[PATCH v2 00/12] Add riscv kvm accel support

[PATCH v2 00/12] Add riscv kvm accel support

[Yifei Jiang]

This series adds both riscv32 and riscv64 kvm support, and implements
migration based on riscv.

Because of RISC-V KVM has been merged into the Linux master, so this
series are changed from RFC to patch.

Several steps to use this:
1. Build emulation
$ ./configure --target-list=riscv64-softmmu
$ make -j$(nproc)

2. Build kernel

3. Build QEMU VM
Cross built in riscv toolchain.
$ PKG_CONFIG_LIBDIR=<toolchain pkgconfig path>
$ export PKG_CONFIG_SYSROOT_DIR=<toolchain sysroot path>
$ ./configure --target-list=riscv64-softmmu --enable-kvm \
--cross-prefix=riscv64-linux-gnu- --disable-libiscsi --disable-glusterfs \
--disable-libusb --disable-usb-redir --audio-drv-list= --disable-opengl \
--disable-libxml2
$ make -j$(nproc)

4. Start emulation
$ ./qemu-system-riscv64 -M virt -m 4096M -cpu rv64,x-h=true -nographic \
        -name guest=riscv-hyp,debug-threads=on \
        -smp 4 \
        -bios ./fw_jump.bin \
        -kernel ./Image \
        -drive file=./hyp.img,format=raw,id=hd0 \
        -device virtio-blk-device,drive=hd0 \
        -append "root=/dev/vda rw console=ttyS0 earlycon=sbi"

5. Start kvm-acceled QEMU VM in emulation
$ ./qemu-system-riscv64 -M virt,accel=kvm -m 1024M -cpu host -nographic \
        -name guest=riscv-guset \
        -smp 2 \
        -bios none \
        -kernel ./Image \
        -drive file=./guest.img,format=raw,id=hd0 \
        -device virtio-blk-device,drive=hd0 \
        -append "root=/dev/vda rw console=ttyS0 earlycon=sbi"

Changes since patch v1
- Rebase on recent commit a216e7cf119c91ffdf5931834a1a030ebea40d70
- Sync-up headers with Linux-5.16-rc4.
- Fixbug in kvm_arch_init_vcpu.
- Create a macro for get and put regs csr.
- Start kernel directly when kvm_enabled.
- Use riscv_cpu_set_irq to inject KVM interrupts.
- Use the Semihosting Console API for RISC-V kvm handle sbi.
- Update vmstate_riscv_cpu version id.
  Placing kvm_timer into a subsection.

Changes since RFC v6
- Rebase on recent commit 8627edfb3f1fca24a96a0954148885c3241c10f8
- Sync-up headers with Linux-5.16-rc1

Changes since RFC v5
- Rebase on QEMU v6.1.0-rc1 and kvm-riscv linux v19.
- Move kvm interrupt setting to riscv_cpu_update_mip().
- Replace __u64 with uint64_t.

Changes since RFC v4
- Rebase on QEMU v6.0.0-rc2 and kvm-riscv linux v17.
- Remove time scaling support as software solution is incomplete.
  Because it will cause unacceptable performance degradation. and
  We will post a better solution.
- Revise according to Alistair's review comments.
  - Remove compile time XLEN checks in kvm_riscv_reg_id
  - Surround TYPE_RISCV_CPU_HOST definition by CONFIG_KVM and share
    it between RV32 and RV64.
  - Add kvm-stub.c for reduce unnecessary compilation checks.
  - Add riscv_setup_direct_kernel() to direct boot kernel for KVM.

Changes since RFC v3
- Rebase on QEMU v5.2.0-rc2 and kvm-riscv linux v15.
- Add time scaling support(New patches 13, 14 and 15).
- Fix the bug that guest vm can't reboot.

Changes since RFC v2
- Fix checkpatch error at target/riscv/sbi_ecall_interface.h.
- Add riscv migration support.

Changes since RFC v1
- Add separate SBI ecall interface header.
- Add riscv32 kvm accel support.

Yifei Jiang (12):
  update-linux-headers: Add asm-riscv/kvm.h
  target/riscv: Add target/riscv/kvm.c to place the public kvm interface
  target/riscv: Implement function kvm_arch_init_vcpu
  target/riscv: Implement kvm_arch_get_registers
  target/riscv: Implement kvm_arch_put_registers
  target/riscv: Support start kernel directly by KVM
  target/riscv: Support setting external interrupt by KVM
  target/riscv: Handle KVM_EXIT_RISCV_SBI exit
  target/riscv: Add host cpu type
  target/riscv: Add kvm_riscv_get/put_regs_timer
  target/riscv: Implement virtual time adjusting with vm state changing
  target/riscv: Support virtual time context synchronization

 hw/riscv/boot.c                    |  11 +
 hw/riscv/virt.c                    |  54 ++-
 include/hw/riscv/boot.h            |   1 +
 linux-headers/asm-riscv/kvm.h      | 128 +++++++
 linux-headers/linux/kvm.h          |   8 +
 meson.build                        |   2 +
 target/riscv/cpu.c                 |  29 +-
 target/riscv/cpu.h                 |  11 +
 target/riscv/kvm-stub.c            |  30 ++
 target/riscv/kvm.c                 | 528 +++++++++++++++++++++++++++++
 target/riscv/kvm_riscv.h           |  25 ++
 target/riscv/machine.c             |  37 +-
 target/riscv/meson.build           |   1 +
 target/riscv/sbi_ecall_interface.h |  72 ++++
 14 files changed, 918 insertions(+), 19 deletions(-)
 create mode 100644 linux-headers/asm-riscv/kvm.h
 create mode 100644 target/riscv/kvm-stub.c
 create mode 100644 target/riscv/kvm.c
 create mode 100644 target/riscv/kvm_riscv.h
 create mode 100644 target/riscv/sbi_ecall_interface.h

-- 
2.19.1

[PATCH v2 01/12] update-linux-headers: Add asm-riscv/kvm.h

[Yifei Jiang]

Add asm-riscv/kvm.h for RISC-V KVM, and update linux/kvm.h

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Anup Patel <anup.patel@wdc.com>
---
 linux-headers/asm-riscv/kvm.h | 128 ++++++++++++++++++++++++++++++++++
 linux-headers/linux/kvm.h     |   8 +++
 2 files changed, 136 insertions(+)
 create mode 100644 linux-headers/asm-riscv/kvm.h

diff --git a/linux-headers/asm-riscv/kvm.h b/linux-headers/asm-riscv/kvm.h
new file mode 100644
index 0000000000..f808ad1ce5
--- /dev/null
+++ b/linux-headers/asm-riscv/kvm.h
@@ -0,0 +1,128 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+/*
+ * Copyright (C) 2019 Western Digital Corporation or its affiliates.
+ *
+ * Authors:
+ *     Anup Patel <anup.patel@wdc.com>
+ */
+
+#ifndef __LINUX_KVM_RISCV_H
+#define __LINUX_KVM_RISCV_H
+
+#ifndef __ASSEMBLY__
+
+#include <linux/types.h>
+#include <asm/ptrace.h>
+
+#define __KVM_HAVE_READONLY_MEM
+
+#define KVM_COALESCED_MMIO_PAGE_OFFSET 1
+
+#define KVM_INTERRUPT_SET	-1U
+#define KVM_INTERRUPT_UNSET	-2U
+
+/* for KVM_GET_REGS and KVM_SET_REGS */
+struct kvm_regs {
+};
+
+/* for KVM_GET_FPU and KVM_SET_FPU */
+struct kvm_fpu {
+};
+
+/* KVM Debug exit structure */
+struct kvm_debug_exit_arch {
+};
+
+/* for KVM_SET_GUEST_DEBUG */
+struct kvm_guest_debug_arch {
+};
+
+/* definition of registers in kvm_run */
+struct kvm_sync_regs {
+};
+
+/* for KVM_GET_SREGS and KVM_SET_SREGS */
+struct kvm_sregs {
+};
+
+/* CONFIG registers for KVM_GET_ONE_REG and KVM_SET_ONE_REG */
+struct kvm_riscv_config {
+	unsigned long isa;
+};
+
+/* CORE registers for KVM_GET_ONE_REG and KVM_SET_ONE_REG */
+struct kvm_riscv_core {
+	struct user_regs_struct regs;
+	unsigned long mode;
+};
+
+/* Possible privilege modes for kvm_riscv_core */
+#define KVM_RISCV_MODE_S	1
+#define KVM_RISCV_MODE_U	0
+
+/* CSR registers for KVM_GET_ONE_REG and KVM_SET_ONE_REG */
+struct kvm_riscv_csr {
+	unsigned long sstatus;
+	unsigned long sie;
+	unsigned long stvec;
+	unsigned long sscratch;
+	unsigned long sepc;
+	unsigned long scause;
+	unsigned long stval;
+	unsigned long sip;
+	unsigned long satp;
+	unsigned long scounteren;
+};
+
+/* TIMER registers for KVM_GET_ONE_REG and KVM_SET_ONE_REG */
+struct kvm_riscv_timer {
+	__u64 frequency;
+	__u64 time;
+	__u64 compare;
+	__u64 state;
+};
+
+/* Possible states for kvm_riscv_timer */
+#define KVM_RISCV_TIMER_STATE_OFF	0
+#define KVM_RISCV_TIMER_STATE_ON	1
+
+#define KVM_REG_SIZE(id)		\
+	(1U << (((id) & KVM_REG_SIZE_MASK) >> KVM_REG_SIZE_SHIFT))
+
+/* If you need to interpret the index values, here is the key: */
+#define KVM_REG_RISCV_TYPE_MASK		0x00000000FF000000
+#define KVM_REG_RISCV_TYPE_SHIFT	24
+
+/* Config registers are mapped as type 1 */
+#define KVM_REG_RISCV_CONFIG		(0x01 << KVM_REG_RISCV_TYPE_SHIFT)
+#define KVM_REG_RISCV_CONFIG_REG(name)	\
+	(offsetof(struct kvm_riscv_config, name) / sizeof(unsigned long))
+
+/* Core registers are mapped as type 2 */
+#define KVM_REG_RISCV_CORE		(0x02 << KVM_REG_RISCV_TYPE_SHIFT)
+#define KVM_REG_RISCV_CORE_REG(name)	\
+		(offsetof(struct kvm_riscv_core, name) / sizeof(unsigned long))
+
+/* Control and status registers are mapped as type 3 */
+#define KVM_REG_RISCV_CSR		(0x03 << KVM_REG_RISCV_TYPE_SHIFT)
+#define KVM_REG_RISCV_CSR_REG(name)	\
+		(offsetof(struct kvm_riscv_csr, name) / sizeof(unsigned long))
+
+/* Timer registers are mapped as type 4 */
+#define KVM_REG_RISCV_TIMER		(0x04 << KVM_REG_RISCV_TYPE_SHIFT)
+#define KVM_REG_RISCV_TIMER_REG(name)	\
+		(offsetof(struct kvm_riscv_timer, name) / sizeof(__u64))
+
+/* F extension registers are mapped as type 5 */
+#define KVM_REG_RISCV_FP_F		(0x05 << KVM_REG_RISCV_TYPE_SHIFT)
+#define KVM_REG_RISCV_FP_F_REG(name)	\
+		(offsetof(struct __riscv_f_ext_state, name) / sizeof(__u32))
+
+/* D extension registers are mapped as type 6 */
+#define KVM_REG_RISCV_FP_D		(0x06 << KVM_REG_RISCV_TYPE_SHIFT)
+#define KVM_REG_RISCV_FP_D_REG(name)	\
+		(offsetof(struct __riscv_d_ext_state, name) / sizeof(__u64))
+
+#endif
+
+#endif /* __LINUX_KVM_RISCV_H */
diff --git a/linux-headers/linux/kvm.h b/linux-headers/linux/kvm.h
index bcaf66cc4d..5e290c3c3e 100644
--- a/linux-headers/linux/kvm.h
+++ b/linux-headers/linux/kvm.h
@@ -269,6 +269,7 @@ struct kvm_xen_exit {
 #define KVM_EXIT_AP_RESET_HOLD    32
 #define KVM_EXIT_X86_BUS_LOCK     33
 #define KVM_EXIT_XEN              34
+#define KVM_EXIT_RISCV_SBI        35
 
 /* For KVM_EXIT_INTERNAL_ERROR */
 /* Emulate instruction failed. */
@@ -469,6 +470,13 @@ struct kvm_run {
 		} msr;
 		/* KVM_EXIT_XEN */
 		struct kvm_xen_exit xen;
+		/* KVM_EXIT_RISCV_SBI */
+		struct {
+			unsigned long extension_id;
+			unsigned long function_id;
+			unsigned long args[6];
+			unsigned long ret[2];
+		} riscv_sbi;
 		/* Fix the size of the union. */
 		char padding[256];
 	};
-- 
2.19.1

[PATCH v2 02/12] target/riscv: Add target/riscv/kvm.c to place the public kvm interface

[Yifei Jiang]

Add target/riscv/kvm.c to place kvm_arch_* function needed by
kvm/kvm-all.c. Meanwhile, add kvm support in meson.build file.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Anup Patel <anup.patel@wdc.com>
---
 meson.build              |   2 +
 target/riscv/kvm.c       | 133 +++++++++++++++++++++++++++++++++++++++
 target/riscv/meson.build |   1 +
 3 files changed, 136 insertions(+)
 create mode 100644 target/riscv/kvm.c

diff --git a/meson.build b/meson.build
index 96de1a6ef9..ae35e76ea4 100644
--- a/meson.build
+++ b/meson.build
@@ -77,6 +77,8 @@ elif cpu in ['ppc', 'ppc64']
   kvm_targets = ['ppc-softmmu', 'ppc64-softmmu']
 elif cpu in ['mips', 'mips64']
   kvm_targets = ['mips-softmmu', 'mipsel-softmmu', 'mips64-softmmu', 'mips64el-softmmu']
+elif cpu in ['riscv']
+  kvm_targets = ['riscv32-softmmu', 'riscv64-softmmu']
 else
   kvm_targets = []
 endif
diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
new file mode 100644
index 0000000000..687dd4b621
--- /dev/null
+++ b/target/riscv/kvm.c
@@ -0,0 +1,133 @@
+/*
+ * RISC-V implementation of KVM hooks
+ *
+ * Copyright (c) 2020 Huawei Technologies Co., Ltd
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "qemu/osdep.h"
+#include <sys/ioctl.h>
+
+#include <linux/kvm.h>
+
+#include "qemu-common.h"
+#include "qemu/timer.h"
+#include "qemu/error-report.h"
+#include "qemu/main-loop.h"
+#include "sysemu/sysemu.h"
+#include "sysemu/kvm.h"
+#include "sysemu/kvm_int.h"
+#include "cpu.h"
+#include "trace.h"
+#include "hw/pci/pci.h"
+#include "exec/memattrs.h"
+#include "exec/address-spaces.h"
+#include "hw/boards.h"
+#include "hw/irq.h"
+#include "qemu/log.h"
+#include "hw/loader.h"
+
+const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
+    KVM_CAP_LAST_INFO
+};
+
+int kvm_arch_get_registers(CPUState *cs)
+{
+    return 0;
+}
+
+int kvm_arch_put_registers(CPUState *cs, int level)
+{
+    return 0;
+}
+
+int kvm_arch_release_virq_post(int virq)
+{
+    return 0;
+}
+
+int kvm_arch_fixup_msi_route(struct kvm_irq_routing_entry *route,
+                             uint64_t address, uint32_t data, PCIDevice *dev)
+{
+    return 0;
+}
+
+int kvm_arch_destroy_vcpu(CPUState *cs)
+{
+    return 0;
+}
+
+unsigned long kvm_arch_vcpu_id(CPUState *cpu)
+{
+    return cpu->cpu_index;
+}
+
+void kvm_arch_init_irq_routing(KVMState *s)
+{
+}
+
+int kvm_arch_init_vcpu(CPUState *cs)
+{
+    return 0;
+}
+
+int kvm_arch_msi_data_to_gsi(uint32_t data)
+{
+    abort();
+}
+
+int kvm_arch_add_msi_route_post(struct kvm_irq_routing_entry *route,
+                                int vector, PCIDevice *dev)
+{
+    return 0;
+}
+
+int kvm_arch_init(MachineState *ms, KVMState *s)
+{
+    return 0;
+}
+
+int kvm_arch_irqchip_create(KVMState *s)
+{
+    return 0;
+}
+
+int kvm_arch_process_async_events(CPUState *cs)
+{
+    return 0;
+}
+
+void kvm_arch_pre_run(CPUState *cs, struct kvm_run *run)
+{
+}
+
+MemTxAttrs kvm_arch_post_run(CPUState *cs, struct kvm_run *run)
+{
+    return MEMTXATTRS_UNSPECIFIED;
+}
+
+bool kvm_arch_stop_on_emulation_error(CPUState *cs)
+{
+    return true;
+}
+
+int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
+{
+    return 0;
+}
+
+bool kvm_arch_cpu_check_are_resettable(void)
+{
+    return true;
+}
diff --git a/target/riscv/meson.build b/target/riscv/meson.build
index d5e0bc93ea..2faf08a941 100644
--- a/target/riscv/meson.build
+++ b/target/riscv/meson.build
@@ -19,6 +19,7 @@ riscv_ss.add(files(
   'bitmanip_helper.c',
   'translate.c',
 ))
+riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'))
 
 riscv_softmmu_ss = ss.source_set()
 riscv_softmmu_ss.add(files(
-- 
2.19.1

[PATCH v2 03/12] target/riscv: Implement function kvm_arch_init_vcpu

[Yifei Jiang]

