From: Boris Derzhavets <bderzhavets@yahoo.com>
To: Bruce Edge <bruce.edge@gmail.com>,
Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>, xen-devel@lists.xensource.com
Subject: Re: Re: 2.6.37-rc1 mainline domU - BUG: unable to handle kernel paging request
Date: Tue, 16 Nov 2010 11:20:11 -0800 (PST) [thread overview]
Message-ID: <535480.36725.qm@web56104.mail.re3.yahoo.com> (raw)
[-- Attachment #1.1: Type: text/plain, Size: 6157 bytes --]
Sorry, patches for 2.6.37-rc1-git8 are attached.
You also would need :-
patch-2.6.37-rc1.bz2
patch-2.6.37-rc1-git8.bz2
All of them should be in /root/rpmbuild/SOURCES folder after src.rpm install on
any F14,F13,F12.
Boris.
--- On Tue, 11/16/10, Boris Derzhavets <bderzhavets@yahoo.com> wrote:
From: Boris Derzhavets <bderzhavets@yahoo.com>
Subject: Re: [Xen-devel] Re: 2.6.37-rc1 mainline domU - BUG: unable to handle kernel paging request
To: "Bruce Edge" <bruce.edge@gmail.com>, "Konrad Rzeszutek Wilk" <konrad.wilk@oracle.com>
Cc: "Jeremy Fitzhardinge" <jeremy@goop.org>, xen-devel@lists.xensource.com
Date: Tuesday, November 16, 2010, 1:43 PM
Today i've built vmlinuz-2.6.37-0.1.rc1.git8.xendom0.fc14.x86_64
via Michael's http://koji.fedoraproject.org/koji/taskinfo?taskID=2598434
and uncommented xen.pcifront.fixes.patch in kernel.spec, i.e.
# Xen patches
ApplyPatch xen.next-2.6.37.patch
# ApplyPatch xen.upstream.core.patch
ApplyPatch xen.pcifront.fixes.patch
# ApplyPatch xen.pvhvm.fixes.patch
as a fesult a got kernel wich runs pretty stable NFS client at Xen 4.0.1
F14 Dom0 (2.6.32.25-172.xendom0.fc14.x86_64).
I was able several times copied from NFS folder F14's ISO image (3.2 GB)
to DomU and scp'ed it back and didn't get any kernel crashing on DomU.
On Ubuntu 10.04 this kernel may be built as 2.6.37-rc1&git8 patched via
xen.next-2.6.37.patch
xen.pcifront.fixes.patch
All required upstream patches may be taken (as
well as 2 above)
from http://koji.fedoraproject.org/koji/taskinfo?taskID=2598434
I believe as soon as xen.pcifront.fixes.patch will be accepted by upstream
NFS client issue on F14 will be gone
Boris
--- On Mon, 11/15/10, Bruce Edge <bruce.edge@gmail.com> wrote:
From: Bruce Edge <bruce.edge@gmail.com>
Subject: Re: [Xen-devel] Re: 2.6.37-rc1 mainline domU - BUG: unable to handle kernel paging request
To: "Sander Eikelenboom" <linux@eikelenboom.it>
Cc: "Boris Derzhavets" <bderzhavets@yahoo.com>, "Konrad Rzeszutek Wilk" <konrad.wilk@oracle.com>, xen-devel@lists.xensource.com, "Jeremy Fitzhardinge" <jeremy@goop.org>
Date: Monday, November 15, 2010, 3:21 PM
On Sun, Nov 14, 2010 at 8:56 AM, Sander Eikelenboom
<linux@eikelenboom.it> wrote:
> Hmmm have you tried do do a lot of I/O with something else as NFS ?
> That would perhaps pinpoint it to NFS doing something not completely compatible with Xen.
>
> I'm not using NFS (I still use file: based guests, and i use glusterfs (fuse based userspace cluster fs) to share diskspace to domU's via ethernet).
Sander,
I took a quick look at glusterfs and it uses the same nfs client:
http://www.gluster.com/community/documentation/index.php/Gluster_3.1:_Manually_Mounting_Volumes_Using_NFS
I'm assuming that this would cause the same problem on the VM as we're
seeing with NFS, unless it's really an Xen/dom0 NFS server problem
that triggering the domU nfs client
crash.
Perhaps the context is different. I'm exporting filesystems to domU
from dom0 via NFS. Is that how you're using glusterfs, or are you
using it to host your file backed VM's storage? In the latter, that
may explain why you're not seeing these problems because you're not
using the nfs client on domU.
> I tried NFS in the past, but had some troubles setting it up, and even more problems with disconnects.
What kind of NFS problems? It was working very well for us until this
problem cropped up.
-Bruce
>
> I haven't seen any "unable to handle page request" problems with my mix of guest kernels, which includes some 2.6.37-rc1 kernels.
>
> --
>
> Sander
>
>
>
>
>
> Sunday, November 14, 2010, 5:37:59 PM, you wrote:
>
>> I've tested F14 DomU (kernel vmlinuz-2.6.37-0.1.rc1.git8.xendom0.fc14.x86_64) as NFS client and
Xen 4.0.1 F14 Dom0 (kernel vmlinuz-2.6.32.25-172.xendom0.fc14.x86_64) as NFS server . Copied 700 MB ISO images from NFS folder at Dom0 to DomU and scp'ed them back to Dom0. During about 30 - 40 min DomU ran pretty stable , regardless kernel crash as "unable to handle page request" was reported once by F14 DomU, but it didn't actually crash DomU. Same excersises with replacement F14 by Ubuntu 10.04 Server results DomU crash in about several minutes. Dom0's instances dual boot on same development box ( Q9500,ASUS P5Q3,8GB)
>
>> Boris.
>
>> --- On Fri, 11/12/10, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> wrote:
>
>> From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
>> Subject: Re: [Xen-devel] Re:
2.6.37-rc1 mainline domU - BUG: unable to handle kernel paging request
>> To: "Sander Eikelenboom" <linux@eikelenboom.it>
>> Cc: "Boris Derzhavets" <bderzhavets@yahoo.com>, xen-devel@lists.xensource.com, "Bruce Edge" <bruce.edge@gmail.com>, "Jeremy Fitzhardinge" <jeremy@goop.org>
>> Date: Friday, November 12, 2010, 12:01 PM
>
>> On Fri, Nov 12, 2010 at 05:27:43PM +0100, Sander Eikelenboom wrote:
>>> Hi
Bruce,
>>>
>>> Perhaps handpick some kernels before and after the pulls of the xen patches (pv-on-hvm etc) to begin with ?
>>> When you let git choose, especially with rc-1 kernels, you will end up with kernels in between patch series, resulting in panics.
>
>> Well, just the bare-bone boot of PV guests with nothing fancy ought to work.
>
>> But that is the theory and ..
>>> > The git bisecting is slow going. I've never tried that before and I'm a git
>>> > rookie.
>>> > I picked 2.6.36 - 2.6.37-rc1 as the bisect range and my first 2 bisects all
>>> > panic at boot so I'm obviously doing something wrong.
>>> > I'll RTFM a bit more and keep at it.
>
>> .. as Bruce experiences this is not the case. Hmm..
>
>> _______________________________________________
>> Xen-devel mailing
list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
>
>
>
>>
>
>
>
> --
> Best regards,
> Sander mailto:linux@eikelenboom.it
>
>
-----Inline Attachment Follows-----
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
[-- Attachment #1.2: Type: text/html, Size: 8211 bytes --]
[-- Attachment #2: xen.next-2.6.37.patch --]
[-- Type: application/octet-stream, Size: 70021 bytes --]
From 8bd6ddfd569309d8e915ffb6f68ad7bf03e53922 Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Fri, 20 Feb 2009 12:58:42 -0800
Subject: [PATCH 01/21] x86: define arch_vm_get_page_prot to set _PAGE_IOMAP on VM_IO vmas
Set _PAGE_IOMAP in ptes mapping a VM_IO vma. This says that the mapping
is of a real piece of physical hardware, and not just system memory.
Xen, in particular, uses to this to inhibit the normal pfn->mfn conversion
that would normally happen - in other words, treat the address directly
as a machine physical address without converting it from pseudo-physical.
[ Impact: make VM_IO mappings map the right thing under Xen ]
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/pgtable.h | 3 +++
arch/x86/mm/pgtable.c | 10 ++++++++++
2 files changed, 13 insertions(+), 0 deletions(-)
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index a34c785..7198bcf 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -399,6 +399,9 @@ static inline unsigned long pages_to_mb(unsigned long npg)
#define io_remap_pfn_range(vma, vaddr, pfn, size, prot) \
remap_pfn_range(vma, vaddr, pfn, size, prot)
+#define arch_vm_get_page_prot arch_vm_get_page_prot
+extern pgprot_t arch_vm_get_page_prot(unsigned vm_flags);
+
#if PAGETABLE_LEVELS > 2
static inline int pud_none(pud_t pud)
{
diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c
index 5c4ee42..5083449 100644
--- a/arch/x86/mm/pgtable.c
+++ b/arch/x86/mm/pgtable.c
@@ -15,6 +15,16 @@
gfp_t __userpte_alloc_gfp = PGALLOC_GFP | PGALLOC_USER_GFP;
+pgprot_t arch_vm_get_page_prot(unsigned vm_flags)
+{
+ pgprot_t ret = __pgprot(0);
+
+ if (vm_flags & VM_IO)
+ ret = __pgprot(_PAGE_IOMAP);
+
+ return ret;
+}
+
pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
{
return (pte_t *)__get_free_page(PGALLOC_GFP);
--
1.7.3.2
From 5527f9bee43a910b019ad77d7de6620b3412759b Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Fri, 2 Oct 2009 09:49:05 -0700
Subject: [PATCH 02/21] drm: recompute vma->vm_page_prot after changing vm_flags
vm_get_page_prot() computes vm_page_prot depending on vm_flags, so
we need to re-call it if we change flags.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/gpu/drm/ttm/ttm_bo_vm.c | 2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index fe6cb77..2d15c5e 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -273,6 +273,7 @@ int ttm_bo_mmap(struct file *filp, struct vm_area_struct *vma,
vma->vm_private_data = bo;
vma->vm_flags |= VM_RESERVED | VM_IO | VM_MIXEDMAP | VM_DONTEXPAND;
+ vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
return 0;
out_unref:
ttm_bo_unref(&bo);
@@ -288,6 +289,7 @@ int ttm_fbdev_mmap(struct vm_area_struct *vma, struct ttm_buffer_object *bo)
vma->vm_ops = &ttm_bo_vm_ops;
vma->vm_private_data = ttm_bo_reference(bo);
vma->vm_flags |= VM_RESERVED | VM_IO | VM_MIXEDMAP | VM_DONTEXPAND;
+ vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
return 0;
}
EXPORT_SYMBOL(ttm_fbdev_mmap);
--
1.7.3.2
From 66aae0dade39748bb7e068c4518c6160394ae3aa Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Thu, 25 Feb 2010 16:38:29 -0800
Subject: [PATCH 03/21] agp: Use PAGE_KERNEL_IO_NOCACHE for AGP mappings
When mapping AGP memory, the offset is a machine address. In Xen we
need to make sure mappings of physical machine addresses have _PAGE_IO
included in the PTE, so use PAGE_KERNEL_IO_NOCACHE for these mappings.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/pgtable_64.h | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index 076052c..3038ef6 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -160,7 +160,7 @@ extern void cleanup_highmap(void);
#define pgtable_cache_init() do { } while (0)
#define check_pgt_cache() do { } while (0)
-#define PAGE_AGP PAGE_KERNEL_NOCACHE
+#define PAGE_AGP PAGE_KERNEL_IO_NOCACHE
#define HAVE_PAGE_AGP 1
/* fs/proc/kcore.c */
--
1.7.3.2
From 4cd35860df16d7d3bd55ce4d4500bfe59c46a849 Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Wed, 24 Feb 2010 11:10:45 -0800
Subject: [PATCH 04/21] agp: use DMA API when compiled for Xen as well
Xen guests need translation between pseudo-physical and real machine
physical addresses when accessing graphics devices, so use the DMA API
in that case too.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/char/agp/intel-gtt.c | 6 +++++-
1 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index 75e0a34..ce63e5c 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -20,8 +20,12 @@
* an Intel IOMMU. So make the correct use of the PCI DMA API contingent
* on the Intel IOMMU support (CONFIG_DMAR).
