> + * \return On success, 0. Otherwise an errno value indicating the
> + * type of failure.
> + *
> + * Intiailizes the softc for this bus, installs an interrupt driven
> + * configuration hook to block boot processing until XenBus devices fully
> + * configure, performs an initial probe/attach of the bus, and registers
> + * a XenStore watch so we are notified when the bus topology changes.
> + */
> +XENSTORE_STATUS
> +XenbusEnumerateBus (
> + XENBUS_DEVICE *Dev
> + );
> +
> +#endif /* _XEN_XENBUS_XENBUSB_H */
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> index 2c85d5e..56225c4 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> @@ -19,6 +19,7 @@
> #include "XenHypercall.h"
> #include "GrantTable.h"
> #include "Xenstore.h"
> +#include "XenBus.h"
>
> ///
> /// Driver Binding Protocol instance
> @@ -295,6 +296,7 @@ XenbusDxeDriverBindingStart (
> EFI_PCI_IO_PROTOCOL *PciIo;
> EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR *BarDesc;
> UINT64 MmioAddr;
> + EFI_DEVICE_PATH_PROTOCOL *DevicePath;
>
> Status = gBS->OpenProtocol (
> ControllerHandle,
> @@ -308,11 +310,26 @@ XenbusDxeDriverBindingStart (
> return Status;
> }
>
> + Status = gBS->OpenProtocol (
> + ControllerHandle,
> + &gEfiDevicePathProtocolGuid,
> + (VOID **) &DevicePath,
> + This->DriverBindingHandle,
> + ControllerHandle,
> + EFI_OPEN_PROTOCOL_BY_DRIVER
> + );
> +
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
> +
> Dev = AllocateZeroPool (sizeof (*Dev));
> Dev->Signature = XENBUS_DEVICE_SIGNATURE;
> Dev->This = This;
> Dev->ControllerHandle = ControllerHandle;
> Dev->PciIo = PciIo;
> + Dev->DevicePath = DevicePath;
> + InitializeListHead (&Dev->ChildList);
>
> Status = PciIo->GetBarAttributes (PciIo, PCI_BAR_IDX1, NULL, (VOID**) &BarDesc);
> ASSERT_EFI_ERROR (Status);
> @@ -334,6 +351,8 @@ XenbusDxeDriverBindingStart (
> Status = XenstoreInit (Dev);
> ASSERT_EFI_ERROR (Status);
>
> + XenbusEnumerateBus (Dev);
> +
> Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
> NotifyExitBoot,
> (VOID*) Dev,
> @@ -379,12 +398,54 @@ XenbusDxeDriverBindingStop (
> IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
> )
> {
> + UINTN Index;
> + XENBUS_PROTOCOL *XenbusIo;
> + XENBUS_PRIVATE_DATA *ChildData;
> + EFI_STATUS Status;
> XENBUS_DEVICE *Dev = mMyDevice;
>
> + for (Index = 0; Index < NumberOfChildren; Index++) {
> + Status = gBS->OpenProtocol (
> + ChildHandleBuffer[Index],
> + &gXenbusProtocolGuid,
> + (VOID **) &XenbusIo,
> + This->DriverBindingHandle,
> + ControllerHandle,
> + EFI_OPEN_PROTOCOL_GET_PROTOCOL);
> + if (EFI_ERROR (Status)) {
> + DEBUG ((EFI_D_ERROR, "XenbusDxe: get childre proto: %r\n", Status));
get children protocol failed: %r
> + continue;
> + }
> + ChildData = XENBUS_PRIVATE_DATA_FROM_THIS(XenbusIo);
> + Status = gBS->DisconnectController(ChildData->Handle, NULL, NULL);
> + if (EFI_ERROR (Status)) {
> + DEBUG ((EFI_D_ERROR, "XenbusDxe: error disconnecting child: %r\n",
> + Status));
> + continue;
> + }
> +
> + Status = gBS->UninstallMultipleProtocolInterfaces (
> + ChildData->Handle,
> + &gEfiDevicePathProtocolGuid, ChildData->DevicePath,
> + &gXenbusProtocolGuid, &ChildData->XenbusIo,
> + NULL);
> + ASSERT_EFI_ERROR (Status);
> +
> + FreePool ((VOID*)ChildData->XenbusIo.Type);
> + FreePool ((VOID*)ChildData->XenbusIo.Node);
> + FreePool ((VOID*)ChildData->XenbusIo.Backend);
> + FreePool (ChildData->DevicePath);
> + RemoveEntryList (&ChildData->Link);
> + FreePool (ChildData);
> + }
> +
> gBS->CloseEvent (Dev->ExitBootEvent);
> + // XXX xenbus cleanup
> // XXX xenstore cleanup
> XenGrantTableDeinit (Dev);
>
> + gBS->CloseProtocol(ControllerHandle, &gEfiDevicePathProtocolGuid,
> + This->DriverBindingHandle, ControllerHandle);
> gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
> This->DriverBindingHandle, ControllerHandle);
> return EFI_SUCCESS;
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> index 4d9db72..3ff08dd 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> @@ -80,6 +80,7 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
> #define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
>
>
> +typedef struct _XENBUS_DEVICE_PATH XENBUS_DEVICE_PATH;
> typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
>
> #define XENBUS_DEVICE_SIGNATURE SIGNATURE_32 ('X','B','b','d')
> @@ -89,11 +90,28 @@ struct _XENBUS_DEVICE {
> EFI_HANDLE ControllerHandle;
> EFI_PCI_IO_PROTOCOL *PciIo;
> EFI_EVENT ExitBootEvent;
> + EFI_DEVICE_PATH_PROTOCOL *DevicePath;
> + LIST_ENTRY ChildList;
>
> VOID *Hyperpage;
> shared_info_t *SharedInfo;
> };
>
> +#define XENBUS_PRIVATE_DATA_SIGNATURE SIGNATURE_32 ('X', 'B', 'u', 's')
> +typedef struct {
> + UINTN Signature;
> + LIST_ENTRY Link;
> + EFI_HANDLE Handle;
> + XENBUS_PROTOCOL XenbusIo;
> + XENBUS_DEVICE *Dev;
> + XENBUS_DEVICE_PATH *DevicePath;
> +} XENBUS_PRIVATE_DATA;
> +
> +#define XENBUS_PRIVATE_DATA_FROM_THIS(a) \
> + CR (a, XENBUS_PRIVATE_DATA, XenbusIo, XENBUS_PRIVATE_DATA_SIGNATURE)
> +#define XENBUS_PRIVATE_DATA_FROM_LINK(a) \
> + CR (a, XENBUS_PRIVATE_DATA, Link, XENBUS_PRIVATE_DATA_SIGNATURE)
> +
> /*
> * Helpers
> */
> diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> index 6c8260f..708fae6 100644
> --- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
> +++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> @@ -40,6 +40,9 @@
> EventChannel.h
> Xenstore.c
> Xenstore.h
> + XenBus.c
> + XenBus.h
> + Helpers.c
>
> [Sources.X64]
> X64/hypercall.S
> --
> Anthony PERARD
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client.
[not found] ` <1405523747-5024-18-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 19:41 ` Konrad Rzeszutek Wilk
2014-07-18 11:58 ` Anthony PERARD
[not found] ` <20140718115817.GG1582@perard.uk.xensource.com>
0 siblings, 2 replies; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 19:41 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel, Samuel Thibault
On Wed, Jul 16, 2014 at 04:15:46PM +0100, Anthony PERARD wrote:
> This is the code that will do the actual communication between OVMF and
> a PV block backend, where the block device lives.
>
> This implementation originally comes from Mini-OS, a part of the Xen
> Project.
>
> Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> ---
> OvmfPkg/XenPvBlkDxe/BlockFront.c | 583 ++++++++++++++++++++++++++++++++++++
> OvmfPkg/XenPvBlkDxe/BlockFront.h | 88 ++++++
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 8 +
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 2 +
> 4 files changed, 681 insertions(+)
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.c
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.h
>
> diff --git a/OvmfPkg/XenPvBlkDxe/BlockFront.c b/OvmfPkg/XenPvBlkDxe/BlockFront.c
> new file mode 100644
> index 0000000..ff5641c
> --- /dev/null
> +++ b/OvmfPkg/XenPvBlkDxe/BlockFront.c
> @@ -0,0 +1,583 @@
> +/* Minimal block driver for Mini-OS.
> + * Copyright (c) 2007-2008 Samuel Thibault.
> + * Based on netfront.c.
> + */
> +
> +#include <Library/PrintLib.h>
> +#include <Library/DebugLib.h>
> +
> +#include "BlockFront.h"
> +
> +#include <IndustryStandard/Xen/io/protocols.h>
> +#include <IndustryStandard/Xen/io/xenbus.h>
> +
> +#include "inttypes.h"
> +
> +STATIC
> +XENSTORE_STATUS
> +XenbusReadInteger (
> + IN XENBUS_PROTOCOL *This,
> + IN CONST CHAR8 *Node,
> + IN BOOLEAN FromBackend,
> + OUT UINT64 *ValuePtr
> + )
> +{
> + XENSTORE_STATUS Status;
> + CHAR8 *p;
> +
> + if (!FromBackend) {
> + Status = This->XsRead (This, XST_NIL, Node, (VOID**)&p);
> + } else {
> + Status = This->XsBackendRead (This, XST_NIL, Node, (VOID**)&p);
> + }
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + return Status;
> + }
> + // XXX Will assert if p overflow UINT64 ...
> + *ValuePtr = AsciiStrDecimalToUint64 (p);
> + FreePool (p);
> + return Status;
> +}
> +
> +STATIC
> +VOID
> +XenPvBlockFree (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev
> + )
> +{
> + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> +
> + if (Dev->RingRef != 0) {
> + XenbusIo->GrantEndAccess (XenbusIo, Dev->RingRef);
> + }
> + if (Dev->Ring.sring != NULL) {
> + FreePages (Dev->Ring.sring, 1);
> + }
> + if (Dev->EventChannel != 0) {
> + XenbusIo->EventChannelClose (XenbusIo, Dev->EventChannel);
> + }
> + FreePool (Dev);
> +}
> +
> +XENSTORE_STATUS
> +XenPvBlkWaitForBackendState (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev,
> + IN XenbusState WaitedState,
ExpectedState
?
> + OUT XenbusState *LastStatePtr OPTIONAL
> + )
> +{
> + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> + XenbusState State;
> + UINT64 Value;
> + XENSTORE_STATUS Status;
> +
> + while (TRUE) {
> + Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + return Status;
> + }
> + if (Value > XenbusStateReconfigured) {
> + return XENSTORE_STATUS_EIO;
> + }
> + State = Value;
> + if (State >= WaitedState) {
Not State != WaitedState ?
> + break;
> + }
> + DEBUG ((EFI_D_INFO,
> + "XenPvBlk: waiting backend state %d, current: %d\n",
> + WaitedState, State));
> + XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
> + }
> +
> + if (LastStatePtr != NULL) {
> + *LastStatePtr = State;
> + }
> +
> + return XENSTORE_STATUS_SUCCESS;
> +}
> +
> +EFI_STATUS
> +XenPvBlockFrontInitialization (
> + IN XENBUS_PROTOCOL *XenbusIo,
> + IN CONST CHAR8 *NodeName,
> + OUT XEN_BLOCK_FRONT_DEVICE **DevPtr
> + )
> +{
> + XENSTORE_TRANSACTION xbt;
> + CHAR8 *DeviceType;
> + blkif_sring_t *SharedRing;
> + XENSTORE_STATUS Status;
> + XEN_BLOCK_FRONT_DEVICE *Dev;
> + XenbusState State;
> + UINT64 Value;
> +
> + ASSERT (NodeName != NULL);
> +
> + Dev = AllocateZeroPool (sizeof (XEN_BLOCK_FRONT_DEVICE));
> + Dev->Signature = XEN_BLOCK_FRONT_SIGNATURE;
> + Dev->NodeName = NodeName;
> + Dev->XenbusIo = XenbusIo;
> + Dev->DeviceId = XenbusIo->DeviceId;
> +
> + XenbusIo->XsRead (XenbusIo, XST_NIL, "device-type", (VOID**)&DeviceType);
> + if (AsciiStrCmp (DeviceType, "cdrom") == 0) {
> + Dev->MediaInfo.CdRom = TRUE;
> + } else {
> + Dev->MediaInfo.CdRom = FALSE;
> + }
> + FreePool (DeviceType);
> +
> + Status = XenbusReadInteger (XenbusIo, "backend-id", FALSE, &Value);
> + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT16_MAX) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to get backend-id (%d)\n",
> + Status));
> + goto Error;
> + }
> + Dev->DomainId = Value;
> + XenbusIo->EventChannelAllocate (XenbusIo, Dev->DomainId, &Dev->EventChannel);
> +
> + SharedRing = (blkif_sring_t*) AllocatePages (1);
> + SHARED_RING_INIT (SharedRing);
> + FRONT_RING_INIT (&Dev->Ring, SharedRing, EFI_PAGE_SIZE);
> + XenbusIo->GrantAccess (XenbusIo,
> + Dev->DomainId,
> + (INTN) SharedRing >> EFI_PAGE_SHIFT,
> + FALSE,
> + &Dev->RingRef);
> +
> +Again:
> + Status = XenbusIo->XsTransactionStart (XenbusIo, &xbt);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_WARN, "XenPvBlk: Failed to start transaction, %d\n", Status));
.. and do you want to error out?
> + }
> +
> + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName, "ring-ref", "%d",
> + Dev->RingRef);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write ring-ref.\n"));
> + goto AbortTransaction;
> + }
> + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
> + "event-channel", "%d", Dev->EventChannel);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write event-channel.\n"));
> + goto AbortTransaction;
> + }
> + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
> + "protocol", "%a", XEN_IO_PROTO_ABI_NATIVE);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write protocol.\n"));
> + goto AbortTransaction;
> + }
> +
> + Status = XenbusIo->SetState (XenbusIo, xbt, XenbusStateConnected);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed switch state.\n"));
> + goto AbortTransaction;
> + }
> +
> + Status = XenbusIo->XsTransactionEnd (XenbusIo, xbt, FALSE);
> + if (Status == XENSTORE_STATUS_EAGAIN) {
> + goto Again;
> + }
> +
> + XenbusIo->RegisterWatchBackend (XenbusIo, "state", &Dev->StateWatchToken);
> +
> + // Waiting backend
> + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateConnected, &State);
> + if (Status != XENSTORE_STATUS_SUCCESS || State != XenbusStateConnected) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: backend for %a/%d not available, rc=%d state=%d\n",
> + XenbusIo->Type, XenbusIo->DeviceId, Status, State));
> + goto Error2;
> + }
> +
> + Status = XenbusReadInteger (XenbusIo, "info", TRUE, &Value);
> + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
> + goto Error2;
> + }
> + Dev->MediaInfo.VDiskInfo = Value;
> + if (Dev->MediaInfo.VDiskInfo & VDISK_READONLY) {
> + Dev->MediaInfo.ReadWrite = FALSE;
> + } else {
> + Dev->MediaInfo.ReadWrite = TRUE;
> + }
> +
> + Status = XenbusReadInteger (XenbusIo, "sectors", TRUE, &Dev->MediaInfo.Sectors);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + goto Error2;
> + }
> +
> + Status = XenbusReadInteger (XenbusIo, "sector-size", TRUE, &Value);
> + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
> + goto Error2;
> + }
> + Dev->MediaInfo.SectorSize = Value;
Why not &Dev->MediaInfo.SectorSize in XenbusReadInteger?
Is it because of types? If so, please add a comment stating that.
> +
> + // Default value
> + Value = 0;
> + Status = XenbusReadInteger (XenbusIo, "feature-barrier", TRUE, &Value);
> + if (Value == 1) {
> + Dev->MediaInfo.FeatureBarrier = TRUE;
> + } else {
> + Dev->MediaInfo.FeatureBarrier = FALSE;
> + }
> +
> + // Default value
> + Value = 0;
> + XenbusReadInteger (XenbusIo, "feature-flush-cache", TRUE, &Value);
> + if (Value == 1) {
> + Dev->MediaInfo.FeatureFlushCache = TRUE;
> + } else {
> + Dev->MediaInfo.FeatureFlushCache = FALSE;
> + }
> +
> + DEBUG ((EFI_D_INFO, "XenPvBlk: New disk with %d sectors of %d bytes\n",
> + Dev->MediaInfo.Sectors, Dev->MediaInfo.SectorSize));
> +
> + *DevPtr = Dev;
> + return EFI_SUCCESS;
> +
> +Error2:
> + XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
> + XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
> + XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
XenbusIo->XsRemove (XenbusIo, XST_NIL, "protocol");
?
> + goto Error;
> +AbortTransaction:
> + XenbusIo->XsTransactionEnd (XenbusIo, xbt, TRUE);
> +Error:
> + XenPvBlockFree (Dev);
> + return EFI_DEVICE_ERROR;
> +}
> +
> +VOID
> +XenPvBlockFrontShutdown (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev
> + )
> +{
> + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> + XENSTORE_STATUS Status;
> + UINT64 Value;
> +
> + XenPvBlockSync (Dev);
> +
> + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosing);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: error while changing state to Closing: %d\n",
> + Status));
> + goto Close;
> + }
> +
> + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosing, NULL);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: error while waiting for closing backend state: %d\n",
> + Status));
> + goto Close;
> + }
> +
> + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosed);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: error while changing state to Closed: %d\n",
> + Status));
> + goto Close;
> + }
> +
> + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosed, NULL);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: error while waiting for closed backend state: %d\n",
> + Status));
> + goto Close;
> + }
> +
> + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateInitialising);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: error while changing state to initialising: %d\n",
> + Status));
> + goto Close;
> + }
> +
> + while (TRUE) {
> + Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
> + if (Status != XENSTORE_STATUS_SUCCESS) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: error while waiting for new backend state: %d\n",
> + Status));
> + goto Close;
> + }
> + if (Value < XenbusStateInitWait || Value >= XenbusStateClosed) {
> + break;
> + }
> + DEBUG ((EFI_D_INFO,
> + "XenPvBlk: waiting backend state %d, current: %d\n",
> + XenbusStateInitWait, Value));
> + XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
> + }
> +
> +Close:
> + XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
> + XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
> + XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
"protocol"
> +
> + XenPvBlockFree (Dev);
> +}
> +
> +STATIC
> +VOID
> +XenPvBlockWaitSlot (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev
> + )
> +{
> + /* Wait for a slot */
> + if (RING_FULL (&Dev->Ring)) {
> + while (TRUE) {
> + XenPvBlockAsyncIoPoll (Dev);
> + if (!RING_FULL(&Dev->Ring)) {
> + break;
> + }
> + /* Really no slot, go to sleep. */
> + // XXX wait for an event on Dev->EventChannel
> + DEBUG ((EFI_D_INFO, "%a %d, waiting\n", __func__, __LINE__));
> + }
> + }
> +}
> +
> +/* Issue an aio */
> +VOID
> +XenPvBlockAsyncIo (
> + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> + IN BOOLEAN IsWrite
> + )
> +{
> + XEN_BLOCK_FRONT_DEVICE *Dev = IoData->Dev;
> + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> + blkif_request_t *Request;
> + RING_IDX RingIndex;
> + BOOLEAN Notify;
> + INT32 NumSegments, Index;
> + UINTN Start, End;
> +
> + // Can't io at non-sector-aligned location
> + ASSERT(!(IoData->Offset & (Dev->MediaInfo.SectorSize - 1)));
> + // Can't io non-sector-sized amounts
> + ASSERT(!(IoData->Size & (Dev->MediaInfo.SectorSize - 1)));
> + // Can't io non-sector-aligned buffer
> + ASSERT(!((UINTN) IoData->Buffer & (Dev->MediaInfo.SectorSize - 1)));
> +
> + Start = (UINTN) IoData->Buffer & ~EFI_PAGE_MASK;
> + End = ((UINTN) IoData->Buffer + IoData->Size + EFI_PAGE_SIZE - 1) & ~EFI_PAGE_MASK;
> + IoData->NumRef = NumSegments = (End - Start) / EFI_PAGE_SIZE;
> +
> + // XXX is this comment still true ?
I think it is outdated, as Linux would do max DMA up to the segment
size that is advertised. Ditto for the SCSI stack.
> + /* qemu's IDE max multsect is 16 (8KB) and SCSI max DMA was set to 32KB,
> + * so max 44KB can't happen */
> + ASSERT (NumSegments <= BLKIF_MAX_SEGMENTS_PER_REQUEST);
But this check is OK as we can't truly fit more on the ring.
> +
> + XenPvBlockWaitSlot (Dev);
What happens if there are multiple threads calling this
function? Don't we need a mutex to only allow
one thread to grab an RingIndex?
> + RingIndex = Dev->Ring.req_prod_pvt;
> + Request = RING_GET_REQUEST (&Dev->Ring, RingIndex);
> +
> + Request->operation = IsWrite ? BLKIF_OP_WRITE : BLKIF_OP_READ;
> + Request->nr_segments = NumSegments;
> + Request->handle = Dev->DeviceId;
> + Request->id = (UINTN) IoData;
> + Request->sector_number = IoData->Offset / 512;
> +
> + for (Index = 0; Index < NumSegments; Index++) {
> + Request->seg[Index].first_sect = 0;
Uh?
> + Request->seg[Index].last_sect = EFI_PAGE_SIZE / 512 - 1;
Uh? I thought you wanted to use the real sector values?
> + }
> + Request->seg[0].first_sect = ((UINTN) IoData->Buffer & EFI_PAGE_MASK) / 512;
> + Request->seg[NumSegments - 1].last_sect =
> + (((UINTN) IoData->Buffer + IoData->Size - 1) & EFI_PAGE_MASK) / 512;
What about in between 0 and NumSegments? You should also
fill out those values with valid entries.
> + for (Index = 0; Index < NumSegments; Index++) {
> + UINTN Data = Start + Index * EFI_PAGE_SIZE;
> + if (!IsWrite) {
> + /* Trigger CoW if needed */
Not sure I understand that reference?
> + *(CHAR8 *)(Data + (Request->seg[Index].first_sect << 9)) = 0;
> + MemoryFence ();
> + }
> + XenbusIo->GrantAccess (XenbusIo, Dev->DomainId,
> + Data >> EFI_PAGE_SHIFT, IsWrite,
> + &Request->seg[Index].gref);
> + IoData->GrantRef[Index] = Request->seg[Index].gref;
> + }
> +
> + Dev->Ring.req_prod_pvt = RingIndex + 1;
> +
> + MemoryFence ();
> + RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
> +
> + if (Notify) {
> + XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
> + }
> +}
> +
> +STATIC
> +VOID
> +XenPvBlockAsyncIoCallback (
> + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> + IN EFI_STATUS Status
> + )
> +{
> + IoData->Data = (VOID *) 1;
> + IoData->Callback = NULL;
> +}
> +
> +VOID
> +XenPvBlockIo (
> + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> + IN BOOLEAN IsWrite
> + )
> +{
> + ASSERT (IoData->Callback == NULL);
> + // XXX Signal/Event instead ?
Maybe for later versions. I think this is OK for right now, thought
it will mean we can only do one I/O .. and have then to wait
until it is completed.
That limititation should be mentioned in the commit description.
> + IoData->Callback = XenPvBlockAsyncIoCallback;
> + XenPvBlockAsyncIo (IoData, IsWrite);
> + IoData->Data = NULL;
> +
> + while (TRUE) {
> + XenPvBlockAsyncIoPoll (IoData->Dev);
> + if (IoData->Data) {
> + break;
> + }
> + }
> +}
> +
> +STATIC
> +VOID
> +XenPvBlockPushOperation (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev,
> + IN UINT8 Operation,
> + IN UINT64 Id
> + )
> +{
> + INT32 Index;
> + blkif_request_t *Request;
> + BOOLEAN Notify;
> +
> + XenPvBlockWaitSlot (Dev);
How do we make sure that there aren't multiple threads
calling this?