Get isa info from kvm while kvm init.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
---
 target/riscv/kvm.c | 32 +++++++++++++++++++++++++++++++-
 1 file changed, 31 insertions(+), 1 deletion(-)

diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index 687dd4b621..ccf3753048 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -38,6 +38,23 @@
 #include "qemu/log.h"
 #include "hw/loader.h"
 
+static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
+{
+    uint64_t id = KVM_REG_RISCV | type | idx;
+
+    switch (riscv_cpu_mxl(env)) {
+    case MXL_RV32:
+        id |= KVM_REG_SIZE_U32;
+        break;
+    case MXL_RV64:
+        id |= KVM_REG_SIZE_U64;
+        break;
+    default:
+        g_assert_not_reached();
+    }
+    return id;
+}
+
 const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
     KVM_CAP_LAST_INFO
 };
@@ -79,7 +96,20 @@ void kvm_arch_init_irq_routing(KVMState *s)
 
 int kvm_arch_init_vcpu(CPUState *cs)
 {
-    return 0;
+    int ret = 0;
+    target_ulong isa;
+    RISCVCPU *cpu = RISCV_CPU(cs);
+    CPURISCVState *env = &cpu->env;
+    uint64_t id;
+
+    id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CONFIG, KVM_REG_RISCV_CONFIG_REG(isa));
+    ret = kvm_get_one_reg(cs, id, &isa);
+    if (ret) {
+        return ret;
+    }
+    env->misa_ext = isa;
+
+    return ret;
 }
 
 int kvm_arch_msi_data_to_gsi(uint32_t data)
-- 
2.19.1

[PATCH v2 04/12] target/riscv: Implement kvm_arch_get_registers

[Yifei Jiang]

Get GPR CSR and FP registers from kvm by KVM_GET_ONE_REG ioctl.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
---
 target/riscv/kvm.c | 112 ++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 111 insertions(+), 1 deletion(-)

diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index ccf3753048..6d4df0ef6d 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -55,13 +55,123 @@ static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx
     return id;
 }
 
+#define RISCV_CORE_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_CORE, \
+                 KVM_REG_RISCV_CORE_REG(name))
+
+#define RISCV_CSR_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_CSR, \
+                 KVM_REG_RISCV_CSR_REG(name))
+
+#define RISCV_FP_F_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_F, idx)
+
+#define RISCV_FP_D_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_D, idx)
+
+#define KVM_RISCV_GET_CSR(cs, env, csr, reg) \
+    do { \
+        int ret = kvm_get_one_reg(cs, RISCV_CSR_REG(env, csr), &reg); \
+        if (ret) { \
+            return ret; \
+        } \
+    } while(0)
+
+static int kvm_riscv_get_regs_core(CPUState *cs)
+{
+    int ret = 0;
+    int i;
+    target_ulong reg;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    ret = kvm_get_one_reg(cs, RISCV_CORE_REG(env, regs.pc), &reg);
+    if (ret) {
+        return ret;
+    }
+    env->pc = reg;
+
+    for (i = 1; i < 32; i++) {
+        uint64_t id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CORE, i);
+        ret = kvm_get_one_reg(cs, id, &reg);
+        if (ret) {
+            return ret;
+        }
+        env->gpr[i] = reg;
+    }
+
+    return ret;
+}
+
+static int kvm_riscv_get_regs_csr(CPUState *cs)
+{
+    int ret = 0;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    KVM_RISCV_GET_CSR(cs, env, sstatus, env->mstatus);
+    KVM_RISCV_GET_CSR(cs, env, sie, env->mie);
+    KVM_RISCV_GET_CSR(cs, env, stvec, env->stvec);
+    KVM_RISCV_GET_CSR(cs, env, sscratch, env->sscratch);
+    KVM_RISCV_GET_CSR(cs, env, sepc, env->sepc);
+    KVM_RISCV_GET_CSR(cs, env, scause, env->scause);
+    KVM_RISCV_GET_CSR(cs, env, stval, env->stval);
+    KVM_RISCV_GET_CSR(cs, env, sip, env->mip);
+    KVM_RISCV_GET_CSR(cs, env, satp, env->satp);
+    return ret;
+}
+
+static int kvm_riscv_get_regs_fp(CPUState *cs)
+{
+    int ret = 0;
+    int i;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    if (riscv_has_ext(env, RVD)) {
+        uint64_t reg;
+        for (i = 0; i < 32; i++) {
+            ret = kvm_get_one_reg(cs, RISCV_FP_D_REG(env, i), &reg);
+            if (ret) {
+                return ret;
+            }
+            env->fpr[i] = reg;
+        }
+        return ret;
+    }
+
+    if (riscv_has_ext(env, RVF)) {
+        uint32_t reg;
+        for (i = 0; i < 32; i++) {
+            ret = kvm_get_one_reg(cs, RISCV_FP_F_REG(env, i), &reg);
+            if (ret) {
+                return ret;
+            }
+            env->fpr[i] = reg;
+        }
+        return ret;
+    }
+
+    return ret;
+}
+
 const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
     KVM_CAP_LAST_INFO
 };
 
 int kvm_arch_get_registers(CPUState *cs)
 {
-    return 0;
+    int ret = 0;
+
+    ret = kvm_riscv_get_regs_core(cs);
+    if (ret) {
+        return ret;
+    }
+
+    ret = kvm_riscv_get_regs_csr(cs);
+    if (ret) {
+        return ret;
+    }
+
+    ret = kvm_riscv_get_regs_fp(cs);
+    if (ret) {
+        return ret;
+    }
+
+    return ret;
 }
 
 int kvm_arch_put_registers(CPUState *cs, int level)
-- 
2.19.1

[PATCH v2 05/12] target/riscv: Implement kvm_arch_put_registers

[Yifei Jiang]

Put GPR CSR and FP registers to kvm by KVM_SET_ONE_REG ioctl

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
---
 target/riscv/kvm.c | 104 ++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 103 insertions(+), 1 deletion(-)

diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index 6d4df0ef6d..e695b91dc7 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -73,6 +73,14 @@ static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx
         } \
     } while(0)
 
+#define KVM_RISCV_SET_CSR(cs, env, csr, reg) \
+    do { \
+        int ret = kvm_set_one_reg(cs, RISCV_CSR_REG(env, csr), &reg); \
+        if (ret) { \
+            return ret; \
+        } \
+    } while(0)
+
 static int kvm_riscv_get_regs_core(CPUState *cs)
 {
     int ret = 0;
@@ -98,6 +106,31 @@ static int kvm_riscv_get_regs_core(CPUState *cs)
     return ret;
 }
 
+static int kvm_riscv_put_regs_core(CPUState *cs)
+{
+    int ret = 0;
+    int i;
+    target_ulong reg;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    reg = env->pc;
+    ret = kvm_set_one_reg(cs, RISCV_CORE_REG(env, regs.pc), &reg);
+    if (ret) {
+        return ret;
+    }
+
+    for (i = 1; i < 32; i++) {
+        uint64_t id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CORE, i);
+        reg = env->gpr[i];
+        ret = kvm_set_one_reg(cs, id, &reg);
+        if (ret) {
+            return ret;
+        }
+    }
+
+    return ret;
+}
+
 static int kvm_riscv_get_regs_csr(CPUState *cs)
 {
     int ret = 0;
@@ -115,6 +148,24 @@ static int kvm_riscv_get_regs_csr(CPUState *cs)
     return ret;
 }
 
+static int kvm_riscv_put_regs_csr(CPUState *cs)
+{
+    int ret = 0;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    KVM_RISCV_SET_CSR(cs, env, sstatus, env->mstatus);
+    KVM_RISCV_SET_CSR(cs, env, sie, env->mie);
+    KVM_RISCV_SET_CSR(cs, env, stvec, env->stvec);
+    KVM_RISCV_SET_CSR(cs, env, sscratch, env->sscratch);
+    KVM_RISCV_SET_CSR(cs, env, sepc, env->sepc);
+    KVM_RISCV_SET_CSR(cs, env, scause, env->scause);
+    KVM_RISCV_SET_CSR(cs, env, stval, env->stval);
+    KVM_RISCV_SET_CSR(cs, env, sip, env->mip);
+    KVM_RISCV_SET_CSR(cs, env, satp, env->satp);
+
+    return ret;
+}
+
 static int kvm_riscv_get_regs_fp(CPUState *cs)
 {
     int ret = 0;
@@ -148,6 +199,40 @@ static int kvm_riscv_get_regs_fp(CPUState *cs)
     return ret;
 }
 
+static int kvm_riscv_put_regs_fp(CPUState *cs)
+{
+    int ret = 0;
+    int i;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    if (riscv_has_ext(env, RVD)) {
+        uint64_t reg;
+        for (i = 0; i < 32; i++) {
+            reg = env->fpr[i];
+            ret = kvm_set_one_reg(cs, RISCV_FP_D_REG(env, i), &reg);
+            if (ret) {
+                return ret;
+            }
+        }
+        return ret;
+    }
+
+    if (riscv_has_ext(env, RVF)) {
+        uint32_t reg;
+        for (i = 0; i < 32; i++) {
+            reg = env->fpr[i];
+            ret = kvm_set_one_reg(cs, RISCV_FP_F_REG(env, i), &reg);
+            if (ret) {
+                return ret;
+            }
+        }
+        return ret;
+    }
+
+    return ret;
+}
+
+
 const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
     KVM_CAP_LAST_INFO
 };
@@ -176,7 +261,24 @@ int kvm_arch_get_registers(CPUState *cs)
 
 int kvm_arch_put_registers(CPUState *cs, int level)
 {
-    return 0;
+    int ret = 0;
+
+    ret = kvm_riscv_put_regs_core(cs);
+    if (ret) {
+        return ret;
+    }
+
+    ret = kvm_riscv_put_regs_csr(cs);
+    if (ret) {
+        return ret;
+    }
+
+    ret = kvm_riscv_put_regs_fp(cs);
+    if (ret) {
+        return ret;
+    }
+
+    return ret;
 }
 
 int kvm_arch_release_virq_post(int virq)
-- 
2.19.1

[PATCH v2 06/12] target/riscv: Support start kernel directly by KVM

[Yifei Jiang]

Get kernel and fdt start address in virt.c, and pass them to KVM
when cpu reset. In addition, add kvm_riscv.h to place riscv specific
interface.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
---
 hw/riscv/boot.c          | 11 ++++++++
 hw/riscv/virt.c          | 54 ++++++++++++++++++++++++++++------------
 include/hw/riscv/boot.h  |  1 +
 target/riscv/cpu.c       |  8 ++++++
 target/riscv/cpu.h       |  3 +++
 target/riscv/kvm-stub.c  | 25 +++++++++++++++++++
 target/riscv/kvm.c       | 14 +++++++++++
 target/riscv/kvm_riscv.h | 24 ++++++++++++++++++
 target/riscv/meson.build |  2 +-
 9 files changed, 125 insertions(+), 17 deletions(-)
 create mode 100644 target/riscv/kvm-stub.c
 create mode 100644 target/riscv/kvm_riscv.h

diff --git a/hw/riscv/boot.c b/hw/riscv/boot.c
index 519fa455a1..00df6d7810 100644
--- a/hw/riscv/boot.c
+++ b/hw/riscv/boot.c
@@ -317,3 +317,14 @@ void riscv_setup_rom_reset_vec(MachineState *machine, RISCVHartArrayState *harts
 
     return;
 }
+
+void riscv_setup_direct_kernel(hwaddr kernel_addr, hwaddr fdt_addr)
+{
+    CPUState *cs;
+
+    for (cs = first_cpu; cs; cs = CPU_NEXT(cs)) {
+        RISCVCPU *riscv_cpu = RISCV_CPU(cs);
+        riscv_cpu->env.kernel_addr = kernel_addr;
+        riscv_cpu->env.fdt_addr = fdt_addr;
+    }
+}
diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c
index 3af074148e..c8bcd9d9e5 100644
--- a/hw/riscv/virt.c
+++ b/hw/riscv/virt.c
@@ -38,6 +38,7 @@
 #include "chardev/char.h"
 #include "sysemu/device_tree.h"
 #include "sysemu/sysemu.h"
+#include "sysemu/kvm.h"
 #include "hw/pci/pci.h"
 #include "hw/pci-host/gpex.h"
 #include "hw/display/ramfb.h"
@@ -801,23 +802,25 @@ static void virt_machine_init(MachineState *machine)
                                 hart_count, &error_abort);
         sysbus_realize(SYS_BUS_DEVICE(&s->soc[i]), &error_abort);
 
-        /* Per-socket CLINT */
-        riscv_aclint_swi_create(
-            memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size,
-            base_hartid, hart_count, false);
-        riscv_aclint_mtimer_create(
-            memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size +
-                RISCV_ACLINT_SWI_SIZE,
-            RISCV_ACLINT_DEFAULT_MTIMER_SIZE, base_hartid, hart_count,
-            RISCV_ACLINT_DEFAULT_MTIMECMP, RISCV_ACLINT_DEFAULT_MTIME,
-            RISCV_ACLINT_DEFAULT_TIMEBASE_FREQ, true);
-
-        /* Per-socket ACLINT SSWI */
-        if (s->have_aclint) {
+        if (!kvm_enabled()) {
+            /* Per-socket CLINT */
             riscv_aclint_swi_create(
-                memmap[VIRT_ACLINT_SSWI].base +
-                    i * memmap[VIRT_ACLINT_SSWI].size,
-                base_hartid, hart_count, true);
+                memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size,
+                base_hartid, hart_count, false);
+            riscv_aclint_mtimer_create(
+                memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size +
+                    RISCV_ACLINT_SWI_SIZE,
+                RISCV_ACLINT_DEFAULT_MTIMER_SIZE, base_hartid, hart_count,
+                RISCV_ACLINT_DEFAULT_MTIMECMP, RISCV_ACLINT_DEFAULT_MTIME,
+                RISCV_ACLINT_DEFAULT_TIMEBASE_FREQ, true);
+
+            /* Per-socket ACLINT SSWI */
+            if (s->have_aclint) {
+                riscv_aclint_swi_create(
+                    memmap[VIRT_ACLINT_SSWI].base +
+                        i * memmap[VIRT_ACLINT_SSWI].size,
+                    base_hartid, hart_count, true);
+            }
         }
 