* Only newer chipsets need to bother with this, of course.
+ *
+ * Xen guests accessing graphics hardware also need proper translation
+ * between pseudo-physical addresses and real machine addresses, which
+ * is also achieved by using the DMA API.
*/
-#ifdef CONFIG_DMAR
+#if defined(CONFIG_DMAR) || defined(CONFIG_XEN)
#define USE_PCI_DMA_API 1
#else
#define USE_PCI_DMA_API 0
--
1.7.3.2
From 31b643ec05a644705d75d6d047917710b414487e Mon Sep 17 00:00:00 2001
From: Ian Campbell <ian.campbell@citrix.com>
Date: Thu, 30 Sep 2010 12:37:26 +0100
Subject: [PATCH 05/21] xen: implement XENMEM_machphys_mapping
This hypercall allows Xen to specify a non-default location for the
machine to physical mapping. This capability is used when running a 32
bit domain 0 on a 64 bit hypervisor to shrink the hypervisor hole to
exactly the size required.
[ Impact: add Xen hypercall definitions ]
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
arch/x86/include/asm/xen/interface.h | 6 +++---
arch/x86/include/asm/xen/interface_32.h | 5 +++++
arch/x86/include/asm/xen/interface_64.h | 13 +------------
arch/x86/include/asm/xen/page.h | 7 ++++---
arch/x86/xen/enlighten.c | 7 +++++++
arch/x86/xen/mmu.c | 14 ++++++++++++++
include/xen/interface/memory.h | 13 +++++++++++++
7 files changed, 47 insertions(+), 18 deletions(-)
diff --git a/arch/x86/include/asm/xen/interface.h b/arch/x86/include/asm/xen/interface.h
index e8506c1..1c10c88 100644
--- a/arch/x86/include/asm/xen/interface.h
+++ b/arch/x86/include/asm/xen/interface.h
@@ -61,9 +61,9 @@ DEFINE_GUEST_HANDLE(void);
#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START)
#endif
-#ifndef machine_to_phys_mapping
-#define machine_to_phys_mapping ((unsigned long *)HYPERVISOR_VIRT_START)
-#endif
+#define MACH2PHYS_VIRT_START mk_unsigned_long(__MACH2PHYS_VIRT_START)
+#define MACH2PHYS_VIRT_END mk_unsigned_long(__MACH2PHYS_VIRT_END)
+#define MACH2PHYS_NR_ENTRIES ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>__MACH2PHYS_SHIFT)
/* Maximum number of virtual CPUs in multi-processor guests. */
#define MAX_VIRT_CPUS 32
diff --git a/arch/x86/include/asm/xen/interface_32.h b/arch/x86/include/asm/xen/interface_32.h
index 42a7e00..8413688 100644
--- a/arch/x86/include/asm/xen/interface_32.h
+++ b/arch/x86/include/asm/xen/interface_32.h
@@ -32,6 +32,11 @@
/* And the trap vector is... */
#define TRAP_INSTR "int $0x82"
+#define __MACH2PHYS_VIRT_START 0xF5800000
+#define __MACH2PHYS_VIRT_END 0xF6800000
+
+#define __MACH2PHYS_SHIFT 2
+
/*
* Virtual addresses beyond this are not modifiable by guest OSes. The
* machine->physical mapping table starts at this address, read-only.
diff --git a/arch/x86/include/asm/xen/interface_64.h b/arch/x86/include/asm/xen/interface_64.h
index 100d266..839a481 100644
--- a/arch/x86/include/asm/xen/interface_64.h
+++ b/arch/x86/include/asm/xen/interface_64.h
@@ -39,18 +39,7 @@
#define __HYPERVISOR_VIRT_END 0xFFFF880000000000
#define __MACH2PHYS_VIRT_START 0xFFFF800000000000
#define __MACH2PHYS_VIRT_END 0xFFFF804000000000
-
-#ifndef HYPERVISOR_VIRT_START
-#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START)
-#define HYPERVISOR_VIRT_END mk_unsigned_long(__HYPERVISOR_VIRT_END)
-#endif
-
-#define MACH2PHYS_VIRT_START mk_unsigned_long(__MACH2PHYS_VIRT_START)
-#define MACH2PHYS_VIRT_END mk_unsigned_long(__MACH2PHYS_VIRT_END)
-#define MACH2PHYS_NR_ENTRIES ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>3)
-#ifndef machine_to_phys_mapping
-#define machine_to_phys_mapping ((unsigned long *)HYPERVISOR_VIRT_START)
-#endif
+#define __MACH2PHYS_SHIFT 3
/*
* int HYPERVISOR_set_segment_base(unsigned int which, unsigned long base)
diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h
index bf5f7d3..d5fb2e8 100644
--- a/arch/x86/include/asm/xen/page.h
+++ b/arch/x86/include/asm/xen/page.h
@@ -5,6 +5,7 @@
#include <linux/types.h>
#include <linux/spinlock.h>
#include <linux/pfn.h>
+#include <linux/mm.h>
#include <asm/uaccess.h>
#include <asm/page.h>
@@ -35,6 +36,8 @@ typedef struct xpaddr {
#define MAX_DOMAIN_PAGES \
((unsigned long)((u64)CONFIG_XEN_MAX_DOMAIN_MEMORY * 1024 * 1024 * 1024 / PAGE_SIZE))
+extern unsigned long *machine_to_phys_mapping;
+extern unsigned int machine_to_phys_order;
extern unsigned long get_phys_to_machine(unsigned long pfn);
extern void set_phys_to_machine(unsigned long pfn, unsigned long mfn);
@@ -62,10 +65,8 @@ static inline unsigned long mfn_to_pfn(unsigned long mfn)
if (xen_feature(XENFEAT_auto_translated_physmap))
return mfn;
-#if 0
if (unlikely((mfn >> machine_to_phys_order) != 0))
- return max_mapnr;
-#endif
+ return ~0;
pfn = 0;
/*
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 7d46c84..adfc2f4 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -74,6 +74,11 @@ DEFINE_PER_CPU(struct vcpu_info, xen_vcpu_info);
enum xen_domain_type xen_domain_type = XEN_NATIVE;
EXPORT_SYMBOL_GPL(xen_domain_type);
+unsigned long *machine_to_phys_mapping = (void *)MACH2PHYS_VIRT_START;
+EXPORT_SYMBOL(machine_to_phys_mapping);
+unsigned int machine_to_phys_order;
+EXPORT_SYMBOL(machine_to_phys_order);
+
struct start_info *xen_start_info;
EXPORT_SYMBOL_GPL(xen_start_info);
@@ -1098,6 +1103,8 @@ asmlinkage void __init xen_start_kernel(void)
xen_domain_type = XEN_PV_DOMAIN;
+ xen_setup_machphys_mapping();
+
/* Install Xen paravirt ops */
pv_info = xen_info;
pv_init_ops = xen_init_ops;
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index 42086ac..9625f2a 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -1737,6 +1737,20 @@ static __init void xen_map_identity_early(pmd_t *pmd, unsigned long max_pfn)
set_page_prot(pmd, PAGE_KERNEL_RO);
}
+void __init xen_setup_machphys_mapping(void)
+{
+ struct xen_machphys_mapping mapping;
+ unsigned long machine_to_phys_nr_ents;
+
+ if (HYPERVISOR_memory_op(XENMEM_machphys_mapping, &mapping) == 0) {
+ machine_to_phys_mapping = (unsigned long *)mapping.v_start;
+ machine_to_phys_nr_ents = mapping.max_mfn + 1;
+ } else {
+ machine_to_phys_nr_ents = MACH2PHYS_NR_ENTRIES;
+ }
+ machine_to_phys_order = fls(machine_to_phys_nr_ents - 1);
+}
+
#ifdef CONFIG_X86_64
static void convert_pfn_mfn(void *v)
{
diff --git a/include/xen/interface/memory.h b/include/xen/interface/memory.h
index d3938d3..134d33d 100644
--- a/include/xen/interface/memory.h
+++ b/include/xen/interface/memory.h
@@ -141,6 +141,19 @@ struct xen_machphys_mfn_list {
DEFINE_GUEST_HANDLE_STRUCT(xen_machphys_mfn_list);
/*
+ * Returns the location in virtual address space of the machine_to_phys
+ * mapping table. Architectures which do not have a m2p table, or which do not
+ * map it by default into guest address space, do not implement this command.
+ * arg == addr of xen_machphys_mapping_t.
+ */
+#define XENMEM_machphys_mapping 12
+struct xen_machphys_mapping {
+ unsigned long v_start, v_end; /* Start and end virtual addresses. */
+ unsigned long max_mfn; /* Maximum MFN that can be looked up. */
+};
+DEFINE_GUEST_HANDLE_STRUCT(xen_machphys_mapping_t);
+
+/*
* Sets the GPFN at which a particular page appears in the specified guest's
* pseudophysical address space.
* arg == addr of xen_add_to_physmap_t.
--
1.7.3.2
From 4ffb5de0f02bad2b0ebe7a9a6a45bc5855c15369 Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Mon, 9 Aug 2010 14:35:36 -0700
Subject: [PATCH 06/21] pvops: make pte_flags() go via pvops
As part of PAT support in Xen we need to fiddle with the page flags.