> + Index = Dev->Ring.req_prod_pvt;
> + Request = RING_GET_REQUEST(&Dev->Ring, Index);
> + Request->operation = Operation;
> + Request->nr_segments = 0;
> + Request->handle = Dev->DeviceId;
> + Request->id = Id;
> + /* Not needed anyway, but the backend will check it */
> + Request->sector_number = 0;
> + Dev->Ring.req_prod_pvt = Index + 1;
> + MemoryFence ();
> + RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
> + if (Notify) {
> + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> + XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
> + }
> +}
> +
> +VOID
> +XenPvBlockSync (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev
> + )
> +{
> + if (Dev->MediaInfo.ReadWrite) {
> + if (Dev->MediaInfo.FeatureBarrier) {
> + XenPvBlockPushOperation (Dev, BLKIF_OP_WRITE_BARRIER, 0);
> + }
> +
> + if (Dev->MediaInfo.FeatureFlushCache) {
> + XenPvBlockPushOperation (Dev, BLKIF_OP_FLUSH_DISKCACHE, 0);
> + }
> + }
> +
> + /* Note: This won't finish if another thread enqueues requests. */
> + while (TRUE) {
> + XenPvBlockAsyncIoPoll (Dev);
> + if (RING_FREE_REQUESTS (&Dev->Ring) == RING_SIZE (&Dev->Ring)) {
> + break;
> + }
> + }
> +}
> +
> +VOID
> +XenPvBlockAsyncIoPoll (
> + IN XEN_BLOCK_FRONT_DEVICE *Dev
> + )
> +{
> + RING_IDX ProducerIndex, ConsumerIndex;
> + blkif_response_t *Response;
> + INT32 More;
> +
> + do {
> + ProducerIndex = Dev->Ring.sring->rsp_prod;
> + /* Ensure we see queued responses up to 'ProducerIndex'. */
> + MemoryFence ();
> + ConsumerIndex = Dev->Ring.rsp_cons;
> +
> + while (ConsumerIndex != ProducerIndex) {
> + XEN_BLOCK_FRONT_IO *IoData;
> + INT16 Status;
> +
> + Response = RING_GET_RESPONSE (&Dev->Ring, ConsumerIndex);
> +
> + IoData = (VOID *) (UINTN) Response->id;
> + Status = Response->status;
> +
> + switch (Response->operation) {
> + case BLKIF_OP_READ:
> + case BLKIF_OP_WRITE:
> + {
> + INT32 Index;
> +
> + if (Status != BLKIF_RSP_OKAY) {
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: "
> + "%a error %d on %a at offset %Ld, num bytes %Ld\n",
> + Response->operation == BLKIF_OP_READ ? "read" : "write",
> + Status, IoData->Dev->NodeName,
> + IoData->Offset,
> + IoData->Size));
> + }
> +
> + for (Index = 0; Index < IoData->NumRef; Index++) {
> + Dev->XenbusIo->GrantEndAccess (Dev->XenbusIo, IoData->GrantRef[Index]);
> + }
> +
> + break;
> + }
> +
> + case BLKIF_OP_WRITE_BARRIER:
> + if (Status != BLKIF_RSP_OKAY) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: write barrier error %d\n", Status));
> + }
> + break;
> + case BLKIF_OP_FLUSH_DISKCACHE:
> + if (Status != BLKIF_RSP_OKAY) {
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: flush error %d\n", Status));
> + }
> + break;
> +
> + default:
> + DEBUG ((EFI_D_ERROR,
> + "XenPvBlk: unrecognized block operation %d response (status %d)\n",
> + Response->operation, Status));
> + break;
> + }
> +
> + Dev->Ring.rsp_cons = ++ConsumerIndex;
> + /* Nota: callback frees IoData itself */
> + if (IoData && IoData->Callback) {
> + IoData->Callback (IoData, Status ? EFI_DEVICE_ERROR : EFI_SUCCESS);
> + }
> + if (Dev->Ring.rsp_cons != ConsumerIndex) {
> + /* We reentered, we must not continue here */
Uh, how does that happen?
> + break;
> + }
> + }
> +
> + RING_FINAL_CHECK_FOR_RESPONSES (&Dev->Ring, More);
> + } while (More != 0);
> +}
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo.
[not found] ` <1405523747-5024-19-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 19:57 ` Konrad Rzeszutek Wilk
2014-07-18 12:10 ` Anthony PERARD
0 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 19:57 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:47PM +0100, Anthony PERARD wrote:
> Install the BlockIo protocol.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> ---
> OvmfPkg/XenPvBlkDxe/BlockIo.c | 256 ++++++++++++++++++++++++++++++++++++
> OvmfPkg/XenPvBlkDxe/BlockIo.h | 109 +++++++++++++++
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 59 +++++++++
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h | 1 +
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 2 +
> 5 files changed, 427 insertions(+)
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.c
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.h
>
> diff --git a/OvmfPkg/XenPvBlkDxe/BlockIo.c b/OvmfPkg/XenPvBlkDxe/BlockIo.c
> new file mode 100644
> index 0000000..4ac2411
> --- /dev/null
> +++ b/OvmfPkg/XenPvBlkDxe/BlockIo.c
> @@ -0,0 +1,256 @@
> +
> +/** @file
> + TODO: Brief Description of UEFI Driver XenPvBlkDxe
> +
> + TODO: Detailed Description of UEFI Driver XenPvBlkDxe
> +
> + TODO: Copyright for UEFI Driver XenPvBlkDxe
> +
> + TODO: License for UEFI Driver XenPvBlkDxe
> +
> +**/
> +
> +#include "XenPvBlkDxe.h"
> +
> +#include "BlockFront.h"
> +
> +///
> +/// Block I/O Media structure
> +///
> +GLOBAL_REMOVE_IF_UNREFERENCED
> +EFI_BLOCK_IO_MEDIA gXenPvBlkDxeBlockIoMedia = {
> + 0, // MediaId
> + FALSE, // RemovableMedia
> + FALSE, // MediaPresent
> + FALSE, // LogicalPartition
> + TRUE, // ReadOnly
> + FALSE, // WriteCaching
> + 512, // BlockSize
> + 512, // IoAlign
> + 0, // LastBlock
> + 0, // LowestAlignedLba
> + 0, // LogicalBlocksPerPhysicalBlock
> + 0 // OptimalTransferLengthGranularity
> +};
> +
> +///
> +/// Block I/O Protocol instance
> +///
> +GLOBAL_REMOVE_IF_UNREFERENCED
> +EFI_BLOCK_IO_PROTOCOL gXenPvBlkDxeBlockIo = {
> + EFI_BLOCK_IO_PROTOCOL_REVISION3, // Revision
> + &gXenPvBlkDxeBlockIoMedia, // Media
> + (EFI_BLOCK_RESET)XenPvBlkDxeBlockIoReset, // Reset
> + XenPvBlkDxeBlockIoReadBlocks, // ReadBlocks
> + XenPvBlkDxeBlockIoWriteBlocks, // WriteBlocks
> + XenPvBlkDxeBlockIoFlushBlocks // FlushBlocks
> +};
> +
> +
> +
> +
> +STATIC
> +EFI_STATUS
> +XenPvBlkDxeBlockIoReadWriteBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This,
> + IN UINT32 MediaId,
> + IN EFI_LBA Lba,
> + IN UINTN BufferSize,
> + IN OUT VOID *Buffer,
> + IN BOOLEAN IsWrite
> + )
> +{
> + XEN_BLOCK_FRONT_IO IoData;
> + EFI_BLOCK_IO_MEDIA *Media;
> + UINTN NextOffset;
> +
> + if (Buffer == NULL) {
> + DEBUG ((EFI_D_ERROR, "%a %d, buffer null\n", __func__, __LINE__));
Heheh.
> + return EFI_INVALID_PARAMETER;
> + }
> + if (BufferSize == 0) {
> + return EFI_SUCCESS;
> + }
> +
> + Media = This->Media;
> +
> + if (BufferSize % Media->BlockSize != 0) {
> + DEBUG ((EFI_D_ERROR, "%a %d, bad buffer size\n", __func__, __LINE__));
> + return EFI_BAD_BUFFER_SIZE;
> + }
> +
> + if (Lba > Media->LastBlock ||
> + (BufferSize / Media->BlockSize) - 1 > Media->LastBlock - Lba) {
> + DEBUG ((EFI_D_ERROR, "%a %d, read too large\n", __func__, __LINE__));
> + return EFI_INVALID_PARAMETER;
> + }
> +
> + if (IsWrite && Media->ReadOnly) {
> + DEBUG ((EFI_D_ERROR, "%a %d, read only\n", __func__, __LINE__));
> + return EFI_WRITE_PROTECTED;
> + }
> +
> + if ((Media->IoAlign > 1) && (UINTN)Buffer & (Media->IoAlign - 1)) {
> + // This happen often when using grub2
> + VOID *NewBuffer;
> + EFI_STATUS Status;
> +
> + /* Try again with a properly aligned buffer. */
> + NewBuffer = AllocateAlignedPages((BufferSize + EFI_PAGE_SIZE) / EFI_PAGE_SIZE,
> + Media->IoAlign);
> + if (!IsWrite) {
> + Status = XenPvBlkDxeBlockIoReadBlocks (This, MediaId,
> + Lba, BufferSize, NewBuffer);
> + CopyMem (Buffer, NewBuffer, BufferSize);
> + } else {
> + CopyMem (NewBuffer, Buffer, BufferSize);
> + Status = XenPvBlkDxeBlockIoWriteBlocks (This, MediaId,
> + Lba, BufferSize, NewBuffer);
> + }
> + FreeAlignedPages (NewBuffer, (BufferSize + EFI_PAGE_SIZE) / EFI_PAGE_SIZE);
> + return Status;
> + }
> +
> +
> + // XXX Check MediaId
> +
> + IoData.Dev = XEN_BLOCK_FRONT_FROM_BLOCK_IO (This);
> + IoData.Offset = Lba * Media->BlockSize;
> + IoData.Callback = NULL;
> + IoData.Data = NULL;
> + IoData.Size = BufferSize;
> + IoData.Buffer = Buffer;
> +
> + NextOffset = Lba * Media->BlockSize;
Or just = IoData.Offset ?
> + while (BufferSize > 0) {
> + // This should work if every Buffer is page aligned
> + // but block accept sector aligned, so FIXME
Isn't the work-around for GRUB2 above doing this?
> + IoData.Size = MIN(EFI_PAGE_SIZE * 8,//BLKIF_MAX_SEGMENTS_PER_REQUEST,
> + BufferSize);
Why 8 ?
> + IoData.Buffer = Buffer;
> + IoData.Offset = NextOffset;
> + BufferSize -= IoData.Size;
> + Buffer = (VOID*) ((UINTN) Buffer + IoData.Size);
> + NextOffset += IoData.Size;
> + XenPvBlockIo (&IoData, IsWrite);
> + if (!IoData.Data) {
Well that wouldn't be possible as the XenPvBlockIo loops forever until that
is done.
However, you should probably have a comment here:
/* XXX If we move to a different mechanism for CallBack this might
go away */
> + DEBUG ((EFI_D_ERROR, "%a %d, XenPvBlockIo no data...\n", __func__, __LINE__));
> + return EFI_DEVICE_ERROR;
> + }
> + }
> + return EFI_SUCCESS;
> +}
> +
> +
> +/**
> + Read BufferSize bytes from Lba into Buffer.
> +
> + @param This Indicates a pointer to the calling context.
> + @param MediaId Id of the media, changes every time the media is replaced.
> + @param Lba The starting Logical Block Address to read from
> + @param BufferSize Size of Buffer, must be a multiple of device block size.
> + @param Buffer A pointer to the destination buffer for the data. The caller is
> + responsible for either having implicit or explicit ownership of the buffer.
> +
> + @retval EFI_SUCCESS The data was read correctly from the device.
> + @retval EFI_DEVICE_ERROR The device reported an error while performing the read.
> + @retval EFI_NO_MEDIA There is no media in the device.
> + @retval EFI_MEDIA_CHANGED The MediaId does not matched the current device.
> + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> + @retval EFI_INVALID_PARAMETER The read request contains LBAs that are not valid,
> + or the buffer is not on proper alignment.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoReadBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This,
> + IN UINT32 MediaId,
> + IN EFI_LBA Lba,
> + IN UINTN BufferSize,
> + OUT VOID *Buffer
> + )
> +{
> + return XenPvBlkDxeBlockIoReadWriteBlocks (This,
> + MediaId, Lba, BufferSize, Buffer, FALSE);
> +}
> +
> +/**
> + Write BufferSize bytes from Lba into Buffer.
> +
> + @param This Indicates a pointer to the calling context.
> + @param MediaId The media ID that the write request is for.
> + @param Lba The starting logical block address to be written. The caller is
> + responsible for writing to only legitimate locations.
> + @param BufferSize Size of Buffer, must be a multiple of device block size.
> + @param Buffer A pointer to the source buffer for the data.
> +
> + @retval EFI_SUCCESS The data was written correctly to the device.
> + @retval EFI_WRITE_PROTECTED The device can not be written to.
> + @retval EFI_DEVICE_ERROR The device reported an error while performing the write.
> + @retval EFI_NO_MEDIA There is no media in the device.
> + @retval EFI_MEDIA_CHNAGED The MediaId does not matched the current device.
> + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> + @retval EFI_INVALID_PARAMETER The write request contains LBAs that are not valid,
> + or the buffer is not on proper alignment.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoWriteBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This,
> + IN UINT32 MediaId,
> + IN EFI_LBA Lba,
> + IN UINTN BufferSize,
> + IN VOID *Buffer
> + )
> +{
> + return XenPvBlkDxeBlockIoReadWriteBlocks (This,
> + MediaId, Lba, BufferSize, Buffer, TRUE);
> +}
> +
> +/**
> + Flush the Block Device.
> +
> + @param This Indicates a pointer to the calling context.
> +
> + @retval EFI_SUCCESS All outstanding data was written to the device
> + @retval EFI_DEVICE_ERROR The device reported an error while writting back the data
> + @retval EFI_NO_MEDIA There is no media in the device.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoFlushBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This
> + )
> +{
> + XenPvBlockSync (XEN_BLOCK_FRONT_FROM_BLOCK_IO (This));
> + return EFI_SUCCESS;
> +}
> +
> +/**
> + Reset the block device hardware.
> +
> + @param[in] This Indicates a pointer to the calling context.
> + @param[in] ExtendedVerification Indicates that the driver may perform a more
> + exhausive verfication operation of the device
> + during reset.
> +
> + @retval EFI_SUCCESS The device was reset.
> + @retval EFI_DEVICE_ERROR The device is not functioning properly and could
> + not be reset.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoReset (
> + IN EFI_BLOCK_IO2_PROTOCOL *This,
> + IN BOOLEAN ExtendedVerification
> + )
> +{
> + // Is there anything to do ?
> + // since initialisation is done in Start (driver binding)
Won't we leak memory? Or will it after an Reset do Stop, then Start?
> + DEBUG ((EFI_D_INFO, "%a %d\n", __func__, __LINE__));
> + return EFI_SUCCESS;
> +}
> diff --git a/OvmfPkg/XenPvBlkDxe/BlockIo.h b/OvmfPkg/XenPvBlkDxe/BlockIo.h
> new file mode 100644
> index 0000000..ec0fe2b
> --- /dev/null
> +++ b/OvmfPkg/XenPvBlkDxe/BlockIo.h
> @@ -0,0 +1,109 @@
> +
> +/** @file
> + TODO: Brief Description of UEFI Driver XenPvBlkDxe
> +
> + TODO: Detailed Description of UEFI Driver XenPvBlkDxe
> +
> + TODO: Copyright for UEFI Driver XenPvBlkDxe
> +
> + TODO: License for UEFI Driver XenPvBlkDxe
> +
> +**/
> +
> +/**
> + Read BufferSize bytes from Lba into Buffer.
> +
> + @param This Indicates a pointer to the calling context.
> + @param MediaId Id of the media, changes every time the media is replaced.
> + @param Lba The starting Logical Block Address to read from
> + @param BufferSize Size of Buffer, must be a multiple of device block size.
> + @param Buffer A pointer to the destination buffer for the data. The caller is
> + responsible for either having implicit or explicit ownership of the buffer.
> +
> + @retval EFI_SUCCESS The data was read correctly from the device.
> + @retval EFI_DEVICE_ERROR The device reported an error while performing the read.
> + @retval EFI_NO_MEDIA There is no media in the device.
> + @retval EFI_MEDIA_CHANGED The MediaId does not matched the current device.
> + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> + @retval EFI_INVALID_PARAMETER The read request contains LBAs that are not valid,
> + or the buffer is not on proper alignment.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoReadBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This,
> + IN UINT32 MediaId,
> + IN EFI_LBA Lba,
> + IN UINTN BufferSize,
> + OUT VOID *Buffer
> + );
> +
> +/**
> + Write BufferSize bytes from Lba into Buffer.
> +
> + @param This Indicates a pointer to the calling context.
> + @param MediaId The media ID that the write request is for.
> + @param Lba The starting logical block address to be written. The caller is
> + responsible for writing to only legitimate locations.
> + @param BufferSize Size of Buffer, must be a multiple of device block size.
> + @param Buffer A pointer to the source buffer for the data.
> +
> + @retval EFI_SUCCESS The data was written correctly to the device.
> + @retval EFI_WRITE_PROTECTED The device can not be written to.
> + @retval EFI_DEVICE_ERROR The device reported an error while performing the write.
> + @retval EFI_NO_MEDIA There is no media in the device.
> + @retval EFI_MEDIA_CHNAGED The MediaId does not matched the current device.
> + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> + @retval EFI_INVALID_PARAMETER The write request contains LBAs that are not valid,
> + or the buffer is not on proper alignment.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoWriteBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This,
> + IN UINT32 MediaId,
> + IN EFI_LBA Lba,
> + IN UINTN BufferSize,
> + IN VOID *Buffer
> + );
> +
> +/**
> + Flush the Block Device.
> +
> + @param This Indicates a pointer to the calling context.
> +
> + @retval EFI_SUCCESS All outstanding data was written to the device
> + @retval EFI_DEVICE_ERROR The device reported an error while writting back the data
> + @retval EFI_NO_MEDIA There is no media in the device.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoFlushBlocks (
> + IN EFI_BLOCK_IO_PROTOCOL *This
> + );
> +
> +/**
> + Reset the block device hardware.
> +
> + @param[in] This Indicates a pointer to the calling context.
> + @param[in] ExtendedVerification Indicates that the driver may perform a more
> + exhausive verfication operation of the device
> + during reset.
> +
> + @retval EFI_SUCCESS The device was reset.
> + @retval EFI_DEVICE_ERROR The device is not functioning properly and could
> + not be reset.
> +
> +**/
> +EFI_STATUS
> +EFIAPI
> +XenPvBlkDxeBlockIoReset (
> + IN EFI_BLOCK_IO2_PROTOCOL *This,
> + IN BOOLEAN ExtendedVerification
> + );
> +
> +extern EFI_BLOCK_IO_MEDIA gXenPvBlkDxeBlockIoMedia;
> +extern EFI_BLOCK_IO_PROTOCOL gXenPvBlkDxeBlockIo;
> diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> index 7ec7d60..3991e83 100644
> --- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> +++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> @@ -266,6 +266,7 @@ XenPvBlkDxeDriverBindingStart (
> EFI_STATUS Status;
> XENBUS_PROTOCOL *XenbusIo;
> XEN_BLOCK_FRONT_DEVICE *Dev;
> + EFI_BLOCK_IO_MEDIA *Media;
>
> Status = gBS->OpenProtocol (
> ControllerHandle,
> @@ -284,6 +285,37 @@ XenPvBlkDxeDriverBindingStart (
> return Status;
> }
>
> + CopyMem (&Dev->BlockIo, &gXenPvBlkDxeBlockIo, sizeof (EFI_BLOCK_IO_PROTOCOL));
> + Media = AllocateCopyPool (sizeof (EFI_BLOCK_IO_MEDIA),
> + &gXenPvBlkDxeBlockIoMedia);
> + if (Dev->MediaInfo.VDiskInfo & VDISK_REMOVABLE) {
> + Media->RemovableMedia = TRUE;
> + }
> + Media->MediaPresent = TRUE;
> + Media->ReadOnly = !Dev->MediaInfo.ReadWrite;
> + if (Dev->MediaInfo.CdRom) {
> + // If it's a cdrom, the blocksize value need to be 2048 for OVMF to
> + // recognize it as a cdrom:
> + // MdeModulePkg/Universal/Disk/PartitionDxe/ElTorito.c
> + Media->BlockSize = 2048;
> + } else {
> + Media->BlockSize = Dev->MediaInfo.SectorSize;
> + }
> + Media->LastBlock = Dev->MediaInfo.Sectors - 1;
> + // XXX What about LowestAlignedLba, LogicalBlocksPerPhysicalBlock and OptimalTransferLengthGranularity?
I think you can skip LogicalBlocksPerPhysicalBlock and OptimalTransferLengthGranularity
as they are used for discard operations.
The LowestAlignedLba .. no idea.
> + Dev->BlockIo.Media = Media;
> +
> + Status = gBS->InstallMultipleProtocolInterfaces (
> + &ControllerHandle,
> + &gEfiBlockIoProtocolGuid, &Dev->BlockIo,
> + NULL
> + );
> + if (EFI_ERROR (Status)) {
> + // XXX: uninit everythings
XenPvBlockFrontShutdown (Dev); ?
FreePool(Media) ?
> + DEBUG ((EFI_D_ERROR, "XenPvBlk: install protocol fail: %r\n", Status));
> + return Status;
> + }
> +
> return EFI_SUCCESS;
> }
>
> @@ -322,6 +354,33 @@ XenPvBlkDxeDriverBindingStop (
> IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
> )
> {
> + EFI_BLOCK_IO_PROTOCOL *BlockIo;
> + XEN_BLOCK_FRONT_DEVICE *Dev;
> + EFI_BLOCK_IO_MEDIA *Media;
> + EFI_STATUS Status;
> +
> + Status = gBS->OpenProtocol (
> + ControllerHandle, &gEfiBlockIoProtocolGuid,
> + (VOID **)&BlockIo,
> + This->DriverBindingHandle, ControllerHandle,
> + EFI_OPEN_PROTOCOL_GET_PROTOCOL
> + );
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
> +
> + Status = gBS->UninstallProtocolInterface (ControllerHandle,
> + &gEfiBlockIoProtocolGuid, BlockIo);
> + if (EFI_ERROR (Status)) {
> + return Status;
> + }
> +
> + Media = BlockIo->Media;
> + Dev = XEN_BLOCK_FRONT_FROM_BLOCK_IO (BlockIo);
> + XenPvBlockFrontShutdown (Dev);
> +
> + FreePool (Media);
> +
> gBS->CloseProtocol (ControllerHandle, &gXenbusProtocolGuid,
> This->DriverBindingHandle, ControllerHandle);
>
> diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> index 186d0ee..3120137 100644
> --- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> +++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> @@ -76,6 +76,7 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenPvBlkDxeComponentName;
> //
> #include "DriverBinding.h"
> #include "ComponentName.h"
> +#include "BlockIo.h"
>
>
>
> diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> index a43d006..b379020 100644
> --- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> +++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> @@ -32,6 +32,8 @@
> ComponentName.h
> BlockFront.c
> BlockFront.h
> + BlockIo.c
> + BlockIo.h
>
>
> [LibraryClasses]
> --
> Anthony PERARD
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 00/18] Introducing Xen PV block driver to OVMF
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
` (26 preceding siblings ...)
[not found] ` <1405523747-5024-19-git-send-email-anthony.perard@citrix.com>
@ 2014-07-16 20:10 ` Konrad Rzeszutek Wilk
2014-07-18 12:12 ` Anthony PERARD
27 siblings, 1 reply; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-16 20:10 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:15:29PM +0100, Anthony PERARD wrote:
> Hi all,
>
> This patch series is implementing the necessary in order to access a PV block
> device. For that, one need a XenStore client, a XenBus client, and the PV block
> driver.
>
> There are two new drivers, XenbusDxe and XenPvBlkDxe. The first one implement a
> bus drivers, and the second is a block drivers.
>
> There are still a bit of work to be done on this series, especially the comment
> in the code, but I'd like your comment on this patch series.
Went through it - I had some questions and spotted some issues that are
pretty easy to fix.
Otherwise I think you just need to flesh it out with more comments, links
to the specs or just copy the relevant parts.
And little puzzled by the usage of 8 pages instead of doing it via 11.
But that is probably not a big deal since you are doing each I/O request
synchronously anyhow.