         /* Per-socket PLIC hart topology configuration string */
@@ -884,6 +887,16 @@ static void virt_machine_init(MachineState *machine)
     memory_region_add_subregion(system_memory, memmap[VIRT_MROM].base,
                                 mask_rom);
 
+    /*
+     * Only direct boot kernel is currently supported for KVM VM,
+     * so the "-bios" parameter is ignored and treated like "-bios none"
+     * when KVM is enabled.
+     */
+    if (kvm_enabled()) {
+        g_free(machine->firmware);
+        machine->firmware = g_strdup("none");
+    }
+
     if (riscv_is_32bit(&s->soc[0])) {
         firmware_end_addr = riscv_find_and_load_firmware(machine,
                                     RISCV32_BIOS_BIN, start_addr, NULL);
@@ -941,6 +954,15 @@ static void virt_machine_init(MachineState *machine)
                               virt_memmap[VIRT_MROM].size, kernel_entry,
                               fdt_load_addr, machine->fdt);
 
+    /*
+     * Only direct boot kernel is currently supported for KVM VM,
+     * So here setup kernel start address and fdt address.
+     * TODO:Support firmware loading and integrate to TCG start
+     */
+    if (kvm_enabled()) {
+        riscv_setup_direct_kernel(kernel_entry, fdt_load_addr);
+    }
+
     /* SiFive Test MMIO device */
     sifive_test_create(memmap[VIRT_TEST].base);
 
diff --git a/include/hw/riscv/boot.h b/include/hw/riscv/boot.h
index baff11dd8a..5834c234aa 100644
--- a/include/hw/riscv/boot.h
+++ b/include/hw/riscv/boot.h
@@ -58,5 +58,6 @@ void riscv_rom_copy_firmware_info(MachineState *machine, hwaddr rom_base,
                                   hwaddr rom_size,
                                   uint32_t reset_vec_size,
                                   uint64_t kernel_entry);
+void riscv_setup_direct_kernel(hwaddr kernel_addr, hwaddr fdt_addr);
 
 #endif /* RISCV_BOOT_H */
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index f812998123..1c944872a3 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -29,6 +29,8 @@
 #include "hw/qdev-properties.h"
 #include "migration/vmstate.h"
 #include "fpu/softfloat-helpers.h"
+#include "sysemu/kvm.h"
+#include "kvm_riscv.h"
 
 /* RISC-V CPU definitions */
 
@@ -380,6 +382,12 @@ static void riscv_cpu_reset(DeviceState *dev)
     cs->exception_index = RISCV_EXCP_NONE;
     env->load_res = -1;
     set_default_nan_mode(1, &env->fp_status);
+
+#ifndef CONFIG_USER_ONLY
+    if (kvm_enabled()) {
+        kvm_riscv_reset_vcpu(cpu);
+    }
+#endif
 }
 
 static void riscv_cpu_disas_set_info(CPUState *s, disassemble_info *info)
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index 0760c0af93..2807eb1bcb 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -255,6 +255,9 @@ struct CPURISCVState {
 
     /* Fields from here on are preserved across CPU reset. */
     QEMUTimer *timer; /* Internal timer */
+
+    hwaddr kernel_addr;
+    hwaddr fdt_addr;
 };
 
 OBJECT_DECLARE_TYPE(RISCVCPU, RISCVCPUClass,
diff --git a/target/riscv/kvm-stub.c b/target/riscv/kvm-stub.c
new file mode 100644
index 0000000000..39b96fe3f4
--- /dev/null
+++ b/target/riscv/kvm-stub.c
@@ -0,0 +1,25 @@
+/*
+ * QEMU KVM RISC-V specific function stubs
+ *
+ * Copyright (c) 2020 Huawei Technologies Co., Ltd
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include "qemu/osdep.h"
+#include "cpu.h"
+#include "kvm_riscv.h"
+
+void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
+{
+    abort();
+}
diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index e695b91dc7..db6d8a5b6e 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -37,6 +37,7 @@
 #include "hw/irq.h"
 #include "qemu/log.h"
 #include "hw/loader.h"
+#include "kvm_riscv.h"
 
 static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
 {
@@ -369,6 +370,19 @@ int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
     return 0;
 }
 
+void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
+{
+    CPURISCVState *env = &cpu->env;
+
+    if (!kvm_enabled()) {
+        return;
+    }
+    env->pc = cpu->env.kernel_addr;
+    env->gpr[10] = kvm_arch_vcpu_id(CPU(cpu)); /* a0 */
+    env->gpr[11] = cpu->env.fdt_addr;          /* a1 */
+    env->satp = 0;
+}
+
 bool kvm_arch_cpu_check_are_resettable(void)
 {
     return true;
diff --git a/target/riscv/kvm_riscv.h b/target/riscv/kvm_riscv.h
new file mode 100644
index 0000000000..f38c82bf59
--- /dev/null
+++ b/target/riscv/kvm_riscv.h
@@ -0,0 +1,24 @@
+/*
+ * QEMU KVM support -- RISC-V specific functions.
+ *
+ * Copyright (c) 2020 Huawei Technologies Co., Ltd
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2 or later, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef QEMU_KVM_RISCV_H
+#define QEMU_KVM_RISCV_H
+
+void kvm_riscv_reset_vcpu(RISCVCPU *cpu);
+
+#endif
diff --git a/target/riscv/meson.build b/target/riscv/meson.build
index 2faf08a941..fe41cc5805 100644
--- a/target/riscv/meson.build
+++ b/target/riscv/meson.build
@@ -19,7 +19,7 @@ riscv_ss.add(files(
   'bitmanip_helper.c',
   'translate.c',
 ))
-riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'))
+riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'), if_false: files('kvm-stub.c'))
 
 riscv_softmmu_ss = ss.source_set()
 riscv_softmmu_ss.add(files(
-- 
2.19.1

[PATCH v2 07/12] target/riscv: Support setting external interrupt by KVM

[Yifei Jiang]

When KVM is enabled, set the S-mode external interrupt through
kvm_riscv_set_irq function.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
---
 target/riscv/cpu.c       |  6 +++++-
 target/riscv/kvm-stub.c  |  5 +++++
 target/riscv/kvm.c       | 17 +++++++++++++++++
 target/riscv/kvm_riscv.h |  1 +
 4 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 1c944872a3..71a7ac6831 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -603,7 +603,11 @@ static void riscv_cpu_set_irq(void *opaque, int irq, int level)
     case IRQ_S_EXT:
     case IRQ_VS_EXT:
     case IRQ_M_EXT:
-        riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
+        if (kvm_enabled() && (irq & IRQ_M_EXT) ) {
+            kvm_riscv_set_irq(cpu, IRQ_S_EXT, level);
+        } else {
+            riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
+        }
         break;
     default:
         g_assert_not_reached();
diff --git a/target/riscv/kvm-stub.c b/target/riscv/kvm-stub.c
index 39b96fe3f4..4e8fc31a21 100644
--- a/target/riscv/kvm-stub.c
+++ b/target/riscv/kvm-stub.c
@@ -23,3 +23,8 @@ void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
 {
     abort();
 }
+
+void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level)
+{
+    abort();
+}
diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index db6d8a5b6e..0027f11f45 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -383,6 +383,23 @@ void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
     env->satp = 0;
 }
 
+void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level)
+{
+    int ret;
+    unsigned virq = level ? KVM_INTERRUPT_SET : KVM_INTERRUPT_UNSET;
+
+    if (irq != IRQ_S_EXT) {
+        perror("kvm riscv set irq != IRQ_S_EXT\n");
+        abort();
+    }
+
+    ret = kvm_vcpu_ioctl(CPU(cpu), KVM_INTERRUPT, &virq);
+    if (ret < 0) {
+        perror("Set irq failed");
+        abort();
+    }
+}
+
 bool kvm_arch_cpu_check_are_resettable(void)
 {
     return true;
diff --git a/target/riscv/kvm_riscv.h b/target/riscv/kvm_riscv.h
index f38c82bf59..ed281bdce0 100644
--- a/target/riscv/kvm_riscv.h
+++ b/target/riscv/kvm_riscv.h
@@ -20,5 +20,6 @@
 #define QEMU_KVM_RISCV_H
 
 void kvm_riscv_reset_vcpu(RISCVCPU *cpu);
+void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level);
 
 #endif
-- 
2.19.1

[PATCH v2 08/12] target/riscv: Handle KVM_EXIT_RISCV_SBI exit

[Yifei Jiang]

Use char-fe to handle console sbi call, which implement early
console io while apply 'earlycon=sbi' into kernel parameters.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
---
 target/riscv/kvm.c                 | 38 +++++++++++++++-
 target/riscv/sbi_ecall_interface.h | 72 ++++++++++++++++++++++++++++++
 2 files changed, 109 insertions(+), 1 deletion(-)
 create mode 100644 target/riscv/sbi_ecall_interface.h

diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index 0027f11f45..171a32adf9 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -38,6 +38,8 @@
 #include "qemu/log.h"
 #include "hw/loader.h"
 #include "kvm_riscv.h"
+#include "sbi_ecall_interface.h"
+#include "semihosting/console.h"
 
 static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
 {
@@ -365,9 +367,43 @@ bool kvm_arch_stop_on_emulation_error(CPUState *cs)
     return true;
 }
 
+static int kvm_riscv_handle_sbi(CPUState *cs, struct kvm_run *run)
+{
+    int ret = 0;
+    unsigned char ch;
+    switch (run->riscv_sbi.extension_id) {
+    case SBI_EXT_0_1_CONSOLE_PUTCHAR:
+        ch = run->riscv_sbi.args[0];
+        qemu_semihosting_log_out((const char *)&ch, sizeof(ch));
+        break;
+    case SBI_EXT_0_1_CONSOLE_GETCHAR:
+        run->riscv_sbi.args[0] =
+            (unsigned long)qemu_semihosting_console_inc(cs->env_ptr);
+        break;
+    default:
+        qemu_log_mask(LOG_UNIMP,
+                      "%s: un-handled SBI EXIT, specific reasons is %lu\n",
+                      __func__, run->riscv_sbi.extension_id);
+        ret = -1;
+        break;
+    }
+    return ret;
+}
+
 int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
 {
-    return 0;
+    int ret = 0;
+    switch (run->exit_reason) {
+    case KVM_EXIT_RISCV_SBI:
+        ret = kvm_riscv_handle_sbi(cs, run);
+        break;
+    default:
+        qemu_log_mask(LOG_UNIMP, "%s: un-handled exit reason %d\n",
+                      __func__, run->exit_reason);
+        ret = -1;
+        break;
+    }
+    return ret;
 }
 
 void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
diff --git a/target/riscv/sbi_ecall_interface.h b/target/riscv/sbi_ecall_interface.h
new file mode 100644
index 0000000000..fb1a3fa8f2
--- /dev/null
+++ b/target/riscv/sbi_ecall_interface.h
@@ -0,0 +1,72 @@
+/*
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2019 Western Digital Corporation or its affiliates.
+ *
+ * Authors:
+ *   Anup Patel <anup.patel@wdc.com>
+ */
+
+#ifndef __SBI_ECALL_INTERFACE_H__
+#define __SBI_ECALL_INTERFACE_H__
+
+/* clang-format off */
+
+/* SBI Extension IDs */
+#define SBI_EXT_0_1_SET_TIMER           0x0
+#define SBI_EXT_0_1_CONSOLE_PUTCHAR     0x1
+#define SBI_EXT_0_1_CONSOLE_GETCHAR     0x2
+#define SBI_EXT_0_1_CLEAR_IPI           0x3
+#define SBI_EXT_0_1_SEND_IPI            0x4
+#define SBI_EXT_0_1_REMOTE_FENCE_I      0x5
+#define SBI_EXT_0_1_REMOTE_SFENCE_VMA   0x6
+#define SBI_EXT_0_1_REMOTE_SFENCE_VMA_ASID 0x7
+#define SBI_EXT_0_1_SHUTDOWN            0x8
+#define SBI_EXT_BASE                    0x10
+#define SBI_EXT_TIME                    0x54494D45
+#define SBI_EXT_IPI                     0x735049
+#define SBI_EXT_RFENCE                  0x52464E43
+#define SBI_EXT_HSM                     0x48534D
+
+/* SBI function IDs for BASE extension*/
+#define SBI_EXT_BASE_GET_SPEC_VERSION   0x0
+#define SBI_EXT_BASE_GET_IMP_ID         0x1
+#define SBI_EXT_BASE_GET_IMP_VERSION    0x2
+#define SBI_EXT_BASE_PROBE_EXT          0x3
+#define SBI_EXT_BASE_GET_MVENDORID      0x4
+#define SBI_EXT_BASE_GET_MARCHID        0x5
+#define SBI_EXT_BASE_GET_MIMPID         0x6
+
+/* SBI function IDs for TIME extension*/
+#define SBI_EXT_TIME_SET_TIMER          0x0
+
+/* SBI function IDs for IPI extension*/
+#define SBI_EXT_IPI_SEND_IPI            0x0
+
+/* SBI function IDs for RFENCE extension*/
+#define SBI_EXT_RFENCE_REMOTE_FENCE_I       0x0
+#define SBI_EXT_RFENCE_REMOTE_SFENCE_VMA    0x1
+#define SBI_EXT_RFENCE_REMOTE_SFENCE_VMA_ASID  0x2
+#define SBI_EXT_RFENCE_REMOTE_HFENCE_GVMA   0x3
+#define SBI_EXT_RFENCE_REMOTE_HFENCE_GVMA_VMID 0x4
+#define SBI_EXT_RFENCE_REMOTE_HFENCE_VVMA   0x5
+#define SBI_EXT_RFENCE_REMOTE_HFENCE_VVMA_ASID 0x6
+
+/* SBI function IDs for HSM extension */
+#define SBI_EXT_HSM_HART_START          0x0
+#define SBI_EXT_HSM_HART_STOP           0x1
+#define SBI_EXT_HSM_HART_GET_STATUS     0x2
+
+#define SBI_HSM_HART_STATUS_STARTED     0x0
+#define SBI_HSM_HART_STATUS_STOPPED     0x1
+#define SBI_HSM_HART_STATUS_START_PENDING   0x2
+#define SBI_HSM_HART_STATUS_STOP_PENDING    0x3
+
+#define SBI_SPEC_VERSION_MAJOR_OFFSET   24
+#define SBI_SPEC_VERSION_MAJOR_MASK     0x7f
+#define SBI_SPEC_VERSION_MINOR_MASK     0xffffff
+#define SBI_EXT_VENDOR_START            0x09000000
+#define SBI_EXT_VENDOR_END              0x09FFFFFF
+/* clang-format on */
+
+#endif
-- 
2.19.1

[PATCH v2 09/12] target/riscv: Add host cpu type

[Yifei Jiang]

'host' type cpu is set isa to RV32 or RV64 simply, more isa info
will obtain from KVM in kvm_arch_init_vcpu()