For consistency, we need to make sure the conversion happens both ways
when converting from kernel<->Xen, so pte_flags() must go via pvops.
Unfortunately this undermines the original rationale for introducing
pte_flags (which avoids a pvop call for a common operation in the
mm code)...
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/pgtable.h | 5 +++++
arch/x86/include/asm/pgtable_types.h | 5 -----
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 7198bcf..ce5b845 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -78,6 +78,11 @@ extern struct list_head pgd_list;
#endif /* CONFIG_PARAVIRT */
+static inline pteval_t pte_flags(pte_t pte)
+{
+ return pte_val(pte) & PTE_FLAGS_MASK;
+}
+
/*
* The following only work if pte_present() is true.
* Undefined behaviour if not..
diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/pgtable_types.h
index d1f4a76..a81b0ed 100644
--- a/arch/x86/include/asm/pgtable_types.h
+++ b/arch/x86/include/asm/pgtable_types.h
@@ -265,11 +265,6 @@ static inline pteval_t native_pte_val(pte_t pte)
return pte.pte;
}
-static inline pteval_t pte_flags(pte_t pte)
-{
- return native_pte_val(pte) & PTE_FLAGS_MASK;
-}
-
#define pgprot_val(x) ((x).pgprot)
#define __pgprot(x) ((pgprot_t) { (x) } )
--
1.7.3.2
From 313e74412105c670ff8900ec8099a3a5df1fa83c Mon Sep 17 00:00:00 2001
From: Vasiliy Kulikov <segooon@gmail.com>
Date: Thu, 28 Oct 2010 15:39:02 +0400
Subject: [PATCH 07/21] xen: xenfs: privcmd: check put_user() return code
put_user() may fail. In this case propagate error code from
privcmd_ioctl_mmap_batch().
Signed-off-by: Vasiliy Kulikov <segooon@gmail.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/xen/xenfs/privcmd.c | 8 ++------
1 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/drivers/xen/xenfs/privcmd.c b/drivers/xen/xenfs/privcmd.c
index f80be7f..2eb04c8 100644
--- a/drivers/xen/xenfs/privcmd.c
+++ b/drivers/xen/xenfs/privcmd.c
@@ -266,9 +266,7 @@ static int mmap_return_errors(void *data, void *state)
xen_pfn_t *mfnp = data;
struct mmap_batch_state *st = state;
- put_user(*mfnp, st->user++);
-
- return 0;
+ return put_user(*mfnp, st->user++);
}
static struct vm_operations_struct privcmd_vm_ops;
@@ -323,10 +321,8 @@ static long privcmd_ioctl_mmap_batch(void __user *udata)
up_write(&mm->mmap_sem);
if (state.err > 0) {
- ret = 0;
-
state.user = m.arr;
- traverse_pages(m.num, sizeof(xen_pfn_t),
+ ret = traverse_pages(m.num, sizeof(xen_pfn_t),
&pagelist,
mmap_return_errors, &state);
}
--
1.7.3.2
From a2d771c036eb8c040683089ca04c36dfb93a0e60 Mon Sep 17 00:00:00 2001
From: Ian Campbell <ian.campbell@citrix.com>
Date: Fri, 29 Oct 2010 16:56:19 +0100
Subject: [PATCH 08/21] xen: correct size of level2_kernel_pgt
sizeof(pmd_t *) is 4 bytes on 32-bit PAE leading to an allocation of
only 2048 bytes. The correct size is sizeof(pmd_t) giving us a full
page allocation.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/xen/mmu.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index c237b81..21ed8d7 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -2126,7 +2126,7 @@ __init pgd_t *xen_setup_kernel_pagetable(pgd_t *pgd,
{
pmd_t *kernel_pmd;
- level2_kernel_pgt = extend_brk(sizeof(pmd_t *) * PTRS_PER_PMD, PAGE_SIZE);
+ level2_kernel_pgt = extend_brk(sizeof(pmd_t) * PTRS_PER_PMD, PAGE_SIZE);
max_pfn_mapped = PFN_DOWN(__pa(xen_start_info->pt_base) +
xen_start_info->nr_pt_frames * PAGE_SIZE +
--
1.7.3.2
From c64e38ea17a81721da0393584fd807f8434050fa Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Mon, 1 Nov 2010 14:32:27 -0400
Subject: [PATCH 09/21] xen/blkfront: map REQ_FLUSH into a full barrier
Implement a flush as a full barrier, since we have nothing weaker.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Acked-by: Christoph Hellwig <hch@lst.de>
---
drivers/block/xen-blkfront.c | 19 +++++--------------
1 files changed, 5 insertions(+), 14 deletions(-)
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 06e2812..3a318d8 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -245,14 +245,11 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
}
/*
- * blkif_queue_request
+ * Generate a Xen blkfront IO request from a blk layer request. Reads
+ * and writes are handled as expected. Since we lack a loose flush
+ * request, we map flushes into a full ordered barrier.
*
- * request block io
- *
- * id: for guest use only.
- * operation: BLKIF_OP_{READ,WRITE,PROBE}
- * buffer: buffer to read/write into. this should be a
- * virtual address in the guest os.
+ * @req: a request struct
*/
static int blkif_queue_request(struct request *req)
{
@@ -289,7 +286,7 @@ static int blkif_queue_request(struct request *req)
ring_req->operation = rq_data_dir(req) ?
BLKIF_OP_WRITE : BLKIF_OP_READ;
- if (req->cmd_flags & REQ_HARDBARRIER)
+ if (req->cmd_flags & REQ_FLUSH)
ring_req->operation = BLKIF_OP_WRITE_BARRIER;
ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg);
@@ -1069,14 +1066,8 @@ static void blkfront_connect(struct blkfront_info *info)
*/
info->feature_flush = 0;
- /*
- * The driver doesn't properly handled empty flushes, so
- * lets disable barrier support for now.
- */
-#if 0
if (!err && barrier)
info->feature_flush = REQ_FLUSH;
-#endif
err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
if (err) {
--
1.7.3.2
From a945b9801a9bfd4a98bcfd9f6656b5027b254e3f Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Mon, 1 Nov 2010 17:03:14 -0400
Subject: [PATCH 10/21] xen/blkfront: change blk_shadow.request to proper pointer
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/block/xen-blkfront.c | 14 ++++++--------
1 files changed, 6 insertions(+), 8 deletions(-)
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 3a318d8..31c8a64 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -65,7 +65,7 @@ enum blkif_state {
struct blk_shadow {
struct blkif_request req;
- unsigned long request;
+ struct request *request;
unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST];
};
@@ -136,7 +136,7 @@ static void add_id_to_freelist(struct blkfront_info *info,
unsigned long id)
{
info->shadow[id].req.id = info->shadow_free;
- info->shadow[id].request = 0;
+ info->shadow[id].request = NULL;
info->shadow_free = id;
}
@@ -278,7 +278,7 @@ static int blkif_queue_request(struct request *req)
/* Fill out a communications ring structure. */
ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
id = get_id_from_freelist(info);
- info->shadow[id].request = (unsigned long)req;
+ info->shadow[id].request = req;
ring_req->id = id;
ring_req->sector_number = (blkif_sector_t)blk_rq_pos(req);
@@ -633,7 +633,7 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
bret = RING_GET_RESPONSE(&info->ring, i);
id = bret->id;
- req = (struct request *)info->shadow[id].request;
+ req = info->shadow[id].request;
blkif_completion(&info->shadow[id]);
@@ -898,7 +898,7 @@ static int blkif_recover(struct blkfront_info *info)
/* Stage 3: Find pending requests and requeue them. */
for (i = 0; i < BLK_RING_SIZE; i++) {
/* Not in use? */
- if (copy[i].request == 0)
+ if (!copy[i].request)
continue;
/* Grab a request slot and copy shadow state into it. */
@@ -915,9 +915,7 @@ static int blkif_recover(struct blkfront_info *info)
req->seg[j].gref,
info->xbdev->otherend_id,
pfn_to_mfn(info->shadow[req->id].frame[j]),
- rq_data_dir(
- (struct request *)
- info->shadow[req->id].request));
+ rq_data_dir(info->shadow[req->id].request));
info->shadow[req->id].req = *req;
info->ring.req_prod_pvt++;
--
1.7.3.2
From be2f8373c188ed1f5d36003c9928e4d695213080 Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Tue, 2 Nov 2010 10:38:33 -0400
Subject: [PATCH 11/21] xen/blkfront: Implement FUA with BLKIF_OP_WRITE_BARRIER
The BLKIF_OP_WRITE_BARRIER is a full ordered barrier, so we can use it
to implement FUA as well as a plain FLUSH.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Acked-by: Christoph Hellwig <hch@lst.de>
---
drivers/block/xen-blkfront.c | 14 ++++++++++++--
1 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 31c8a64..76b874a 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -286,8 +286,18 @@ static int blkif_queue_request(struct request *req)
ring_req->operation = rq_data_dir(req) ?
BLKIF_OP_WRITE : BLKIF_OP_READ;
- if (req->cmd_flags & REQ_FLUSH)
+
+ if (req->cmd_flags & (REQ_FLUSH | REQ_FUA)) {
+ /*
+ * Ideally we could just do an unordered
+ * flush-to-disk, but all we have is a full write
+ * barrier at the moment. However, a barrier write is
+ * a superset of FUA, so we can implement it the same
+ * way. (It's also a FLUSH+FUA, since it is
+ * guaranteed ordered WRT previous writes.)