Thank you for posting this and developing it!
>
> Anthony PERARD (18):
> OvmfPkg: Add public headers from Xen Project.
> OvmfPkg: Add basic skeleton for the Xenbus driver.
> OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot
> services event.
> OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls.
> OvmfPkg/XenbusDxe: Open PciIo protocol.
> OvmfPkg: Introduce Xenbus Protocol.
> OvmfPkg/XenbusDxe: Add InterlockedCompareExchange16.
> OvmfPkg/XenbusDxe: Add Grant Table functions.
> OvmfPkg/XenbusDxe: Add Event Channel Notify.
> OvmfPkg/XenbusDxe: Add TestAndClearBit.
> OvmfPkg/XenbusDxe: Add XenStore client implementation
> OvmfPkg/XenbusDxe: Add an helper AsciiStrDup.
> OvmfPkg/XenbusDxe: Add Xenstore function into the Xenbus protocol
> OvmfPkg/XenbusDxe: Indroduce XenBus support itself.
> OvmfPkg/XenbusDxe: Add Event Channel into XenBus protocol.
> OvmfPkg/XenPvBlkDxe: Xen PV Block device, initial skeleton
> OvmfPkg/XenPvBlkDxe: Add BlockFront client.
> OvmfPkg/XenPvBlkDxe: Add BlockIo.
>
> .../IndustryStandard/Xen/arch-x86/xen-x86_32.h | 171 +++
> .../IndustryStandard/Xen/arch-x86/xen-x86_64.h | 202 +++
> .../Include/IndustryStandard/Xen/arch-x86/xen.h | 273 ++++
> .../Include/IndustryStandard/Xen/event_channel.h | 381 +++++
> OvmfPkg/Include/IndustryStandard/Xen/grant_table.h | 662 +++++++++
> OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h | 275 ++++
> OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h | 150 ++
> OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h | 608 ++++++++
> .../Include/IndustryStandard/Xen/io/protocols.h | 40 +
> OvmfPkg/Include/IndustryStandard/Xen/io/ring.h | 312 +++++
> OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h | 80 ++
> OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h | 138 ++
> OvmfPkg/Include/IndustryStandard/Xen/memory.h | 480 +++++++
> OvmfPkg/Include/IndustryStandard/Xen/sched.h | 174 +++
> OvmfPkg/Include/IndustryStandard/Xen/trace.h | 310 +++++
> OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h | 44 +
> OvmfPkg/Include/IndustryStandard/Xen/xen.h | 897 ++++++++++++
> OvmfPkg/Include/Protocol/Xenbus.h | 231 +++
> OvmfPkg/OvmfPkg.dec | 2 +
> OvmfPkg/OvmfPkgX64.dsc | 2 +
> OvmfPkg/OvmfPkgX64.fdf | 2 +
> OvmfPkg/XenPvBlkDxe/BlockFront.c | 583 ++++++++
> OvmfPkg/XenPvBlkDxe/BlockFront.h | 88 ++
> OvmfPkg/XenPvBlkDxe/BlockIo.c | 256 ++++
> OvmfPkg/XenPvBlkDxe/BlockIo.h | 109 ++
> OvmfPkg/XenPvBlkDxe/ComponentName.c | 160 +++
> OvmfPkg/XenPvBlkDxe/ComponentName.h | 91 ++
> OvmfPkg/XenPvBlkDxe/DriverBinding.h | 139 ++
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 388 ++++++
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h | 83 ++
> OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 61 +
> OvmfPkg/XenbusDxe/ComponentName.c | 160 +++
> OvmfPkg/XenbusDxe/ComponentName.h | 91 ++
> OvmfPkg/XenbusDxe/DriverBinding.h | 139 ++
> OvmfPkg/XenbusDxe/EventChannel.c | 71 +
> OvmfPkg/XenbusDxe/EventChannel.h | 40 +
> OvmfPkg/XenbusDxe/GrantTable.c | 204 +++
> OvmfPkg/XenbusDxe/GrantTable.h | 34 +
> OvmfPkg/XenbusDxe/Helpers.c | 9 +
> OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h | 7 +
> .../XenbusDxe/X64/InterlockedCompareExchange16.c | 41 +
> OvmfPkg/XenbusDxe/X64/TestAndClearBit.S | 10 +
> OvmfPkg/XenbusDxe/X64/hypercall.S | 16 +
> OvmfPkg/XenbusDxe/X64/hypercall.asm | 19 +
> OvmfPkg/XenbusDxe/XenBus.c | 363 +++++
> OvmfPkg/XenbusDxe/XenBus.h | 76 +
> OvmfPkg/XenbusDxe/XenHypercall.c | 101 ++
> OvmfPkg/XenbusDxe/XenHypercall.h | 44 +
> OvmfPkg/XenbusDxe/XenbusDxe.c | 452 ++++++
> OvmfPkg/XenbusDxe/XenbusDxe.h | 131 ++
> OvmfPkg/XenbusDxe/XenbusDxe.inf | 75 +
> OvmfPkg/XenbusDxe/Xenstore.c | 1472 ++++++++++++++++++++
> OvmfPkg/XenbusDxe/Xenstore.h | 367 +++++
> 53 files changed, 11314 insertions(+)
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_32.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen-x86_64.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/arch-x86/xen.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/event_channel.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/grant_table.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/hvm/hvm_op.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/hvm/params.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/blkif.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/protocols.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/ring.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/xenbus.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/io/xs_wire.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/memory.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/sched.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/trace.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/xen-compat.h
> create mode 100644 OvmfPkg/Include/IndustryStandard/Xen/xen.h
> create mode 100644 OvmfPkg/Include/Protocol/Xenbus.h
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.c
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.h
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.c
> create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.h
> create mode 100644 OvmfPkg/XenPvBlkDxe/ComponentName.c
> create mode 100644 OvmfPkg/XenPvBlkDxe/ComponentName.h
> create mode 100644 OvmfPkg/XenPvBlkDxe/DriverBinding.h
> create mode 100644 OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> create mode 100644 OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> create mode 100644 OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> create mode 100644 OvmfPkg/XenbusDxe/ComponentName.c
> create mode 100644 OvmfPkg/XenbusDxe/ComponentName.h
> create mode 100644 OvmfPkg/XenbusDxe/DriverBinding.h
> create mode 100644 OvmfPkg/XenbusDxe/EventChannel.c
> create mode 100644 OvmfPkg/XenbusDxe/EventChannel.h
> create mode 100644 OvmfPkg/XenbusDxe/GrantTable.c
> create mode 100644 OvmfPkg/XenbusDxe/GrantTable.h
> create mode 100644 OvmfPkg/XenbusDxe/Helpers.c
> create mode 100644 OvmfPkg/XenbusDxe/InterlockedCompareExchange16.h
> create mode 100644 OvmfPkg/XenbusDxe/X64/InterlockedCompareExchange16.c
> create mode 100644 OvmfPkg/XenbusDxe/X64/TestAndClearBit.S
> create mode 100644 OvmfPkg/XenbusDxe/X64/hypercall.S
> create mode 100644 OvmfPkg/XenbusDxe/X64/hypercall.asm
> create mode 100644 OvmfPkg/XenbusDxe/XenBus.c
> create mode 100644 OvmfPkg/XenbusDxe/XenBus.h
> create mode 100644 OvmfPkg/XenbusDxe/XenHypercall.c
> create mode 100644 OvmfPkg/XenbusDxe/XenHypercall.h
> create mode 100644 OvmfPkg/XenbusDxe/XenbusDxe.c
> create mode 100644 OvmfPkg/XenbusDxe/XenbusDxe.h
> create mode 100644 OvmfPkg/XenbusDxe/XenbusDxe.inf
> create mode 100644 OvmfPkg/XenbusDxe/Xenstore.c
> create mode 100644 OvmfPkg/XenbusDxe/Xenstore.h
>
> --
> Anthony PERARD
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls.
2014-07-16 17:37 ` [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls Konrad Rzeszutek Wilk
@ 2014-07-17 9:43 ` Wei Liu
0 siblings, 0 replies; 39+ messages in thread
From: Wei Liu @ 2014-07-17 9:43 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: Anthony PERARD, EDK2 devel, wei.liu2, Xen Devel
On Wed, Jul 16, 2014 at 01:37:05PM -0400, Konrad Rzeszutek Wilk wrote:
[...]
> > +
> > +EFI_STATUS
> > +XenHyperpageInit (
> > + XENBUS_DEVICE *Dev
> > + )
> > +{
> > + EFI_HOB_GUID_TYPE *GuidHob;
> > + EFI_XEN_INFO *XenInfo;
> > +
> > + GuidHob = GetFirstGuidHob (&gEfiXenInfoGuid);
> > + if (GuidHob == NULL) {
> > + DEBUG ((EFI_D_INFO, "XenbusInit: No xeninfo ?\n"));
> > + return EFI_NOT_FOUND;
> > + }
> > + XenInfo = (EFI_XEN_INFO*)GET_GUID_HOB_DATA (GuidHob);
> > + Dev->Hyperpage = XenInfo->HyperPages;
> > + return EFI_SUCCESS;
> > +}
>
> Could you describe in the git commit a bit on how/where in hvmloader
> this EfiXenInfoGuid is set and how the OVMF gets a handle on it?
>
I don't think hvmloader creates this info. OVMF creates this by itself.
See OvmfPkg/PlatformPei/Xen.c:XenConnect.
Wei.
> Thank you.
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver.
2014-07-16 17:25 ` [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver Konrad Rzeszutek Wilk
@ 2014-07-18 10:30 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 10:30 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 01:25:03PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:31PM +0100, Anthony PERARD wrote:
> > This includes Component Name and Driver Binding.
>
> The title says driver, but the cover letter said 'bus driver'.
> Should it have bus driver in it?
Do you mean having "Xenbus bus driver" in the title? It feels a bit
redondant, but might be better.
> .. snip..
> > +/**
> > + Retrieves a Unicode string that is the user readable name of the controller
> > + that is being managed by an EFI Driver.
> > +
> > + @param This A pointer to the EFI_COMPONENT_NAME_PROTOCOL instance.
> > + @param ControllerHandle The handle of a controller that the driver specified by
> > + This is managing. This handle specifies the controller
> > + whose name is to be returned.
> > + @param ChildHandle The handle of the child controller to retrieve the name
> > + of. This is an optional parameter that may be NULL. It
> > + will be NULL for device drivers. It will also be NULL
> > + for a bus drivers that wish to retrieve the name of the
> > + bus controller. It will not be NULL for a bus driver
> > + that wishes to retrieve the name of a child controller.
> .. snip..
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusDxeComponentNameGetControllerName (
> > + IN EFI_COMPONENT_NAME2_PROTOCOL *This,
> > + IN EFI_HANDLE ControllerHandle,
> > + IN EFI_HANDLE ChildHandle OPTIONAL,
> > + IN CHAR8 *Language,
> > + OUT CHAR16 **ControllerName
> > + )
> > +{
> > + EFI_STATUS Status;
> > +
> > + //
> > + // ChildHandle must be NULL for a Device Driver
>
> OK, but this is a bus driver in which case ChildHandle can be NULL or not (at
> least that is what the comment says) ?
Yes. Most (all?) of the code in this files (and this patch) comes from a
template. I haven't been through the comment yet. But yes this bus
driver can have childs.
> > + //
> > + if (ChildHandle != NULL) {
> > + return EFI_UNSUPPORTED;
> > + }
> > +
> > + //
> > + // Lookup name of controller specified by ControllerHandle
> > + //
> > + Status = EFI_UNSUPPORTED;
> > +
> > + return Status;
> > +}
> .. snip..
>
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > new file mode 100644
> > index 0000000..14113ad
> > --- /dev/null
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > @@ -0,0 +1,314 @@
> > +
> > +/** @file
> > + TODO: Brief Description of UEFI Driver XenbusDxe
> > +
> > + TODO: Detailed Description of UEFI Driver XenbusDxe
> > +
> > + TODO: Copyright for UEFI Driver XenbusDxe
> > +
> > + TODO: License for UEFI Driver XenbusDxe
> > +
> > +**/
> > +
> > +#include <IndustryStandard/Pci.h>
> > +#include <IndustryStandard/Acpi.h>
> > +#include <Library/DebugLib.h>
> > +
> > +#include "XenbusDxe.h"
> > +
> > +
> > +///
> > +/// Driver Binding Protocol instance
> > +///
> > +EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding = {
> > + XenbusDxeDriverBindingSupported,
> > + XenbusDxeDriverBindingStart,
> > + XenbusDxeDriverBindingStop,
> > + XENBUS_DXE_VERSION,
> > + NULL,
> > + NULL
> > +};
> > +
> > +
> > +/**
> > + Unloads an image.
> > +
> > + @param ImageHandle Handle that identifies the image to be unloaded.
> > +
> > + @retval EFI_SUCCESS The image has been unloaded.
> > + @retval EFI_INVALID_PARAMETER ImageHandle is not a valid image handle.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusDxeUnload (
> > + IN EFI_HANDLE ImageHandle
> > + )
> > +{
> > + EFI_STATUS Status;
> > +
> > + EFI_HANDLE *HandleBuffer;
> > + UINTN HandleCount;
> > + UINTN Index;
> > +
> > +
> > + Status = EFI_SUCCESS;
>
> Is that neccessary considering the next line you overwrite it?
No, it is not.
> > +
> > + //
> > + // Retrieve array of all handles in the handle database
> > + //
> > + Status = gBS->LocateHandleBuffer (
> > + AllHandles,
> > + NULL,
> > + NULL,
> > + &HandleCount,
> > + &HandleBuffer
> > + );
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> > +
> > + //
> > + // Disconnect the current driver from handles in the handle database
> > + //
> > + for (Index = 0; Index < HandleCount; Index++) {
> > + Status = gBS->DisconnectController (HandleBuffer[Index], gImageHandle, NULL);
> > + }
>
> Should you check the Status?
> I guess it doesn't matter at all. In which case why don't you just
> (void)gBS->DisconnectController ..
Will do.
> > +
> > + //
> > + // Free the array of handles
> > + //
> > + FreePool (HandleBuffer);
> > +
> > +
> > + //
> > + // Uninstall protocols installed in the driver entry point
> > + //
> > + Status = gBS->UninstallMultipleProtocolInterfaces (
> > + ImageHandle,
> > + &gEfiDriverBindingProtocolGuid, &gXenbusDxeDriverBinding,
> > + &gEfiComponentNameProtocolGuid, &gXenbusDxeComponentName,
> > + &gEfiComponentName2ProtocolGuid, &gXenbusDxeComponentName2,
> > + NULL
> > + );
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> > +
> > +
> > + //
> > + // Do any additional cleanup that is required for this driver
> > + //
> > +
> > + return EFI_SUCCESS;
> > +}
> > +
> > +/**
> > + This is the declaration of an EFI image entry point. This entry point is
> > + the same for UEFI Applications, UEFI OS Loaders, and UEFI Drivers including
> > + both device drivers and bus drivers.
> > +
> > + @param ImageHandle The firmware allocated handle for the UEFI image.
> > + @param SystemTable A pointer to the EFI System Table.
> > +
> > + @retval EFI_SUCCESS The operation completed successfully.
> > + @retval Others An unexpected error occurred.
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusDxeDriverEntryPoint (
> > + IN EFI_HANDLE ImageHandle,
> > + IN EFI_SYSTEM_TABLE *SystemTable
> > + )
> > +{
> > + EFI_STATUS Status;
> > +
> > +
> > + Status = EFI_SUCCESS;
>
> Not needed.
>
> > +
> > +
> > + //
> > + // Install UEFI Driver Model protocol(s).
> > + //
> > + Status = EfiLibInstallDriverBindingComponentName2 (
> > + ImageHandle,
> > + SystemTable,
> > + &gXenbusDxeDriverBinding,
> > + ImageHandle,
> > + &gXenbusDxeComponentName,
> > + &gXenbusDxeComponentName2
> > + );
> > + ASSERT_EFI_ERROR (Status);
> > +
> > +
> > + return Status;
> > +}
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event.
2014-07-16 17:32 ` [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event Konrad Rzeszutek Wilk
@ 2014-07-18 10:32 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 10:32 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 01:32:54PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:32PM +0100, Anthony PERARD wrote:
> > The ExitBoot event is used to disconnect from the device before the
> > next operating system start using them.
> >
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> > ---
> > OvmfPkg/XenbusDxe/XenbusDxe.c | 39 ++++++++++++++++++++++++++++++++++++---
> > OvmfPkg/XenbusDxe/XenbusDxe.h | 10 ++++++++++
> > 2 files changed, 46 insertions(+), 3 deletions(-)
> >
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > index 14113ad..2c3d9b8 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > @@ -16,7 +16,6 @@
> >
> > #include "XenbusDxe.h"
> >
> > -
> > ///
> > /// Driver Binding Protocol instance
> > ///
> > @@ -30,6 +29,8 @@ EFI_DRIVER_BINDING_PROTOCOL gXenbusDxeDriverBinding = {
> > };
> >
> >
> > +XENBUS_DEVICE *mMyDevice;
> > +
> > /**
> > Unloads an image.
> >
> > @@ -229,6 +230,19 @@ XenbusDxeDriverBindingSupported (
> > return Status;
> > }
> >
> > +VOID
> > +EFIAPI
> > +NotifyExitBoot (
> > + IN EFI_EVENT Event,
> > + IN VOID *Context
> > + )
> > +{
> > + XENBUS_DEVICE *Dev = Context;
> > +
> > + gBS->DisconnectController(Dev->ControllerHandle,
> > + Dev->This->DriverBindingHandle, NULL);
> > +}
> > +
> > /**
> > Starts a device controller or a bus controller.
> >
> > @@ -272,7 +286,22 @@ XenbusDxeDriverBindingStart (
> > IN EFI_DEVICE_PATH_PROTOCOL *RemainingDevicePath OPTIONAL
> > )
> > {
> > - return EFI_UNSUPPORTED;
> > + EFI_STATUS Status;
> > + XENBUS_DEVICE *Dev;
> > +
> > + Dev = AllocateZeroPool (sizeof (*Dev));
> > + Dev->Signature = XENBUS_DEVICE_SIGNATURE;
> > + Dev->This = This;
> > + Dev->ControllerHandle = ControllerHandle;
> > +
> > + Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
> > + NotifyExitBoot,
> > + (VOID*) Dev,
> > + &Dev->ExitBootEvent);
> > + ASSERT_EFI_ERROR (Status);
> > +
> > + mMyDevice = Dev;
> > + return EFI_SUCCESS;
> > }
> >
> > /**
> > @@ -310,5 +339,9 @@ XenbusDxeDriverBindingStop (
> > IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
> > )
> > {
> > - return EFI_UNSUPPORTED;
> > + XENBUS_DEVICE *Dev = mMyDevice;
> > +
> > + gBS->CloseEvent (Dev->ExitBootEvent);
> > +
> > + return EFI_SUCCESS;
> > }
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > index 6f22732..c1ca87a 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > @@ -75,4 +75,14 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
> > #define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
> >
> >
> > +typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
> > +
> > +#define XENBUS_DEVICE_SIGNATURE SIGNATURE_32 ('X','B','b','d')
>
> I get the 'XB', but 'bd' ? Should it be 'XenB' ? Or perhaps
> 'XBus' ?
I don't remember about 'bd', maybus "bus driver". I think I'll go with
'XBus'.
> > +struct _XENBUS_DEVICE {
> > + UINT32 Signature;
> > + EFI_DRIVER_BINDING_PROTOCOL *This;
> > + EFI_HANDLE ControllerHandle;
> > + EFI_EVENT ExitBootEvent;
> > +};
> > +
> > #endif
> > --
> > Anthony PERARD
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol.
2014-07-16 17:39 ` [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol Konrad Rzeszutek Wilk
@ 2014-07-18 10:36 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 10:36 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 01:39:36PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:34PM +0100, Anthony PERARD wrote:
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
>
> Could you describe a bit why you need to use it?
>
> I think I know why - that is you need to poke at its IO ports
> but it would be good to mention that in here.
I thinks it's need to be open so that, OVMF will not be able to start
the drivers a second time (OpenProtocol would return ALREADY_STARTED or
something equivalent).
So far I did not do anything with the IO ports, I probably should. But I
don't think it will be needed to explain that before it's actually done.
> > ---
> > OvmfPkg/XenbusDxe/XenbusDxe.c | 16 ++++++++++++++++
> > OvmfPkg/XenbusDxe/XenbusDxe.h | 1 +
> > 2 files changed, 17 insertions(+)
> >
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > index 63ea31b..ba5e8f4 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > @@ -290,11 +290,25 @@ XenbusDxeDriverBindingStart (
> > {
> > EFI_STATUS Status;
> > XENBUS_DEVICE *Dev;
> > + EFI_PCI_IO_PROTOCOL *PciIo;
> > +
> > + Status = gBS->OpenProtocol (
> > + ControllerHandle,
> > + &gEfiPciIoProtocolGuid,
> > + (VOID **)&PciIo,
> > + This->DriverBindingHandle,
> > + ControllerHandle,
> > + EFI_OPEN_PROTOCOL_BY_DRIVER
> > + );
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> >
> > Dev = AllocateZeroPool (sizeof (*Dev));
> > Dev->Signature = XENBUS_DEVICE_SIGNATURE;
> > Dev->This = This;
> > Dev->ControllerHandle = ControllerHandle;
> > + Dev->PciIo = PciIo;
> >
> > Status = XenHyperpageInit (Dev);
> > ASSERT_EFI_ERROR (Status);
> > @@ -351,5 +365,7 @@ XenbusDxeDriverBindingStop (
> >
> > gBS->CloseEvent (Dev->ExitBootEvent);
> >
> > + gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
> > + This->DriverBindingHandle, ControllerHandle);
> > return EFI_SUCCESS;
> > }
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > index 6140f94..d57e3c8 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > @@ -84,6 +84,7 @@ struct _XENBUS_DEVICE {
> > UINT32 Signature;
> > EFI_DRIVER_BINDING_PROTOCOL *This;
> > EFI_HANDLE ControllerHandle;
> > + EFI_PCI_IO_PROTOCOL *PciIo;
> > EFI_EVENT ExitBootEvent;
> >
> > VOID *Hyperpage;
> > --
> > Anthony PERARD
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol.
2014-07-16 17:42 ` [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol Konrad Rzeszutek Wilk
@ 2014-07-18 10:40 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 10:40 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 01:42:09PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:35PM +0100, Anthony PERARD wrote:
> > This protocol will be used for communication between a PV driver (like a
> > PV block driver) and the Xenbus/Xenstore.
> >
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> > ---
> > OvmfPkg/Include/Protocol/Xenbus.h | 62 +++++++++++++++++++++++++++++++++++++++
> > OvmfPkg/OvmfPkg.dec | 2 ++
> > OvmfPkg/XenbusDxe/ComponentName.c | 12 ++++++++
> > OvmfPkg/XenbusDxe/XenbusDxe.h | 3 ++
> > OvmfPkg/XenbusDxe/XenbusDxe.inf | 1 +
> > 5 files changed, 80 insertions(+)
> > create mode 100644 OvmfPkg/Include/Protocol/Xenbus.h
> >
> > diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
> > new file mode 100644
> > index 0000000..191cee1
> > --- /dev/null
> > +++ b/OvmfPkg/Include/Protocol/Xenbus.h
> > @@ -0,0 +1,62 @@
> > +
> > +/** @file
> > + TODO: Brief Description of Protocol Xenbus
> > +
> > + TODO: Detailed Description of Protocol Xenbus
> > +
> > + TODO: Copyright for Protocol Xenbus
> > +
> > + TODO: License for Protocol Xenbus
> > +
> > +**/
> > +
> > +#ifndef __PROTOCOL_XENBUS_H__
> > +#define __PROTOCOL_XENBUS_H__
> > +
> > +#define XENBUS_PROTOCOL_GUID \
> > + {0x3d3ca290, 0xb9a5, 0x11e3, {0xb7, 0x5d, 0xb8, 0xac, 0x6f, 0x7d, 0x65, 0xe6}}
>
> For those who are not familiar with these GUIDs and live in the Linux world - how
> are they constructed? Is there a chance for collision ?