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Anup Patel <anup.patel@wdc.com>
---
 target/riscv/cpu.c | 15 +++++++++++++++
 target/riscv/cpu.h |  1 +
 2 files changed, 16 insertions(+)

diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 71a7ac6831..a137fd3d2b 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -213,6 +213,18 @@ static void rv32_imafcu_nommu_cpu_init(Object *obj)
 }
 #endif
 
+#if defined(CONFIG_KVM)
+static void riscv_host_cpu_init(Object *obj)
+{
+    CPURISCVState *env = &RISCV_CPU(obj)->env;
+#if defined(TARGET_RISCV32)
+    set_misa(env, MXL_RV32, 0);
+#elif defined(TARGET_RISCV64)
+    set_misa(env, MXL_RV64, 0);
+#endif
+}
+#endif
+
 static ObjectClass *riscv_cpu_class_by_name(const char *cpu_model)
 {
     ObjectClass *oc;
@@ -814,6 +826,9 @@ static const TypeInfo riscv_cpu_type_infos[] = {
         .class_init = riscv_cpu_class_init,
     },
     DEFINE_CPU(TYPE_RISCV_CPU_ANY,              riscv_any_cpu_init),
+#if defined(CONFIG_KVM)
+    DEFINE_CPU(TYPE_RISCV_CPU_HOST,             riscv_host_cpu_init),
+#endif
 #if defined(TARGET_RISCV32)
     DEFINE_CPU(TYPE_RISCV_CPU_BASE32,           rv32_base_cpu_init),
     DEFINE_CPU(TYPE_RISCV_CPU_IBEX,             rv32_ibex_cpu_init),
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index 2807eb1bcb..e7dba35acb 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -45,6 +45,7 @@
 #define TYPE_RISCV_CPU_SIFIVE_E51       RISCV_CPU_TYPE_NAME("sifive-e51")
 #define TYPE_RISCV_CPU_SIFIVE_U34       RISCV_CPU_TYPE_NAME("sifive-u34")
 #define TYPE_RISCV_CPU_SIFIVE_U54       RISCV_CPU_TYPE_NAME("sifive-u54")
+#define TYPE_RISCV_CPU_HOST             RISCV_CPU_TYPE_NAME("host")
 
 #if defined(TARGET_RISCV32)
 # define TYPE_RISCV_CPU_BASE            TYPE_RISCV_CPU_BASE32
-- 
2.19.1

[PATCH v2 10/12] target/riscv: Add kvm_riscv_get/put_regs_timer

[Yifei Jiang]

Add kvm_riscv_get/put_regs_timer to synchronize virtual time context
from KVM.

To set register of RISCV_TIMER_REG(state) will occur a error from KVM
on kvm_timer_state == 0. It's better to adapt in KVM, but it doesn't matter
that adaping in QEMU.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
---
 target/riscv/cpu.h |  7 +++++
 target/riscv/kvm.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 79 insertions(+)

diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index e7dba35acb..c892a2c8b7 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -259,6 +259,13 @@ struct CPURISCVState {
 
     hwaddr kernel_addr;
     hwaddr fdt_addr;
+
+    /* kvm timer */
+    bool kvm_timer_dirty;
+    uint64_t kvm_timer_time;
+    uint64_t kvm_timer_compare;
+    uint64_t kvm_timer_state;
+    uint64_t kvm_timer_frequency;
 };
 
 OBJECT_DECLARE_TYPE(RISCVCPU, RISCVCPUClass,
diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index 171a32adf9..802c076b22 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -64,6 +64,9 @@ static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx
 #define RISCV_CSR_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_CSR, \
                  KVM_REG_RISCV_CSR_REG(name))
 
+#define RISCV_TIMER_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_TIMER, \
+                 KVM_REG_RISCV_TIMER_REG(name))
+
 #define RISCV_FP_F_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_F, idx)
 
 #define RISCV_FP_D_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_D, idx)
@@ -235,6 +238,75 @@ static int kvm_riscv_put_regs_fp(CPUState *cs)
     return ret;
 }
 
+static void kvm_riscv_get_regs_timer(CPUState *cs)
+{
+    int ret;
+    uint64_t reg;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    if (env->kvm_timer_dirty) {
+        return;
+    }
+
+    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, time), &reg);
+    if (ret) {
+        abort();
+    }
+    env->kvm_timer_time = reg;
+
+    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, compare), &reg);
+    if (ret) {
+        abort();
+    }
+    env->kvm_timer_compare = reg;
+
+    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, state), &reg);
+    if (ret) {
+        abort();
+    }
+    env->kvm_timer_state = reg;
+
+    env->kvm_timer_dirty = true;
+}
+
+static void kvm_riscv_put_regs_timer(CPUState *cs)
+{
+    int ret;
+    uint64_t reg;
+    CPURISCVState *env = &RISCV_CPU(cs)->env;
+
+    if (!env->kvm_timer_dirty) {
+        return;
+    }
+
+    reg = env->kvm_timer_time;
+    ret = kvm_set_one_reg(cs, RISCV_TIMER_REG(env, time), &reg);
+    if (ret) {
+        abort();
+    }
+
+    reg = env->kvm_timer_compare;
+    ret = kvm_set_one_reg(cs, RISCV_TIMER_REG(env, compare), &reg);
+    if (ret) {
+        abort();
+    }
+
+    /*
+     * To set register of RISCV_TIMER_REG(state) will occur a error from KVM
+     * on env->kvm_timer_state == 0, It's better to adapt in KVM, but it
+     * doesn't matter that adaping in QEMU now.
+     * TODO If KVM changes, adapt here.
+     */
+    if (env->kvm_timer_state) {
+        reg = env->kvm_timer_state;
+        ret = kvm_set_one_reg(cs, RISCV_TIMER_REG(env, state), &reg);
+        if (ret) {
+            abort();
+        }
+    }
+
+    env->kvm_timer_dirty = false;
+}
 
 const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
     KVM_CAP_LAST_INFO
-- 
2.19.1

[PATCH v2 11/12] target/riscv: Implement virtual time adjusting with vm state changing

[Yifei Jiang]

We hope that virtual time adjusts with vm state changing. When a vm
is stopped, guest virtual time should stop counting and kvm_timer
should be stopped. When the vm is resumed, guest virtual time should
continue to count and kvm_timer should be restored.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
---
 target/riscv/kvm.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
index 802c076b22..be95dbc3f3 100644
--- a/target/riscv/kvm.c
+++ b/target/riscv/kvm.c
@@ -40,6 +40,7 @@
 #include "kvm_riscv.h"
 #include "sbi_ecall_interface.h"
 #include "semihosting/console.h"
+#include "sysemu/runstate.h"
 
 static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
 {
@@ -377,6 +378,17 @@ unsigned long kvm_arch_vcpu_id(CPUState *cpu)
     return cpu->cpu_index;
 }
 
+static void kvm_riscv_vm_state_change(void *opaque, bool running, RunState state)
+{
+    CPUState *cs = opaque;
+
+    if (running) {
+        kvm_riscv_put_regs_timer(cs);
+    } else {
+        kvm_riscv_get_regs_timer(cs);
+    }
+}
+
 void kvm_arch_init_irq_routing(KVMState *s)
 {
 }
@@ -389,6 +401,8 @@ int kvm_arch_init_vcpu(CPUState *cs)
     CPURISCVState *env = &cpu->env;
     uint64_t id;
 
+    qemu_add_vm_change_state_handler(kvm_riscv_vm_state_change, cs);
+
     id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CONFIG, KVM_REG_RISCV_CONFIG_REG(isa));
     ret = kvm_get_one_reg(cs, id, &isa);
     if (ret) {
-- 
2.19.1

[PATCH v2 12/12] target/riscv: Support virtual time context synchronization

[Yifei Jiang]

Add virtual time context description to vmstate_kvmtimer. After cpu being
loaded, virtual time context is updated to KVM.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Mingwang Li <limingwang@huawei.com>
---
 target/riscv/machine.c | 37 +++++++++++++++++++++++++++++++++++--
 1 file changed, 35 insertions(+), 2 deletions(-)

diff --git a/target/riscv/machine.c b/target/riscv/machine.c
index ad8248ebfd..f46443c316 100644
--- a/target/riscv/machine.c
+++ b/target/riscv/machine.c
@@ -164,10 +164,42 @@ static const VMStateDescription vmstate_pointermasking = {
     }
 };
 
+static bool kvmtimer_needed(void *opaque)
+{
+    return kvm_enabled();
+}
+
+
+static const VMStateDescription vmstate_kvmtimer = {
+    .name = "cpu/kvmtimer",
+    .version_id = 1,
+    .minimum_version_id = 1,
+    .needed = kvmtimer_needed,
+    .fields = (VMStateField[]) {
+        VMSTATE_UINT64(env.kvm_timer_time, RISCVCPU),
+        VMSTATE_UINT64(env.kvm_timer_compare, RISCVCPU),
+        VMSTATE_UINT64(env.kvm_timer_state, RISCVCPU),
+
+        VMSTATE_END_OF_LIST()
+    }
+};
+
+static int cpu_post_load(void *opaque, int version_id)
+{
+    RISCVCPU *cpu = opaque;
+    CPURISCVState *env = &cpu->env;
+
+    if (kvm_enabled()) {
+        env->kvm_timer_dirty = true;
+    }
+    return 0;
+}
+
 const VMStateDescription vmstate_riscv_cpu = {
     .name = "cpu",
-    .version_id = 3,
-    .minimum_version_id = 3,
+    .version_id = 4,
+    .minimum_version_id = 4,
+    .post_load = cpu_post_load,
     .fields = (VMStateField[]) {
         VMSTATE_UINTTL_ARRAY(env.gpr, RISCVCPU, 32),
         VMSTATE_UINT64_ARRAY(env.fpr, RISCVCPU, 32),
@@ -218,6 +250,7 @@ const VMStateDescription vmstate_riscv_cpu = {
         &vmstate_hyper,
         &vmstate_vector,
         &vmstate_pointermasking,
+        &vmstate_kvmtimer,
         NULL
     }
 };
-- 
2.19.1

Re: [PATCH v2 03/12] target/riscv: Implement function kvm_arch_init_vcpu

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Get isa info from kvm while kvm init.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Looks good to me.

Reviewed-by: Anup Patel <anup.patel@wdc.com>

Regards,
Anup

> ---
>  target/riscv/kvm.c | 32 +++++++++++++++++++++++++++++++-
>  1 file changed, 31 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index 687dd4b621..ccf3753048 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -38,6 +38,23 @@
>  #include "qemu/log.h"
>  #include "hw/loader.h"
>
> +static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
> +{
> +    uint64_t id = KVM_REG_RISCV | type | idx;
> +
> +    switch (riscv_cpu_mxl(env)) {
> +    case MXL_RV32:
> +        id |= KVM_REG_SIZE_U32;
> +        break;
> +    case MXL_RV64:
> +        id |= KVM_REG_SIZE_U64;
> +        break;
> +    default:
> +        g_assert_not_reached();
> +    }
> +    return id;
> +}
> +
>  const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
>      KVM_CAP_LAST_INFO
>  };
> @@ -79,7 +96,20 @@ void kvm_arch_init_irq_routing(KVMState *s)
>
>  int kvm_arch_init_vcpu(CPUState *cs)
>  {
> -    return 0;
> +    int ret = 0;
> +    target_ulong isa;
> +    RISCVCPU *cpu = RISCV_CPU(cs);
> +    CPURISCVState *env = &cpu->env;
> +    uint64_t id;
> +
> +    id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CONFIG, KVM_REG_RISCV_CONFIG_REG(isa));
> +    ret = kvm_get_one_reg(cs, id, &isa);
> +    if (ret) {
> +        return ret;
> +    }
> +    env->misa_ext = isa;
> +
> +    return ret;
>  }
>
>  int kvm_arch_msi_data_to_gsi(uint32_t data)
> --
> 2.19.1
>

Re: [PATCH v2 04/12] target/riscv: Implement kvm_arch_get_registers

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Get GPR CSR and FP registers from kvm by KVM_GET_ONE_REG ioctl.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Looks good to me.

Reviewed-by: Anup Patel <anup.patel@wdc.com>

Regards,
Anup

> ---
>  target/riscv/kvm.c | 112 ++++++++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 111 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index ccf3753048..6d4df0ef6d 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -55,13 +55,123 @@ static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx
>      return id;
>  }
>
> +#define RISCV_CORE_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_CORE, \
> +                 KVM_REG_RISCV_CORE_REG(name))
> +
> +#define RISCV_CSR_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_CSR, \
> +                 KVM_REG_RISCV_CSR_REG(name))
> +
> +#define RISCV_FP_F_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_F, idx)
> +
> +#define RISCV_FP_D_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_D, idx)
> +
> +#define KVM_RISCV_GET_CSR(cs, env, csr, reg) \
> +    do { \
> +        int ret = kvm_get_one_reg(cs, RISCV_CSR_REG(env, csr), &reg); \
> +        if (ret) { \
> +            return ret; \
> +        } \
> +    } while(0)
> +
> +static int kvm_riscv_get_regs_core(CPUState *cs)
> +{
> +    int ret = 0;
> +    int i;
> +    target_ulong reg;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    ret = kvm_get_one_reg(cs, RISCV_CORE_REG(env, regs.pc), &reg);
> +    if (ret) {
> +        return ret;
> +    }
> +    env->pc = reg;
> +
> +    for (i = 1; i < 32; i++) {
> +        uint64_t id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CORE, i);
> +        ret = kvm_get_one_reg(cs, id, &reg);
> +        if (ret) {
> +            return ret;
> +        }
> +        env->gpr[i] = reg;
> +    }
> +
> +    return ret;
> +}
> +
> +static int kvm_riscv_get_regs_csr(CPUState *cs)
> +{
> +    int ret = 0;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    KVM_RISCV_GET_CSR(cs, env, sstatus, env->mstatus);
> +    KVM_RISCV_GET_CSR(cs, env, sie, env->mie);
> +    KVM_RISCV_GET_CSR(cs, env, stvec, env->stvec);
> +    KVM_RISCV_GET_CSR(cs, env, sscratch, env->sscratch);
> +    KVM_RISCV_GET_CSR(cs, env, sepc, env->sepc);
> +    KVM_RISCV_GET_CSR(cs, env, scause, env->scause);
> +    KVM_RISCV_GET_CSR(cs, env, stval, env->stval);
> +    KVM_RISCV_GET_CSR(cs, env, sip, env->mip);
> +    KVM_RISCV_GET_CSR(cs, env, satp, env->satp);
> +    return ret;
> +}
> +
> +static int kvm_riscv_get_regs_fp(CPUState *cs)
> +{
> +    int ret = 0;
> +    int i;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    if (riscv_has_ext(env, RVD)) {
> +        uint64_t reg;
> +        for (i = 0; i < 32; i++) {
> +            ret = kvm_get_one_reg(cs, RISCV_FP_D_REG(env, i), &reg);
> +            if (ret) {
> +                return ret;
> +            }
> +            env->fpr[i] = reg;
> +        }
> +        return ret;
> +    }
> +
> +    if (riscv_has_ext(env, RVF)) {
> +        uint32_t reg;
> +        for (i = 0; i < 32; i++) {
> +            ret = kvm_get_one_reg(cs, RISCV_FP_F_REG(env, i), &reg);
> +            if (ret) {
> +                return ret;
> +            }
> +            env->fpr[i] = reg;
> +        }
> +        return ret;
> +    }
> +
> +    return ret;
> +}
> +
>  const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
>      KVM_CAP_LAST_INFO
>  };
>
>  int kvm_arch_get_registers(CPUState *cs)
>  {
> -    return 0;
> +    int ret = 0;
> +
> +    ret = kvm_riscv_get_regs_core(cs);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    ret = kvm_riscv_get_regs_csr(cs);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    ret = kvm_riscv_get_regs_fp(cs);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    return ret;
>  }
>
>  int kvm_arch_put_registers(CPUState *cs, int level)
> --
> 2.19.1
>

Re: [PATCH v2 05/12] target/riscv: Implement kvm_arch_put_registers

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Put GPR CSR and FP registers to kvm by KVM_SET_ONE_REG ioctl
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Looks good to me.