+ */
ring_req->operation = BLKIF_OP_WRITE_BARRIER;
+ }
ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg);
BUG_ON(ring_req->nr_segments > BLKIF_MAX_SEGMENTS_PER_REQUEST);
@@ -1065,7 +1075,7 @@ static void blkfront_connect(struct blkfront_info *info)
info->feature_flush = 0;
if (!err && barrier)
- info->feature_flush = REQ_FLUSH;
+ info->feature_flush = REQ_FLUSH | REQ_FUA;
err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
if (err) {
--
1.7.3.2
From dcb8baeceaa1c629bbd06f472cea023ad08a0c33 Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Tue, 2 Nov 2010 11:55:58 -0400
Subject: [PATCH 12/21] xen/blkfront: cope with backend that fail empty BLKIF_OP_WRITE_BARRIER requests
Some(?) Xen block backends fail BLKIF_OP_WRITE_BARRIER requests, which
Linux uses as a cache flush operation. In that case, disable use
of FLUSH.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Daniel Stodden <daniel.stodden@citrix.com>
---
drivers/block/xen-blkfront.c | 10 ++++++++++
1 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 76b874a..4f9e22f 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -656,6 +656,16 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
printk(KERN_WARNING "blkfront: %s: write barrier op failed\n",
info->gd->disk_name);
error = -EOPNOTSUPP;
+ }
+ if (unlikely(bret->status == BLKIF_RSP_ERROR &&
+ info->shadow[id].req.nr_segments == 0)) {
+ printk(KERN_WARNING "blkfront: %s: empty write barrier op failed\n",
+ info->gd->disk_name);
+ error = -EOPNOTSUPP;
+ }
+ if (unlikely(error)) {
+ if (error == -EOPNOTSUPP)
+ error = 0;
info->feature_flush = 0;
xlvbd_flush(info);
}
--
1.7.3.2
From aed8ff456bd7847683776e5c4d0dd4e4abc5087e Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Wed, 10 Nov 2010 12:28:57 -0800
Subject: [PATCH 13/21] x86: demacro set_iopl_mask()
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/processor.h | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index cae9c3c..32f75be 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -591,7 +591,10 @@ static inline void load_sp0(struct tss_struct *tss,
native_load_sp0(tss, thread);
}
-#define set_iopl_mask native_set_iopl_mask
+static inline void set_iopl_mask(unsigned mask)
+{
+ native_set_iopl_mask(mask);
+}
#endif /* CONFIG_PARAVIRT */
/*
--
1.7.3.2
From 640524e029f5b5cb93462c7bccdd93141f53ae65 Mon Sep 17 00:00:00 2001
From: Christophe Saout <chtephan@leto.intern.saout.de>
Date: Sat, 17 Jan 2009 17:30:17 +0100
Subject: [PATCH 14/21] x86/paravirt: paravirtualize IO permission bitmap
Paravirtualized x86 systems don't have an exposed TSS, as it is only
directly visible in ring 0. The IO permission bitmap is part of
the TSS, so with out a TSS, it must be paravirtualized separately,
like the iopl mask.
[ Impact: make ioperm bitmap work under Xen ]
Signed-off-by: Christophe Saout <chtephan@leto.intern.saout.de>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/include/asm/paravirt.h | 7 +++++++
arch/x86/include/asm/paravirt_types.h | 2 ++
arch/x86/include/asm/processor.h | 9 +++++++++
arch/x86/kernel/ioport.c | 29 ++++++++++++++++++++++-------
arch/x86/kernel/paravirt.c | 1 +
arch/x86/kernel/process.c | 28 ++++++++--------------------
6 files changed, 49 insertions(+), 27 deletions(-)
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index 18e3b8a..1764f4a 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -330,11 +330,18 @@ static inline void write_idt_entry(gate_desc *dt, int entry, const gate_desc *g)
{
PVOP_VCALL3(pv_cpu_ops.write_idt_entry, dt, entry, g);
}
+
static inline void set_iopl_mask(unsigned mask)
{
PVOP_VCALL1(pv_cpu_ops.set_iopl_mask, mask);
}
+static inline void set_io_bitmap(struct thread_struct *thread,
+ unsigned long bytes_updated)
+{
+ PVOP_VCALL2(pv_cpu_ops.set_io_bitmap, thread, bytes_updated);
+}
+
/* The paravirtualized I/O functions */
static inline void slow_down_io(void)
{
diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index b82bac9..dcb0c44 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -135,6 +135,8 @@ struct pv_cpu_ops {
void (*load_sp0)(struct tss_struct *tss, struct thread_struct *t);
void (*set_iopl_mask)(unsigned mask);
+ void (*set_io_bitmap)(struct thread_struct *thread,
+ unsigned long bytes_updated);
void (*wbinvd)(void);
void (*io_delay)(void);
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index 32f75be..9c5528e 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -551,6 +551,9 @@ static inline void native_set_iopl_mask(unsigned mask)
#endif
}
+extern void native_set_io_bitmap(struct thread_struct *thread,
+ unsigned long updated_bytes);
+
static inline void
native_load_sp0(struct tss_struct *tss, struct thread_struct *thread)
{
@@ -595,6 +598,12 @@ static inline void set_iopl_mask(unsigned mask)
{
native_set_iopl_mask(mask);
}
+
+static inline void set_io_bitmap(struct thread_struct *thread,
+ unsigned long updated_bytes)
+{
+ native_set_io_bitmap(thread, updated_bytes);
+}
#endif /* CONFIG_PARAVIRT */
/*
diff --git a/arch/x86/kernel/ioport.c b/arch/x86/kernel/ioport.c
index 8eec0ec..8ead1f0 100644
--- a/arch/x86/kernel/ioport.c
+++ b/arch/x86/kernel/ioport.c
@@ -30,13 +30,29 @@ static void set_bitmap(unsigned long *bitmap, unsigned int base,
}
}
+void native_set_io_bitmap(struct thread_struct *t,
+ unsigned long bytes_updated)
+{
+ struct tss_struct *tss;
+
+ if (!bytes_updated)
+ return;
+
+ tss = &__get_cpu_var(init_tss);
+
+ /* Update the TSS: */
+ if (t->io_bitmap_ptr)
+ memcpy(tss->io_bitmap, t->io_bitmap_ptr, bytes_updated);
+ else
+ memset(tss->io_bitmap, 0xff, bytes_updated);
+}
+
/*
* this changes the io permissions bitmap in the current task.
*/
asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on)
{
struct thread_struct *t = ¤t->thread;
- struct tss_struct *tss;
unsigned int i, max_long, bytes, bytes_updated;
if ((from + num <= from) || (from + num > IO_BITMAP_BITS))
@@ -61,13 +77,13 @@ asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on)
}
/*
- * do it in the per-thread copy and in the TSS ...
+ * do it in the per-thread copy
*
- * Disable preemption via get_cpu() - we must not switch away
+ * Disable preemption - we must not switch away
* because the ->io_bitmap_max value must match the bitmap
* contents:
*/
- tss = &per_cpu(init_tss, get_cpu());
+ preempt_disable();
set_bitmap(t->io_bitmap_ptr, from, num, !turn_on);
@@ -85,10 +101,9 @@ asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on)
t->io_bitmap_max = bytes;
- /* Update the TSS: */
- memcpy(tss->io_bitmap, t->io_bitmap_ptr, bytes_updated);
+ set_io_bitmap(t, bytes_updated);
- put_cpu();
+ preempt_enable();
return 0;
}
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index c5b2500..febd851 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -376,6 +376,7 @@ struct pv_cpu_ops pv_cpu_ops = {
.swapgs = native_swapgs,
.set_iopl_mask = native_set_iopl_mask,
+ .set_io_bitmap = native_set_io_bitmap,
.io_delay = native_io_delay,
.start_context_switch = paravirt_nop,
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 57d1868..a48e82a 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -74,16 +74,12 @@ void exit_thread(void)
unsigned long *bp = t->io_bitmap_ptr;
if (bp) {
- struct tss_struct *tss = &per_cpu(init_tss, get_cpu());
-
+ preempt_disable();
t->io_bitmap_ptr = NULL;
clear_thread_flag(TIF_IO_BITMAP);
- /*
- * Careful, clear this in the TSS too:
- */
- memset(tss->io_bitmap, 0xff, t->io_bitmap_max);
+ set_io_bitmap(t, t->io_bitmap_max);
t->io_bitmap_max = 0;
- put_cpu();
+ preempt_enable();
kfree(bp);
}
}
@@ -214,19 +210,11 @@ void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p,
hard_enable_TSC();
}
- if (test_tsk_thread_flag(next_p, TIF_IO_BITMAP)) {
- /*
- * Copy the relevant range of the IO bitmap.
- * Normally this is 128 bytes or less:
- */
- memcpy(tss->io_bitmap, next->io_bitmap_ptr,
- max(prev->io_bitmap_max, next->io_bitmap_max));
- } else if (test_tsk_thread_flag(prev_p, TIF_IO_BITMAP)) {
- /*
- * Clear any possible leftover bits:
- */
- memset(tss->io_bitmap, 0xff, prev->io_bitmap_max);
- }
+ if (test_tsk_thread_flag(next_p, TIF_IO_BITMAP) ||
+ test_tsk_thread_flag(prev_p, TIF_IO_BITMAP))
+ set_io_bitmap(next,
+ max(prev->io_bitmap_max, next->io_bitmap_max));
+
propagate_user_return_notify(prev_p, next_p);
}
--
1.7.3.2
From 950952701cf9e218c2269d13b8538f3c07cff762 Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Thu, 18 Jun 2009 15:04:16 -0700
Subject: [PATCH 15/21] xen: implement IO permission bitmap
Add Xen implementation of IO permission bitmap pvop.
[ Impact: allow guests to set usermode IO permission bitmaps. ]
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/xen/enlighten.c | 13 +++++++++++++
1 files changed, 13 insertions(+), 0 deletions(-)
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 235c0f4..4ad88fd 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -700,6 +700,18 @@ static void xen_set_iopl_mask(unsigned mask)
HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl);
}
+static void xen_set_io_bitmap(struct thread_struct *thread,
+ unsigned long bytes_updated)
+{
+ struct physdev_set_iobitmap set_iobitmap;
+
+ set_xen_guest_handle(set_iobitmap.bitmap,
+ (char *)thread->io_bitmap_ptr);
+ set_iobitmap.nr_ports = thread->io_bitmap_ptr ? IO_BITMAP_BITS : 0;
+ WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iobitmap,
+ &set_iobitmap));
+}
+
static void xen_io_delay(void)
{
}
@@ -997,6 +1009,7 @@ static const struct pv_cpu_ops xen_cpu_ops __initdata = {
.load_sp0 = xen_load_sp0,
.set_iopl_mask = xen_set_iopl_mask,
+ .set_io_bitmap = xen_set_io_bitmap,
.io_delay = xen_io_delay,
/* Xen takes care of %gs when switching to usermode for us */
--
1.7.3.2
From 6903591f314b8947d0e362bda7715e90eb9df75e Mon Sep 17 00:00:00 2001
From: Ian Campbell <ian.campbell@citrix.com>
Date: Mon, 1 Nov 2010 16:30:09 +0000
Subject: [PATCH 16/21] xen: events: do not unmask event channels on resume
The IRQ core code will take care of disabling and reenabling
interrupts over suspend resume automatically, therefore we do not need
to do this in the Xen event channel code.
The only exception is those event channels marked IRQF_NO_SUSPEND
which the IRQ core ignores. We must unmask these ourselves, taking
care to obey the current IRQ_DISABLED status. Failure check for
IRQ_DISABLED leads to enabling polled only event channels, such as
that associated with the pv spinlocks, which must never be enabled:
[ 21.970432] ------------[ cut here ]------------
[ 21.970432] kernel BUG at arch/x86/xen/spinlock.c:343!