Well, I actually don't know. To get the GUID, I've used UefiDriverWizard
which gave me this GUID and a bunch of files to actually code in. That's
way there is so many comment written TODO or that may no make sense.
> > +
> > +///
> > +/// Forward declaration
> > +///
> > +typedef struct _XENBUS_PROTOCOL XENBUS_PROTOCOL;
> > +
> > +
> > +#include <IndustryStandard/Xen/grant_table.h>
> > +
> > +///
> > +/// Function prototypes
> > +///
> > +
> > +typedef
> > +EFI_STATUS
> > +(EFIAPI *XENBUS_GRANT_ACCESS)(
> > + IN XENBUS_PROTOCOL *This,
> > + IN domid_t DomainId,
> > + IN UINTN Frame,
> > + IN BOOLEAN ReadOnly,
> > + OUT grant_ref_t *refp
> > + );
> > +
> > +typedef
> > +EFI_STATUS
> > +(EFIAPI *XENBUS_GRANT_END_ACCESS)(
> > + IN XENBUS_PROTOCOL *This,
> > + IN grant_ref_t Ref
> > + );
> > +
> > +
> > +///
> > +/// Protocol structure
> > +///
> > +struct _XENBUS_PROTOCOL {
> > + XENBUS_GRANT_ACCESS GrantAccess;
> > + XENBUS_GRANT_END_ACCESS GrantEndAccess;
> > + //
> > + // Place protocol data fields here
> > + //
> > +};
> > +
> > +extern EFI_GUID gXenbusProtocolGuid;
> > +
> > +#endif
> > diff --git a/OvmfPkg/OvmfPkg.dec b/OvmfPkg/OvmfPkg.dec
> > index f829247..095d40a 100644
> > --- a/OvmfPkg/OvmfPkg.dec
> > +++ b/OvmfPkg/OvmfPkg.dec
> > @@ -47,6 +47,8 @@
> > [Protocols]
> > gVirtioDeviceProtocolGuid = {0xfa920010, 0x6785, 0x4941, {0xb6, 0xec, 0x49, 0x8c, 0x57, 0x9f, 0x16, 0x0a}}
> > gBlockMmioProtocolGuid = {0x6b558ce3, 0x69e5, 0x4c67, {0xa6, 0x34, 0xf7, 0xfe, 0x72, 0xad, 0xbe, 0x84}}
> > + ## Include/Protocol/Xenbus.h
> > + gXenbusProtocolGuid = {0x3d3ca290, 0xb9a5, 0x11e3, {0xb7, 0x5d, 0xb8, 0xac, 0x6f, 0x7d, 0x65, 0xe6}}
>
> Formatting? Looks like the others use tabs?
Probably, yes, I'll change that.
> >
> > [PcdsFixedAtBuild]
> > gUefiOvmfPkgTokenSpaceGuid.PcdOvmfPeiMemFvBase|0x0|UINT32|0
> > diff --git a/OvmfPkg/XenbusDxe/ComponentName.c b/OvmfPkg/XenbusDxe/ComponentName.c
> > index 2cf11b5..fd46cc8 100644
> > --- a/OvmfPkg/XenbusDxe/ComponentName.c
> > +++ b/OvmfPkg/XenbusDxe/ComponentName.c
> > @@ -140,6 +140,18 @@ XenbusDxeComponentNameGetControllerName (
> > }
> >
> > //
> > + // Make sure this driver is currently managing ControllerHandle
> > + //
> > + Status = EfiTestManagedDevice (
> > + ControllerHandle,
> > + gXenbusDxeDriverBinding.DriverBindingHandle,
> > + &gXenbusProtocolGuid
> > + );
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> > +
> > + //
> > // Lookup name of controller specified by ControllerHandle
> > //
> > Status = EFI_UNSUPPORTED;
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > index d57e3c8..975fb6b 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > @@ -42,6 +42,9 @@
> > //
> > // Produced Protocols
> > //
> > +// Xen interface version used:
> > +#define __XEN_INTERFACE_VERSION__ 0x00040400
> > +#include <Protocol/Xenbus.h>
> >
> >
> > //
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> > index 8b69f93..1a6b131 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> > @@ -55,6 +55,7 @@
> > gEfiPciIoProtocolGuid
> > gEfiComponentName2ProtocolGuid
> > gEfiComponentNameProtocolGuid
> > + gXenbusProtocolGuid
> >
> >
> > [Guids]
> > --
> > Anthony PERARD
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions.
2014-07-16 17:48 ` [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions Konrad Rzeszutek Wilk
@ 2014-07-18 10:53 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 10:53 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Steven Smith, Grzegorz Milos, Xen Devel
On Wed, Jul 16, 2014 at 01:48:22PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:37PM +0100, Anthony PERARD wrote:
> > There are used to grant access of pages to other Xen domains.
> >
> > This code originaly comes from the Xen Project, and more precisely from
> > MiniOS.
>
> Should you mention something about its license?
Just in case, maybe. But both mini-os and ovmf are mostly BSD, so it's
probably not a big issue either way. I'll make sure to better describe
the original licenses in the commit messages.
> >
> > Signed-off-by: Steven Smith <sos22@cam.ac.uk>
> > Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> > ---
> > OvmfPkg/XenbusDxe/GrantTable.c | 204 ++++++++++++++++++++++++++++++++++++++++
> > OvmfPkg/XenbusDxe/GrantTable.h | 34 +++++++
> > OvmfPkg/XenbusDxe/XenbusDxe.c | 15 +++
> > OvmfPkg/XenbusDxe/XenbusDxe.inf | 2 +
> > 4 files changed, 255 insertions(+)
> > create mode 100644 OvmfPkg/XenbusDxe/GrantTable.c
> > create mode 100644 OvmfPkg/XenbusDxe/GrantTable.h
> >
> > diff --git a/OvmfPkg/XenbusDxe/GrantTable.c b/OvmfPkg/XenbusDxe/GrantTable.c
> > new file mode 100644
> > index 0000000..97bd15a
> > --- /dev/null
> > +++ b/OvmfPkg/XenbusDxe/GrantTable.c
> > @@ -0,0 +1,204 @@
> > +/*
> > + ****************************************************************************
> > + * (C) 2006 - Cambridge University
> > + ****************************************************************************
> > + *
> > + * File: GrantTable.c
> > + * Author: Steven Smith (sos22@cam.ac.uk)
> > + * Changes: Grzegorz Milos (gm281@cam.ac.uk)
> > + *
> > + * Date: July 2006
> > + *
> > + * Environment: Xen Minimal OS
> > + * Description: Simple grant tables implementation. About as stupid as it's
> > + * possible to be and still work.
> > + *
> > + ****************************************************************************
> > + */
> > +#include "XenbusDxe.h"
> > +
> > +#include <IndustryStandard/Xen/memory.h>
> > +
> > +#include "XenHypercall.h"
> > +
> > +#include "GrantTable.h"
> > +#include "InterlockedCompareExchange16.h"
> > +
> > +#define NR_RESERVED_ENTRIES 8
> > +
> > +/* NR_GRANT_FRAMES must be less than or equal to that configured in Xen */
> > +#define NR_GRANT_FRAMES 4
> > +#define NR_GRANT_ENTRIES (NR_GRANT_FRAMES * EFI_PAGE_SIZE / sizeof(grant_entry_v1_t))
> > +
> > +STATIC grant_entry_v1_t *GrantTable = NULL;
> > +STATIC grant_ref_t GrantList[NR_GRANT_ENTRIES];
> > +#ifdef GNT_DEBUG
> > +STATIC BOOLEAN GrantInUseList[NR_GRANT_ENTRIES];
> > +#endif
> > +
> > +STATIC
> > +VOID
> > +XenGrantTablePutFreeEntry (
> > + grant_ref_t Ref
> > + )
> > +{
> > + //local_irq_save(flags);
> > + /* MemoryFence (); */
> > + // lock ?
>
> You just need some form of locking. If EFI has some simple Mutex() you can
> use that.
No, no mutex available, but there are some form of lock. It increase the
priority of the current task, so nothing else can run.
> > +#ifdef GNT_DEBUG
> > + ASSERT (GrantInUseList[Ref]);
> > + GrantInUseList[Ref] = FALSE;
> > +#endif
> > + GrantList[Ref] = GrantList[0];
> > + GrantList[0] = Ref;
> > + //local_irq_restore(flags);
> > + /* MemoryFence (); */
> > +}
> > +
> > +STATIC
> > +grant_ref_t
> > +XenGrantTableGetFreeEntry (
> > + VOID
> > + )
> > +{
> > + UINTN Ref;
> > + /* local_irq_save(flags); */
> > + // lock ??
> > + Ref = GrantList[0];
> > + ASSERT (Ref >= NR_RESERVED_ENTRIES && Ref < NR_GRANT_ENTRIES);
> > + GrantList[0] = GrantList[Ref];
> > +#ifdef GNT_DEBUG
> > + ASSERT (!GrantInUseList[Ref]);
> > + GrantInUseList[Ref] = TRUE;
> > +#endif
> > + /* local_irq_restore(flags); */
> > + return Ref;
> > +}
> > +
> > +STATIC
> > +grant_ref_t
> > +XenGrantTableGrantAccess (
> > + IN domid_t DomainId,
> > + IN UINTN Frame,
> > + IN BOOLEAN ReadOnly
> > + )
> > +{
> > + grant_ref_t Ref;
> > + UINT32 Flags;
> > +
> > + ASSERT (GrantTable != NULL);
> > + Ref = XenGrantTableGetFreeEntry ();
> > + GrantTable[Ref].frame = Frame;
> > + GrantTable[Ref].domid = DomainId;
> > + MemoryFence ();
> > + Flags = GTF_permit_access;
> > + if (ReadOnly) {
> > + Flags |= GTF_readonly;
> > + }
> > + GrantTable[Ref].flags = Flags;
> > +
> > + return Ref;
> > +}
> > +
> > +STATIC
> > +EFI_STATUS
> > +XenGrantTableEndAccess (
> > + grant_ref_t Ref
> > + )
> > +{
> > + UINT16 Flags, OldFlags;
> > +
> > + ASSERT (GrantTable != NULL);
> > + ASSERT (Ref >= NR_RESERVED_ENTRIES && Ref < NR_GRANT_ENTRIES);
> > +
> > + OldFlags = GrantTable[Ref].flags;
> > + do {
> > + if ((Flags = OldFlags) & (GTF_reading | GTF_writing)) {
> > + DEBUG ((EFI_D_WARN, "WARNING: g.e. still in use! (%x)\n", Flags));
> > + return EFI_NOT_READY;
> > + }
> > + OldFlags = InterlockedCompareExchange16 (&GrantTable[Ref].flags, Flags, 0);
> > + } while (OldFlags != Flags);
> > +
> > + XenGrantTablePutFreeEntry (Ref);
> > + return EFI_SUCCESS;
> > +}
> > +
> > +VOID
> > +XenGrantTableInit (
> > + IN XENBUS_DEVICE *Dev,
> > + IN UINT64 MmioAddr
> > + )
> > +{
> > + xen_add_to_physmap_t Args;
> > + INTN Index;
> > + INTN ReturnCode;
> > +
> > +#ifdef GNT_DEBUG
> > + SetMem(GrantInUseList, sizeof (GrantInUseList), 1);
> > +#endif
> > + for (Index = NR_RESERVED_ENTRIES; Index < NR_GRANT_ENTRIES; Index++) {
> > + XenGrantTablePutFreeEntry (Index);
> > + }
> > +
> > + GrantTable = (VOID*)(UINTN) MmioAddr;
> > + for (Index = 0; Index < NR_GRANT_FRAMES; Index++) {
> > + Args.domid = DOMID_SELF;
> > + Args.idx = Index;
> > + Args.space = XENMAPSPACE_grant_table;
> > + Args.gpfn = (((xen_pfn_t) GrantTable) >> EFI_PAGE_SHIFT) + Index;
> > + ReturnCode = XenHypercallMemoryOp (Dev, XENMEM_add_to_physmap, &Args);
> > + if (ReturnCode != 0) {
> > + DEBUG ((EFI_D_ERROR, "Xen GrantTable, add_to_physmap hypercall error: %d\n", ReturnCode));
> > + }
> > + }
> > +}
> > +
> > +VOID
> > +XenGrantTableDeinit (
> > + XENBUS_DEVICE *Dev
> > + )
> > +{
> > + INTN ReturnCode, Index;
> > + xen_remove_from_physmap_t Args;
> > +
> > + if (GrantTable == NULL) {
> > + return;
> > + }
> > +
> > + for (Index = NR_GRANT_FRAMES - 1; Index >= 0; Index--) {
> > + Args.domid = DOMID_SELF;
> > + Args.gpfn = (((xen_pfn_t) GrantTable) >> EFI_PAGE_SHIFT) + Index;
> > + DEBUG ((EFI_D_INFO, "%a %d, removing %X\n", __func__, __LINE__, Args.gpfn));
> > + ReturnCode = XenHypercallMemoryOp (Dev, XENMEM_remove_from_physmap, &Args);
> > + if (ReturnCode != 0) {
> > + DEBUG ((EFI_D_ERROR, "Xen GrantTable, remove_from_physmap hypercall error: %d\n", ReturnCode));
> > + }
> > + }
> > + // XXX remove it from the physmap?
>
> Didn't you just do that?
Yes, a comment to remove...
> > + GrantTable = NULL;
> > +}
> > +
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusGrantAccess (
> > + IN XENBUS_PROTOCOL *This,
> > + IN domid_t DomainId,
> > + IN UINTN Frame, // MFN
> > + IN BOOLEAN ReadOnly,
> > + OUT grant_ref_t *RefPtr
> > + )
> > +{
> > + *RefPtr = XenGrantTableGrantAccess (DomainId, Frame, ReadOnly);
> > + return EFI_SUCCESS;
> > +}
> > +
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusGrantEndAccess (
> > + IN XENBUS_PROTOCOL *This,
> > + IN grant_ref_t Ref
> > + )
> > +{
> > + return XenGrantTableEndAccess (Ref);
> > +}
> > diff --git a/OvmfPkg/XenbusDxe/GrantTable.h b/OvmfPkg/XenbusDxe/GrantTable.h
> > new file mode 100644
> > index 0000000..05ab4be
> > --- /dev/null
> > +++ b/OvmfPkg/XenbusDxe/GrantTable.h
> > @@ -0,0 +1,34 @@
> > +#ifndef __GNTTAB_H__
> > +#define __GNTTAB_H__
> > +
> > +#include <IndustryStandard/Xen/grant_table.h>
> > +
> > +VOID
> > +XenGrantTableInit (
> > + IN XENBUS_DEVICE *Dev,
> > + IN UINT64 MmioAddr
> > + );
> > +
> > +VOID
> > +XenGrantTableDeinit (
> > + IN XENBUS_DEVICE *Dev
> > + );
> > +
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusGrantAccess (
> > + IN XENBUS_PROTOCOL *This,
> > + IN domid_t DomainId,
> > + IN UINTN Frame, // MFN
> > + IN BOOLEAN ReadOnly,
> > + OUT grant_ref_t *RefPtr
> > + );
> > +
> > +EFI_STATUS
> > +EFIAPI
> > +XenbusGrantEndAccess (
> > + IN XENBUS_PROTOCOL *This,
> > + IN grant_ref_t Ref
> > + );
> > +
> > +#endif /* !__GNTTAB_H__ */
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > index ba5e8f4..b19055d 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > @@ -17,6 +17,7 @@
> > #include "XenbusDxe.h"
> >
> > #include "XenHypercall.h"
> > +#include "GrantTable.h"
> >
> > ///
> > /// Driver Binding Protocol instance
> > @@ -291,6 +292,8 @@ XenbusDxeDriverBindingStart (
> > EFI_STATUS Status;
> > XENBUS_DEVICE *Dev;
> > EFI_PCI_IO_PROTOCOL *PciIo;
> > + EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR *BarDesc;
> > + UINT64 MmioAddr;
> >
> > Status = gBS->OpenProtocol (
> > ControllerHandle,
> > @@ -310,12 +313,23 @@ XenbusDxeDriverBindingStart (
> > Dev->ControllerHandle = ControllerHandle;
> > Dev->PciIo = PciIo;
> >
> > + Status = PciIo->GetBarAttributes (PciIo, PCI_BAR_IDX1, NULL, (VOID**) &BarDesc);
>
> I think we have a spec somewhere for this device. Do you think
> it might make sense to reference it here?
Probably, yes. But I'm not sure yet if it's the right things to do here.
I'm just use the BAR address to map the grant table in.
> > + ASSERT_EFI_ERROR (Status);
> > + ASSERT (BarDesc->ResType == ACPI_ADDRESS_SPACE_TYPE_MEM);
>
> I sooo hope the spec mandates that it is always this type of BAR.
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself.
2014-07-16 19:04 ` [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself Konrad Rzeszutek Wilk
@ 2014-07-18 11:04 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 11:04 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 03:04:50PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:43PM +0100, Anthony PERARD wrote:
> > This is a bus-like on top of Xenstore. It will look for advertised
> > ParaVirtualized devices and initialize them by producing XenBus
> > protocol.
> >
> > Origin: FreeBSD 10.0
> >
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> > ---
> > OvmfPkg/Include/Protocol/Xenbus.h | 9 +
> > OvmfPkg/XenbusDxe/XenBus.c | 360 ++++++++++++++++++++++++++++++++++++++
> > OvmfPkg/XenbusDxe/XenBus.h | 76 ++++++++
> > OvmfPkg/XenbusDxe/XenbusDxe.c | 61 +++++++
> > OvmfPkg/XenbusDxe/XenbusDxe.h | 18 ++
> > OvmfPkg/XenbusDxe/XenbusDxe.inf | 3 +
> > 6 files changed, 527 insertions(+)
> > create mode 100644 OvmfPkg/XenbusDxe/XenBus.c
> > create mode 100644 OvmfPkg/XenbusDxe/XenBus.h
> >
> > diff --git a/OvmfPkg/Include/Protocol/Xenbus.h b/OvmfPkg/Include/Protocol/Xenbus.h
> > index ef4d0c2..25408c1 100644
> > --- a/OvmfPkg/Include/Protocol/Xenbus.h
> > +++ b/OvmfPkg/Include/Protocol/Xenbus.h
> > @@ -113,6 +113,14 @@ XENSTORE_STATUS
> > );
> >
> > typedef
> > +XENSTORE_STATUS
> > +(EFIAPI *XENBUS_SET_STATE)(
> > + IN XENBUS_PROTOCOL *This,
> > + IN XENSTORE_TRANSACTION Transaction,
> > + IN XenbusState State
> > + );
> > +
> > +typedef
> > EFI_STATUS
> > (EFIAPI *XENBUS_GRANT_ACCESS)(
> > IN XENBUS_PROTOCOL *This,
> > @@ -171,6 +179,7 @@ struct _XENBUS_PROTOCOL {
> > XENBUS_XS_REMOVE XsRemove;
> > XENBUS_XS_TRANSACTION_START XsTransactionStart;
> > XENBUS_XS_TRANSACTION_END XsTransactionEnd;
> > + XENBUS_SET_STATE SetState;
> >
> > XENBUS_GRANT_ACCESS GrantAccess;
> > XENBUS_GRANT_END_ACCESS GrantEndAccess;
> > diff --git a/OvmfPkg/XenbusDxe/XenBus.c b/OvmfPkg/XenbusDxe/XenBus.c
> > new file mode 100644
> > index 0000000..b0cf1ba
> > --- /dev/null
> > +++ b/OvmfPkg/XenbusDxe/XenBus.c
> > @@ -0,0 +1,360 @@
> > +/******************************************************************************
> > + * Copyright (C) 2010 Spectra Logic Corporation
> > + * Copyright (C) 2008 Doug Rabson
> > + * Copyright (C) 2005 Rusty Russell, IBM Corporation
> > + * Copyright (C) 2005 Mike Wray, Hewlett-Packard
> > + * Copyright (C) 2005 XenSource Ltd
> > + *
> > + * This file may be distributed separately from the Linux kernel, or
> > + * incorporated into other software packages, subject to the following license:
> > + *
> > + * Permission is hereby granted, free of charge, to any person obtaining a copy
> > + * of this source file (the "Software"), to deal in the Software without
> > + * restriction, including without limitation the rights to use, copy, modify,
> > + * merge, publish, distribute, sublicense, and/or sell copies of the Software,
> > + * and to permit persons to whom the Software is furnished to do so, subject to
> > + * the following conditions:
> > + *
> > + * The above copyright notice and this permission notice shall be included in
> > + * all copies or substantial portions of the Software.
> > + *
> > + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> > + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> > + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
> > + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> > + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> > + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
> > + * IN THE SOFTWARE.
> > + */
> > +
> > +#include "XenbusDxe.h"
> > +
> > +#include <Library/PrintLib.h>
> > +
> > +#include "XenBus.h"
> > +#include "GrantTable.h"
> > +#include "Xenstore.h"
> > +#include "EventChannel.h"
> > +
> > +#include <IndustryStandard/Xen/io/xenbus.h>
> > +
> > +STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData;
> > +
> > +STATIC XENBUS_DEVICE_PATH gXenbusDevicePathTemplate = {
> > + .Vendor.Header.Type = HARDWARE_DEVICE_PATH,
> > + .Vendor.Header.SubType = HW_VENDOR_DP,
> > + .Vendor.Header.Length[0] = (UINT8) sizeof (XENBUS_DEVICE_PATH),
> > + .Vendor.Header.Length[1] = (UINT8) (sizeof (XENBUS_DEVICE_PATH) >> 8),
> > + .Vendor.Guid = XENBUS_PROTOCOL_GUID,
> > + .Type = 0,
> > + .DeviceId = 0
> > +};
> > +
> > +
> > +/**
> > + * Search our internal record of configured devices (not the XenStore)
> > + * to determine if the XenBus device indicated by \a node is known to
> > + * the system.
> > + *
> > + * \param Dev The XenBus bus instance to search for device children.
> > + * \param node The XenStore node path for the device to find.
> > + *
> > + * \return The device_t of the found device if any, or NULL.
> > + *
> > + * \note device_t is a pointer type, so it can be compared against
> > + * NULL for validity.
> > + */
> > +STATIC
> > +XENBUS_PRIVATE_DATA *
> > +XenbusDeviceInitialized (
> > + IN XENBUS_DEVICE *Dev,
> > + IN CONST CHAR8 *Node
> > + )
> > +{
> > + LIST_ENTRY *Entry;
> > + XENBUS_PRIVATE_DATA *Child;
> > + XENBUS_PRIVATE_DATA *Result;
> > +
> > + if (IsListEmpty (&Dev->ChildList)) {
> > + return NULL;
> > + }
> > +
> > + Result = NULL;
> > + for (Entry = GetFirstNode (&Dev->ChildList);
> > + !IsNodeAtEnd (&Dev->ChildList, Entry);
> > + Entry = GetNextNode (&Dev->ChildList, Entry)) {
> > + Child = XENBUS_PRIVATE_DATA_FROM_LINK (Entry);
> > + if (!AsciiStrCmp (Child->XenbusIo.Node, Node)) {
> > + Result = Child;
> > + break;
> > + }
> > + }
> > +
> > + return (Result);
> > +}
> > +
> > +STATIC
> > +XenbusState
> > +XenbusReadDriverState (
> > + IN CONST CHAR8 *Path
> > + )
> > +{
> > + XenbusState State;
> > + CHAR8 *Ptr;
> > + XENSTORE_STATUS Status;
> > +
> > + Status = XenstoreRead (XST_NIL, Path, "state", NULL, (VOID **)&Ptr);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + State = XenbusStateClosed;
> > + } else {
> > + State = AsciiStrDecimalToUintn (Ptr);
> > + FreePool (Ptr);
> > + }
>
> I think the FreePool(Ptr) should be here. You want to it
> irregardless of what the error might be (thought the implemenation
> in XenstoreRead et.all is pretty clear that we won't initialize
> *Ptr to anything if !Status. But just in case this changes in
> the future..
OK, will do. But I might have to check for NULL.