Reviewed-by: Anup Patel <anup.patel@wdc.com>

Regards,
Anup

> ---
>  target/riscv/kvm.c | 104 ++++++++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 103 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index 6d4df0ef6d..e695b91dc7 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -73,6 +73,14 @@ static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx
>          } \
>      } while(0)
>
> +#define KVM_RISCV_SET_CSR(cs, env, csr, reg) \
> +    do { \
> +        int ret = kvm_set_one_reg(cs, RISCV_CSR_REG(env, csr), &reg); \
> +        if (ret) { \
> +            return ret; \
> +        } \
> +    } while(0)
> +
>  static int kvm_riscv_get_regs_core(CPUState *cs)
>  {
>      int ret = 0;
> @@ -98,6 +106,31 @@ static int kvm_riscv_get_regs_core(CPUState *cs)
>      return ret;
>  }
>
> +static int kvm_riscv_put_regs_core(CPUState *cs)
> +{
> +    int ret = 0;
> +    int i;
> +    target_ulong reg;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    reg = env->pc;
> +    ret = kvm_set_one_reg(cs, RISCV_CORE_REG(env, regs.pc), &reg);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    for (i = 1; i < 32; i++) {
> +        uint64_t id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CORE, i);
> +        reg = env->gpr[i];
> +        ret = kvm_set_one_reg(cs, id, &reg);
> +        if (ret) {
> +            return ret;
> +        }
> +    }
> +
> +    return ret;
> +}
> +
>  static int kvm_riscv_get_regs_csr(CPUState *cs)
>  {
>      int ret = 0;
> @@ -115,6 +148,24 @@ static int kvm_riscv_get_regs_csr(CPUState *cs)
>      return ret;
>  }
>
> +static int kvm_riscv_put_regs_csr(CPUState *cs)
> +{
> +    int ret = 0;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    KVM_RISCV_SET_CSR(cs, env, sstatus, env->mstatus);
> +    KVM_RISCV_SET_CSR(cs, env, sie, env->mie);
> +    KVM_RISCV_SET_CSR(cs, env, stvec, env->stvec);
> +    KVM_RISCV_SET_CSR(cs, env, sscratch, env->sscratch);
> +    KVM_RISCV_SET_CSR(cs, env, sepc, env->sepc);
> +    KVM_RISCV_SET_CSR(cs, env, scause, env->scause);
> +    KVM_RISCV_SET_CSR(cs, env, stval, env->stval);
> +    KVM_RISCV_SET_CSR(cs, env, sip, env->mip);
> +    KVM_RISCV_SET_CSR(cs, env, satp, env->satp);
> +
> +    return ret;
> +}
> +
>  static int kvm_riscv_get_regs_fp(CPUState *cs)
>  {
>      int ret = 0;
> @@ -148,6 +199,40 @@ static int kvm_riscv_get_regs_fp(CPUState *cs)
>      return ret;
>  }
>
> +static int kvm_riscv_put_regs_fp(CPUState *cs)
> +{
> +    int ret = 0;
> +    int i;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    if (riscv_has_ext(env, RVD)) {
> +        uint64_t reg;
> +        for (i = 0; i < 32; i++) {
> +            reg = env->fpr[i];
> +            ret = kvm_set_one_reg(cs, RISCV_FP_D_REG(env, i), &reg);
> +            if (ret) {
> +                return ret;
> +            }
> +        }
> +        return ret;
> +    }
> +
> +    if (riscv_has_ext(env, RVF)) {
> +        uint32_t reg;
> +        for (i = 0; i < 32; i++) {
> +            reg = env->fpr[i];
> +            ret = kvm_set_one_reg(cs, RISCV_FP_F_REG(env, i), &reg);
> +            if (ret) {
> +                return ret;
> +            }
> +        }
> +        return ret;
> +    }
> +
> +    return ret;
> +}
> +
> +
>  const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
>      KVM_CAP_LAST_INFO
>  };
> @@ -176,7 +261,24 @@ int kvm_arch_get_registers(CPUState *cs)
>
>  int kvm_arch_put_registers(CPUState *cs, int level)
>  {
> -    return 0;
> +    int ret = 0;
> +
> +    ret = kvm_riscv_put_regs_core(cs);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    ret = kvm_riscv_put_regs_csr(cs);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    ret = kvm_riscv_put_regs_fp(cs);
> +    if (ret) {
> +        return ret;
> +    }
> +
> +    return ret;
>  }
>
>  int kvm_arch_release_virq_post(int virq)
> --
> 2.19.1
>

Re: [PATCH v2 06/12] target/riscv: Support start kernel directly by KVM

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Get kernel and fdt start address in virt.c, and pass them to KVM
> when cpu reset. In addition, add kvm_riscv.h to place riscv specific
> interface.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
> ---
>  hw/riscv/boot.c          | 11 ++++++++
>  hw/riscv/virt.c          | 54 ++++++++++++++++++++++++++++------------
>  include/hw/riscv/boot.h  |  1 +
>  target/riscv/cpu.c       |  8 ++++++
>  target/riscv/cpu.h       |  3 +++
>  target/riscv/kvm-stub.c  | 25 +++++++++++++++++++
>  target/riscv/kvm.c       | 14 +++++++++++
>  target/riscv/kvm_riscv.h | 24 ++++++++++++++++++
>  target/riscv/meson.build |  2 +-
>  9 files changed, 125 insertions(+), 17 deletions(-)
>  create mode 100644 target/riscv/kvm-stub.c
>  create mode 100644 target/riscv/kvm_riscv.h
>
> diff --git a/hw/riscv/boot.c b/hw/riscv/boot.c
> index 519fa455a1..00df6d7810 100644
> --- a/hw/riscv/boot.c
> +++ b/hw/riscv/boot.c
> @@ -317,3 +317,14 @@ void riscv_setup_rom_reset_vec(MachineState *machine, RISCVHartArrayState *harts
>
>      return;
>  }
> +
> +void riscv_setup_direct_kernel(hwaddr kernel_addr, hwaddr fdt_addr)
> +{
> +    CPUState *cs;
> +
> +    for (cs = first_cpu; cs; cs = CPU_NEXT(cs)) {
> +        RISCVCPU *riscv_cpu = RISCV_CPU(cs);
> +        riscv_cpu->env.kernel_addr = kernel_addr;
> +        riscv_cpu->env.fdt_addr = fdt_addr;
> +    }
> +}
> diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c
> index 3af074148e..c8bcd9d9e5 100644
> --- a/hw/riscv/virt.c
> +++ b/hw/riscv/virt.c
> @@ -38,6 +38,7 @@
>  #include "chardev/char.h"
>  #include "sysemu/device_tree.h"
>  #include "sysemu/sysemu.h"
> +#include "sysemu/kvm.h"
>  #include "hw/pci/pci.h"
>  #include "hw/pci-host/gpex.h"
>  #include "hw/display/ramfb.h"
> @@ -801,23 +802,25 @@ static void virt_machine_init(MachineState *machine)
>                                  hart_count, &error_abort);
>          sysbus_realize(SYS_BUS_DEVICE(&s->soc[i]), &error_abort);
>
> -        /* Per-socket CLINT */
> -        riscv_aclint_swi_create(
> -            memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size,
> -            base_hartid, hart_count, false);
> -        riscv_aclint_mtimer_create(
> -            memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size +
> -                RISCV_ACLINT_SWI_SIZE,
> -            RISCV_ACLINT_DEFAULT_MTIMER_SIZE, base_hartid, hart_count,
> -            RISCV_ACLINT_DEFAULT_MTIMECMP, RISCV_ACLINT_DEFAULT_MTIME,
> -            RISCV_ACLINT_DEFAULT_TIMEBASE_FREQ, true);
> -
> -        /* Per-socket ACLINT SSWI */
> -        if (s->have_aclint) {
> +        if (!kvm_enabled()) {
> +            /* Per-socket CLINT */
>              riscv_aclint_swi_create(
> -                memmap[VIRT_ACLINT_SSWI].base +
> -                    i * memmap[VIRT_ACLINT_SSWI].size,
> -                base_hartid, hart_count, true);
> +                memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size,
> +                base_hartid, hart_count, false);
> +            riscv_aclint_mtimer_create(
> +                memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size +
> +                    RISCV_ACLINT_SWI_SIZE,
> +                RISCV_ACLINT_DEFAULT_MTIMER_SIZE, base_hartid, hart_count,
> +                RISCV_ACLINT_DEFAULT_MTIMECMP, RISCV_ACLINT_DEFAULT_MTIME,
> +                RISCV_ACLINT_DEFAULT_TIMEBASE_FREQ, true);
> +
> +            /* Per-socket ACLINT SSWI */
> +            if (s->have_aclint) {
> +                riscv_aclint_swi_create(
> +                    memmap[VIRT_ACLINT_SSWI].base +
> +                        i * memmap[VIRT_ACLINT_SSWI].size,
> +                    base_hartid, hart_count, true);
> +            }

Along with this, we should not generate FDT nodes of ACLINT (or SiFive CLINT)
when KVM is enabled.

>          }
>
>          /* Per-socket PLIC hart topology configuration string */
> @@ -884,6 +887,16 @@ static void virt_machine_init(MachineState *machine)
>      memory_region_add_subregion(system_memory, memmap[VIRT_MROM].base,
>                                  mask_rom);
>
> +    /*
> +     * Only direct boot kernel is currently supported for KVM VM,
> +     * so the "-bios" parameter is ignored and treated like "-bios none"
> +     * when KVM is enabled.
> +     */
> +    if (kvm_enabled()) {
> +        g_free(machine->firmware);
> +        machine->firmware = g_strdup("none");
> +    }
> +
>      if (riscv_is_32bit(&s->soc[0])) {
>          firmware_end_addr = riscv_find_and_load_firmware(machine,
>                                      RISCV32_BIOS_BIN, start_addr, NULL);
> @@ -941,6 +954,15 @@ static void virt_machine_init(MachineState *machine)
>                                virt_memmap[VIRT_MROM].size, kernel_entry,
>                                fdt_load_addr, machine->fdt);
>
> +    /*
> +     * Only direct boot kernel is currently supported for KVM VM,
> +     * So here setup kernel start address and fdt address.
> +     * TODO:Support firmware loading and integrate to TCG start
> +     */
> +    if (kvm_enabled()) {
> +        riscv_setup_direct_kernel(kernel_entry, fdt_load_addr);
> +    }
> +
>      /* SiFive Test MMIO device */
>      sifive_test_create(memmap[VIRT_TEST].base);
>
> diff --git a/include/hw/riscv/boot.h b/include/hw/riscv/boot.h
> index baff11dd8a..5834c234aa 100644
> --- a/include/hw/riscv/boot.h
> +++ b/include/hw/riscv/boot.h
> @@ -58,5 +58,6 @@ void riscv_rom_copy_firmware_info(MachineState *machine, hwaddr rom_base,
>                                    hwaddr rom_size,
>                                    uint32_t reset_vec_size,
>                                    uint64_t kernel_entry);
> +void riscv_setup_direct_kernel(hwaddr kernel_addr, hwaddr fdt_addr);
>
>  #endif /* RISCV_BOOT_H */
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index f812998123..1c944872a3 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -29,6 +29,8 @@
>  #include "hw/qdev-properties.h"
>  #include "migration/vmstate.h"
>  #include "fpu/softfloat-helpers.h"
> +#include "sysemu/kvm.h"
> +#include "kvm_riscv.h"
>
>  /* RISC-V CPU definitions */
>
> @@ -380,6 +382,12 @@ static void riscv_cpu_reset(DeviceState *dev)
>      cs->exception_index = RISCV_EXCP_NONE;
>      env->load_res = -1;
>      set_default_nan_mode(1, &env->fp_status);
> +
> +#ifndef CONFIG_USER_ONLY
> +    if (kvm_enabled()) {
> +        kvm_riscv_reset_vcpu(cpu);
> +    }
> +#endif
>  }
>
>  static void riscv_cpu_disas_set_info(CPUState *s, disassemble_info *info)
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index 0760c0af93..2807eb1bcb 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -255,6 +255,9 @@ struct CPURISCVState {
>
>      /* Fields from here on are preserved across CPU reset. */
>      QEMUTimer *timer; /* Internal timer */
> +
> +    hwaddr kernel_addr;
> +    hwaddr fdt_addr;
>  };
>
>  OBJECT_DECLARE_TYPE(RISCVCPU, RISCVCPUClass,
> diff --git a/target/riscv/kvm-stub.c b/target/riscv/kvm-stub.c
> new file mode 100644
> index 0000000000..39b96fe3f4
> --- /dev/null
> +++ b/target/riscv/kvm-stub.c
> @@ -0,0 +1,25 @@
> +/*
> + * QEMU KVM RISC-V specific function stubs
> + *
> + * Copyright (c) 2020 Huawei Technologies Co., Ltd
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2 or later, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include "qemu/osdep.h"
> +#include "cpu.h"
> +#include "kvm_riscv.h"
> +
> +void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
> +{
> +    abort();
> +}
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index e695b91dc7..db6d8a5b6e 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -37,6 +37,7 @@
>  #include "hw/irq.h"
>  #include "qemu/log.h"
>  #include "hw/loader.h"
> +#include "kvm_riscv.h"
>
>  static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
>  {
> @@ -369,6 +370,19 @@ int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
>      return 0;
>  }
>
> +void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
> +{
> +    CPURISCVState *env = &cpu->env;
> +
> +    if (!kvm_enabled()) {
> +        return;
> +    }
> +    env->pc = cpu->env.kernel_addr;
> +    env->gpr[10] = kvm_arch_vcpu_id(CPU(cpu)); /* a0 */
> +    env->gpr[11] = cpu->env.fdt_addr;          /* a1 */
> +    env->satp = 0;
> +}
> +
>  bool kvm_arch_cpu_check_are_resettable(void)
>  {
>      return true;
> diff --git a/target/riscv/kvm_riscv.h b/target/riscv/kvm_riscv.h
> new file mode 100644
> index 0000000000..f38c82bf59
> --- /dev/null
> +++ b/target/riscv/kvm_riscv.h
> @@ -0,0 +1,24 @@
> +/*
> + * QEMU KVM support -- RISC-V specific functions.
> + *
> + * Copyright (c) 2020 Huawei Technologies Co., Ltd
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2 or later, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#ifndef QEMU_KVM_RISCV_H
> +#define QEMU_KVM_RISCV_H
> +
> +void kvm_riscv_reset_vcpu(RISCVCPU *cpu);
> +
> +#endif
> diff --git a/target/riscv/meson.build b/target/riscv/meson.build
> index 2faf08a941..fe41cc5805 100644
> --- a/target/riscv/meson.build
> +++ b/target/riscv/meson.build
> @@ -19,7 +19,7 @@ riscv_ss.add(files(
>    'bitmanip_helper.c',
>    'translate.c',
>  ))
> -riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'))
> +riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'), if_false: files('kvm-stub.c'))
>
>  riscv_softmmu_ss = ss.source_set()
>  riscv_softmmu_ss.add(files(
> --
> 2.19.1
>

Regards,
Anup

Re: [PATCH v2 07/12] target/riscv: Support setting external interrupt by KVM

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> When KVM is enabled, set the S-mode external interrupt through
> kvm_riscv_set_irq function.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
> ---
>  target/riscv/cpu.c       |  6 +++++-
>  target/riscv/kvm-stub.c  |  5 +++++
>  target/riscv/kvm.c       | 17 +++++++++++++++++
>  target/riscv/kvm_riscv.h |  1 +
>  4 files changed, 28 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 1c944872a3..71a7ac6831 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -603,7 +603,11 @@ static void riscv_cpu_set_irq(void *opaque, int irq, int level)
>      case IRQ_S_EXT:
>      case IRQ_VS_EXT:
>      case IRQ_M_EXT:
> -        riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
> +        if (kvm_enabled() && (irq & IRQ_M_EXT) ) {
> +            kvm_riscv_set_irq(cpu, IRQ_S_EXT, level);
> +        } else {
> +            riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
> +        }

This does not look right.