[ 21.970432] invalid opcode: 0000 [#1] SMP
[ 21.970432] last sysfs file: /sys/devices/virtual/net/lo/operstate
[ 21.970432] Modules linked in:
[ 21.970432]
[ 21.970432] Pid: 0, comm: swapper Not tainted (2.6.32.24-x86_32p-xen-01034-g787c727 #34)
[ 21.970432] EIP: 0061:[<c102e209>] EFLAGS: 00010046 CPU: 3
[ 21.970432] EIP is at dummy_handler+0x3/0x7
[ 21.970432] EAX: 0000021c EBX: dfc16880 ECX: 0000001a EDX: 00000000
[ 21.970432] ESI: dfc02c00 EDI: 00000001 EBP: dfc47e10 ESP: dfc47e10
[ 21.970432] DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0069
[ 21.970432] Process swapper (pid: 0, ti=dfc46000 task=dfc39440 task.ti=dfc46000)
[ 21.970432] Stack:
[ 21.970432] dfc47e30 c10a39f0 0000021c 00000000 00000000 dfc16880 0000021c 00000001
[ 21.970432] <0> dfc47e40 c10a4f08 0000021c 00000000 dfc47e78 c12240a7 c1839284 c1839284
[ 21.970432] <0> 00000200 00000000 00000000 f5720000 c1f3d028 c1f3d02c 00000180 dfc47e90
[ 21.970432] Call Trace:
[ 21.970432] [<c10a39f0>] ? handle_IRQ_event+0x5f/0x122
[ 21.970432] [<c10a4f08>] ? handle_percpu_irq+0x2f/0x55
[ 21.970432] [<c12240a7>] ? __xen_evtchn_do_upcall+0xdb/0x15f
[ 21.970432] [<c122481e>] ? xen_evtchn_do_upcall+0x20/0x30
[ 21.970432] [<c1030d47>] ? xen_do_upcall+0x7/0xc
[ 21.970432] [<c102007b>] ? apic_reg_read+0xd3/0x22d
[ 21.970432] [<c1002227>] ? hypercall_page+0x227/0x1005
[ 21.970432] [<c102d30b>] ? xen_force_evtchn_callback+0xf/0x14
[ 21.970432] [<c102da7c>] ? check_events+0x8/0xc
[ 21.970432] [<c102da3b>] ? xen_irq_enable_direct_end+0x0/0x1
[ 21.970432] [<c105e485>] ? finish_task_switch+0x62/0xba
[ 21.970432] [<c14e3f84>] ? schedule+0x808/0x89d
[ 21.970432] [<c1084dc5>] ? hrtimer_start_expires+0x1a/0x22
[ 21.970432] [<c1085154>] ? tick_nohz_restart_sched_tick+0x15a/0x162
[ 21.970432] [<c102f43a>] ? cpu_idle+0x6d/0x6f
[ 21.970432] [<c14db29e>] ? cpu_bringup_and_idle+0xd/0xf
[ 21.970432] Code: 5d 0f 95 c0 0f b6 c0 c3 55 66 83 78 02 00 89 e5 5d 0f 95 \
c0 0f b6 c0 c3 55 b2 01 86 10 31 c0 84 d2 89 e5 0f 94 c0 5d c3 55 89 e5 <0f> 0b \
eb fe 55 80 3d 4c ce 84 c1 00 89 e5 57 56 89 c6 53 74 15
[ 21.970432] EIP: [<c102e209>] dummy_handler+0x3/0x7 SS:ESP 0069:dfc47e10
[ 21.970432] ---[ end trace c0b71f7e12cf3011 ]---
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/xen/events.c | 25 ++++++++++++++++++-------
1 files changed, 18 insertions(+), 7 deletions(-)
diff --git a/drivers/xen/events.c b/drivers/xen/events.c
index 97612f5..321a0c8 100644
--- a/drivers/xen/events.c
+++ b/drivers/xen/events.c
@@ -1299,9 +1299,6 @@ static void restore_cpu_virqs(unsigned int cpu)
evtchn_to_irq[evtchn] = irq;
irq_info[irq] = mk_virq_info(evtchn, virq);
bind_evtchn_to_cpu(evtchn, cpu);
-
- /* Ready for use. */
- unmask_evtchn(evtchn);
}
}
@@ -1327,10 +1324,6 @@ static void restore_cpu_ipis(unsigned int cpu)
evtchn_to_irq[evtchn] = irq;
irq_info[irq] = mk_ipi_info(evtchn, ipi);
bind_evtchn_to_cpu(evtchn, cpu);
-
- /* Ready for use. */
- unmask_evtchn(evtchn);
-
}
}
@@ -1390,6 +1383,7 @@ void xen_poll_irq(int irq)
void xen_irq_resume(void)
{
unsigned int cpu, irq, evtchn;
+ struct irq_desc *desc;
init_evtchn_cpu_bindings();
@@ -1408,6 +1402,23 @@ void xen_irq_resume(void)
restore_cpu_virqs(cpu);
restore_cpu_ipis(cpu);
}
+
+ /*
+ * Unmask any IRQF_NO_SUSPEND IRQs which are enabled. These
+ * are not handled by the IRQ core.
+ */
+ for_each_irq_desc(irq, desc) {
+ if (!desc->action || !(desc->action->flags & IRQF_NO_SUSPEND))
+ continue;
+ if (desc->status & IRQ_DISABLED)
+ continue;
+
+ evtchn = evtchn_from_irq(irq);
+ if (evtchn == -1)
+ continue;
+
+ unmask_evtchn(evtchn);
+ }
}
static struct irq_chip xen_dynamic_chip __read_mostly = {
--
1.7.3.2
From 9ec23a7f6d2537faf14368e066e307c06812c4ca Mon Sep 17 00:00:00 2001
From: Ian Campbell <ian.campbell@citrix.com>
Date: Thu, 28 Oct 2010 11:32:29 -0700
Subject: [PATCH 17/21] xen: do not release any memory under 1M in domain 0
We already deliberately setup a 1-1 P2M for the region up to 1M in
order to allow code which assumes this region is already mapped to
work without having to convert everything to ioremap.
Domain 0 should not return any apparently unused memory regions
(reserved or otherwise) in this region to Xen since the e820 may not
accurately reflect what the BIOS has stashed in this region.
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/xen/setup.c | 18 +++++++++++-------
1 files changed, 11 insertions(+), 7 deletions(-)
diff --git a/arch/x86/xen/setup.c b/arch/x86/xen/setup.c
index b1dbdaa..769c4b0 100644
--- a/arch/x86/xen/setup.c
+++ b/arch/x86/xen/setup.c
@@ -118,16 +118,18 @@ static unsigned long __init xen_return_unused_memory(unsigned long max_pfn,
const struct e820map *e820)
{
phys_addr_t max_addr = PFN_PHYS(max_pfn);
- phys_addr_t last_end = 0;
+ phys_addr_t last_end = ISA_END_ADDRESS;
unsigned long released = 0;
int i;
+ /* Free any unused memory above the low 1Mbyte. */
for (i = 0; i < e820->nr_map && last_end < max_addr; i++) {
phys_addr_t end = e820->map[i].addr;
end = min(max_addr, end);
- released += xen_release_chunk(last_end, end);
- last_end = e820->map[i].addr + e820->map[i].size;
+ if (last_end < end)
+ released += xen_release_chunk(last_end, end);
+ last_end = max(last_end, e820->map[i].addr + e820->map[i].size);
}
if (last_end < max_addr)
@@ -164,6 +166,7 @@ char * __init xen_memory_setup(void)
XENMEM_memory_map;
rc = HYPERVISOR_memory_op(op, &memmap);
if (rc == -ENOSYS) {
+ BUG_ON(xen_initial_domain());
memmap.nr_entries = 1;
map[0].addr = 0ULL;
map[0].size = mem_end;
@@ -201,12 +204,13 @@ char * __init xen_memory_setup(void)
}
/*
- * Even though this is normal, usable memory under Xen, reserve
- * ISA memory anyway because too many things think they can poke
+ * In domU, the ISA region is normal, usable memory, but we
+ * reserve ISA memory anyway because too many things poke
* about in there.
*
- * In a dom0 kernel, this region is identity mapped with the
- * hardware ISA area, so it really is out of bounds.
+ * In Dom0, the host E820 information can leave gaps in the
+ * ISA range, which would cause us to release those pages. To
+ * avoid this, we unconditionally reserve them here.
*/
e820_add_region(ISA_START_ADDRESS, ISA_END_ADDRESS - ISA_START_ADDRESS,
E820_RESERVED);
--
1.7.3.2
From e060e7af98182494b764d002eba7fa022fe91bdf Mon Sep 17 00:00:00 2001
From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Date: Thu, 11 Nov 2010 12:37:43 -0800
Subject: [PATCH 18/21] xen: set vma flag VM_PFNMAP in the privcmd mmap file_op
Set VM_PFNMAP in the privcmd mmap file_op, rather than later in
xen_remap_domain_mfn_range when it is too late because
vma_wants_writenotify has already been called and vm_page_prot has
already been modified.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
arch/x86/xen/mmu.c | 3 ++-
drivers/xen/xenfs/privcmd.c | 5 +++--
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index f08ea04..792de43 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -2299,7 +2299,8 @@ int xen_remap_domain_mfn_range(struct vm_area_struct *vma,
prot = __pgprot(pgprot_val(prot) | _PAGE_IOMAP);
- vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;
+ BUG_ON(!((vma->vm_flags & (VM_PFNMAP | VM_RESERVED | VM_IO)) ==
+ (VM_PFNMAP | VM_RESERVED | VM_IO)));
rmd.mfn = mfn;
rmd.prot = prot;
diff --git a/drivers/xen/xenfs/privcmd.c b/drivers/xen/xenfs/privcmd.c
index 2eb04c8..88474d4 100644
--- a/drivers/xen/xenfs/privcmd.c
+++ b/drivers/xen/xenfs/privcmd.c
@@ -380,8 +380,9 @@ static int privcmd_mmap(struct file *file, struct vm_area_struct *vma)
if (xen_feature(XENFEAT_auto_translated_physmap))
return -ENOSYS;
- /* DONTCOPY is essential for Xen as copy_page_range is broken. */
- vma->vm_flags |= VM_RESERVED | VM_IO | VM_DONTCOPY;
+ /* DONTCOPY is essential for Xen because copy_page_range doesn't know
+ * how to recreate these mappings */
+ vma->vm_flags |= VM_RESERVED | VM_IO | VM_DONTCOPY | VM_PFNMAP;
vma->vm_ops = &privcmd_vm_ops;
vma->vm_private_data = NULL;
--
1.7.3.2
From a188301f0e78daed011dde56139630d88299a954 Mon Sep 17 00:00:00 2001
From: Ian Campbell <ian.campbell@citrix.com>
Date: Mon, 9 Feb 2009 12:05:49 -0800
Subject: [PATCH 19/21] xen: define gnttab_set_map_op/unmap_op
Impact: hypercall definitions
These functions populate the gnttab data structures used by the
granttab map and unmap ops and are used in the backend drivers.