> > +
> > + return State;
> > +}
> > +
> > +STATIC
> > +EFI_STATUS
> > +XenbusAddDevice (
> > + XENBUS_DEVICE *Dev,
> > + CONST CHAR8 *Type,
> > + CONST CHAR8 *Id)
> > +{
> > + CHAR8 DevicePath[XENSTORE_ABS_PATH_MAX];
> > + XENSTORE_STATUS StatusXenstore;
> > + XENBUS_PRIVATE_DATA *Private;
> > + EFI_STATUS Status;
> > + XENBUS_DEVICE_PATH *TempXenbusPath;
> > + VOID *ChildPciIo;
> > +
> > + AsciiSPrint (DevicePath, sizeof (DevicePath),
> > + XENBUS_XENSTORE_NODE "/%a/%a", Type, Id);
> > +
> > + Status = EFI_NOT_STARTED; //ENXIO; // Device not configured
> > +
> > + if (XenstorePathExists (XST_NIL, DevicePath, "")) {
> > + XENBUS_PRIVATE_DATA *Child;
> > + enum xenbus_state State;
> > + CHAR8 *BackendPath;
> > +
> > + Child = XenbusDeviceInitialized (Dev, DevicePath);
> > + if (Child != NULL) {
> > + /*
> > + * We are already tracking this node
> > + */
> > + Status = EFI_SUCCESS;
> > + goto out;
> > + }
> > +
> > + State = XenbusReadDriverState (DevicePath);
> > + if (State != XenbusStateInitialising) {
> > + /*
> > + * Device is not new, so ignore it. This can
> > + * happen if a device is going away after
> > + * switching to Closed.
> > + */
> > + DEBUG ((EFI_D_INFO, "Xenbus: Device %a ignored. "
> > + "State %d\n", DevicePath, State));
> > + Status = EFI_SUCCESS;
> > + goto out;
> > + }
> > +
> > + StatusXenstore = XenstoreRead (XST_NIL, DevicePath, "backend",
> > + NULL, (VOID **) &BackendPath);
> > + if (StatusXenstore != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR, "xenbus: %a no backend path.\n", DevicePath));
> > + Status = EFI_NOT_FOUND;
> > + goto out;
> > + }
> > +
> > + Private = AllocateCopyPool (sizeof *Private, &gXenbusPrivateData);
> > + InsertTailList (&Dev->ChildList, &Private->Link);
> > +
> > + Private->XenbusIo.Type = AsciiStrDup (Type);
> > + Private->XenbusIo.Node = AsciiStrDup (DevicePath);
> > + Private->XenbusIo.Backend = BackendPath;
> > + Private->XenbusIo.DeviceId = AsciiStrDecimalToUintn (Id);
> > + Private->Dev = Dev;
> > +
> > + TempXenbusPath = AllocateCopyPool (sizeof (XENBUS_DEVICE_PATH),
> > + &gXenbusDevicePathTemplate);
> > + if (!AsciiStrCmp (Private->XenbusIo.Type, "vbd")) {
> > + TempXenbusPath->Type = XENBUS_DEVICE_PATH_TYPE_VBD;
> > + }
> > + TempXenbusPath->DeviceId = Private->XenbusIo.DeviceId;
> > + Private->DevicePath = (XENBUS_DEVICE_PATH *)AppendDevicePathNode (
> > + Dev->DevicePath,
> > + &TempXenbusPath->Vendor.Header);
>
> Could you provide a comment explaining why you do that please?
Will do.
I think bus drivers need to do that. It is used to specify wich device
to boot on. Without it done here, it would be hard to differentiate two
PV block drivers from each other.
> > +
> > + Status = gBS->InstallMultipleProtocolInterfaces (
> > + &Private->Handle,
> > + &gEfiDevicePathProtocolGuid, Private->DevicePath,
> > + &gXenbusProtocolGuid, &Private->XenbusIo,
> > + NULL);
> > + if (EFI_ERROR (Status)) {
>
> Shouldn't you free 'Private' and 'BackendPath' ? And
> also de-link yourself from 'Dev->ChildList' ?
Yes, I need to fix the error handling.
> > + return Status;
> > + }
> > +
> > + Status = gBS->OpenProtocol (Dev->ControllerHandle,
> > + &gEfiPciIoProtocolGuid,
> > + &ChildPciIo, Dev->This->DriverBindingHandle,
> > + Private->Handle,
> > + EFI_OPEN_PROTOCOL_BY_CHILD_CONTROLLER);
> > + if (EFI_ERROR (Status)) {
> > + DEBUG ((EFI_D_ERROR, "open by child controller fail (%r)\n",
> > + Status));
> > + }
> > + } else {
> > + DEBUG ((EFI_D_INFO, "Xenbus: does not exist: %a\n", DevicePath));
> > + }
> > +
> > +out:
> > + return Status;
> > +}
> > +
> > +/**
> > + * \brief Enumerate all devices of the given type on this bus.
> > + *
> > + * \param Dev NewBus device_t for this XenBus front bus instance.
> > + * \param type String indicating the device sub-tree (e.g. "vfb", "vif")
> > + * to enumerate.
> > + *
> > + * Devices that are found are entered into the NewBus hierarchy via
> > + * xenbusb_add_device(). xenbusb_add_device() ignores duplicate detects
> > + * and ignores duplicate devices, so it can be called unconditionally
> > + * for any device found in the XenStore.
> > + */
> > +STATIC
> > +VOID
> > +XenbusEnumerateDeviceType (
> > + XENBUS_DEVICE *Dev,
> > + CONST CHAR8 *Type
> > + )
> > +{
> > + CONST CHAR8 **Directory;
> > + UINTN Index;
> > + UINT32 Count;
> > + XENSTORE_STATUS Status;
> > +
> > + Status = XenstoreListDirectory (XST_NIL,
> > + XENBUS_XENSTORE_NODE, Type,
> > + &Count, &Directory);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + return;
> > + }
> > + for (Index = 0; Index < Count; Index++) {
> > + XenbusAddDevice (Dev, Type, Directory[Index]);
> > + }
> > +
> > + FreePool (Directory);
> > +}
> > +
> > +
> > +/**
> > + * \brief Enumerate the devices on a XenBus bus and register them with
> > + * the NewBus device tree.
> > + *
> > + * xenbusb_enumerate_bus() will create entries (in state DS_NOTPRESENT)
> > + * for nodes that appear in the XenStore, but will not invoke probe/attach
> > + * operations on drivers. Probe/Attach processing must be separately
> > + * performed via an invocation of xenbusb_probe_children(). This is usually
> > + * done via the xbs_probe_children task.
> > + *
> > + * \param Dev XenBus Bus device softc of the owner of the bus to enumerate.
> > + *
> > + * \return On success, 0. Otherwise an errno value indicating the
> > + * type of failure.
> > + */
> > +XENSTORE_STATUS
> > +XenbusEnumerateBus (
> > + XENBUS_DEVICE *Dev
> > + )
> > +{
> > + CONST CHAR8 **Types;
> > + UINTN Index;
> > + UINT32 Count;
> > + XENSTORE_STATUS Status;
> > +
> > + Status = XenstoreListDirectory (XST_NIL,
> > + XENBUS_XENSTORE_NODE, "",
> > + &Count, &Types);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + return Status;
> > + }
> > +
> > + for (Index = 0; Index < Count; Index++) {
> > + XenbusEnumerateDeviceType (Dev, Types[Index]);
> > + }
> > +
> > + FreePool (Types);
> > +
> > + return XENSTORE_STATUS_SUCCESS;
> > +}
> > +
> > +STATIC
> > +XENSTORE_STATUS
> > +EFIAPI
> > +XenbusSetState (
> > + IN XENBUS_PROTOCOL *This,
> > + IN XENSTORE_TRANSACTION Transaction,
> > + IN enum xenbus_state NewState
> > + )
> > +{
> > + XENBUS_PRIVATE_DATA *Private;
> > + enum xenbus_state CurrentState;
> > + XENSTORE_STATUS Status;
> > + CHAR8 *Temp;
> > +
> > + DEBUG ((EFI_D_INFO, "Xenbus: Set state to %d\n", NewState));
> > +
> > + Private = XENBUS_PRIVATE_DATA_FROM_THIS (This);
> > +
> > + Status = XenstoreRead (Transaction, This->Node, "state", NULL, (VOID **)&Temp);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + goto Error;
> > + }
> > + CurrentState = AsciiStrDecimalToUintn (Temp);
> > + FreePool (Temp);
> > + if (CurrentState == NewState) {
> > + DEBUG ((EFI_D_INFO, "%a %d, same state\n", __func__, __LINE__));
> > + goto Error;
> > + }
> > +
> > + do {
> > + Status = XenstoreSPrint (Transaction, This->Node, "state", "%d", NewState);
> > + } while (Status == XENSTORE_STATUS_EAGAIN);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR, "Xenbus: fail to writing new state\n"));
>
> failed to write new state
OK.
> > + goto Error;
> > + }
> > + DEBUG ((EFI_D_INFO, "Xenbus: Set state to %d, done\n", NewState));
> > +
> > +Error:
>
> s/Error/Out/
>
> as it is OK for to set to the same state twice.
OK.
> > + return Status;
> > +}
> > +
> > +STATIC XENBUS_PRIVATE_DATA gXenbusPrivateData = {
> > + .Signature = XENBUS_PRIVATE_DATA_SIGNATURE,
> > +
> > + .XenbusIo.XsRead = XenbusXenstoreRead,
> > + .XenbusIo.XsBackendRead = XenbusXenstoreBackendRead,
> > + .XenbusIo.XsPrintf = XenbusXenstoreSPrint,
> > + .XenbusIo.XsRemove = XenbusXenstoreRemove,
> > + .XenbusIo.XsTransactionStart = XenbusXenstoreTransactionStart,
> > + .XenbusIo.XsTransactionEnd = XenbusXenstoreTransactionEnd,
> > + .XenbusIo.SetState = XenbusSetState,
> > + .XenbusIo.GrantAccess = XenbusGrantAccess,
> > + .XenbusIo.GrantEndAccess = XenbusGrantEndAccess,
> > + .XenbusIo.RegisterWatch = XenbusRegisterWatch,
> > + .XenbusIo.RegisterWatchBackend = XenbusRegisterWatchBackend,
> > + .XenbusIo.UnregisterWatch = XenbusUnregisterWatch,
> > + .XenbusIo.WaitForWatch = XenbusWaitForWatch,
> > +
> > + .XenbusIo.Type = NULL,
> > + .XenbusIo.Node = NULL,
> > + .XenbusIo.Backend = NULL,
> > +
> > + .Dev = NULL
> > +};
> > diff --git a/OvmfPkg/XenbusDxe/XenBus.h b/OvmfPkg/XenbusDxe/XenBus.h
> > new file mode 100644
> > index 0000000..63d0815
> > --- /dev/null
> > +++ b/OvmfPkg/XenbusDxe/XenBus.h
> > @@ -0,0 +1,76 @@
> > +/*-
> > + * Core definitions and data structures shareable across OS platforms.
> > + *
> > + * Copyright (c) 2010 Spectra Logic Corporation
> > + * Copyright (C) 2008 Doug Rabson
> > + * All rights reserved.
> > + *
> > + * Redistribution and use in source and binary forms, with or without
> > + * modification, are permitted provided that the following conditions
> > + * are met:
> > + * 1. Redistributions of source code must retain the above copyright
> > + * notice, this list of conditions, and the following disclaimer,
> > + * without modification.
> > + * 2. Redistributions in binary form must reproduce at minimum a disclaimer
> > + * substantially similar to the "NO WARRANTY" disclaimer below
> > + * ("Disclaimer") and any redistribution must be conditioned upon
> > + * including a substantially similar Disclaimer requirement for further
> > + * binary redistribution.
> > + *
> > + * NO WARRANTY
> > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
> > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> > + * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
> > + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
> > + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
> > + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
> > + * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
> > + * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
> > + * POSSIBILITY OF SUCH DAMAGES.
> > + *
> > + * $FreeBSD: release/10.0.0/sys/xen/xenbus/xenbusb.h 222975 2011-06-11 04:59:01Z gibbs $
> > + */
> > +#ifndef _XEN_XENBUS_XENBUSB_H
> > +#define _XEN_XENBUS_XENBUSB_H
> > +
> > +#include "XenbusDxe.h"
> > +
> > +#define XENBUS_DEVICE_PATH_TYPE_VBD 0x1
> > +struct _XENBUS_DEVICE_PATH {
> > + VENDOR_DEVICE_PATH Vendor;
> > + UINT8 Type;
> > + UINT16 DeviceId;
> > +};
> > +
> > +/**
> > + * The VM relative path to the XenStore subtree this
> > + * bus attachment manages.
> > + *
> > + * OR
> > + *
> > + * The XenStore path to the XenStore subtree for this XenBus bus.
> > + */
> > +#define XENBUS_XENSTORE_NODE "device"
> > +
> > +
> > +/**
> > + * \brief Perform common XenBus bus attach processing.
> > + *
> > + * \param Dev The NewBus device representing this XenBus bus.
> > + *
> > + * \return On success, 0. Otherwise an errno value indicating the
> > + * type of failure.
> > + *
> > + * Intiailizes the softc for this bus, installs an interrupt driven
> > + * configuration hook to block boot processing until XenBus devices fully
> > + * configure, performs an initial probe/attach of the bus, and registers
> > + * a XenStore watch so we are notified when the bus topology changes.
> > + */
> > +XENSTORE_STATUS
> > +XenbusEnumerateBus (
> > + XENBUS_DEVICE *Dev
> > + );
> > +
> > +#endif /* _XEN_XENBUS_XENBUSB_H */
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.c b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > index 2c85d5e..56225c4 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.c
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.c
> > @@ -19,6 +19,7 @@
> > #include "XenHypercall.h"
> > #include "GrantTable.h"
> > #include "Xenstore.h"
> > +#include "XenBus.h"
> >
> > ///
> > /// Driver Binding Protocol instance
> > @@ -295,6 +296,7 @@ XenbusDxeDriverBindingStart (
> > EFI_PCI_IO_PROTOCOL *PciIo;
> > EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR *BarDesc;
> > UINT64 MmioAddr;
> > + EFI_DEVICE_PATH_PROTOCOL *DevicePath;
> >
> > Status = gBS->OpenProtocol (
> > ControllerHandle,
> > @@ -308,11 +310,26 @@ XenbusDxeDriverBindingStart (
> > return Status;
> > }
> >
> > + Status = gBS->OpenProtocol (
> > + ControllerHandle,
> > + &gEfiDevicePathProtocolGuid,
> > + (VOID **) &DevicePath,
> > + This->DriverBindingHandle,
> > + ControllerHandle,
> > + EFI_OPEN_PROTOCOL_BY_DRIVER
> > + );
> > +
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> > +
> > Dev = AllocateZeroPool (sizeof (*Dev));
> > Dev->Signature = XENBUS_DEVICE_SIGNATURE;
> > Dev->This = This;
> > Dev->ControllerHandle = ControllerHandle;
> > Dev->PciIo = PciIo;
> > + Dev->DevicePath = DevicePath;
> > + InitializeListHead (&Dev->ChildList);
> >
> > Status = PciIo->GetBarAttributes (PciIo, PCI_BAR_IDX1, NULL, (VOID**) &BarDesc);
> > ASSERT_EFI_ERROR (Status);
> > @@ -334,6 +351,8 @@ XenbusDxeDriverBindingStart (
> > Status = XenstoreInit (Dev);
> > ASSERT_EFI_ERROR (Status);
> >
> > + XenbusEnumerateBus (Dev);
> > +
> > Status = gBS->CreateEvent (EVT_SIGNAL_EXIT_BOOT_SERVICES, TPL_CALLBACK,
> > NotifyExitBoot,
> > (VOID*) Dev,
> > @@ -379,12 +398,54 @@ XenbusDxeDriverBindingStop (
> > IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
> > )
> > {
> > + UINTN Index;
> > + XENBUS_PROTOCOL *XenbusIo;
> > + XENBUS_PRIVATE_DATA *ChildData;
> > + EFI_STATUS Status;
> > XENBUS_DEVICE *Dev = mMyDevice;
> >
> > + for (Index = 0; Index < NumberOfChildren; Index++) {
> > + Status = gBS->OpenProtocol (
> > + ChildHandleBuffer[Index],
> > + &gXenbusProtocolGuid,
> > + (VOID **) &XenbusIo,
> > + This->DriverBindingHandle,
> > + ControllerHandle,
> > + EFI_OPEN_PROTOCOL_GET_PROTOCOL);
> > + if (EFI_ERROR (Status)) {
> > + DEBUG ((EFI_D_ERROR, "XenbusDxe: get childre proto: %r\n", Status));
>
> get children protocol failed: %r
Will fix.
> > + continue;
> > + }
> > + ChildData = XENBUS_PRIVATE_DATA_FROM_THIS(XenbusIo);
> > + Status = gBS->DisconnectController(ChildData->Handle, NULL, NULL);
> > + if (EFI_ERROR (Status)) {
> > + DEBUG ((EFI_D_ERROR, "XenbusDxe: error disconnecting child: %r\n",
> > + Status));
> > + continue;
> > + }
> > +
> > + Status = gBS->UninstallMultipleProtocolInterfaces (
> > + ChildData->Handle,
> > + &gEfiDevicePathProtocolGuid, ChildData->DevicePath,
> > + &gXenbusProtocolGuid, &ChildData->XenbusIo,
> > + NULL);
> > + ASSERT_EFI_ERROR (Status);
> > +
> > + FreePool ((VOID*)ChildData->XenbusIo.Type);
> > + FreePool ((VOID*)ChildData->XenbusIo.Node);
> > + FreePool ((VOID*)ChildData->XenbusIo.Backend);
> > + FreePool (ChildData->DevicePath);
> > + RemoveEntryList (&ChildData->Link);
> > + FreePool (ChildData);
> > + }
> > +
> > gBS->CloseEvent (Dev->ExitBootEvent);
> > + // XXX xenbus cleanup
> > // XXX xenstore cleanup
> > XenGrantTableDeinit (Dev);
> >
> > + gBS->CloseProtocol(ControllerHandle, &gEfiDevicePathProtocolGuid,
> > + This->DriverBindingHandle, ControllerHandle);
> > gBS->CloseProtocol(ControllerHandle, &gEfiPciIoProtocolGuid,
> > This->DriverBindingHandle, ControllerHandle);
> > return EFI_SUCCESS;
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.h b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > index 4d9db72..3ff08dd 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.h
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.h
> > @@ -80,6 +80,7 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenbusDxeComponentName;
> > #define PCI_DEVICE_ID_XEN_PLATFORM 0x0001
> >
> >
> > +typedef struct _XENBUS_DEVICE_PATH XENBUS_DEVICE_PATH;
> > typedef struct _XENBUS_DEVICE XENBUS_DEVICE;
> >
> > #define XENBUS_DEVICE_SIGNATURE SIGNATURE_32 ('X','B','b','d')
> > @@ -89,11 +90,28 @@ struct _XENBUS_DEVICE {
> > EFI_HANDLE ControllerHandle;
> > EFI_PCI_IO_PROTOCOL *PciIo;
> > EFI_EVENT ExitBootEvent;
> > + EFI_DEVICE_PATH_PROTOCOL *DevicePath;
> > + LIST_ENTRY ChildList;
> >
> > VOID *Hyperpage;
> > shared_info_t *SharedInfo;
> > };
> >
> > +#define XENBUS_PRIVATE_DATA_SIGNATURE SIGNATURE_32 ('X', 'B', 'u', 's')
> > +typedef struct {
> > + UINTN Signature;
> > + LIST_ENTRY Link;
> > + EFI_HANDLE Handle;
> > + XENBUS_PROTOCOL XenbusIo;
> > + XENBUS_DEVICE *Dev;
> > + XENBUS_DEVICE_PATH *DevicePath;
> > +} XENBUS_PRIVATE_DATA;
> > +
> > +#define XENBUS_PRIVATE_DATA_FROM_THIS(a) \
> > + CR (a, XENBUS_PRIVATE_DATA, XenbusIo, XENBUS_PRIVATE_DATA_SIGNATURE)
> > +#define XENBUS_PRIVATE_DATA_FROM_LINK(a) \
> > + CR (a, XENBUS_PRIVATE_DATA, Link, XENBUS_PRIVATE_DATA_SIGNATURE)
> > +
> > /*
> > * Helpers
> > */
> > diff --git a/OvmfPkg/XenbusDxe/XenbusDxe.inf b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> > index 6c8260f..708fae6 100644
> > --- a/OvmfPkg/XenbusDxe/XenbusDxe.inf
> > +++ b/OvmfPkg/XenbusDxe/XenbusDxe.inf
> > @@ -40,6 +40,9 @@
> > EventChannel.h
> > Xenstore.c
> > Xenstore.h
> > + XenBus.c
> > + XenBus.h
> > + Helpers.c
> >
> > [Sources.X64]
> > X64/hypercall.S
> > --
> > Anthony PERARD
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client.
2014-07-16 19:41 ` [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client Konrad Rzeszutek Wilk
@ 2014-07-18 11:58 ` Anthony PERARD
[not found] ` <20140718115817.GG1582@perard.uk.xensource.com>
1 sibling, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 11:58 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel, Samuel Thibault
On Wed, Jul 16, 2014 at 03:41:17PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:46PM +0100, Anthony PERARD wrote:
> > This is the code that will do the actual communication between OVMF and
> > a PV block backend, where the block device lives.
> >
> > This implementation originally comes from Mini-OS, a part of the Xen
> > Project.
> >
> > Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com>
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> > ---
> > OvmfPkg/XenPvBlkDxe/BlockFront.c | 583 ++++++++++++++++++++++++++++++++++++
> > OvmfPkg/XenPvBlkDxe/BlockFront.h | 88 ++++++
> > OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 8 +
> > OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 2 +
> > 4 files changed, 681 insertions(+)
> > create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.c
> > create mode 100644 OvmfPkg/XenPvBlkDxe/BlockFront.h
> >
> > diff --git a/OvmfPkg/XenPvBlkDxe/BlockFront.c b/OvmfPkg/XenPvBlkDxe/BlockFront.c
> > new file mode 100644
> > index 0000000..ff5641c
> > --- /dev/null
> > +++ b/OvmfPkg/XenPvBlkDxe/BlockFront.c
> > @@ -0,0 +1,583 @@
> > +/* Minimal block driver for Mini-OS.
> > + * Copyright (c) 2007-2008 Samuel Thibault.
> > + * Based on netfront.c.
> > + */
> > +
> > +#include <Library/PrintLib.h>
> > +#include <Library/DebugLib.h>
> > +
> > +#include "BlockFront.h"
> > +
> > +#include <IndustryStandard/Xen/io/protocols.h>
> > +#include <IndustryStandard/Xen/io/xenbus.h>
> > +
> > +#include "inttypes.h"
> > +
> > +STATIC
> > +XENSTORE_STATUS
> > +XenbusReadInteger (
> > + IN XENBUS_PROTOCOL *This,
> > + IN CONST CHAR8 *Node,
> > + IN BOOLEAN FromBackend,
> > + OUT UINT64 *ValuePtr
> > + )
> > +{
> > + XENSTORE_STATUS Status;
> > + CHAR8 *p;
> > +
> > + if (!FromBackend) {
> > + Status = This->XsRead (This, XST_NIL, Node, (VOID**)&p);
> > + } else {
> > + Status = This->XsBackendRead (This, XST_NIL, Node, (VOID**)&p);
> > + }
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + return Status;
> > + }
> > + // XXX Will assert if p overflow UINT64 ...
> > + *ValuePtr = AsciiStrDecimalToUint64 (p);
> > + FreePool (p);
> > + return Status;
> > +}
> > +
> > +STATIC
> > +VOID
> > +XenPvBlockFree (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > + )
> > +{
> > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > +
> > + if (Dev->RingRef != 0) {
> > + XenbusIo->GrantEndAccess (XenbusIo, Dev->RingRef);
> > + }
> > + if (Dev->Ring.sring != NULL) {
> > + FreePages (Dev->Ring.sring, 1);
> > + }
> > + if (Dev->EventChannel != 0) {
> > + XenbusIo->EventChannelClose (XenbusIo, Dev->EventChannel);
> > + }
> > + FreePool (Dev);
> > +}
> > +
> > +XENSTORE_STATUS
> > +XenPvBlkWaitForBackendState (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev,
> > + IN XenbusState WaitedState,
>
> ExpectedState
>
> ?
Yes, will change.