I suggest the following:

if (kvm_enabled()) {
    kvm_riscv_set_irq(cpu, irq, level);
} else {
   riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
}

>          break;
>      default:
>          g_assert_not_reached();
> diff --git a/target/riscv/kvm-stub.c b/target/riscv/kvm-stub.c
> index 39b96fe3f4..4e8fc31a21 100644
> --- a/target/riscv/kvm-stub.c
> +++ b/target/riscv/kvm-stub.c
> @@ -23,3 +23,8 @@ void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
>  {
>      abort();
>  }
> +
> +void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level)
> +{
> +    abort();
> +}
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index db6d8a5b6e..0027f11f45 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -383,6 +383,23 @@ void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
>      env->satp = 0;
>  }
>
> +void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level)
> +{
> +    int ret;
> +    unsigned virq = level ? KVM_INTERRUPT_SET : KVM_INTERRUPT_UNSET;
> +
> +    if (irq != IRQ_S_EXT) {
> +        perror("kvm riscv set irq != IRQ_S_EXT\n");
> +        abort();
> +    }
> +
> +    ret = kvm_vcpu_ioctl(CPU(cpu), KVM_INTERRUPT, &virq);
> +    if (ret < 0) {
> +        perror("Set irq failed");
> +        abort();
> +    }
> +}
> +
>  bool kvm_arch_cpu_check_are_resettable(void)
>  {
>      return true;
> diff --git a/target/riscv/kvm_riscv.h b/target/riscv/kvm_riscv.h
> index f38c82bf59..ed281bdce0 100644
> --- a/target/riscv/kvm_riscv.h
> +++ b/target/riscv/kvm_riscv.h
> @@ -20,5 +20,6 @@
>  #define QEMU_KVM_RISCV_H
>
>  void kvm_riscv_reset_vcpu(RISCVCPU *cpu);
> +void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level);
>
>  #endif
> --
> 2.19.1
>

Regards,
Anup

Re: [PATCH v2 08/12] target/riscv: Handle KVM_EXIT_RISCV_SBI exit

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Use char-fe to handle console sbi call, which implement early
> console io while apply 'earlycon=sbi' into kernel parameters.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>

Looks good to me.

Reviewed-by: Anup Patel <anup.patel@wdc.com>

Regards,
Anup

> ---
>  target/riscv/kvm.c                 | 38 +++++++++++++++-
>  target/riscv/sbi_ecall_interface.h | 72 ++++++++++++++++++++++++++++++
>  2 files changed, 109 insertions(+), 1 deletion(-)
>  create mode 100644 target/riscv/sbi_ecall_interface.h
>
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index 0027f11f45..171a32adf9 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -38,6 +38,8 @@
>  #include "qemu/log.h"
>  #include "hw/loader.h"
>  #include "kvm_riscv.h"
> +#include "sbi_ecall_interface.h"
> +#include "semihosting/console.h"
>
>  static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
>  {
> @@ -365,9 +367,43 @@ bool kvm_arch_stop_on_emulation_error(CPUState *cs)
>      return true;
>  }
>
> +static int kvm_riscv_handle_sbi(CPUState *cs, struct kvm_run *run)
> +{
> +    int ret = 0;
> +    unsigned char ch;
> +    switch (run->riscv_sbi.extension_id) {
> +    case SBI_EXT_0_1_CONSOLE_PUTCHAR:
> +        ch = run->riscv_sbi.args[0];
> +        qemu_semihosting_log_out((const char *)&ch, sizeof(ch));
> +        break;
> +    case SBI_EXT_0_1_CONSOLE_GETCHAR:
> +        run->riscv_sbi.args[0] =
> +            (unsigned long)qemu_semihosting_console_inc(cs->env_ptr);
> +        break;
> +    default:
> +        qemu_log_mask(LOG_UNIMP,
> +                      "%s: un-handled SBI EXIT, specific reasons is %lu\n",
> +                      __func__, run->riscv_sbi.extension_id);
> +        ret = -1;
> +        break;
> +    }
> +    return ret;
> +}
> +
>  int kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
>  {
> -    return 0;
> +    int ret = 0;
> +    switch (run->exit_reason) {
> +    case KVM_EXIT_RISCV_SBI:
> +        ret = kvm_riscv_handle_sbi(cs, run);
> +        break;
> +    default:
> +        qemu_log_mask(LOG_UNIMP, "%s: un-handled exit reason %d\n",
> +                      __func__, run->exit_reason);
> +        ret = -1;
> +        break;
> +    }
> +    return ret;
>  }
>
>  void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
> diff --git a/target/riscv/sbi_ecall_interface.h b/target/riscv/sbi_ecall_interface.h
> new file mode 100644
> index 0000000000..fb1a3fa8f2
> --- /dev/null
> +++ b/target/riscv/sbi_ecall_interface.h
> @@ -0,0 +1,72 @@
> +/*
> + * SPDX-License-Identifier: BSD-2-Clause
> + *
> + * Copyright (c) 2019 Western Digital Corporation or its affiliates.
> + *
> + * Authors:
> + *   Anup Patel <anup.patel@wdc.com>
> + */
> +
> +#ifndef __SBI_ECALL_INTERFACE_H__
> +#define __SBI_ECALL_INTERFACE_H__
> +
> +/* clang-format off */
> +
> +/* SBI Extension IDs */
> +#define SBI_EXT_0_1_SET_TIMER           0x0
> +#define SBI_EXT_0_1_CONSOLE_PUTCHAR     0x1
> +#define SBI_EXT_0_1_CONSOLE_GETCHAR     0x2
> +#define SBI_EXT_0_1_CLEAR_IPI           0x3
> +#define SBI_EXT_0_1_SEND_IPI            0x4
> +#define SBI_EXT_0_1_REMOTE_FENCE_I      0x5
> +#define SBI_EXT_0_1_REMOTE_SFENCE_VMA   0x6
> +#define SBI_EXT_0_1_REMOTE_SFENCE_VMA_ASID 0x7
> +#define SBI_EXT_0_1_SHUTDOWN            0x8
> +#define SBI_EXT_BASE                    0x10
> +#define SBI_EXT_TIME                    0x54494D45
> +#define SBI_EXT_IPI                     0x735049
> +#define SBI_EXT_RFENCE                  0x52464E43
> +#define SBI_EXT_HSM                     0x48534D
> +
> +/* SBI function IDs for BASE extension*/
> +#define SBI_EXT_BASE_GET_SPEC_VERSION   0x0
> +#define SBI_EXT_BASE_GET_IMP_ID         0x1
> +#define SBI_EXT_BASE_GET_IMP_VERSION    0x2
> +#define SBI_EXT_BASE_PROBE_EXT          0x3
> +#define SBI_EXT_BASE_GET_MVENDORID      0x4
> +#define SBI_EXT_BASE_GET_MARCHID        0x5
> +#define SBI_EXT_BASE_GET_MIMPID         0x6
> +
> +/* SBI function IDs for TIME extension*/
> +#define SBI_EXT_TIME_SET_TIMER          0x0
> +
> +/* SBI function IDs for IPI extension*/
> +#define SBI_EXT_IPI_SEND_IPI            0x0
> +
> +/* SBI function IDs for RFENCE extension*/
> +#define SBI_EXT_RFENCE_REMOTE_FENCE_I       0x0
> +#define SBI_EXT_RFENCE_REMOTE_SFENCE_VMA    0x1
> +#define SBI_EXT_RFENCE_REMOTE_SFENCE_VMA_ASID  0x2
> +#define SBI_EXT_RFENCE_REMOTE_HFENCE_GVMA   0x3
> +#define SBI_EXT_RFENCE_REMOTE_HFENCE_GVMA_VMID 0x4
> +#define SBI_EXT_RFENCE_REMOTE_HFENCE_VVMA   0x5
> +#define SBI_EXT_RFENCE_REMOTE_HFENCE_VVMA_ASID 0x6
> +
> +/* SBI function IDs for HSM extension */
> +#define SBI_EXT_HSM_HART_START          0x0
> +#define SBI_EXT_HSM_HART_STOP           0x1
> +#define SBI_EXT_HSM_HART_GET_STATUS     0x2
> +
> +#define SBI_HSM_HART_STATUS_STARTED     0x0
> +#define SBI_HSM_HART_STATUS_STOPPED     0x1
> +#define SBI_HSM_HART_STATUS_START_PENDING   0x2
> +#define SBI_HSM_HART_STATUS_STOP_PENDING    0x3
> +
> +#define SBI_SPEC_VERSION_MAJOR_OFFSET   24
> +#define SBI_SPEC_VERSION_MAJOR_MASK     0x7f
> +#define SBI_SPEC_VERSION_MINOR_MASK     0xffffff
> +#define SBI_EXT_VENDOR_START            0x09000000
> +#define SBI_EXT_VENDOR_END              0x09FFFFFF
> +/* clang-format on */
> +
> +#endif
> --
> 2.19.1
>

Re: [PATCH v2 11/12] target/riscv: Implement virtual time adjusting with vm state changing

[Anup Patel]

On Fri, Dec 10, 2021 at 3:38 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> We hope that virtual time adjusts with vm state changing. When a vm
> is stopped, guest virtual time should stop counting and kvm_timer
> should be stopped. When the vm is resumed, guest virtual time should
> continue to count and kvm_timer should be restored.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>

Looks good to me.

Reviewed-by: Anup Patel <anup.patel@wdc.com>

Regards,
Anup

> ---
>  target/riscv/kvm.c | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
>
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index 802c076b22..be95dbc3f3 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -40,6 +40,7 @@
>  #include "kvm_riscv.h"
>  #include "sbi_ecall_interface.h"
>  #include "semihosting/console.h"
> +#include "sysemu/runstate.h"
>
>  static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx)
>  {
> @@ -377,6 +378,17 @@ unsigned long kvm_arch_vcpu_id(CPUState *cpu)
>      return cpu->cpu_index;
>  }
>
> +static void kvm_riscv_vm_state_change(void *opaque, bool running, RunState state)
> +{
> +    CPUState *cs = opaque;
> +
> +    if (running) {
> +        kvm_riscv_put_regs_timer(cs);
> +    } else {
> +        kvm_riscv_get_regs_timer(cs);
> +    }
> +}
> +
>  void kvm_arch_init_irq_routing(KVMState *s)
>  {
>  }
> @@ -389,6 +401,8 @@ int kvm_arch_init_vcpu(CPUState *cs)
>      CPURISCVState *env = &cpu->env;
>      uint64_t id;
>
> +    qemu_add_vm_change_state_handler(kvm_riscv_vm_state_change, cs);
> +
>      id = kvm_riscv_reg_id(env, KVM_REG_RISCV_CONFIG, KVM_REG_RISCV_CONFIG_REG(isa));
>      ret = kvm_get_one_reg(cs, id, &isa);
>      if (ret) {
> --
> 2.19.1
>

Re: [PATCH v2 12/12] target/riscv: Support virtual time context synchronization

[Anup Patel]

On Fri, Dec 10, 2021 at 3:38 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Add virtual time context description to vmstate_kvmtimer. After cpu being
> loaded, virtual time context is updated to KVM.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> ---
>  target/riscv/machine.c | 37 +++++++++++++++++++++++++++++++++++--
>  1 file changed, 35 insertions(+), 2 deletions(-)
>
> diff --git a/target/riscv/machine.c b/target/riscv/machine.c
> index ad8248ebfd..f46443c316 100644
> --- a/target/riscv/machine.c
> +++ b/target/riscv/machine.c
> @@ -164,10 +164,42 @@ static const VMStateDescription vmstate_pointermasking = {
>      }
>  };
>
> +static bool kvmtimer_needed(void *opaque)
> +{
> +    return kvm_enabled();
> +}
> +
> +

Remove extra newline from here.

> +static const VMStateDescription vmstate_kvmtimer = {
> +    .name = "cpu/kvmtimer",
> +    .version_id = 1,
> +    .minimum_version_id = 1,
> +    .needed = kvmtimer_needed,
> +    .fields = (VMStateField[]) {
> +        VMSTATE_UINT64(env.kvm_timer_time, RISCVCPU),
> +        VMSTATE_UINT64(env.kvm_timer_compare, RISCVCPU),
> +        VMSTATE_UINT64(env.kvm_timer_state, RISCVCPU),
> +
> +        VMSTATE_END_OF_LIST()
> +    }
> +};
> +
> +static int cpu_post_load(void *opaque, int version_id)
> +{
> +    RISCVCPU *cpu = opaque;
> +    CPURISCVState *env = &cpu->env;
> +
> +    if (kvm_enabled()) {
> +        env->kvm_timer_dirty = true;
> +    }
> +    return 0;
> +}
> +
>  const VMStateDescription vmstate_riscv_cpu = {
>      .name = "cpu",
> -    .version_id = 3,
> -    .minimum_version_id = 3,
> +    .version_id = 4,
> +    .minimum_version_id = 4,
> +    .post_load = cpu_post_load,
>      .fields = (VMStateField[]) {
>          VMSTATE_UINTTL_ARRAY(env.gpr, RISCVCPU, 32),
>          VMSTATE_UINT64_ARRAY(env.fpr, RISCVCPU, 32),
> @@ -218,6 +250,7 @@ const VMStateDescription vmstate_riscv_cpu = {
>          &vmstate_hyper,
>          &vmstate_vector,
>          &vmstate_pointermasking,
> +        &vmstate_kvmtimer,
>          NULL
>      }
>  };
> --
> 2.19.1
>
>
> --
> kvm-riscv mailing list
> kvm-riscv@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kvm-riscv

Otherwise, it looks good to me.