Originally xen-unstable.hg 9625:c3bb51c443a7
[ Include Stefano's fix for phys_addr_t ]
Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
include/xen/grant_table.h | 39 ++++++++++++++++++++++++++++++++++++++-
1 files changed, 38 insertions(+), 1 deletions(-)
diff --git a/include/xen/grant_table.h b/include/xen/grant_table.h
index 9a73170..1821aa1 100644
--- a/include/xen/grant_table.h
+++ b/include/xen/grant_table.h
@@ -37,10 +37,16 @@
#ifndef __ASM_GNTTAB_H__
#define __ASM_GNTTAB_H__
-#include <asm/xen/hypervisor.h>
+#include <asm/page.h>
+
+#include <xen/interface/xen.h>
#include <xen/interface/grant_table.h>
+
+#include <asm/xen/hypervisor.h>
#include <asm/xen/grant_table.h>
+#include <xen/features.h>
+
/* NR_GRANT_FRAMES must be less than or equal to that configured in Xen */
#define NR_GRANT_FRAMES 4
@@ -107,6 +113,37 @@ void gnttab_grant_foreign_access_ref(grant_ref_t ref, domid_t domid,
void gnttab_grant_foreign_transfer_ref(grant_ref_t, domid_t domid,
unsigned long pfn);
+static inline void
+gnttab_set_map_op(struct gnttab_map_grant_ref *map, phys_addr_t addr,
+ uint32_t flags, grant_ref_t ref, domid_t domid)
+{
+ if (flags & GNTMAP_contains_pte)
+ map->host_addr = addr;
+ else if (xen_feature(XENFEAT_auto_translated_physmap))
+ map->host_addr = __pa(addr);
+ else
+ map->host_addr = addr;
+
+ map->flags = flags;
+ map->ref = ref;
+ map->dom = domid;
+}
+
+static inline void
+gnttab_set_unmap_op(struct gnttab_unmap_grant_ref *unmap, phys_addr_t addr,
+ uint32_t flags, grant_handle_t handle)
+{
+ if (flags & GNTMAP_contains_pte)
+ unmap->host_addr = addr;
+ else if (xen_feature(XENFEAT_auto_translated_physmap))
+ unmap->host_addr = __pa(addr);
+ else
+ unmap->host_addr = addr;
+
+ unmap->handle = handle;
+ unmap->dev_bus_addr = 0;
+}
+
int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes,
unsigned long max_nr_gframes,
struct grant_entry **__shared);
--
1.7.3.2
From 56385560d6d8fd4c89c4f328d3ff0ecc9c44c52d Mon Sep 17 00:00:00 2001
From: Gerd Hoffmann <kraxel@redhat.com>
Date: Tue, 3 Mar 2009 12:27:55 -0800
Subject: [PATCH 20/21] xen/gntdev: allow usermode to map granted pages
The gntdev driver allows usermode to map granted pages from other
domains. This is typically used to implement a Xen backend driver
in user mode.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Stefano Stabellini <Stefano.Stabellini@eu.citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/xen/Kconfig | 7 +
drivers/xen/Makefile | 4 +
drivers/xen/gntdev.c | 646 ++++++++++++++++++++++++++++++++++++++++++++++++++
include/xen/gntdev.h | 119 +++++++++
4 files changed, 776 insertions(+), 0 deletions(-)
create mode 100644 drivers/xen/gntdev.c
create mode 100644 include/xen/gntdev.h
diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig
index 6e6180c..0c6d2a1 100644
--- a/drivers/xen/Kconfig
+++ b/drivers/xen/Kconfig
@@ -62,6 +62,13 @@ config XEN_SYS_HYPERVISOR
virtual environment, /sys/hypervisor will still be present,
but will have no xen contents.
+config XEN_GNTDEV
+ tristate "userspace grant access device driver"
+ depends on XEN
+ select MMU_NOTIFIER
+ help
+ Allows userspace processes use grants.
+
config XEN_PLATFORM_PCI
tristate "xen platform pci device driver"
depends on XEN_PVHVM
diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
index eb8a78d..7ed8418 100644
--- a/drivers/xen/Makefile
+++ b/drivers/xen/Makefile
@@ -9,8 +9,12 @@ obj-$(CONFIG_HOTPLUG_CPU) += cpu_hotplug.o
obj-$(CONFIG_XEN_XENCOMM) += xencomm.o
obj-$(CONFIG_XEN_BALLOON) += balloon.o
obj-$(CONFIG_XEN_DEV_EVTCHN) += evtchn.o
+obj-$(CONFIG_XEN_GNTDEV) += xen-gntdev.o
obj-$(CONFIG_XENFS) += xenfs/
obj-$(CONFIG_XEN_SYS_HYPERVISOR) += sys-hypervisor.o
obj-$(CONFIG_XEN_PLATFORM_PCI) += platform-pci.o
obj-$(CONFIG_SWIOTLB_XEN) += swiotlb-xen.o
obj-$(CONFIG_XEN_DOM0) += pci.o
+
+xen-gntdev-y := gntdev.o
+
diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
new file mode 100644
index 0000000..45898d4
--- /dev/null
+++ b/drivers/xen/gntdev.c
@@ -0,0 +1,646 @@
+/******************************************************************************
+ * gntdev.c
+ *
+ * Device for accessing (in user-space) pages that have been granted by other
+ * domains.
+ *
+ * Copyright (c) 2006-2007, D G Murray.
+ * (c) 2009 Gerd Hoffmann <kraxel@redhat.com>
+ *
+ * This program is distributed in the hope that 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, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/miscdevice.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/mmu_notifier.h>
+#include <linux/types.h>
+#include <linux/uaccess.h>
+#include <linux/sched.h>
+#include <linux/spinlock.h>
+#include <linux/slab.h>
+
+#include <xen/xen.h>
+#include <xen/grant_table.h>
+#include <xen/gntdev.h>
+#include <asm/xen/hypervisor.h>
+#include <asm/xen/hypercall.h>
+#include <asm/xen/page.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Derek G. Murray <Derek.Murray@cl.cam.ac.uk>, "
+ "Gerd Hoffmann <kraxel@redhat.com>");
+MODULE_DESCRIPTION("User-space granted page access driver");
+
+static int debug = 0;
+module_param(debug, int, 0644);
+static int limit = 1024;
+module_param(limit, int, 0644);
+
+struct gntdev_priv {
+ struct list_head maps;
+ uint32_t used;
+ uint32_t limit;
+ spinlock_t lock;
+ struct mm_struct *mm;
+ struct mmu_notifier mn;
+};
+
+struct grant_map {
+ struct list_head next;
+ struct gntdev_priv *priv;
+ struct vm_area_struct *vma;
+ int index;
+ int count;
+ int flags;
+ int is_mapped;
+ struct ioctl_gntdev_grant_ref *grants;
+ struct gnttab_map_grant_ref *map_ops;
+ struct gnttab_unmap_grant_ref *unmap_ops;
+};
+
+/* ------------------------------------------------------------------ */
+
+static void gntdev_print_maps(struct gntdev_priv *priv,
+ char *text, int text_index)
+{
+ struct grant_map *map;
+
+ printk("%s: maps list (priv %p, usage %d/%d)\n",
+ __FUNCTION__, priv, priv->used, priv->limit);
+ list_for_each_entry(map, &priv->maps, next)
+ printk(" index %2d, count %2d %s\n",
+ map->index, map->count,
+ map->index == text_index && text ? text : "");
+}
+
+static struct grant_map *gntdev_alloc_map(struct gntdev_priv *priv, int count)
+{
+ struct grant_map *add;
+
+ add = kzalloc(sizeof(struct grant_map), GFP_KERNEL);
+ if (NULL == add)
+ return NULL;
+
+ add->grants = kzalloc(sizeof(add->grants[0]) * count, GFP_KERNEL);
+ add->map_ops = kzalloc(sizeof(add->map_ops[0]) * count, GFP_KERNEL);
+ add->unmap_ops = kzalloc(sizeof(add->unmap_ops[0]) * count, GFP_KERNEL);
+ if (NULL == add->grants ||
+ NULL == add->map_ops ||
+ NULL == add->unmap_ops)
+ goto err;
+
+ add->index = 0;
+ add->count = count;
+ add->priv = priv;
+
+ if (add->count + priv->used > priv->limit)
+ goto err;
+
+ return add;
+
+err:
+ kfree(add->grants);
+ kfree(add->map_ops);
+ kfree(add->unmap_ops);
+ kfree(add);
+ return NULL;
+}
+
+static void gntdev_add_map(struct gntdev_priv *priv, struct grant_map *add)
+{
+ struct grant_map *map;
+
+ list_for_each_entry(map, &priv->maps, next) {
+ if (add->index + add->count < map->index) {
+ list_add_tail(&add->next, &map->next);
+ goto done;
+ }
+ add->index = map->index + map->count;
+ }
+ list_add_tail(&add->next, &priv->maps);
+
+done:
+ priv->used += add->count;
+ if (debug)
+ gntdev_print_maps(priv, "[new]", add->index);
+}
+
+static struct grant_map *gntdev_find_map_index(struct gntdev_priv *priv, int index,
+ int count)
+{
+ struct grant_map *map;
+
+ list_for_each_entry(map, &priv->maps, next) {
+ if (map->index != index)
+ continue;
+ if (map->count != count)
+ continue;
+ return map;
+ }
+ return NULL;
+}
+
+static struct grant_map *gntdev_find_map_vaddr(struct gntdev_priv *priv,
+ unsigned long vaddr)
+{
+ struct grant_map *map;
+
+ list_for_each_entry(map, &priv->maps, next) {
+ if (!map->vma)
+ continue;
+ if (vaddr < map->vma->vm_start)
+ continue;
+ if (vaddr >= map->vma->vm_end)
+ continue;
+ return map;
+ }
+ return NULL;
+}
+
+static int gntdev_del_map(struct grant_map *map)
+{
+ int i;
+
+ if (map->vma)
+ return -EBUSY;
+ for (i = 0; i < map->count; i++)
+ if (map->unmap_ops[i].handle)
+ return -EBUSY;
+
+ map->priv->used -= map->count;
+ list_del(&map->next);
+ return 0;
+}
+
+static void gntdev_free_map(struct grant_map *map)
+{
+ if (!map)
+ return;
+ kfree(map->grants);
+ kfree(map->map_ops);
+ kfree(map->unmap_ops);
+ kfree(map);
+}
+
+/* ------------------------------------------------------------------ */
+