> > + OUT XenbusState *LastStatePtr OPTIONAL
> > + )
> > +{
> > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > + XenbusState State;
> > + UINT64 Value;
> > + XENSTORE_STATUS Status;
> > +
> > + while (TRUE) {
> > + Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + return Status;
> > + }
> > + if (Value > XenbusStateReconfigured) {
> > + return XENSTORE_STATUS_EIO;
> > + }
> > + State = Value;
> > + if (State >= WaitedState) {
>
> Not State != WaitedState ?
I don't think so, the function is actually waiting for the State to be
equal to WaitedState. At least, that how I wrote it. There might be a
better algorithm.
> > + break;
> > + }
> > + DEBUG ((EFI_D_INFO,
> > + "XenPvBlk: waiting backend state %d, current: %d\n",
> > + WaitedState, State));
> > + XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
> > + }
> > +
> > + if (LastStatePtr != NULL) {
> > + *LastStatePtr = State;
> > + }
> > +
> > + return XENSTORE_STATUS_SUCCESS;
> > +}
> > +
> > +EFI_STATUS
> > +XenPvBlockFrontInitialization (
> > + IN XENBUS_PROTOCOL *XenbusIo,
> > + IN CONST CHAR8 *NodeName,
> > + OUT XEN_BLOCK_FRONT_DEVICE **DevPtr
> > + )
> > +{
> > + XENSTORE_TRANSACTION xbt;
> > + CHAR8 *DeviceType;
> > + blkif_sring_t *SharedRing;
> > + XENSTORE_STATUS Status;
> > + XEN_BLOCK_FRONT_DEVICE *Dev;
> > + XenbusState State;
> > + UINT64 Value;
> > +
> > + ASSERT (NodeName != NULL);
> > +
> > + Dev = AllocateZeroPool (sizeof (XEN_BLOCK_FRONT_DEVICE));
> > + Dev->Signature = XEN_BLOCK_FRONT_SIGNATURE;
> > + Dev->NodeName = NodeName;
> > + Dev->XenbusIo = XenbusIo;
> > + Dev->DeviceId = XenbusIo->DeviceId;
> > +
> > + XenbusIo->XsRead (XenbusIo, XST_NIL, "device-type", (VOID**)&DeviceType);
> > + if (AsciiStrCmp (DeviceType, "cdrom") == 0) {
> > + Dev->MediaInfo.CdRom = TRUE;
> > + } else {
> > + Dev->MediaInfo.CdRom = FALSE;
> > + }
> > + FreePool (DeviceType);
> > +
> > + Status = XenbusReadInteger (XenbusIo, "backend-id", FALSE, &Value);
> > + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT16_MAX) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to get backend-id (%d)\n",
> > + Status));
> > + goto Error;
> > + }
> > + Dev->DomainId = Value;
> > + XenbusIo->EventChannelAllocate (XenbusIo, Dev->DomainId, &Dev->EventChannel);
> > +
> > + SharedRing = (blkif_sring_t*) AllocatePages (1);
> > + SHARED_RING_INIT (SharedRing);
> > + FRONT_RING_INIT (&Dev->Ring, SharedRing, EFI_PAGE_SIZE);
> > + XenbusIo->GrantAccess (XenbusIo,
> > + Dev->DomainId,
> > + (INTN) SharedRing >> EFI_PAGE_SHIFT,
> > + FALSE,
> > + &Dev->RingRef);
> > +
> > +Again:
> > + Status = XenbusIo->XsTransactionStart (XenbusIo, &xbt);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_WARN, "XenPvBlk: Failed to start transaction, %d\n", Status));
>
> .. and do you want to error out?
The original code (mini-os) is continuing in thes case, so I thought it
was possible to do the initialisation without transaction. But maybe I'm
wrong and this should be an error.
> > + }
> > +
> > + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName, "ring-ref", "%d",
> > + Dev->RingRef);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write ring-ref.\n"));
> > + goto AbortTransaction;
> > + }
> > + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
> > + "event-channel", "%d", Dev->EventChannel);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write event-channel.\n"));
> > + goto AbortTransaction;
> > + }
> > + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
> > + "protocol", "%a", XEN_IO_PROTO_ABI_NATIVE);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write protocol.\n"));
> > + goto AbortTransaction;
> > + }
> > +
> > + Status = XenbusIo->SetState (XenbusIo, xbt, XenbusStateConnected);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed switch state.\n"));
> > + goto AbortTransaction;
> > + }
> > +
> > + Status = XenbusIo->XsTransactionEnd (XenbusIo, xbt, FALSE);
> > + if (Status == XENSTORE_STATUS_EAGAIN) {
> > + goto Again;
> > + }
> > +
> > + XenbusIo->RegisterWatchBackend (XenbusIo, "state", &Dev->StateWatchToken);
> > +
> > + // Waiting backend
> > + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateConnected, &State);
> > + if (Status != XENSTORE_STATUS_SUCCESS || State != XenbusStateConnected) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: backend for %a/%d not available, rc=%d state=%d\n",
> > + XenbusIo->Type, XenbusIo->DeviceId, Status, State));
> > + goto Error2;
> > + }
> > +
> > + Status = XenbusReadInteger (XenbusIo, "info", TRUE, &Value);
> > + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
> > + goto Error2;
> > + }
> > + Dev->MediaInfo.VDiskInfo = Value;
> > + if (Dev->MediaInfo.VDiskInfo & VDISK_READONLY) {
> > + Dev->MediaInfo.ReadWrite = FALSE;
> > + } else {
> > + Dev->MediaInfo.ReadWrite = TRUE;
> > + }
> > +
> > + Status = XenbusReadInteger (XenbusIo, "sectors", TRUE, &Dev->MediaInfo.Sectors);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + goto Error2;
> > + }
> > +
> > + Status = XenbusReadInteger (XenbusIo, "sector-size", TRUE, &Value);
> > + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
> > + goto Error2;
> > + }
> > + Dev->MediaInfo.SectorSize = Value;
>
> Why not &Dev->MediaInfo.SectorSize in XenbusReadInteger?
>
> Is it because of types? If so, please add a comment stating that.
Yes, because of the type, uint64 vs uint32.
Maybe I could rename the XenbusReadInteger to XenbusReadUint64, and it
would be more obvious that the function require a specific type.
> > +
> > + // Default value
> > + Value = 0;
> > + Status = XenbusReadInteger (XenbusIo, "feature-barrier", TRUE, &Value);
> > + if (Value == 1) {
> > + Dev->MediaInfo.FeatureBarrier = TRUE;
> > + } else {
> > + Dev->MediaInfo.FeatureBarrier = FALSE;
> > + }
> > +
> > + // Default value
> > + Value = 0;
> > + XenbusReadInteger (XenbusIo, "feature-flush-cache", TRUE, &Value);
> > + if (Value == 1) {
> > + Dev->MediaInfo.FeatureFlushCache = TRUE;
> > + } else {
> > + Dev->MediaInfo.FeatureFlushCache = FALSE;
> > + }
> > +
> > + DEBUG ((EFI_D_INFO, "XenPvBlk: New disk with %d sectors of %d bytes\n",
> > + Dev->MediaInfo.Sectors, Dev->MediaInfo.SectorSize));
> > +
> > + *DevPtr = Dev;
> > + return EFI_SUCCESS;
> > +
> > +Error2:
> > + XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
> > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
> > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
>
> XenbusIo->XsRemove (XenbusIo, XST_NIL, "protocol");
>
> ?
I don't know, but I can do that.
> > + goto Error;
> > +AbortTransaction:
> > + XenbusIo->XsTransactionEnd (XenbusIo, xbt, TRUE);
> > +Error:
> > + XenPvBlockFree (Dev);
> > + return EFI_DEVICE_ERROR;
> > +}
> > +
> > +VOID
> > +XenPvBlockFrontShutdown (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > + )
> > +{
> > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > + XENSTORE_STATUS Status;
> > + UINT64 Value;
> > +
> > + XenPvBlockSync (Dev);
> > +
> > + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosing);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: error while changing state to Closing: %d\n",
> > + Status));
> > + goto Close;
> > + }
> > +
> > + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosing, NULL);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: error while waiting for closing backend state: %d\n",
> > + Status));
> > + goto Close;
> > + }
> > +
> > + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosed);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: error while changing state to Closed: %d\n",
> > + Status));
> > + goto Close;
> > + }
> > +
> > + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosed, NULL);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: error while waiting for closed backend state: %d\n",
> > + Status));
> > + goto Close;
> > + }
> > +
> > + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateInitialising);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: error while changing state to initialising: %d\n",
> > + Status));
> > + goto Close;
> > + }
> > +
> > + while (TRUE) {
> > + Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
> > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: error while waiting for new backend state: %d\n",
> > + Status));
> > + goto Close;
> > + }
> > + if (Value < XenbusStateInitWait || Value >= XenbusStateClosed) {
> > + break;
> > + }
> > + DEBUG ((EFI_D_INFO,
> > + "XenPvBlk: waiting backend state %d, current: %d\n",
> > + XenbusStateInitWait, Value));
> > + XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
> > + }
> > +
> > +Close:
> > + XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
> > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
> > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
>
> "protocol"
> > +
> > + XenPvBlockFree (Dev);
> > +}
> > +
> > +STATIC
> > +VOID
> > +XenPvBlockWaitSlot (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > + )
> > +{
> > + /* Wait for a slot */
> > + if (RING_FULL (&Dev->Ring)) {
> > + while (TRUE) {
> > + XenPvBlockAsyncIoPoll (Dev);
> > + if (!RING_FULL(&Dev->Ring)) {
> > + break;
> > + }
> > + /* Really no slot, go to sleep. */
> > + // XXX wait for an event on Dev->EventChannel
> > + DEBUG ((EFI_D_INFO, "%a %d, waiting\n", __func__, __LINE__));
> > + }
> > + }
> > +}
> > +
> > +/* Issue an aio */
> > +VOID
> > +XenPvBlockAsyncIo (
> > + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> > + IN BOOLEAN IsWrite
> > + )
> > +{
> > + XEN_BLOCK_FRONT_DEVICE *Dev = IoData->Dev;
> > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > + blkif_request_t *Request;
> > + RING_IDX RingIndex;
> > + BOOLEAN Notify;
> > + INT32 NumSegments, Index;
> > + UINTN Start, End;
> > +
> > + // Can't io at non-sector-aligned location
> > + ASSERT(!(IoData->Offset & (Dev->MediaInfo.SectorSize - 1)));
> > + // Can't io non-sector-sized amounts
> > + ASSERT(!(IoData->Size & (Dev->MediaInfo.SectorSize - 1)));
> > + // Can't io non-sector-aligned buffer
> > + ASSERT(!((UINTN) IoData->Buffer & (Dev->MediaInfo.SectorSize - 1)));
> > +
> > + Start = (UINTN) IoData->Buffer & ~EFI_PAGE_MASK;
> > + End = ((UINTN) IoData->Buffer + IoData->Size + EFI_PAGE_SIZE - 1) & ~EFI_PAGE_MASK;
> > + IoData->NumRef = NumSegments = (End - Start) / EFI_PAGE_SIZE;
> > +
> > + // XXX is this comment still true ?
>
> I think it is outdated, as Linux would do max DMA up to the segment
> size that is advertised. Ditto for the SCSI stack.
OK, thanks.
> > + /* qemu's IDE max multsect is 16 (8KB) and SCSI max DMA was set to 32KB,
> > + * so max 44KB can't happen */
> > + ASSERT (NumSegments <= BLKIF_MAX_SEGMENTS_PER_REQUEST);
>
> But this check is OK as we can't truly fit more on the ring.
OK.
> > +
> > + XenPvBlockWaitSlot (Dev);
>
> What happens if there are multiple threads calling this
> function? Don't we need a mutex to only allow
> one thread to grab an RingIndex?
So, OVMF does not have threads but the are "events", called by a timer.
This code does not support it, yet, so I don't need any form of locking,
yet.
> > + RingIndex = Dev->Ring.req_prod_pvt;
> > + Request = RING_GET_REQUEST (&Dev->Ring, RingIndex);
> > +
> > + Request->operation = IsWrite ? BLKIF_OP_WRITE : BLKIF_OP_READ;
> > + Request->nr_segments = NumSegments;
> > + Request->handle = Dev->DeviceId;
> > + Request->id = (UINTN) IoData;
> > + Request->sector_number = IoData->Offset / 512;
> > +
> > + for (Index = 0; Index < NumSegments; Index++) {
> > + Request->seg[Index].first_sect = 0;
>
> Uh?
> > + Request->seg[Index].last_sect = EFI_PAGE_SIZE / 512 - 1;
>
> Uh? I thought you wanted to use the real sector values?
> > + }
> > + Request->seg[0].first_sect = ((UINTN) IoData->Buffer & EFI_PAGE_MASK) / 512;
> > + Request->seg[NumSegments - 1].last_sect =
> > + (((UINTN) IoData->Buffer + IoData->Size - 1) & EFI_PAGE_MASK) / 512;
>
> What about in between 0 and NumSegments? You should also
> fill out those values with valid entries.
It's done by the previous for loop. Maybe this part is hard to
understand without the interface explain.
include io/blkif.h says:
@first_sect: first sector in frame to transfer (inclusive).
@last_sect: last sector in frame to transfer (inclusive).
but does not explain what a frame is, beyond this comment on:
grant_ref_t gref; /* reference to I/O buffer frame */
(all this from the struct blkif_request_segment)
Or maybe frame == page.
Anyway, the way the segments are initialized feel weirds now, but I
don't think it's incorrect, otherwise, I would not be able to boot any
guest.
> > + for (Index = 0; Index < NumSegments; Index++) {
> > + UINTN Data = Start + Index * EFI_PAGE_SIZE;
> > + if (!IsWrite) {
> > + /* Trigger CoW if needed */
>
> Not sure I understand that reference?
Maybe a reference to the original commit will help to understand:
minios: Fix bug when blkfront reading into zero-mapped buffer by just poking the page.
No need to use virtual_to_mfn() for the ring since that is a real page.
7b1257946080b63421a8dc59c7b3dfd67997c643 (xen.git)
But I still don't undertand this.
> > + *(CHAR8 *)(Data + (Request->seg[Index].first_sect << 9)) = 0;
> > + MemoryFence ();
> > + }
> > + XenbusIo->GrantAccess (XenbusIo, Dev->DomainId,
> > + Data >> EFI_PAGE_SHIFT, IsWrite,
> > + &Request->seg[Index].gref);
> > + IoData->GrantRef[Index] = Request->seg[Index].gref;
> > + }
> > +
> > + Dev->Ring.req_prod_pvt = RingIndex + 1;
> > +
> > + MemoryFence ();
> > + RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
> > +
> > + if (Notify) {
> > + XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
> > + }
> > +}
> > +
> > +STATIC
> > +VOID
> > +XenPvBlockAsyncIoCallback (
> > + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> > + IN EFI_STATUS Status
> > + )
> > +{
> > + IoData->Data = (VOID *) 1;
> > + IoData->Callback = NULL;
> > +}
> > +
> > +VOID
> > +XenPvBlockIo (
> > + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> > + IN BOOLEAN IsWrite
> > + )
> > +{
> > + ASSERT (IoData->Callback == NULL);
> > + // XXX Signal/Event instead ?
>
> Maybe for later versions. I think this is OK for right now, thought
> it will mean we can only do one I/O .. and have then to wait
> until it is completed.
>
> That limititation should be mentioned in the commit description.
That right, right now, it's only one IO at a time. I think my comment is
more to tell me that Callback is not really one, and it could be replace
by something more Uefi, an event.
> > + IoData->Callback = XenPvBlockAsyncIoCallback;
> > + XenPvBlockAsyncIo (IoData, IsWrite);
> > + IoData->Data = NULL;
> > +
> > + while (TRUE) {
> > + XenPvBlockAsyncIoPoll (IoData->Dev);
> > + if (IoData->Data) {
> > + break;
> > + }
> > + }
> > +}
> > +
> > +STATIC
> > +VOID
> > +XenPvBlockPushOperation (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev,
> > + IN UINT8 Operation,
> > + IN UINT64 Id
> > + )
> > +{
> > + INT32 Index;
> > + blkif_request_t *Request;
> > + BOOLEAN Notify;
> > +
> > + XenPvBlockWaitSlot (Dev);
>
> How do we make sure that there aren't multiple threads
> calling this?
Not possible yet.
> > + Index = Dev->Ring.req_prod_pvt;
> > + Request = RING_GET_REQUEST(&Dev->Ring, Index);
> > + Request->operation = Operation;
> > + Request->nr_segments = 0;
> > + Request->handle = Dev->DeviceId;
> > + Request->id = Id;
> > + /* Not needed anyway, but the backend will check it */
> > + Request->sector_number = 0;
> > + Dev->Ring.req_prod_pvt = Index + 1;
> > + MemoryFence ();
> > + RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
> > + if (Notify) {
> > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > + XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
> > + }
> > +}
> > +
> > +VOID
> > +XenPvBlockSync (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > + )
> > +{
> > + if (Dev->MediaInfo.ReadWrite) {
> > + if (Dev->MediaInfo.FeatureBarrier) {
> > + XenPvBlockPushOperation (Dev, BLKIF_OP_WRITE_BARRIER, 0);
> > + }
> > +
> > + if (Dev->MediaInfo.FeatureFlushCache) {
> > + XenPvBlockPushOperation (Dev, BLKIF_OP_FLUSH_DISKCACHE, 0);
> > + }
> > + }
> > +
> > + /* Note: This won't finish if another thread enqueues requests. */
> > + while (TRUE) {
> > + XenPvBlockAsyncIoPoll (Dev);
> > + if (RING_FREE_REQUESTS (&Dev->Ring) == RING_SIZE (&Dev->Ring)) {
> > + break;
> > + }
> > + }
> > +}
> > +
> > +VOID
> > +XenPvBlockAsyncIoPoll (
> > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > + )
> > +{
> > + RING_IDX ProducerIndex, ConsumerIndex;
> > + blkif_response_t *Response;
> > + INT32 More;
> > +
> > + do {
> > + ProducerIndex = Dev->Ring.sring->rsp_prod;
> > + /* Ensure we see queued responses up to 'ProducerIndex'. */
> > + MemoryFence ();
> > + ConsumerIndex = Dev->Ring.rsp_cons;
> > +
> > + while (ConsumerIndex != ProducerIndex) {
> > + XEN_BLOCK_FRONT_IO *IoData;
> > + INT16 Status;
> > +
> > + Response = RING_GET_RESPONSE (&Dev->Ring, ConsumerIndex);
> > +
> > + IoData = (VOID *) (UINTN) Response->id;
> > + Status = Response->status;
> > +
> > + switch (Response->operation) {
> > + case BLKIF_OP_READ:
> > + case BLKIF_OP_WRITE:
> > + {
> > + INT32 Index;
> > +
> > + if (Status != BLKIF_RSP_OKAY) {
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: "
> > + "%a error %d on %a at offset %Ld, num bytes %Ld\n",
> > + Response->operation == BLKIF_OP_READ ? "read" : "write",
> > + Status, IoData->Dev->NodeName,
> > + IoData->Offset,
> > + IoData->Size));
> > + }
> > +
> > + for (Index = 0; Index < IoData->NumRef; Index++) {
> > + Dev->XenbusIo->GrantEndAccess (Dev->XenbusIo, IoData->GrantRef[Index]);
> > + }
> > +
> > + break;
> > + }
> > +
> > + case BLKIF_OP_WRITE_BARRIER:
> > + if (Status != BLKIF_RSP_OKAY) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: write barrier error %d\n", Status));
> > + }
> > + break;
> > + case BLKIF_OP_FLUSH_DISKCACHE:
> > + if (Status != BLKIF_RSP_OKAY) {
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: flush error %d\n", Status));
> > + }
> > + break;
> > +
> > + default:
> > + DEBUG ((EFI_D_ERROR,
> > + "XenPvBlk: unrecognized block operation %d response (status %d)\n",
> > + Response->operation, Status));
> > + break;
> > + }
> > +
> > + Dev->Ring.rsp_cons = ++ConsumerIndex;
> > + /* Nota: callback frees IoData itself */
> > + if (IoData && IoData->Callback) {
> > + IoData->Callback (IoData, Status ? EFI_DEVICE_ERROR : EFI_SUCCESS);
> > + }
> > + if (Dev->Ring.rsp_cons != ConsumerIndex) {
> > + /* We reentered, we must not continue here */
>
> Uh, how does that happen?
That an original comment, I don't think it's possible yet in ovmf.
> > + break;
> > + }
> > + }
> > +
> > + RING_FINAL_CHECK_FOR_RESPONSES (&Dev->Ring, More);
> > + } while (More != 0);
> > +}
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo.
2014-07-16 19:57 ` [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo Konrad Rzeszutek Wilk
@ 2014-07-18 12:10 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 12:10 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 03:57:23PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:47PM +0100, Anthony PERARD wrote:
> > Install the BlockIo protocol.
> >
> > Contributed-under: TianoCore Contribution Agreement 1.0
> > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
> > ---
> > OvmfPkg/XenPvBlkDxe/BlockIo.c | 256 ++++++++++++++++++++++++++++++++++++
> > OvmfPkg/XenPvBlkDxe/BlockIo.h | 109 +++++++++++++++
> > OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c | 59 +++++++++
> > OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h | 1 +
> > OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf | 2 +
> > 5 files changed, 427 insertions(+)
> > create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.c
> > create mode 100644 OvmfPkg/XenPvBlkDxe/BlockIo.h
> >
> > diff --git a/OvmfPkg/XenPvBlkDxe/BlockIo.c b/OvmfPkg/XenPvBlkDxe/BlockIo.c
> > new file mode 100644
> > index 0000000..4ac2411
> > --- /dev/null
> > +++ b/OvmfPkg/XenPvBlkDxe/BlockIo.c
> > @@ -0,0 +1,256 @@
> > +
> > +/** @file
> > + TODO: Brief Description of UEFI Driver XenPvBlkDxe
> > +
> > + TODO: Detailed Description of UEFI Driver XenPvBlkDxe
> > +
> > + TODO: Copyright for UEFI Driver XenPvBlkDxe
> > +
> > + TODO: License for UEFI Driver XenPvBlkDxe
> > +
> > +**/
> > +
> > +#include "XenPvBlkDxe.h"
> > +
> > +#include "BlockFront.h"
> > +
> > +///
> > +/// Block I/O Media structure
> > +///
> > +GLOBAL_REMOVE_IF_UNREFERENCED
> > +EFI_BLOCK_IO_MEDIA gXenPvBlkDxeBlockIoMedia = {
> > + 0, // MediaId
> > + FALSE, // RemovableMedia
> > + FALSE, // MediaPresent
> > + FALSE, // LogicalPartition
> > + TRUE, // ReadOnly
> > + FALSE, // WriteCaching
> > + 512, // BlockSize
> > + 512, // IoAlign
> > + 0, // LastBlock
> > + 0, // LowestAlignedLba
> > + 0, // LogicalBlocksPerPhysicalBlock
> > + 0 // OptimalTransferLengthGranularity
> > +};
> > +
> > +///
> > +/// Block I/O Protocol instance
> > +///
> > +GLOBAL_REMOVE_IF_UNREFERENCED
> > +EFI_BLOCK_IO_PROTOCOL gXenPvBlkDxeBlockIo = {
> > + EFI_BLOCK_IO_PROTOCOL_REVISION3, // Revision
> > + &gXenPvBlkDxeBlockIoMedia, // Media
> > + (EFI_BLOCK_RESET)XenPvBlkDxeBlockIoReset, // Reset
> > + XenPvBlkDxeBlockIoReadBlocks, // ReadBlocks
> > + XenPvBlkDxeBlockIoWriteBlocks, // WriteBlocks
> > + XenPvBlkDxeBlockIoFlushBlocks // FlushBlocks
> > +};
> > +
> > +
> > +
> > +
> > +STATIC
> > +EFI_STATUS
> > +XenPvBlkDxeBlockIoReadWriteBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This,
> > + IN UINT32 MediaId,
> > + IN EFI_LBA Lba,
> > + IN UINTN BufferSize,
> > + IN OUT VOID *Buffer,
> > + IN BOOLEAN IsWrite
> > + )
> > +{
> > + XEN_BLOCK_FRONT_IO IoData;
> > + EFI_BLOCK_IO_MEDIA *Media;
> > + UINTN NextOffset;
> > +
> > + if (Buffer == NULL) {
> > + DEBUG ((EFI_D_ERROR, "%a %d, buffer null\n", __func__, __LINE__));
>
> Heheh.