Reviewed-by: Anup Patel <anup.patel@wdc.com>

Regards,
Anup

Re: [PATCH v2 10/12] target/riscv: Add kvm_riscv_get/put_regs_timer

[Anup Patel]

On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
>
> Add kvm_riscv_get/put_regs_timer to synchronize virtual time context
> from KVM.
>
> To set register of RISCV_TIMER_REG(state) will occur a error from KVM
> on kvm_timer_state == 0. It's better to adapt in KVM, but it doesn't matter
> that adaping in QEMU.
>
> Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> Signed-off-by: Mingwang Li <limingwang@huawei.com>
> ---
>  target/riscv/cpu.h |  7 +++++
>  target/riscv/kvm.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 79 insertions(+)
>
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index e7dba35acb..c892a2c8b7 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -259,6 +259,13 @@ struct CPURISCVState {
>
>      hwaddr kernel_addr;
>      hwaddr fdt_addr;
> +
> +    /* kvm timer */
> +    bool kvm_timer_dirty;
> +    uint64_t kvm_timer_time;
> +    uint64_t kvm_timer_compare;
> +    uint64_t kvm_timer_state;
> +    uint64_t kvm_timer_frequency;
>  };
>
>  OBJECT_DECLARE_TYPE(RISCVCPU, RISCVCPUClass,
> diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c
> index 171a32adf9..802c076b22 100644
> --- a/target/riscv/kvm.c
> +++ b/target/riscv/kvm.c
> @@ -64,6 +64,9 @@ static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type, uint64_t idx
>  #define RISCV_CSR_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_CSR, \
>                   KVM_REG_RISCV_CSR_REG(name))
>
> +#define RISCV_TIMER_REG(env, name)  kvm_riscv_reg_id(env, KVM_REG_RISCV_TIMER, \
> +                 KVM_REG_RISCV_TIMER_REG(name))
> +
>  #define RISCV_FP_F_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_F, idx)
>
>  #define RISCV_FP_D_REG(env, idx)  kvm_riscv_reg_id(env, KVM_REG_RISCV_FP_D, idx)
> @@ -235,6 +238,75 @@ static int kvm_riscv_put_regs_fp(CPUState *cs)
>      return ret;
>  }
>
> +static void kvm_riscv_get_regs_timer(CPUState *cs)
> +{
> +    int ret;
> +    uint64_t reg;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    if (env->kvm_timer_dirty) {
> +        return;
> +    }
> +
> +    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, time), &reg);
> +    if (ret) {
> +        abort();
> +    }
> +    env->kvm_timer_time = reg;
> +
> +    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, compare), &reg);
> +    if (ret) {
> +        abort();
> +    }
> +    env->kvm_timer_compare = reg;
> +
> +    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, state), &reg);
> +    if (ret) {
> +        abort();
> +    }
> +    env->kvm_timer_state = reg;

Please read the timer frequency here.

> +
> +    env->kvm_timer_dirty = true;
> +}
> +
> +static void kvm_riscv_put_regs_timer(CPUState *cs)
> +{
> +    int ret;
> +    uint64_t reg;
> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> +
> +    if (!env->kvm_timer_dirty) {
> +        return;
> +    }

Over here, we should get the timer frequency and abort() with an
error message if it does not match env->kvm_timer_frequency

For now, migration will not work between Hosts with different
timer frequency.

Regards,
Anup

> +
> +    reg = env->kvm_timer_time;
> +    ret = kvm_set_one_reg(cs, RISCV_TIMER_REG(env, time), &reg);
> +    if (ret) {
> +        abort();
> +    }
> +
> +    reg = env->kvm_timer_compare;
> +    ret = kvm_set_one_reg(cs, RISCV_TIMER_REG(env, compare), &reg);
> +    if (ret) {
> +        abort();
> +    }
> +
> +    /*
> +     * To set register of RISCV_TIMER_REG(state) will occur a error from KVM
> +     * on env->kvm_timer_state == 0, It's better to adapt in KVM, but it
> +     * doesn't matter that adaping in QEMU now.
> +     * TODO If KVM changes, adapt here.
> +     */
> +    if (env->kvm_timer_state) {
> +        reg = env->kvm_timer_state;
> +        ret = kvm_set_one_reg(cs, RISCV_TIMER_REG(env, state), &reg);
> +        if (ret) {
> +            abort();
> +        }
> +    }
> +
> +    env->kvm_timer_dirty = false;
> +}
>
>  const KVMCapabilityInfo kvm_arch_required_capabilities[] = {
>      KVM_CAP_LAST_INFO
> --
> 2.19.1
>

Re: [PATCH v2 10/12] target/riscv: Add kvm_riscv_get/put_regs_timer

[Richard Henderson]

On 12/12/21 9:05 PM, Anup Patel wrote:
>> +    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, state), &reg);
>> +    if (ret) {
>> +        abort();
>> +    }
>> +    env->kvm_timer_state = reg;
> 
> Please read the timer frequency here.

Yep.

>> +
>> +    env->kvm_timer_dirty = true;
>> +}
>> +
>> +static void kvm_riscv_put_regs_timer(CPUState *cs)
>> +{
>> +    int ret;
>> +    uint64_t reg;
>> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
>> +
>> +    if (!env->kvm_timer_dirty) {
>> +        return;
>> +    }
> 
> Over here, we should get the timer frequency and abort() with an
> error message if it does not match env->kvm_timer_frequency
> 
> For now, migration will not work between Hosts with different
> timer frequency.

You shouldn't have to do this every "put", only on migration, at which point you can 
actually signal a migration error rather than aborting directly.


r~

Re: [PATCH v2 12/12] target/riscv: Support virtual time context synchronization

[Richard Henderson]

On 12/10/21 2:07 AM, Yifei Jiang via wrote:
> +static bool kvmtimer_needed(void *opaque)
> +{
> +    return kvm_enabled();
> +}
> +
> +
> +static const VMStateDescription vmstate_kvmtimer = {
> +    .name = "cpu/kvmtimer",
> +    .version_id = 1,
> +    .minimum_version_id = 1,
> +    .needed = kvmtimer_needed,
> +    .fields = (VMStateField[]) {
> +        VMSTATE_UINT64(env.kvm_timer_time, RISCVCPU),
> +        VMSTATE_UINT64(env.kvm_timer_compare, RISCVCPU),
> +        VMSTATE_UINT64(env.kvm_timer_state, RISCVCPU),
> +
> +        VMSTATE_END_OF_LIST()
> +    }
> +};
> +
> +static int cpu_post_load(void *opaque, int version_id)
> +{
> +    RISCVCPU *cpu = opaque;
> +    CPURISCVState *env = &cpu->env;
> +
> +    if (kvm_enabled()) {
> +        env->kvm_timer_dirty = true;
> +    }
> +    return 0;
> +}

This post-load belongs on the vmstate_kvmtimer struct, so that you need not re-check 
kvm_enabled().

>   const VMStateDescription vmstate_riscv_cpu = {
>       .name = "cpu",
> -    .version_id = 3,
> -    .minimum_version_id = 3,
> +    .version_id = 4,
> +    .minimum_version_id = 4,
> +    .post_load = cpu_post_load,

No need for version change.


r~

RE: [PATCH v2 06/12] target/riscv: Support start kernel directly by KVM

[Jiangyifei]

> -----Original Message-----
> From: Anup Patel [mailto:anup@brainfault.org]
> Sent: Monday, December 13, 2021 12:21 PM
> To: Jiangyifei <jiangyifei@huawei.com>
> Cc: QEMU Developers <qemu-devel@nongnu.org>; open list:RISC-V
> <qemu-riscv@nongnu.org>; kvm-riscv@lists.infradead.org; KVM General
> <kvm@vger.kernel.org>; libvir-list@redhat.com; Anup Patel
> <anup.patel@wdc.com>; Palmer Dabbelt <palmer@dabbelt.com>; Alistair
> Francis <Alistair.Francis@wdc.com>; Bin Meng <bin.meng@windriver.com>;
> Fanliang (EulerOS) <fanliang@huawei.com>; Wubin (H)
> <wu.wubin@huawei.com>; Wanghaibin (D) <wanghaibin.wang@huawei.com>;
> wanbo (G) <wanbo13@huawei.com>; limingwang (A)
> <limingwang@huawei.com>
> Subject: Re: [PATCH v2 06/12] target/riscv: Support start kernel directly by KVM
> 
> On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
> >
> > Get kernel and fdt start address in virt.c, and pass them to KVM when
> > cpu reset. In addition, add kvm_riscv.h to place riscv specific
> > interface.
> >
> > Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> > Signed-off-by: Mingwang Li <limingwang@huawei.com>
> > Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
> > ---
> >  hw/riscv/boot.c          | 11 ++++++++
> >  hw/riscv/virt.c          | 54 ++++++++++++++++++++++++++++------------
> >  include/hw/riscv/boot.h  |  1 +
> >  target/riscv/cpu.c       |  8 ++++++
> >  target/riscv/cpu.h       |  3 +++
> >  target/riscv/kvm-stub.c  | 25 +++++++++++++++++++
> >  target/riscv/kvm.c       | 14 +++++++++++
> >  target/riscv/kvm_riscv.h | 24 ++++++++++++++++++
> > target/riscv/meson.build |  2 +-
> >  9 files changed, 125 insertions(+), 17 deletions(-)  create mode
> > 100644 target/riscv/kvm-stub.c  create mode 100644
> > target/riscv/kvm_riscv.h
> >
> > diff --git a/hw/riscv/boot.c b/hw/riscv/boot.c index
> > 519fa455a1..00df6d7810 100644
> > --- a/hw/riscv/boot.c
> > +++ b/hw/riscv/boot.c
> > @@ -317,3 +317,14 @@ void riscv_setup_rom_reset_vec(MachineState
> > *machine, RISCVHartArrayState *harts
> >
> >      return;
> >  }
> > +
> > +void riscv_setup_direct_kernel(hwaddr kernel_addr, hwaddr fdt_addr) {
> > +    CPUState *cs;
> > +
> > +    for (cs = first_cpu; cs; cs = CPU_NEXT(cs)) {
> > +        RISCVCPU *riscv_cpu = RISCV_CPU(cs);
> > +        riscv_cpu->env.kernel_addr = kernel_addr;
> > +        riscv_cpu->env.fdt_addr = fdt_addr;
> > +    }
> > +}
> > diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c index
> > 3af074148e..c8bcd9d9e5 100644
> > --- a/hw/riscv/virt.c
> > +++ b/hw/riscv/virt.c
> > @@ -38,6 +38,7 @@
> >  #include "chardev/char.h"
> >  #include "sysemu/device_tree.h"
> >  #include "sysemu/sysemu.h"
> > +#include "sysemu/kvm.h"
> >  #include "hw/pci/pci.h"
> >  #include "hw/pci-host/gpex.h"
> >  #include "hw/display/ramfb.h"
> > @@ -801,23 +802,25 @@ static void virt_machine_init(MachineState
> *machine)
> >                                  hart_count, &error_abort);
> >          sysbus_realize(SYS_BUS_DEVICE(&s->soc[i]), &error_abort);
> >
> > -        /* Per-socket CLINT */
> > -        riscv_aclint_swi_create(
> > -            memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size,
> > -            base_hartid, hart_count, false);
> > -        riscv_aclint_mtimer_create(
> > -            memmap[VIRT_CLINT].base + i * memmap[VIRT_CLINT].size
> +
> > -                RISCV_ACLINT_SWI_SIZE,
> > -            RISCV_ACLINT_DEFAULT_MTIMER_SIZE, base_hartid,
> hart_count,
> > -            RISCV_ACLINT_DEFAULT_MTIMECMP,
> RISCV_ACLINT_DEFAULT_MTIME,
> > -            RISCV_ACLINT_DEFAULT_TIMEBASE_FREQ, true);
> > -
> > -        /* Per-socket ACLINT SSWI */
> > -        if (s->have_aclint) {
> > +        if (!kvm_enabled()) {
> > +            /* Per-socket CLINT */
> >              riscv_aclint_swi_create(
> > -                memmap[VIRT_ACLINT_SSWI].base +
> > -                    i * memmap[VIRT_ACLINT_SSWI].size,
> > -                base_hartid, hart_count, true);
> > +                memmap[VIRT_CLINT].base + i *
> memmap[VIRT_CLINT].size,
> > +                base_hartid, hart_count, false);
> > +            riscv_aclint_mtimer_create(
> > +                memmap[VIRT_CLINT].base + i *
> memmap[VIRT_CLINT].size +
> > +                    RISCV_ACLINT_SWI_SIZE,
> > +                RISCV_ACLINT_DEFAULT_MTIMER_SIZE, base_hartid,
> hart_count,
> > +                RISCV_ACLINT_DEFAULT_MTIMECMP,
> RISCV_ACLINT_DEFAULT_MTIME,
> > +                RISCV_ACLINT_DEFAULT_TIMEBASE_FREQ, true);
> > +
> > +            /* Per-socket ACLINT SSWI */
> > +            if (s->have_aclint) {
> > +                riscv_aclint_swi_create(
> > +                    memmap[VIRT_ACLINT_SSWI].base +
> > +                        i * memmap[VIRT_ACLINT_SSWI].size,
> > +                    base_hartid, hart_count, true);
> > +            }
> 
> Along with this, we should not generate FDT nodes of ACLINT (or SiFive CLINT)
> when KVM is enabled.
> 

OK, and I will remove M-mode PLIC contexts when kvm is enabled in next series.