>
> > + return EFI_INVALID_PARAMETER;
> > + }
> > + if (BufferSize == 0) {
> > + return EFI_SUCCESS;
> > + }
> > +
> > + Media = This->Media;
> > +
> > + if (BufferSize % Media->BlockSize != 0) {
> > + DEBUG ((EFI_D_ERROR, "%a %d, bad buffer size\n", __func__, __LINE__));
> > + return EFI_BAD_BUFFER_SIZE;
> > + }
> > +
> > + if (Lba > Media->LastBlock ||
> > + (BufferSize / Media->BlockSize) - 1 > Media->LastBlock - Lba) {
> > + DEBUG ((EFI_D_ERROR, "%a %d, read too large\n", __func__, __LINE__));
> > + return EFI_INVALID_PARAMETER;
> > + }
> > +
> > + if (IsWrite && Media->ReadOnly) {
> > + DEBUG ((EFI_D_ERROR, "%a %d, read only\n", __func__, __LINE__));
> > + return EFI_WRITE_PROTECTED;
> > + }
> > +
> > + if ((Media->IoAlign > 1) && (UINTN)Buffer & (Media->IoAlign - 1)) {
> > + // This happen often when using grub2
> > + VOID *NewBuffer;
> > + EFI_STATUS Status;
> > +
> > + /* Try again with a properly aligned buffer. */
> > + NewBuffer = AllocateAlignedPages((BufferSize + EFI_PAGE_SIZE) / EFI_PAGE_SIZE,
> > + Media->IoAlign);
> > + if (!IsWrite) {
> > + Status = XenPvBlkDxeBlockIoReadBlocks (This, MediaId,
> > + Lba, BufferSize, NewBuffer);
> > + CopyMem (Buffer, NewBuffer, BufferSize);
> > + } else {
> > + CopyMem (NewBuffer, Buffer, BufferSize);
> > + Status = XenPvBlkDxeBlockIoWriteBlocks (This, MediaId,
> > + Lba, BufferSize, NewBuffer);
> > + }
> > + FreeAlignedPages (NewBuffer, (BufferSize + EFI_PAGE_SIZE) / EFI_PAGE_SIZE);
> > + return Status;
> > + }
> > +
> > +
> > + // XXX Check MediaId
> > +
> > + IoData.Dev = XEN_BLOCK_FRONT_FROM_BLOCK_IO (This);
> > + IoData.Offset = Lba * Media->BlockSize;
> > + IoData.Callback = NULL;
> > + IoData.Data = NULL;
> > + IoData.Size = BufferSize;
> > + IoData.Buffer = Buffer;
> > +
> > + NextOffset = Lba * Media->BlockSize;
>
> Or just = IoData.Offset ?
Yes.
> > + while (BufferSize > 0) {
> > + // This should work if every Buffer is page aligned
> > + // but block accept sector aligned, so FIXME
>
> Isn't the work-around for GRUB2 above doing this?
No, GRUB2 is worth, it's not even sector aligned... I think my comment
is about why a Size of EFI_PAGE_SIZE * BLKIF_MAX_SEGMENTS_PER_REQUEST
did not work and that I should find a better value than page_size * 8.
> > + IoData.Size = MIN(EFI_PAGE_SIZE * 8,//BLKIF_MAX_SEGMENTS_PER_REQUEST,
> > + BufferSize);
>
> Why 8 ?
I don't remember, but maybe the first value that works, or just a random
value that I've seen somewhere else ... anyway, that need to be change.
> > + IoData.Buffer = Buffer;
> > + IoData.Offset = NextOffset;
> > + BufferSize -= IoData.Size;
> > + Buffer = (VOID*) ((UINTN) Buffer + IoData.Size);
> > + NextOffset += IoData.Size;
> > + XenPvBlockIo (&IoData, IsWrite);
> > + if (!IoData.Data) {
>
> Well that wouldn't be possible as the XenPvBlockIo loops forever until that
> is done.
>
> However, you should probably have a comment here:
>
> /* XXX If we move to a different mechanism for CallBack this might
> go away */
That right.
> > + DEBUG ((EFI_D_ERROR, "%a %d, XenPvBlockIo no data...\n", __func__, __LINE__));
> > + return EFI_DEVICE_ERROR;
> > + }
> > + }
> > + return EFI_SUCCESS;
> > +}
> > +
> > +
> > +/**
> > + Read BufferSize bytes from Lba into Buffer.
> > +
> > + @param This Indicates a pointer to the calling context.
> > + @param MediaId Id of the media, changes every time the media is replaced.
> > + @param Lba The starting Logical Block Address to read from
> > + @param BufferSize Size of Buffer, must be a multiple of device block size.
> > + @param Buffer A pointer to the destination buffer for the data. The caller is
> > + responsible for either having implicit or explicit ownership of the buffer.
> > +
> > + @retval EFI_SUCCESS The data was read correctly from the device.
> > + @retval EFI_DEVICE_ERROR The device reported an error while performing the read.
> > + @retval EFI_NO_MEDIA There is no media in the device.
> > + @retval EFI_MEDIA_CHANGED The MediaId does not matched the current device.
> > + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> > + @retval EFI_INVALID_PARAMETER The read request contains LBAs that are not valid,
> > + or the buffer is not on proper alignment.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoReadBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This,
> > + IN UINT32 MediaId,
> > + IN EFI_LBA Lba,
> > + IN UINTN BufferSize,
> > + OUT VOID *Buffer
> > + )
> > +{
> > + return XenPvBlkDxeBlockIoReadWriteBlocks (This,
> > + MediaId, Lba, BufferSize, Buffer, FALSE);
> > +}
> > +
> > +/**
> > + Write BufferSize bytes from Lba into Buffer.
> > +
> > + @param This Indicates a pointer to the calling context.
> > + @param MediaId The media ID that the write request is for.
> > + @param Lba The starting logical block address to be written. The caller is
> > + responsible for writing to only legitimate locations.
> > + @param BufferSize Size of Buffer, must be a multiple of device block size.
> > + @param Buffer A pointer to the source buffer for the data.
> > +
> > + @retval EFI_SUCCESS The data was written correctly to the device.
> > + @retval EFI_WRITE_PROTECTED The device can not be written to.
> > + @retval EFI_DEVICE_ERROR The device reported an error while performing the write.
> > + @retval EFI_NO_MEDIA There is no media in the device.
> > + @retval EFI_MEDIA_CHNAGED The MediaId does not matched the current device.
> > + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> > + @retval EFI_INVALID_PARAMETER The write request contains LBAs that are not valid,
> > + or the buffer is not on proper alignment.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoWriteBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This,
> > + IN UINT32 MediaId,
> > + IN EFI_LBA Lba,
> > + IN UINTN BufferSize,
> > + IN VOID *Buffer
> > + )
> > +{
> > + return XenPvBlkDxeBlockIoReadWriteBlocks (This,
> > + MediaId, Lba, BufferSize, Buffer, TRUE);
> > +}
> > +
> > +/**
> > + Flush the Block Device.
> > +
> > + @param This Indicates a pointer to the calling context.
> > +
> > + @retval EFI_SUCCESS All outstanding data was written to the device
> > + @retval EFI_DEVICE_ERROR The device reported an error while writting back the data
> > + @retval EFI_NO_MEDIA There is no media in the device.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoFlushBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This
> > + )
> > +{
> > + XenPvBlockSync (XEN_BLOCK_FRONT_FROM_BLOCK_IO (This));
> > + return EFI_SUCCESS;
> > +}
> > +
> > +/**
> > + Reset the block device hardware.
> > +
> > + @param[in] This Indicates a pointer to the calling context.
> > + @param[in] ExtendedVerification Indicates that the driver may perform a more
> > + exhausive verfication operation of the device
> > + during reset.
> > +
> > + @retval EFI_SUCCESS The device was reset.
> > + @retval EFI_DEVICE_ERROR The device is not functioning properly and could
> > + not be reset.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoReset (
> > + IN EFI_BLOCK_IO2_PROTOCOL *This,
> > + IN BOOLEAN ExtendedVerification
> > + )
> > +{
> > + // Is there anything to do ?
> > + // since initialisation is done in Start (driver binding)
>
> Won't we leak memory? Or will it after an Reset do Stop, then Start?
>
No, I don't think so. Reset is probably to reset the hardware. That a
function that I need to take a closer look at.
> > + DEBUG ((EFI_D_INFO, "%a %d\n", __func__, __LINE__));
> > + return EFI_SUCCESS;
> > +}
> > diff --git a/OvmfPkg/XenPvBlkDxe/BlockIo.h b/OvmfPkg/XenPvBlkDxe/BlockIo.h
> > new file mode 100644
> > index 0000000..ec0fe2b
> > --- /dev/null
> > +++ b/OvmfPkg/XenPvBlkDxe/BlockIo.h
> > @@ -0,0 +1,109 @@
> > +
> > +/** @file
> > + TODO: Brief Description of UEFI Driver XenPvBlkDxe
> > +
> > + TODO: Detailed Description of UEFI Driver XenPvBlkDxe
> > +
> > + TODO: Copyright for UEFI Driver XenPvBlkDxe
> > +
> > + TODO: License for UEFI Driver XenPvBlkDxe
> > +
> > +**/
> > +
> > +/**
> > + Read BufferSize bytes from Lba into Buffer.
> > +
> > + @param This Indicates a pointer to the calling context.
> > + @param MediaId Id of the media, changes every time the media is replaced.
> > + @param Lba The starting Logical Block Address to read from
> > + @param BufferSize Size of Buffer, must be a multiple of device block size.
> > + @param Buffer A pointer to the destination buffer for the data. The caller is
> > + responsible for either having implicit or explicit ownership of the buffer.
> > +
> > + @retval EFI_SUCCESS The data was read correctly from the device.
> > + @retval EFI_DEVICE_ERROR The device reported an error while performing the read.
> > + @retval EFI_NO_MEDIA There is no media in the device.
> > + @retval EFI_MEDIA_CHANGED The MediaId does not matched the current device.
> > + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> > + @retval EFI_INVALID_PARAMETER The read request contains LBAs that are not valid,
> > + or the buffer is not on proper alignment.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoReadBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This,
> > + IN UINT32 MediaId,
> > + IN EFI_LBA Lba,
> > + IN UINTN BufferSize,
> > + OUT VOID *Buffer
> > + );
> > +
> > +/**
> > + Write BufferSize bytes from Lba into Buffer.
> > +
> > + @param This Indicates a pointer to the calling context.
> > + @param MediaId The media ID that the write request is for.
> > + @param Lba The starting logical block address to be written. The caller is
> > + responsible for writing to only legitimate locations.
> > + @param BufferSize Size of Buffer, must be a multiple of device block size.
> > + @param Buffer A pointer to the source buffer for the data.
> > +
> > + @retval EFI_SUCCESS The data was written correctly to the device.
> > + @retval EFI_WRITE_PROTECTED The device can not be written to.
> > + @retval EFI_DEVICE_ERROR The device reported an error while performing the write.
> > + @retval EFI_NO_MEDIA There is no media in the device.
> > + @retval EFI_MEDIA_CHNAGED The MediaId does not matched the current device.
> > + @retval EFI_BAD_BUFFER_SIZE The Buffer was not a multiple of the block size of the device.
> > + @retval EFI_INVALID_PARAMETER The write request contains LBAs that are not valid,
> > + or the buffer is not on proper alignment.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoWriteBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This,
> > + IN UINT32 MediaId,
> > + IN EFI_LBA Lba,
> > + IN UINTN BufferSize,
> > + IN VOID *Buffer
> > + );
> > +
> > +/**
> > + Flush the Block Device.
> > +
> > + @param This Indicates a pointer to the calling context.
> > +
> > + @retval EFI_SUCCESS All outstanding data was written to the device
> > + @retval EFI_DEVICE_ERROR The device reported an error while writting back the data
> > + @retval EFI_NO_MEDIA There is no media in the device.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoFlushBlocks (
> > + IN EFI_BLOCK_IO_PROTOCOL *This
> > + );
> > +
> > +/**
> > + Reset the block device hardware.
> > +
> > + @param[in] This Indicates a pointer to the calling context.
> > + @param[in] ExtendedVerification Indicates that the driver may perform a more
> > + exhausive verfication operation of the device
> > + during reset.
> > +
> > + @retval EFI_SUCCESS The device was reset.
> > + @retval EFI_DEVICE_ERROR The device is not functioning properly and could
> > + not be reset.
> > +
> > +**/
> > +EFI_STATUS
> > +EFIAPI
> > +XenPvBlkDxeBlockIoReset (
> > + IN EFI_BLOCK_IO2_PROTOCOL *This,
> > + IN BOOLEAN ExtendedVerification
> > + );
> > +
> > +extern EFI_BLOCK_IO_MEDIA gXenPvBlkDxeBlockIoMedia;
> > +extern EFI_BLOCK_IO_PROTOCOL gXenPvBlkDxeBlockIo;
> > diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> > index 7ec7d60..3991e83 100644
> > --- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> > +++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.c
> > @@ -266,6 +266,7 @@ XenPvBlkDxeDriverBindingStart (
> > EFI_STATUS Status;
> > XENBUS_PROTOCOL *XenbusIo;
> > XEN_BLOCK_FRONT_DEVICE *Dev;
> > + EFI_BLOCK_IO_MEDIA *Media;
> >
> > Status = gBS->OpenProtocol (
> > ControllerHandle,
> > @@ -284,6 +285,37 @@ XenPvBlkDxeDriverBindingStart (
> > return Status;
> > }
> >
> > + CopyMem (&Dev->BlockIo, &gXenPvBlkDxeBlockIo, sizeof (EFI_BLOCK_IO_PROTOCOL));
> > + Media = AllocateCopyPool (sizeof (EFI_BLOCK_IO_MEDIA),
> > + &gXenPvBlkDxeBlockIoMedia);
> > + if (Dev->MediaInfo.VDiskInfo & VDISK_REMOVABLE) {
> > + Media->RemovableMedia = TRUE;
> > + }
> > + Media->MediaPresent = TRUE;
> > + Media->ReadOnly = !Dev->MediaInfo.ReadWrite;
> > + if (Dev->MediaInfo.CdRom) {
> > + // If it's a cdrom, the blocksize value need to be 2048 for OVMF to
> > + // recognize it as a cdrom:
> > + // MdeModulePkg/Universal/Disk/PartitionDxe/ElTorito.c
> > + Media->BlockSize = 2048;
> > + } else {
> > + Media->BlockSize = Dev->MediaInfo.SectorSize;
> > + }
> > + Media->LastBlock = Dev->MediaInfo.Sectors - 1;
> > + // XXX What about LowestAlignedLba, LogicalBlocksPerPhysicalBlock and OptimalTransferLengthGranularity?
>
> I think you can skip LogicalBlocksPerPhysicalBlock and OptimalTransferLengthGranularity
> as they are used for discard operations.
Ok.
> The LowestAlignedLba .. no idea.
I think in pv block case, LowestAlignedLba will always be 0.
> > + Dev->BlockIo.Media = Media;
> > +
> > + Status = gBS->InstallMultipleProtocolInterfaces (
> > + &ControllerHandle,
> > + &gEfiBlockIoProtocolGuid, &Dev->BlockIo,
> > + NULL
> > + );
> > + if (EFI_ERROR (Status)) {
> > + // XXX: uninit everythings
>
> XenPvBlockFrontShutdown (Dev); ?
> FreePool(Media) ?
Yes.
> > + DEBUG ((EFI_D_ERROR, "XenPvBlk: install protocol fail: %r\n", Status));
> > + return Status;
> > + }
> > +
> > return EFI_SUCCESS;
> > }
> >
> > @@ -322,6 +354,33 @@ XenPvBlkDxeDriverBindingStop (
> > IN EFI_HANDLE *ChildHandleBuffer OPTIONAL
> > )
> > {
> > + EFI_BLOCK_IO_PROTOCOL *BlockIo;
> > + XEN_BLOCK_FRONT_DEVICE *Dev;
> > + EFI_BLOCK_IO_MEDIA *Media;
> > + EFI_STATUS Status;
> > +
> > + Status = gBS->OpenProtocol (
> > + ControllerHandle, &gEfiBlockIoProtocolGuid,
> > + (VOID **)&BlockIo,
> > + This->DriverBindingHandle, ControllerHandle,
> > + EFI_OPEN_PROTOCOL_GET_PROTOCOL
> > + );
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> > +
> > + Status = gBS->UninstallProtocolInterface (ControllerHandle,
> > + &gEfiBlockIoProtocolGuid, BlockIo);
> > + if (EFI_ERROR (Status)) {
> > + return Status;
> > + }
> > +
> > + Media = BlockIo->Media;
> > + Dev = XEN_BLOCK_FRONT_FROM_BLOCK_IO (BlockIo);
> > + XenPvBlockFrontShutdown (Dev);
> > +
> > + FreePool (Media);
> > +
> > gBS->CloseProtocol (ControllerHandle, &gXenbusProtocolGuid,
> > This->DriverBindingHandle, ControllerHandle);
> >
> > diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> > index 186d0ee..3120137 100644
> > --- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> > +++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.h
> > @@ -76,6 +76,7 @@ extern EFI_COMPONENT_NAME_PROTOCOL gXenPvBlkDxeComponentName;
> > //
> > #include "DriverBinding.h"
> > #include "ComponentName.h"
> > +#include "BlockIo.h"
> >
> >
> >
> > diff --git a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> > index a43d006..b379020 100644
> > --- a/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> > +++ b/OvmfPkg/XenPvBlkDxe/XenPvBlkDxe.inf
> > @@ -32,6 +32,8 @@
> > ComponentName.h
> > BlockFront.c
> > BlockFront.h
> > + BlockIo.c
> > + BlockIo.h
> >
> >
> > [LibraryClasses]
> > --
> > Anthony PERARD
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 00/18] Introducing Xen PV block driver to OVMF
2014-07-16 20:10 ` [PATCH RFC 00/18] Introducing Xen PV block driver to OVMF Konrad Rzeszutek Wilk
@ 2014-07-18 12:12 ` Anthony PERARD
0 siblings, 0 replies; 39+ messages in thread
From: Anthony PERARD @ 2014-07-18 12:12 UTC (permalink / raw)
To: Konrad Rzeszutek Wilk; +Cc: EDK2 devel, Xen Devel
On Wed, Jul 16, 2014 at 04:10:49PM -0400, Konrad Rzeszutek Wilk wrote:
> On Wed, Jul 16, 2014 at 04:15:29PM +0100, Anthony PERARD wrote:
> > Hi all,
> >
> > This patch series is implementing the necessary in order to access a PV block
> > device. For that, one need a XenStore client, a XenBus client, and the PV block
> > driver.
> >
> > There are two new drivers, XenbusDxe and XenPvBlkDxe. The first one implement a
> > bus drivers, and the second is a block drivers.
> >
> > There are still a bit of work to be done on this series, especially the comment
> > in the code, but I'd like your comment on this patch series.
>
> Went through it - I had some questions and spotted some issues that are
> pretty easy to fix.
>
> Otherwise I think you just need to flesh it out with more comments, links
> to the specs or just copy the relevant parts.
>
> And little puzzled by the usage of 8 pages instead of doing it via 11.
> But that is probably not a big deal since you are doing each I/O request
> synchronously anyhow.
>
> Thank you for posting this and developing it!
Thanks for your review.
--
Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
* Re: [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client.
[not found] ` <20140718115817.GG1582@perard.uk.xensource.com>
@ 2014-07-18 18:36 ` Konrad Rzeszutek Wilk
0 siblings, 0 replies; 39+ messages in thread
From: Konrad Rzeszutek Wilk @ 2014-07-18 18:36 UTC (permalink / raw)
To: Anthony PERARD; +Cc: EDK2 devel, Xen Devel, Samuel Thibault
> > > + OUT XenbusState *LastStatePtr OPTIONAL
> > > + )
> > > +{
> > > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > > + XenbusState State;
> > > + UINT64 Value;
> > > + XENSTORE_STATUS Status;
> > > +
> > > + while (TRUE) {
> > > + Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + return Status;
> > > + }
> > > + if (Value > XenbusStateReconfigured) {
> > > + return XENSTORE_STATUS_EIO;
> > > + }
> > > + State = Value;
> > > + if (State >= WaitedState) {
> >
> > Not State != WaitedState ?
>
> I don't think so, the function is actually waiting for the State to be
> equal to WaitedState. At least, that how I wrote it. There might be a
> better algorithm.
Then State == WaitedState
seems the proper way?
>
> > > + break;
> > > + }
> > > + DEBUG ((EFI_D_INFO,
> > > + "XenPvBlk: waiting backend state %d, current: %d\n",
> > > + WaitedState, State));
> > > + XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
> > > + }
> > > +
> > > + if (LastStatePtr != NULL) {
> > > + *LastStatePtr = State;
> > > + }
> > > +
> > > + return XENSTORE_STATUS_SUCCESS;
> > > +}
> > > +
> > > +EFI_STATUS
> > > +XenPvBlockFrontInitialization (
> > > + IN XENBUS_PROTOCOL *XenbusIo,
> > > + IN CONST CHAR8 *NodeName,
> > > + OUT XEN_BLOCK_FRONT_DEVICE **DevPtr
> > > + )
> > > +{
> > > + XENSTORE_TRANSACTION xbt;
> > > + CHAR8 *DeviceType;
> > > + blkif_sring_t *SharedRing;
> > > + XENSTORE_STATUS Status;
> > > + XEN_BLOCK_FRONT_DEVICE *Dev;
> > > + XenbusState State;
> > > + UINT64 Value;
> > > +
> > > + ASSERT (NodeName != NULL);
> > > +
> > > + Dev = AllocateZeroPool (sizeof (XEN_BLOCK_FRONT_DEVICE));
> > > + Dev->Signature = XEN_BLOCK_FRONT_SIGNATURE;
> > > + Dev->NodeName = NodeName;
> > > + Dev->XenbusIo = XenbusIo;
> > > + Dev->DeviceId = XenbusIo->DeviceId;
> > > +
> > > + XenbusIo->XsRead (XenbusIo, XST_NIL, "device-type", (VOID**)&DeviceType);
> > > + if (AsciiStrCmp (DeviceType, "cdrom") == 0) {
> > > + Dev->MediaInfo.CdRom = TRUE;
> > > + } else {
> > > + Dev->MediaInfo.CdRom = FALSE;
> > > + }
> > > + FreePool (DeviceType);
> > > +
> > > + Status = XenbusReadInteger (XenbusIo, "backend-id", FALSE, &Value);
> > > + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT16_MAX) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to get backend-id (%d)\n",
> > > + Status));
> > > + goto Error;
> > > + }
> > > + Dev->DomainId = Value;
> > > + XenbusIo->EventChannelAllocate (XenbusIo, Dev->DomainId, &Dev->EventChannel);
> > > +
> > > + SharedRing = (blkif_sring_t*) AllocatePages (1);
> > > + SHARED_RING_INIT (SharedRing);
> > > + FRONT_RING_INIT (&Dev->Ring, SharedRing, EFI_PAGE_SIZE);
> > > + XenbusIo->GrantAccess (XenbusIo,
> > > + Dev->DomainId,
> > > + (INTN) SharedRing >> EFI_PAGE_SHIFT,
> > > + FALSE,
> > > + &Dev->RingRef);
> > > +
> > > +Again:
> > > + Status = XenbusIo->XsTransactionStart (XenbusIo, &xbt);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_WARN, "XenPvBlk: Failed to start transaction, %d\n", Status));
> >
> > .. and do you want to error out?
>
> The original code (mini-os) is continuing in thes case, so I thought it
> was possible to do the initialisation without transaction. But maybe I'm
> wrong and this should be an error.
No explanation in the commit for that?
>
> > > + }
> > > +
> > > + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName, "ring-ref", "%d",
> > > + Dev->RingRef);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write ring-ref.\n"));
> > > + goto AbortTransaction;
> > > + }
> > > + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
> > > + "event-channel", "%d", Dev->EventChannel);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write event-channel.\n"));
> > > + goto AbortTransaction;
> > > + }
> > > + Status = XenbusIo->XsPrintf (XenbusIo, xbt, NodeName,
> > > + "protocol", "%a", XEN_IO_PROTO_ABI_NATIVE);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed to write protocol.\n"));
> > > + goto AbortTransaction;
> > > + }
> > > +
> > > + Status = XenbusIo->SetState (XenbusIo, xbt, XenbusStateConnected);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: Failed switch state.\n"));
> > > + goto AbortTransaction;
> > > + }
> > > +
> > > + Status = XenbusIo->XsTransactionEnd (XenbusIo, xbt, FALSE);
> > > + if (Status == XENSTORE_STATUS_EAGAIN) {
> > > + goto Again;
> > > + }
> > > +
> > > + XenbusIo->RegisterWatchBackend (XenbusIo, "state", &Dev->StateWatchToken);
> > > +
> > > + // Waiting backend
> > > + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateConnected, &State);
> > > + if (Status != XENSTORE_STATUS_SUCCESS || State != XenbusStateConnected) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: backend for %a/%d not available, rc=%d state=%d\n",
> > > + XenbusIo->Type, XenbusIo->DeviceId, Status, State));
> > > + goto Error2;
> > > + }
> > > +
> > > + Status = XenbusReadInteger (XenbusIo, "info", TRUE, &Value);
> > > + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
> > > + goto Error2;
> > > + }
> > > + Dev->MediaInfo.VDiskInfo = Value;
> > > + if (Dev->MediaInfo.VDiskInfo & VDISK_READONLY) {
> > > + Dev->MediaInfo.ReadWrite = FALSE;
> > > + } else {
> > > + Dev->MediaInfo.ReadWrite = TRUE;
> > > + }
> > > +
> > > + Status = XenbusReadInteger (XenbusIo, "sectors", TRUE, &Dev->MediaInfo.Sectors);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + goto Error2;
> > > + }
> > > +
> > > + Status = XenbusReadInteger (XenbusIo, "sector-size", TRUE, &Value);
> > > + if (Status != XENSTORE_STATUS_SUCCESS || Value > UINT32_MAX) {
> > > + goto Error2;
> > > + }
> > > + Dev->MediaInfo.SectorSize = Value;
> >
> > Why not &Dev->MediaInfo.SectorSize in XenbusReadInteger?
> >
> > Is it because of types? If so, please add a comment stating that.
>
> Yes, because of the type, uint64 vs uint32.
>
> Maybe I could rename the XenbusReadInteger to XenbusReadUint64, and it
> would be more obvious that the function require a specific type.
That would be much better. Thank you
>
> > > +
> > > + // Default value
> > > + Value = 0;
> > > + Status = XenbusReadInteger (XenbusIo, "feature-barrier", TRUE, &Value);
> > > + if (Value == 1) {
> > > + Dev->MediaInfo.FeatureBarrier = TRUE;
> > > + } else {
> > > + Dev->MediaInfo.FeatureBarrier = FALSE;
> > > + }
> > > +
> > > + // Default value
> > > + Value = 0;
> > > + XenbusReadInteger (XenbusIo, "feature-flush-cache", TRUE, &Value);
> > > + if (Value == 1) {
> > > + Dev->MediaInfo.FeatureFlushCache = TRUE;
> > > + } else {
> > > + Dev->MediaInfo.FeatureFlushCache = FALSE;
> > > + }
> > > +
> > > + DEBUG ((EFI_D_INFO, "XenPvBlk: New disk with %d sectors of %d bytes\n",
> > > + Dev->MediaInfo.Sectors, Dev->MediaInfo.SectorSize));
> > > +
> > > + *DevPtr = Dev;
> > > + return EFI_SUCCESS;
> > > +
> > > +Error2:
> > > + XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
> > > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
> > > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
> >
> > XenbusIo->XsRemove (XenbusIo, XST_NIL, "protocol");
> >
> > ?
>
> I don't know, but I can do that.
>
> > > + goto Error;
> > > +AbortTransaction:
> > > + XenbusIo->XsTransactionEnd (XenbusIo, xbt, TRUE);
> > > +Error:
> > > + XenPvBlockFree (Dev);
> > > + return EFI_DEVICE_ERROR;
> > > +}
> > > +
> > > +VOID
> > > +XenPvBlockFrontShutdown (
> > > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > > + )
> > > +{
> > > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > > + XENSTORE_STATUS Status;
> > > + UINT64 Value;
> > > +
> > > + XenPvBlockSync (Dev);
> > > +
> > > + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosing);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: error while changing state to Closing: %d\n",
> > > + Status));
> > > + goto Close;
> > > + }
> > > +
> > > + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosing, NULL);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: error while waiting for closing backend state: %d\n",
> > > + Status));
> > > + goto Close;
> > > + }
> > > +
> > > + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateClosed);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: error while changing state to Closed: %d\n",
> > > + Status));
> > > + goto Close;
> > > + }
> > > +
> > > + Status = XenPvBlkWaitForBackendState (Dev, XenbusStateClosed, NULL);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: error while waiting for closed backend state: %d\n",
> > > + Status));
> > > + goto Close;
> > > + }
> > > +
> > > + Status = XenbusIo->SetState (XenbusIo, XST_NIL, XenbusStateInitialising);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: error while changing state to initialising: %d\n",
> > > + Status));
> > > + goto Close;
> > > + }
> > > +
> > > + while (TRUE) {
> > > + Status = XenbusReadInteger (XenbusIo, "state", TRUE, &Value);
> > > + if (Status != XENSTORE_STATUS_SUCCESS) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: error while waiting for new backend state: %d\n",
> > > + Status));
> > > + goto Close;
> > > + }
> > > + if (Value < XenbusStateInitWait || Value >= XenbusStateClosed) {
> > > + break;
> > > + }
> > > + DEBUG ((EFI_D_INFO,
> > > + "XenPvBlk: waiting backend state %d, current: %d\n",
> > > + XenbusStateInitWait, Value));
> > > + XenbusIo->WaitForWatch (XenbusIo, Dev->StateWatchToken);
> > > + }
> > > +
> > > +Close:
> > > + XenbusIo->UnregisterWatch (XenbusIo, Dev->StateWatchToken);
> > > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "ring-ref");
> > > + XenbusIo->XsRemove (XenbusIo, XST_NIL, "event-channel");
> >
> > "protocol"
As in, also do the XenbusIo->XsRemove(...)
> > > +
> > > + XenPvBlockFree (Dev);
> > > +}
> > > +
> > > +STATIC
> > > +VOID
> > > +XenPvBlockWaitSlot (
> > > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > > + )
> > > +{
> > > + /* Wait for a slot */
> > > + if (RING_FULL (&Dev->Ring)) {
> > > + while (TRUE) {
> > > + XenPvBlockAsyncIoPoll (Dev);
> > > + if (!RING_FULL(&Dev->Ring)) {
> > > + break;
> > > + }
> > > + /* Really no slot, go to sleep. */
> > > + // XXX wait for an event on Dev->EventChannel
> > > + DEBUG ((EFI_D_INFO, "%a %d, waiting\n", __func__, __LINE__));
> > > + }
> > > + }
> > > +}
> > > +
> > > +/* Issue an aio */
> > > +VOID
> > > +XenPvBlockAsyncIo (
> > > + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> > > + IN BOOLEAN IsWrite
> > > + )
> > > +{
> > > + XEN_BLOCK_FRONT_DEVICE *Dev = IoData->Dev;
> > > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > > + blkif_request_t *Request;
> > > + RING_IDX RingIndex;
> > > + BOOLEAN Notify;
> > > + INT32 NumSegments, Index;
> > > + UINTN Start, End;
> > > +
> > > + // Can't io at non-sector-aligned location
> > > + ASSERT(!(IoData->Offset & (Dev->MediaInfo.SectorSize - 1)));
> > > + // Can't io non-sector-sized amounts
> > > + ASSERT(!(IoData->Size & (Dev->MediaInfo.SectorSize - 1)));
> > > + // Can't io non-sector-aligned buffer
> > > + ASSERT(!((UINTN) IoData->Buffer & (Dev->MediaInfo.SectorSize - 1)));
> > > +
> > > + Start = (UINTN) IoData->Buffer & ~EFI_PAGE_MASK;
> > > + End = ((UINTN) IoData->Buffer + IoData->Size + EFI_PAGE_SIZE - 1) & ~EFI_PAGE_MASK;
> > > + IoData->NumRef = NumSegments = (End - Start) / EFI_PAGE_SIZE;
> > > +
> > > + // XXX is this comment still true ?
> >
> > I think it is outdated, as Linux would do max DMA up to the segment
> > size that is advertised. Ditto for the SCSI stack.
>
> OK, thanks.
>
> > > + /* qemu's IDE max multsect is 16 (8KB) and SCSI max DMA was set to 32KB,
> > > + * so max 44KB can't happen */
> > > + ASSERT (NumSegments <= BLKIF_MAX_SEGMENTS_PER_REQUEST);
> >
> > But this check is OK as we can't truly fit more on the ring.
>
> OK.
>
> > > +
> > > + XenPvBlockWaitSlot (Dev);
> >
> > What happens if there are multiple threads calling this
> > function? Don't we need a mutex to only allow
> > one thread to grab an RingIndex?
>
> So, OVMF does not have threads but the are "events", called by a timer.
> This code does not support it, yet, so I don't need any form of locking,
> yet.
Ok, why don't you mention that in the commit description.
>
> > > + RingIndex = Dev->Ring.req_prod_pvt;
> > > + Request = RING_GET_REQUEST (&Dev->Ring, RingIndex);
> > > +
> > > + Request->operation = IsWrite ? BLKIF_OP_WRITE : BLKIF_OP_READ;
> > > + Request->nr_segments = NumSegments;
> > > + Request->handle = Dev->DeviceId;
> > > + Request->id = (UINTN) IoData;
> > > + Request->sector_number = IoData->Offset / 512;
> > > +
> > > + for (Index = 0; Index < NumSegments; Index++) {
> > > + Request->seg[Index].first_sect = 0;
> >
> > Uh?
> > > + Request->seg[Index].last_sect = EFI_PAGE_SIZE / 512 - 1;
> >
> > Uh? I thought you wanted to use the real sector values?
> > > + }
> > > + Request->seg[0].first_sect = ((UINTN) IoData->Buffer & EFI_PAGE_MASK) / 512;
> > > + Request->seg[NumSegments - 1].last_sect =
> > > + (((UINTN) IoData->Buffer + IoData->Size - 1) & EFI_PAGE_MASK) / 512;
> >
> > What about in between 0 and NumSegments? You should also
> > fill out those values with valid entries.
>
> It's done by the previous for loop. Maybe this part is hard to
> understand without the interface explain.
> include io/blkif.h says:
> @first_sect: first sector in frame to transfer (inclusive).
> @last_sect: last sector in frame to transfer (inclusive).
>
> but does not explain what a frame is, beyond this comment on:
> grant_ref_t gref; /* reference to I/O buffer frame */
> (all this from the struct blkif_request_segment)
> Or maybe frame == page.
That is what it means. If you have a page, with 4 sectors
of data (so half-page) and they are discontingious you
have to enumerate the sector values. For example:
First sector = lba + 0 -> lba + 1; #2 sector = lba + 5-> lba + 6;
#3 sector = lba + 55 -> lba + 56; #4 sector = lba + 32768 -> lba + 32769
>
> Anyway, the way the segments are initialized feel weirds now, but I
> don't think it's incorrect, otherwise, I would not be able to boot any
> guest.
Well, you could because you might be reading only one sector per
I/O request. And then the first_sect and last_sect get filled
with proper values.
>
> > > + for (Index = 0; Index < NumSegments; Index++) {
> > > + UINTN Data = Start + Index * EFI_PAGE_SIZE;
> > > + if (!IsWrite) {
> > > + /* Trigger CoW if needed */
> >
> > Not sure I understand that reference?
>
> Maybe a reference to the original commit will help to understand:
> minios: Fix bug when blkfront reading into zero-mapped buffer by just poking the page.
> No need to use virtual_to_mfn() for the ring since that is a real page.
> 7b1257946080b63421a8dc59c7b3dfd67997c643 (xen.git)
>
> But I still don't undertand this.
I think that means the MiniOS would hit a page fault and allocate
a page. I don't know if EFI drivers have this concept?
>
> > > + *(CHAR8 *)(Data + (Request->seg[Index].first_sect << 9)) = 0;
> > > + MemoryFence ();
If the EFI 'things' that use the BlockIO drivers allocate an page
for read, then there is no need for this.
> > > + }
> > > + XenbusIo->GrantAccess (XenbusIo, Dev->DomainId,
> > > + Data >> EFI_PAGE_SHIFT, IsWrite,
> > > + &Request->seg[Index].gref);
> > > + IoData->GrantRef[Index] = Request->seg[Index].gref;
> > > + }
> > > +
> > > + Dev->Ring.req_prod_pvt = RingIndex + 1;
> > > +
> > > + MemoryFence ();
> > > + RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
> > > +
> > > + if (Notify) {
> > > + XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
> > > + }
> > > +}
> > > +
> > > +STATIC
> > > +VOID
> > > +XenPvBlockAsyncIoCallback (
> > > + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> > > + IN EFI_STATUS Status
> > > + )
> > > +{
> > > + IoData->Data = (VOID *) 1;
> > > + IoData->Callback = NULL;
> > > +}
> > > +
> > > +VOID
> > > +XenPvBlockIo (
> > > + IN OUT XEN_BLOCK_FRONT_IO *IoData,
> > > + IN BOOLEAN IsWrite
> > > + )
> > > +{
> > > + ASSERT (IoData->Callback == NULL);
> > > + // XXX Signal/Event instead ?
> >
> > Maybe for later versions. I think this is OK for right now, thought
> > it will mean we can only do one I/O .. and have then to wait
> > until it is completed.
> >
> > That limititation should be mentioned in the commit description.
>
> That right, right now, it's only one IO at a time. I think my comment is
> more to tell me that Callback is not really one, and it could be replace
> by something more Uefi, an event.
>
> > > + IoData->Callback = XenPvBlockAsyncIoCallback;
> > > + XenPvBlockAsyncIo (IoData, IsWrite);
> > > + IoData->Data = NULL;
> > > +
> > > + while (TRUE) {
> > > + XenPvBlockAsyncIoPoll (IoData->Dev);
> > > + if (IoData->Data) {
> > > + break;
> > > + }
> > > + }
> > > +}
> > > +
> > > +STATIC
> > > +VOID
> > > +XenPvBlockPushOperation (
> > > + IN XEN_BLOCK_FRONT_DEVICE *Dev,
> > > + IN UINT8 Operation,
> > > + IN UINT64 Id
> > > + )
> > > +{
> > > + INT32 Index;
> > > + blkif_request_t *Request;
> > > + BOOLEAN Notify;
> > > +
> > > + XenPvBlockWaitSlot (Dev);
> >
> > How do we make sure that there aren't multiple threads
> > calling this?
>
> Not possible yet.
>
> > > + Index = Dev->Ring.req_prod_pvt;
> > > + Request = RING_GET_REQUEST(&Dev->Ring, Index);
> > > + Request->operation = Operation;
> > > + Request->nr_segments = 0;
> > > + Request->handle = Dev->DeviceId;
> > > + Request->id = Id;
> > > + /* Not needed anyway, but the backend will check it */
> > > + Request->sector_number = 0;
> > > + Dev->Ring.req_prod_pvt = Index + 1;
> > > + MemoryFence ();
> > > + RING_PUSH_REQUESTS_AND_CHECK_NOTIFY (&Dev->Ring, Notify);
> > > + if (Notify) {
> > > + XENBUS_PROTOCOL *XenbusIo = Dev->XenbusIo;
> > > + XenbusIo->EventChannelNotify (XenbusIo, Dev->EventChannel);
> > > + }
> > > +}
> > > +
> > > +VOID
> > > +XenPvBlockSync (
> > > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > > + )
> > > +{
> > > + if (Dev->MediaInfo.ReadWrite) {
> > > + if (Dev->MediaInfo.FeatureBarrier) {
> > > + XenPvBlockPushOperation (Dev, BLKIF_OP_WRITE_BARRIER, 0);
> > > + }
> > > +
> > > + if (Dev->MediaInfo.FeatureFlushCache) {
> > > + XenPvBlockPushOperation (Dev, BLKIF_OP_FLUSH_DISKCACHE, 0);
> > > + }
> > > + }
> > > +
> > > + /* Note: This won't finish if another thread enqueues requests. */
> > > + while (TRUE) {
> > > + XenPvBlockAsyncIoPoll (Dev);
> > > + if (RING_FREE_REQUESTS (&Dev->Ring) == RING_SIZE (&Dev->Ring)) {
> > > + break;
> > > + }
> > > + }
> > > +}
> > > +
> > > +VOID
> > > +XenPvBlockAsyncIoPoll (
> > > + IN XEN_BLOCK_FRONT_DEVICE *Dev
> > > + )
> > > +{
> > > + RING_IDX ProducerIndex, ConsumerIndex;
> > > + blkif_response_t *Response;
> > > + INT32 More;
> > > +
> > > + do {
> > > + ProducerIndex = Dev->Ring.sring->rsp_prod;
> > > + /* Ensure we see queued responses up to 'ProducerIndex'. */
> > > + MemoryFence ();
> > > + ConsumerIndex = Dev->Ring.rsp_cons;
> > > +
> > > + while (ConsumerIndex != ProducerIndex) {
> > > + XEN_BLOCK_FRONT_IO *IoData;
> > > + INT16 Status;
> > > +
> > > + Response = RING_GET_RESPONSE (&Dev->Ring, ConsumerIndex);
> > > +
> > > + IoData = (VOID *) (UINTN) Response->id;
> > > + Status = Response->status;
> > > +
> > > + switch (Response->operation) {
> > > + case BLKIF_OP_READ:
> > > + case BLKIF_OP_WRITE:
> > > + {
> > > + INT32 Index;
> > > +
> > > + if (Status != BLKIF_RSP_OKAY) {
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: "
> > > + "%a error %d on %a at offset %Ld, num bytes %Ld\n",
> > > + Response->operation == BLKIF_OP_READ ? "read" : "write",
> > > + Status, IoData->Dev->NodeName,
> > > + IoData->Offset,
> > > + IoData->Size));
> > > + }
> > > +
> > > + for (Index = 0; Index < IoData->NumRef; Index++) {
> > > + Dev->XenbusIo->GrantEndAccess (Dev->XenbusIo, IoData->GrantRef[Index]);
> > > + }
> > > +
> > > + break;
> > > + }
> > > +
> > > + case BLKIF_OP_WRITE_BARRIER:
> > > + if (Status != BLKIF_RSP_OKAY) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: write barrier error %d\n", Status));
> > > + }
> > > + break;
> > > + case BLKIF_OP_FLUSH_DISKCACHE:
> > > + if (Status != BLKIF_RSP_OKAY) {
> > > + DEBUG ((EFI_D_ERROR, "XenPvBlk: flush error %d\n", Status));
> > > + }
> > > + break;
> > > +
> > > + default:
> > > + DEBUG ((EFI_D_ERROR,
> > > + "XenPvBlk: unrecognized block operation %d response (status %d)\n",
> > > + Response->operation, Status));
> > > + break;
> > > + }
> > > +
> > > + Dev->Ring.rsp_cons = ++ConsumerIndex;
> > > + /* Nota: callback frees IoData itself */
> > > + if (IoData && IoData->Callback) {
> > > + IoData->Callback (IoData, Status ? EFI_DEVICE_ERROR : EFI_SUCCESS);
> > > + }
> > > + if (Dev->Ring.rsp_cons != ConsumerIndex) {
> > > + /* We reentered, we must not continue here */
> >
> > Uh, how does that happen?
>
> That an original comment, I don't think it's possible yet in ovmf.
>
> > > + break;
> > > + }
> > > + }
> > > +
> > > + RING_FINAL_CHECK_FOR_RESPONSES (&Dev->Ring, More);
> > > + } while (More != 0);
> > > +}
>
> --
> Anthony PERARD
^ permalink raw reply [flat|nested] 39+ messages in thread
end of thread, other threads:[~2014-07-18 18:36 UTC | newest]
Thread overview: 39+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <1405523747-5024-1-git-send-email-anthony.perard@citrix.com>
2014-07-16 15:15 ` [PATCH RFC 01/18] OvmfPkg: Add public headers from Xen Project Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 07/18] OvmfPkg/XenbusDxe: Add InterlockedCompareExchange16 Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 09/18] OvmfPkg/XenbusDxe: Add Event Channel Notify Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 10/18] OvmfPkg/XenbusDxe: Add TestAndClearBit Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 11/18] OvmfPkg/XenbusDxe: Add XenStore client implementation Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 12/18] OvmfPkg/XenbusDxe: Add an helper AsciiStrDup Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 13/18] OvmfPkg/XenbusDxe: Add Xenstore function into the Xenbus protocol Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 15/18] OvmfPkg/XenbusDxe: Add Event Channel into XenBus protocol Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 16/18] OvmfPkg/XenPvBlkDxe: Xen PV Block device, initial skeleton Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client Anthony PERARD
2014-07-16 15:15 ` [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo Anthony PERARD
[not found] ` <1405523747-5024-3-git-send-email-anthony.perard@citrix.com>
2014-07-16 17:25 ` [PATCH RFC 02/18] OvmfPkg: Add basic skeleton for the Xenbus driver Konrad Rzeszutek Wilk
2014-07-18 10:30 ` Anthony PERARD
[not found] ` <1405523747-5024-4-git-send-email-anthony.perard@citrix.com>
2014-07-16 17:32 ` [PATCH RFC 03/18] OvmfPkg/XenbusDxe: Add device state struct and create an ExitBoot services event Konrad Rzeszutek Wilk
2014-07-18 10:32 ` Anthony PERARD
[not found] ` <1405523747-5024-5-git-send-email-anthony.perard@citrix.com>
2014-07-16 17:37 ` [PATCH RFC 04/18] OvmfPkg/XenbusDxe: Add support to make Xen Hypercalls Konrad Rzeszutek Wilk
2014-07-17 9:43 ` Wei Liu
[not found] ` <1405523747-5024-6-git-send-email-anthony.perard@citrix.com>
2014-07-16 17:39 ` [PATCH RFC 05/18] OvmfPkg/XenbusDxe: Open PciIo protocol Konrad Rzeszutek Wilk
2014-07-18 10:36 ` Anthony PERARD
[not found] ` <1405523747-5024-7-git-send-email-anthony.perard@citrix.com>
2014-07-16 17:42 ` [PATCH RFC 06/18] OvmfPkg: Introduce Xenbus Protocol Konrad Rzeszutek Wilk
2014-07-18 10:40 ` Anthony PERARD
[not found] ` <1405523747-5024-9-git-send-email-anthony.perard@citrix.com>
2014-07-16 17:48 ` [PATCH RFC 08/18] OvmfPkg/XenbusDxe: Add Grant Table functions Konrad Rzeszutek Wilk
2014-07-18 10:53 ` Anthony PERARD
[not found] ` <1405523747-5024-15-git-send-email-anthony.perard@citrix.com>
2014-07-16 19:04 ` [PATCH RFC 14/18] OvmfPkg/XenbusDxe: Indroduce XenBus support itself Konrad Rzeszutek Wilk
2014-07-18 11:04 ` Anthony PERARD
[not found] ` <1405523747-5024-18-git-send-email-anthony.perard@citrix.com>
2014-07-16 19:41 ` [PATCH RFC 17/18] OvmfPkg/XenPvBlkDxe: Add BlockFront client Konrad Rzeszutek Wilk
2014-07-18 11:58 ` Anthony PERARD
[not found] ` <20140718115817.GG1582@perard.uk.xensource.com>
2014-07-18 18:36 ` Konrad Rzeszutek Wilk
[not found] ` <1405523747-5024-19-git-send-email-anthony.perard@citrix.com>
2014-07-16 19:57 ` [PATCH RFC 18/18] OvmfPkg/XenPvBlkDxe: Add BlockIo Konrad Rzeszutek Wilk
2014-07-18 12:10 ` Anthony PERARD
2014-07-16 20:10 ` [PATCH RFC 00/18] Introducing Xen PV block driver to OVMF Konrad Rzeszutek Wilk
2014-07-18 12:12 ` Anthony PERARD
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.