Yifei

> >          }
> >
> >          /* Per-socket PLIC hart topology configuration string */ @@
> > -884,6 +887,16 @@ static void virt_machine_init(MachineState *machine)
> >      memory_region_add_subregion(system_memory,
> memmap[VIRT_MROM].base,
> >                                  mask_rom);
> >
> > +    /*
> > +     * Only direct boot kernel is currently supported for KVM VM,
> > +     * so the "-bios" parameter is ignored and treated like "-bios none"
> > +     * when KVM is enabled.
> > +     */
> > +    if (kvm_enabled()) {
> > +        g_free(machine->firmware);
> > +        machine->firmware = g_strdup("none");
> > +    }
> > +
> >      if (riscv_is_32bit(&s->soc[0])) {
> >          firmware_end_addr = riscv_find_and_load_firmware(machine,
> >                                      RISCV32_BIOS_BIN,
> start_addr,
> > NULL); @@ -941,6 +954,15 @@ static void virt_machine_init(MachineState
> *machine)
> >                                virt_memmap[VIRT_MROM].size,
> kernel_entry,
> >                                fdt_load_addr, machine->fdt);
> >
> > +    /*
> > +     * Only direct boot kernel is currently supported for KVM VM,
> > +     * So here setup kernel start address and fdt address.
> > +     * TODO:Support firmware loading and integrate to TCG start
> > +     */
> > +    if (kvm_enabled()) {
> > +        riscv_setup_direct_kernel(kernel_entry, fdt_load_addr);
> > +    }
> > +
> >      /* SiFive Test MMIO device */
> >      sifive_test_create(memmap[VIRT_TEST].base);
> >
> > diff --git a/include/hw/riscv/boot.h b/include/hw/riscv/boot.h index
> > baff11dd8a..5834c234aa 100644
> > --- a/include/hw/riscv/boot.h
> > +++ b/include/hw/riscv/boot.h
> > @@ -58,5 +58,6 @@ void riscv_rom_copy_firmware_info(MachineState
> *machine, hwaddr rom_base,
> >                                    hwaddr rom_size,
> >                                    uint32_t reset_vec_size,
> >                                    uint64_t kernel_entry);
> > +void riscv_setup_direct_kernel(hwaddr kernel_addr, hwaddr fdt_addr);
> >
> >  #endif /* RISCV_BOOT_H */
> > diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c index
> > f812998123..1c944872a3 100644
> > --- a/target/riscv/cpu.c
> > +++ b/target/riscv/cpu.c
> > @@ -29,6 +29,8 @@
> >  #include "hw/qdev-properties.h"
> >  #include "migration/vmstate.h"
> >  #include "fpu/softfloat-helpers.h"
> > +#include "sysemu/kvm.h"
> > +#include "kvm_riscv.h"
> >
> >  /* RISC-V CPU definitions */
> >
> > @@ -380,6 +382,12 @@ static void riscv_cpu_reset(DeviceState *dev)
> >      cs->exception_index = RISCV_EXCP_NONE;
> >      env->load_res = -1;
> >      set_default_nan_mode(1, &env->fp_status);
> > +
> > +#ifndef CONFIG_USER_ONLY
> > +    if (kvm_enabled()) {
> > +        kvm_riscv_reset_vcpu(cpu);
> > +    }
> > +#endif
> >  }
> >
> >  static void riscv_cpu_disas_set_info(CPUState *s, disassemble_info
> > *info) diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index
> > 0760c0af93..2807eb1bcb 100644
> > --- a/target/riscv/cpu.h
> > +++ b/target/riscv/cpu.h
> > @@ -255,6 +255,9 @@ struct CPURISCVState {
> >
> >      /* Fields from here on are preserved across CPU reset. */
> >      QEMUTimer *timer; /* Internal timer */
> > +
> > +    hwaddr kernel_addr;
> > +    hwaddr fdt_addr;
> >  };
> >
> >  OBJECT_DECLARE_TYPE(RISCVCPU, RISCVCPUClass, diff --git
> > a/target/riscv/kvm-stub.c b/target/riscv/kvm-stub.c new file mode
> > 100644 index 0000000000..39b96fe3f4
> > --- /dev/null
> > +++ b/target/riscv/kvm-stub.c
> > @@ -0,0 +1,25 @@
> > +/*
> > + * QEMU KVM RISC-V specific function stubs
> > + *
> > + * Copyright (c) 2020 Huawei Technologies Co., Ltd
> > + *
> > + * This program is free software; you can redistribute it and/or
> > +modify it
> > + * under the terms and conditions of the GNU General Public License,
> > + * version 2 or later, as published by the Free Software Foundation.
> > + *
> > + * This program is distributed in the hope it will be useful, but
> > +WITHOUT
> > + * ANY WARRANTY; without even the implied warranty of
> MERCHANTABILITY
> > +or
> > + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public
> > +License for
> > + * more details.
> > + *
> > + * You should have received a copy of the GNU General Public License
> > +along with
> > + * this program.  If not, see <http://www.gnu.org/licenses/>.
> > + */
> > +#include "qemu/osdep.h"
> > +#include "cpu.h"
> > +#include "kvm_riscv.h"
> > +
> > +void kvm_riscv_reset_vcpu(RISCVCPU *cpu) {
> > +    abort();
> > +}
> > diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c index
> > e695b91dc7..db6d8a5b6e 100644
> > --- a/target/riscv/kvm.c
> > +++ b/target/riscv/kvm.c
> > @@ -37,6 +37,7 @@
> >  #include "hw/irq.h"
> >  #include "qemu/log.h"
> >  #include "hw/loader.h"
> > +#include "kvm_riscv.h"
> >
> >  static uint64_t kvm_riscv_reg_id(CPURISCVState *env, uint64_t type,
> > uint64_t idx)  { @@ -369,6 +370,19 @@ int
> > kvm_arch_handle_exit(CPUState *cs, struct kvm_run *run)
> >      return 0;
> >  }
> >
> > +void kvm_riscv_reset_vcpu(RISCVCPU *cpu) {
> > +    CPURISCVState *env = &cpu->env;
> > +
> > +    if (!kvm_enabled()) {
> > +        return;
> > +    }
> > +    env->pc = cpu->env.kernel_addr;
> > +    env->gpr[10] = kvm_arch_vcpu_id(CPU(cpu)); /* a0 */
> > +    env->gpr[11] = cpu->env.fdt_addr;          /* a1 */
> > +    env->satp = 0;
> > +}
> > +
> >  bool kvm_arch_cpu_check_are_resettable(void)
> >  {
> >      return true;
> > diff --git a/target/riscv/kvm_riscv.h b/target/riscv/kvm_riscv.h new
> > file mode 100644 index 0000000000..f38c82bf59
> > --- /dev/null
> > +++ b/target/riscv/kvm_riscv.h
> > @@ -0,0 +1,24 @@
> > +/*
> > + * QEMU KVM support -- RISC-V specific functions.
> > + *
> > + * Copyright (c) 2020 Huawei Technologies Co., Ltd
> > + *
> > + * This program is free software; you can redistribute it and/or
> > +modify it
> > + * under the terms and conditions of the GNU General Public License,
> > + * version 2 or later, as published by the Free Software Foundation.
> > + *
> > + * This program is distributed in the hope it will be useful, but
> > +WITHOUT
> > + * ANY WARRANTY; without even the implied warranty of
> MERCHANTABILITY
> > +or
> > + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public
> > +License for
> > + * more details.
> > + *
> > + * You should have received a copy of the GNU General Public License
> > +along with
> > + * this program.  If not, see <http://www.gnu.org/licenses/>.
> > + */
> > +
> > +#ifndef QEMU_KVM_RISCV_H
> > +#define QEMU_KVM_RISCV_H
> > +
> > +void kvm_riscv_reset_vcpu(RISCVCPU *cpu);
> > +
> > +#endif
> > diff --git a/target/riscv/meson.build b/target/riscv/meson.build index
> > 2faf08a941..fe41cc5805 100644
> > --- a/target/riscv/meson.build
> > +++ b/target/riscv/meson.build
> > @@ -19,7 +19,7 @@ riscv_ss.add(files(
> >    'bitmanip_helper.c',
> >    'translate.c',
> >  ))
> > -riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'))
> > +riscv_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'), if_false:
> > +files('kvm-stub.c'))
> >
> >  riscv_softmmu_ss = ss.source_set()
> >  riscv_softmmu_ss.add(files(
> > --
> > 2.19.1
> >
> 
> Regards,
> Anup

RE: [PATCH v2 07/12] target/riscv: Support setting external interrupt by KVM

[Jiangyifei]

> -----Original Message-----
> From: Anup Patel [mailto:anup@brainfault.org]
> Sent: Monday, December 13, 2021 12:33 PM
> To: Jiangyifei <jiangyifei@huawei.com>
> Cc: QEMU Developers <qemu-devel@nongnu.org>; open list:RISC-V
> <qemu-riscv@nongnu.org>; kvm-riscv@lists.infradead.org; KVM General
> <kvm@vger.kernel.org>; libvir-list@redhat.com; Anup Patel
> <anup.patel@wdc.com>; Palmer Dabbelt <palmer@dabbelt.com>; Alistair
> Francis <Alistair.Francis@wdc.com>; Bin Meng <bin.meng@windriver.com>;
> Fanliang (EulerOS) <fanliang@huawei.com>; Wubin (H)
> <wu.wubin@huawei.com>; Wanghaibin (D) <wanghaibin.wang@huawei.com>;
> wanbo (G) <wanbo13@huawei.com>; limingwang (A)
> <limingwang@huawei.com>
> Subject: Re: [PATCH v2 07/12] target/riscv: Support setting external interrupt
> by KVM
> 
> On Fri, Dec 10, 2021 at 3:37 PM Yifei Jiang <jiangyifei@huawei.com> wrote:
> >
> > When KVM is enabled, set the S-mode external interrupt through
> > kvm_riscv_set_irq function.
> >
> > Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
> > Signed-off-by: Mingwang Li <limingwang@huawei.com>
> > Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
> > ---
> >  target/riscv/cpu.c       |  6 +++++-
> >  target/riscv/kvm-stub.c  |  5 +++++
> >  target/riscv/kvm.c       | 17 +++++++++++++++++
> >  target/riscv/kvm_riscv.h |  1 +
> >  4 files changed, 28 insertions(+), 1 deletion(-)
> >
> > diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c index
> > 1c944872a3..71a7ac6831 100644
> > --- a/target/riscv/cpu.c
> > +++ b/target/riscv/cpu.c
> > @@ -603,7 +603,11 @@ static void riscv_cpu_set_irq(void *opaque, int irq,
> int level)
> >      case IRQ_S_EXT:
> >      case IRQ_VS_EXT:
> >      case IRQ_M_EXT:
> > -        riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
> > +        if (kvm_enabled() && (irq & IRQ_M_EXT) ) {
> > +            kvm_riscv_set_irq(cpu, IRQ_S_EXT, level);
> > +        } else {
> > +            riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level));
> > +        }
> 
> This does not look right.
> 
> I suggest the following:
> 
> if (kvm_enabled()) {
>     kvm_riscv_set_irq(cpu, irq, level);
> } else {
>    riscv_cpu_update_mip(cpu, 1 << irq, BOOL_TO_MASK(level)); }
> 

The M-mode PLIC context is not deleted. Therefore, IRQ_M_EXT needs to
be converted to IRQ_S_EXT. The M-mode PLIC contexts will be removed in
the next series, so it will be modified as suggested.

Yifei

> >          break;
> >      default:
> >          g_assert_not_reached();
> > diff --git a/target/riscv/kvm-stub.c b/target/riscv/kvm-stub.c index
> > 39b96fe3f4..4e8fc31a21 100644
> > --- a/target/riscv/kvm-stub.c
> > +++ b/target/riscv/kvm-stub.c
> > @@ -23,3 +23,8 @@ void kvm_riscv_reset_vcpu(RISCVCPU *cpu)  {
> >      abort();
> >  }
> > +
> > +void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level) {
> > +    abort();
> > +}
> > diff --git a/target/riscv/kvm.c b/target/riscv/kvm.c index
> > db6d8a5b6e..0027f11f45 100644
> > --- a/target/riscv/kvm.c
> > +++ b/target/riscv/kvm.c
> > @@ -383,6 +383,23 @@ void kvm_riscv_reset_vcpu(RISCVCPU *cpu)
> >      env->satp = 0;
> >  }
> >
> > +void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level) {
> > +    int ret;
> > +    unsigned virq = level ? KVM_INTERRUPT_SET :
> KVM_INTERRUPT_UNSET;
> > +
> > +    if (irq != IRQ_S_EXT) {
> > +        perror("kvm riscv set irq != IRQ_S_EXT\n");
> > +        abort();
> > +    }
> > +
> > +    ret = kvm_vcpu_ioctl(CPU(cpu), KVM_INTERRUPT, &virq);
> > +    if (ret < 0) {
> > +        perror("Set irq failed");
> > +        abort();
> > +    }
> > +}
> > +
> >  bool kvm_arch_cpu_check_are_resettable(void)
> >  {
> >      return true;
> > diff --git a/target/riscv/kvm_riscv.h b/target/riscv/kvm_riscv.h index
> > f38c82bf59..ed281bdce0 100644
> > --- a/target/riscv/kvm_riscv.h
> > +++ b/target/riscv/kvm_riscv.h
> > @@ -20,5 +20,6 @@
> >  #define QEMU_KVM_RISCV_H
> >
> >  void kvm_riscv_reset_vcpu(RISCVCPU *cpu);
> > +void kvm_riscv_set_irq(RISCVCPU *cpu, int irq, int level);
> >
> >  #endif
> > --
> > 2.19.1
> >
> 
> Regards,
> Anup

RE: [PATCH v2 10/12] target/riscv: Add kvm_riscv_get/put_regs_timer

[Jiangyifei]

> -----Original Message-----
> From: Richard Henderson [mailto:richard.henderson@linaro.org]
> Sent: Monday, December 13, 2021 11:20 PM
> To: Anup Patel <anup@brainfault.org>; Jiangyifei <jiangyifei@huawei.com>
> Cc: Bin Meng <bin.meng@windriver.com>; open list:RISC-V
> <qemu-riscv@nongnu.org>; limingwang (A) <limingwang@huawei.com>; KVM
> General <kvm@vger.kernel.org>; libvir-list@redhat.com; Anup Patel
> <anup.patel@wdc.com>; QEMU Developers <qemu-devel@nongnu.org>;
> wanbo (G) <wanbo13@huawei.com>; Palmer Dabbelt <palmer@dabbelt.com>;
> kvm-riscv@lists.infradead.org; Wanghaibin (D)
> <wanghaibin.wang@huawei.com>; Alistair Francis
> <Alistair.Francis@wdc.com>; Fanliang (EulerOS) <fanliang@huawei.com>;
> Wubin (H) <wu.wubin@huawei.com>
> Subject: Re: [PATCH v2 10/12] target/riscv: Add kvm_riscv_get/put_regs_timer
> 
> On 12/12/21 9:05 PM, Anup Patel wrote:
> >> +    ret = kvm_get_one_reg(cs, RISCV_TIMER_REG(env, state), &reg);
> >> +    if (ret) {
> >> +        abort();
> >> +    }
> >> +    env->kvm_timer_state = reg;
> >
> > Please read the timer frequency here.
> 
> Yep.
> 
> >> +
> >> +    env->kvm_timer_dirty = true;
> >> +}
> >> +
> >> +static void kvm_riscv_put_regs_timer(CPUState *cs) {
> >> +    int ret;
> >> +    uint64_t reg;
> >> +    CPURISCVState *env = &RISCV_CPU(cs)->env;
> >> +
> >> +    if (!env->kvm_timer_dirty) {
> >> +        return;
> >> +    }
> >
> > Over here, we should get the timer frequency and abort() with an error
> > message if it does not match env->kvm_timer_frequency
> >
> > For now, migration will not work between Hosts with different timer
> > frequency.
> 
> You shouldn't have to do this every "put", only on migration, at which point you
> can actually signal a migration error rather than aborting directly.
> 
> 
> r~

Yes, it will be modified in the next series.

Yifei

RE: [PATCH v2 12/12] target/riscv: Support virtual time context synchronization

[Jiangyifei]

> -----Original Message-----
> From: Richard Henderson [mailto:richard.henderson@linaro.org]
> Sent: Monday, December 13, 2021 11:22 PM
> To: Jiangyifei <jiangyifei@huawei.com>; qemu-devel@nongnu.org;
> qemu-riscv@nongnu.org
> Cc: kvm-riscv@lists.infradead.org; kvm@vger.kernel.org;
> libvir-list@redhat.com; anup.patel@wdc.com; palmer@dabbelt.com;
> Alistair.Francis@wdc.com; bin.meng@windriver.com; Fanliang (EulerOS)
> <fanliang@huawei.com>; Wubin (H) <wu.wubin@huawei.com>; Wanghaibin (D)
> <wanghaibin.wang@huawei.com>; wanbo (G) <wanbo13@huawei.com>;
> limingwang (A) <limingwang@huawei.com>
> Subject: Re: [PATCH v2 12/12] target/riscv: Support virtual time context
> synchronization
> 
> On 12/10/21 2:07 AM, Yifei Jiang via wrote:
> > +static bool kvmtimer_needed(void *opaque) {
> > +    return kvm_enabled();
> > +}
> > +
> > +
> > +static const VMStateDescription vmstate_kvmtimer = {
> > +    .name = "cpu/kvmtimer",
> > +    .version_id = 1,
> > +    .minimum_version_id = 1,
> > +    .needed = kvmtimer_needed,
> > +    .fields = (VMStateField[]) {
> > +        VMSTATE_UINT64(env.kvm_timer_time, RISCVCPU),
> > +        VMSTATE_UINT64(env.kvm_timer_compare, RISCVCPU),
> > +        VMSTATE_UINT64(env.kvm_timer_state, RISCVCPU),
> > +
> > +        VMSTATE_END_OF_LIST()
> > +    }
> > +};
> > +
> > +static int cpu_post_load(void *opaque, int version_id) {
> > +    RISCVCPU *cpu = opaque;
> > +    CPURISCVState *env = &cpu->env;
> > +
> > +    if (kvm_enabled()) {
> > +        env->kvm_timer_dirty = true;
> > +    }
> > +    return 0;
> > +}
> 
> This post-load belongs on the vmstate_kvmtimer struct, so that you need not
> re-check kvm_enabled().
> 
> >   const VMStateDescription vmstate_riscv_cpu = {
> >       .name = "cpu",
> > -    .version_id = 3,
> > -    .minimum_version_id = 3,
> > +    .version_id = 4,
> > +    .minimum_version_id = 4,
> > +    .post_load = cpu_post_load,
> 
> No need for version change.
> 
> 
> r~

Ok, it will be modified in the next series.

Yifei