+static int find_grant_ptes(pte_t *pte, pgtable_t token, unsigned long addr, void *data)
+{
+ struct grant_map *map = data;
+ unsigned int pgnr = (addr - map->vma->vm_start) >> PAGE_SHIFT;
+ u64 pte_maddr;
+
+ BUG_ON(pgnr >= map->count);
+ pte_maddr = (u64)pfn_to_mfn(page_to_pfn(token)) << PAGE_SHIFT;
+ pte_maddr += (unsigned long)pte & ~PAGE_MASK;
+ gnttab_set_map_op(&map->map_ops[pgnr], pte_maddr, map->flags,
+ map->grants[pgnr].ref,
+ map->grants[pgnr].domid);
+ gnttab_set_unmap_op(&map->unmap_ops[pgnr], pte_maddr, map->flags,
+ 0 /* handle */);
+ return 0;
+}
+
+static int map_grant_pages(struct grant_map *map)
+{
+ int i, err = 0;
+
+ if (debug)
+ printk("%s: map %d+%d\n", __FUNCTION__, map->index, map->count);
+ err = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref,
+ map->map_ops, map->count);
+ if (WARN_ON(err))
+ return err;
+
+ for (i = 0; i < map->count; i++) {
+ if (map->map_ops[i].status)
+ err = -EINVAL;
+ map->unmap_ops[i].handle = map->map_ops[i].handle;
+ }
+ return err;
+}
+
+static int unmap_grant_pages(struct grant_map *map, int offset, int pages)
+{
+ int i, err = 0;
+
+ if (debug)
+ printk("%s: map %d+%d [%d+%d]\n", __FUNCTION__,
+ map->index, map->count, offset, pages);
+ err = HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref,
+ map->unmap_ops + offset, pages);
+ if (WARN_ON(err))
+ return err;
+
+ for (i = 0; i < pages; i++) {
+ if (map->unmap_ops[offset+i].status)
+ err = -EINVAL;
+ map->unmap_ops[offset+i].handle = 0;
+ }
+ return err;
+}
+
+/* ------------------------------------------------------------------ */
+
+static void gntdev_vma_close(struct vm_area_struct *vma)
+{
+ struct grant_map *map = vma->vm_private_data;
+
+ if (debug)
+ printk("%s\n", __FUNCTION__);
+ map->is_mapped = 0;
+ map->vma = NULL;
+ vma->vm_private_data = NULL;
+}
+
+static int gntdev_vma_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+ if (debug)
+ printk("%s: vaddr %p, pgoff %ld (shouldn't happen)\n",
+ __FUNCTION__, vmf->virtual_address, vmf->pgoff);
+ vmf->flags = VM_FAULT_ERROR;
+ return 0;
+}
+
+static struct vm_operations_struct gntdev_vmops = {
+ .close = gntdev_vma_close,
+ .fault = gntdev_vma_fault,
+};
+
+/* ------------------------------------------------------------------ */
+
+static void mn_invl_range_start(struct mmu_notifier *mn,
+ struct mm_struct *mm,
+ unsigned long start, unsigned long end)
+{
+ struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
+ struct grant_map *map;
+ unsigned long mstart, mend;
+ int err;
+
+ spin_lock(&priv->lock);
+ list_for_each_entry(map, &priv->maps, next) {
+ if (!map->vma)
+ continue;
+ if (!map->is_mapped)
+ continue;
+ if (map->vma->vm_start >= end)
+ continue;
+ if (map->vma->vm_end <= start)
+ continue;
+ mstart = max(start, map->vma->vm_start);
+ mend = min(end, map->vma->vm_end);
+ if (debug)
+ printk("%s: map %d+%d (%lx %lx), range %lx %lx, mrange %lx %lx\n",
+ __FUNCTION__, map->index, map->count,
+ map->vma->vm_start, map->vma->vm_end,
+ start, end, mstart, mend);
+ err = unmap_grant_pages(map,
+ (mstart - map->vma->vm_start) >> PAGE_SHIFT,
+ (mend - mstart) >> PAGE_SHIFT);
+ WARN_ON(err);
+ }
+ spin_unlock(&priv->lock);
+}
+
+static void mn_invl_page(struct mmu_notifier *mn,
+ struct mm_struct *mm,
+ unsigned long address)
+{
+ mn_invl_range_start(mn, mm, address, address + PAGE_SIZE);
+}
+
+static void mn_release(struct mmu_notifier *mn,
+ struct mm_struct *mm)
+{
+ struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
+ struct grant_map *map;
+ int err;
+
+ spin_lock(&priv->lock);
+ list_for_each_entry(map, &priv->maps, next) {
+ if (!map->vma)
+ continue;
+ if (debug)
+ printk("%s: map %d+%d (%lx %lx)\n",
+ __FUNCTION__, map->index, map->count,
+ map->vma->vm_start, map->vma->vm_end);
+ err = unmap_grant_pages(map, 0, map->count);
+ WARN_ON(err);
+ }
+ spin_unlock(&priv->lock);
+}
+
+struct mmu_notifier_ops gntdev_mmu_ops = {
+ .release = mn_release,
+ .invalidate_page = mn_invl_page,
+ .invalidate_range_start = mn_invl_range_start,
+};
+
+/* ------------------------------------------------------------------ */
+
+static int gntdev_open(struct inode *inode, struct file *flip)
+{
+ struct gntdev_priv *priv;
+
+ priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+ if (!priv)
+ return -ENOMEM;
+
+ INIT_LIST_HEAD(&priv->maps);
+ spin_lock_init(&priv->lock);
+ priv->limit = limit;
+
+ priv->mm = get_task_mm(current);
+ if (!priv->mm) {
+ kfree(priv);
+ return -ENOMEM;
+ }
+ priv->mn.ops = &gntdev_mmu_ops;
+ mmu_notifier_register(&priv->mn, priv->mm);
+ mmput(priv->mm);
+
+ flip->private_data = priv;
+ if (debug)
+ printk("%s: priv %p\n", __FUNCTION__, priv);
+
+ return 0;
+}
+
+static int gntdev_release(struct inode *inode, struct file *flip)
+{
+ struct gntdev_priv *priv = flip->private_data;
+ struct grant_map *map;
+ int err;
+
+ if (debug)
+ printk("%s: priv %p\n", __FUNCTION__, priv);
+
+ spin_lock(&priv->lock);
+ while (!list_empty(&priv->maps)) {
+ map = list_entry(priv->maps.next, struct grant_map, next);
+ err = gntdev_del_map(map);
+ if (WARN_ON(err))
+ gntdev_free_map(map);
+
+ }
+ spin_unlock(&priv->lock);
+
+ mmu_notifier_unregister(&priv->mn, priv->mm);
+ kfree(priv);
+ return 0;
+}
+
+static long gntdev_ioctl_map_grant_ref(struct gntdev_priv *priv,
+ struct ioctl_gntdev_map_grant_ref __user *u)
+{
+ struct ioctl_gntdev_map_grant_ref op;
+ struct grant_map *map;
+ int err;
+
+ if (copy_from_user(&op, u, sizeof(op)) != 0)
+ return -EFAULT;
+ if (debug)
+ printk("%s: priv %p, add %d\n", __FUNCTION__, priv,
+ op.count);
+ if (unlikely(op.count <= 0))
+ return -EINVAL;
+ if (unlikely(op.count > priv->limit))
+ return -EINVAL;
+
+ err = -ENOMEM;
+ map = gntdev_alloc_map(priv, op.count);
+ if (!map)
+ return err;
+ if (copy_from_user(map->grants, &u->refs,
+ sizeof(map->grants[0]) * op.count) != 0) {
+ gntdev_free_map(map);
+ return err;
+ }
+
+ spin_lock(&priv->lock);
+ gntdev_add_map(priv, map);
+ op.index = map->index << PAGE_SHIFT;
+ spin_unlock(&priv->lock);
+
+ if (copy_to_user(u, &op, sizeof(op)) != 0) {
+ spin_lock(&priv->lock);
+ gntdev_del_map(map);
+ spin_unlock(&priv->lock);
+ gntdev_free_map(map);
+ return err;
+ }
+ return 0;
+}
+
+static long gntdev_ioctl_unmap_grant_ref(struct gntdev_priv *priv,
+ struct ioctl_gntdev_unmap_grant_ref __user *u)
+{
+ struct ioctl_gntdev_unmap_grant_ref op;
+ struct grant_map *map;
+ int err = -EINVAL;
+
+ if (copy_from_user(&op, u, sizeof(op)) != 0)
+ return -EFAULT;
+ if (debug)
+ printk("%s: priv %p, del %d+%d\n", __FUNCTION__, priv,
+ (int)op.index, (int)op.count);
+
+ spin_lock(&priv->lock);
+ map = gntdev_find_map_index(priv, op.index >> PAGE_SHIFT, op.count);
+ if (map)
+ err = gntdev_del_map(map);
+ spin_unlock(&priv->lock);
+ if (!err)
+ gntdev_free_map(map);
+ return err;
+}
+
+static long gntdev_ioctl_get_offset_for_vaddr(struct gntdev_priv *priv,
+ struct ioctl_gntdev_get_offset_for_vaddr __user *u)
+{
+ struct ioctl_gntdev_get_offset_for_vaddr op;
+ struct grant_map *map;
+
+ if (copy_from_user(&op, u, sizeof(op)) != 0)
+ return -EFAULT;
+ if (debug)
+ printk("%s: priv %p, offset for vaddr %lx\n", __FUNCTION__, priv,
+ (unsigned long)op.vaddr);
+
+ spin_lock(&priv->lock);
+ map = gntdev_find_map_vaddr(priv, op.vaddr);
+ if (map == NULL ||
+ map->vma->vm_start != op.vaddr) {
+ spin_unlock(&priv->lock);
+ return -EINVAL;
+ }
+ op.offset = map->index << PAGE_SHIFT;
+ op.count = map->count;
+ spin_unlock(&priv->lock);
+
+ if (copy_to_user(u, &op, sizeof(op)) != 0)
+ return -EFAULT;
+ return 0;
+}
+
+static long gntdev_ioctl_set_max_grants(struct gntdev_priv *priv,
+ struct ioctl_gntdev_set_max_grants __user *u)
+{
+ struct ioctl_gntdev_set_max_grants op;
+
+ if (copy_from_user(&op, u, sizeof(op)) != 0)
+ return -EFAULT;
+ if (debug)
+ printk("%s: priv %p, limit %d\n", __FUNCTION__, priv, op.count);
+ if (op.count > limit)
+ return -EINVAL;
+
+ spin_lock(&priv->lock);
+ priv->limit = op.count;
+ spin_unlock(&priv->lock);
+ return 0;
+}
+
+static long gntdev_ioctl(struct file *flip,
+ unsigned int cmd, unsigned long arg)
+{
+ struct gntdev_priv *priv = flip->private_data;
+ void __user *ptr = (void __user *)arg;
+
+ switch (cmd) {
+ case IOCTL_GNTDEV_MAP_GRANT_REF:
+ return gntdev_ioctl_map_grant_ref(priv, ptr);
+
+ case IOCTL_GNTDEV_UNMAP_GRANT_REF:
+ return gntdev_ioctl_unmap_grant_ref(priv, ptr);
+
+ case IOCTL_GNTDEV_GET_OFFSET_FOR_VADDR:
+ return gntdev_ioctl_get_offset_for_vaddr(priv, ptr);
+
+ case IOCTL_GNTDEV_SET_MAX_GRANTS:
+ return gntdev_ioctl_set_max_grants(priv, ptr);
+
+ default:
+ if (debug)
+ printk("%s: priv %p, unknown cmd %x\n",
+ __FUNCTION__, priv, cmd);
+ return -ENOIOCTLCMD;
+ }
+
+ return 0;
+}
+
+static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma)
+{
+ struct gntdev_priv *priv = flip->private_data;
+ int index = vma->vm_pgoff;
+ int count = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
+ struct grant_map *map;
+ int err = -EINVAL;
+
+ if ((vma->vm_flags & VM_WRITE) && !(vma->vm_flags & VM_SHARED))
+ return -EINVAL;
+
+ if (debug)
+ printk("%s: map %d+%d at %lx (pgoff %lx)\n", __FUNCTION__,
+ index, count, vma->vm_start, vma->vm_pgoff);
+
+ spin_lock(&priv->lock);
+ map = gntdev_find_map_index(priv, index, count);
+ if (!map)
+ goto unlock_out;
+ if (map->vma)
+ goto unlock_out;
+ if (priv->mm != vma->vm_mm) {
+ printk("%s: Huh? Other mm?\n", __FUNCTION__);
+ goto unlock_out;
+ }
+
+ vma->vm_ops = &gntdev_vmops;
+
+ vma->vm_flags |= VM_RESERVED;
+ vma->vm_flags |= VM_DONTCOPY;
+ vma->vm_flags |= VM_DONTEXPAND;
+
+ vma->vm_private_data = map;
+ map->vma = vma;
+
+ map->flags = GNTMAP_host_map | GNTMAP_application_map | GNTMAP_contains_pte;
+ if (!(vma->vm_flags & VM_WRITE))
+ map->flags |= GNTMAP_readonly;
+
+ err = apply_to_page_range(vma->vm_mm, vma->vm_start,
+ vma->vm_end - vma->vm_start,
+ find_grant_ptes, map);
+ if (err) {
+ goto unlock_out;
+ if (debug)
+ printk("%s: find_grant_ptes() failure.\n", __FUNCTION__);
+ }
+
+ err = map_grant_pages(map);
+ if (err) {
+ goto unlock_out;
+ if (debug)
+ printk("%s: map_grant_pages() failure.\n", __FUNCTION__);
+ }
+ map->is_mapped = 1;
+
+unlock_out:
+ spin_unlock(&priv->lock);
+ return err;
+}
+
+static const struct file_operations gntdev_fops = {
+ .owner = THIS_MODULE,
+ .open = gntdev_open,
+ .release = gntdev_release,
+ .mmap = gntdev_mmap,
+ .unlocked_ioctl = gntdev_ioctl
+};
+
+static struct miscdevice gntdev_miscdev = {
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = "xen/gntdev",
+ .fops = &gntdev_fops,
+};
+
+/* ------------------------------------------------------------------ */
+
+static int __init gntdev_init(void)
+{
+ int err;
+
+ if (!xen_domain())
+ return -ENODEV;
+
+ err = misc_register(&gntdev_miscdev);
+ if (err != 0) {
+ printk(KERN_ERR "Could not register gntdev device\n");
+ return err;
+ }
+ return 0;
+}
+
+static void __exit gntdev_exit(void)
+{
+ misc_deregister(&gntdev_miscdev);
+}
+
+module_init(gntdev_init);
+module_exit(gntdev_exit);
+
+/* ------------------------------------------------------------------ */
diff --git a/include/xen/gntdev.h b/include/xen/gntdev.h
new file mode 100644
index 0000000..8bd1467
--- /dev/null
+++ b/include/xen/gntdev.h
@@ -0,0 +1,119 @@
+/******************************************************************************
+ * gntdev.h
+ *
+ * Interface to /dev/xen/gntdev.
+ *
+ * Copyright (c) 2007, D G Murray
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __LINUX_PUBLIC_GNTDEV_H__
+#define __LINUX_PUBLIC_GNTDEV_H__
+
+struct ioctl_gntdev_grant_ref {
+ /* The domain ID of the grant to be mapped. */
+ uint32_t domid;
+ /* The grant reference of the grant to be mapped. */
+ uint32_t ref;
+};
+
+/*
+ * Inserts the grant references into the mapping table of an instance
+ * of gntdev. N.B. This does not perform the mapping, which is deferred
+ * until mmap() is called with @index as the offset.
+ */
+#define IOCTL_GNTDEV_MAP_GRANT_REF \
+_IOC(_IOC_NONE, 'G', 0, sizeof(struct ioctl_gntdev_map_grant_ref))
+struct ioctl_gntdev_map_grant_ref {
+ /* IN parameters */
+ /* The number of grants to be mapped. */
+ uint32_t count;
+ uint32_t pad;
+ /* OUT parameters */
+ /* The offset to be used on a subsequent call to mmap(). */
+ uint64_t index;
+ /* Variable IN parameter. */
+ /* Array of grant references, of size @count. */
+ struct ioctl_gntdev_grant_ref refs[1];
+};
+
+/*
+ * Removes the grant references from the mapping table of an instance of
+ * of gntdev. N.B. munmap() must be called on the relevant virtual address(es)
+ * before this ioctl is called, or an error will result.
+ */
+#define IOCTL_GNTDEV_UNMAP_GRANT_REF \
+_IOC(_IOC_NONE, 'G', 1, sizeof(struct ioctl_gntdev_unmap_grant_ref))
+struct ioctl_gntdev_unmap_grant_ref {
+ /* IN parameters */
+ /* The offset was returned by the corresponding map operation. */
+ uint64_t index;
+ /* The number of pages to be unmapped. */
+ uint32_t count;
+ uint32_t pad;
+};
+
+/*
+ * Returns the offset in the driver's address space that corresponds
+ * to @vaddr. This can be used to perform a munmap(), followed by an
+ * UNMAP_GRANT_REF ioctl, where no state about the offset is retained by
+ * the caller. The number of pages that were allocated at the same time as
+ * @vaddr is returned in @count.
+ *
+ * N.B. Where more than one page has been mapped into a contiguous range, the
+ * supplied @vaddr must correspond to the start of the range; otherwise
+ * an error will result. It is only possible to munmap() the entire
+ * contiguously-allocated range at once, and not any subrange thereof.
+ */
+#define IOCTL_GNTDEV_GET_OFFSET_FOR_VADDR \
+_IOC(_IOC_NONE, 'G', 2, sizeof(struct ioctl_gntdev_get_offset_for_vaddr))
+struct ioctl_gntdev_get_offset_for_vaddr {
+ /* IN parameters */
+ /* The virtual address of the first mapped page in a range. */
+ uint64_t vaddr;
+ /* OUT parameters */
+ /* The offset that was used in the initial mmap() operation. */
+ uint64_t offset;
+ /* The number of pages mapped in the VM area that begins at @vaddr. */
+ uint32_t count;
+ uint32_t pad;
+};
+
+/*
+ * Sets the maximum number of grants that may mapped at once by this gntdev
+ * instance.
+ *
+ * N.B. This must be called before any other ioctl is performed on the device.
+ */
+#define IOCTL_GNTDEV_SET_MAX_GRANTS \
+_IOC(_IOC_NONE, 'G', 3, sizeof(struct ioctl_gntdev_set_max_grants))
+struct ioctl_gntdev_set_max_grants {
+ /* IN parameter */
+ /* The maximum number of grants that may be mapped at once. */
+ uint32_t count;
+};
+
+#endif /* __LINUX_PUBLIC_GNTDEV_H__ */
--
1.7.3.2
From 8c79aad3bd8a379bd0bd144895b55098aca2ccea Mon Sep 17 00:00:00 2001
From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Date: Thu, 11 Nov 2010 14:39:12 -0800
Subject: [PATCH 21/21] xen/gntdev: add VM_PFNMAP to vma
These pages are from other domains, so don't have any local PFN.
VM_PFNMAP is the closest concept Linux has to this.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
---
drivers/xen/gntdev.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
index 45898d4..cf61c7d 100644
--- a/drivers/xen/gntdev.c
+++ b/drivers/xen/gntdev.c
@@ -574,6 +574,7 @@ static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma)
vma->vm_flags |= VM_RESERVED;
vma->vm_flags |= VM_DONTCOPY;
vma->vm_flags |= VM_DONTEXPAND;
+ vma->vm_flags |= VM_PFNMAP;
vma->vm_private_data = map;
map->vma = vma;
--
1.7.3.2
[-- Attachment #3: xen.pcifront.fixes.patch --]
[-- Type: application/octet-stream, Size: 5894 bytes --]
From 1688c3d6000b1183bcb604c8c85f742a579990e5 Mon Sep 17 00:00:00 2001
From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Date: Sun, 31 Oct 2010 17:04:09 -0400
Subject: [PATCH 1/5] MAINTAINERS: Update mailing list name for Xen pieces.
While the 'xen-devel@lists.xen.org' is more apt, it is not yet
ready. Revert the name back to the old lists.xensource.com for
right now.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
---
MAINTAINERS | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/MAINTAINERS b/MAINTAINERS
index 0094224..acf13f2 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -6613,7 +6613,7 @@ F: drivers/xen/*swiotlb*
XEN HYPERVISOR INTERFACE
M: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
-L: xen-devel@lists.xen.org
+L: xen-devel@lists.xensource.com
L: virtualization@lists.osdl.org
S: Supported
F: arch/x86/xen/
--
1.7.3.2
From 07cf2a64c2ad3408a0e12aa4cd6040b30c09381d Mon Sep 17 00:00:00 2001
From: Jiri Slaby <jslaby@suse.cz>
Date: Sat, 6 Nov 2010 10:06:49 +0100
Subject: [PATCH 2/5] xen: fix memory leak in Xen PCI MSI/MSI-X allocator.
Stanse found that xen_setup_msi_irqs leaks memory when
xen_allocate_pirq fails. Free the memory in that fail path.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: xen-devel@lists.xensource.com
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
---
arch/x86/pci/xen.c | 8 +++++---
1 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/arch/x86/pci/xen.c b/arch/x86/pci/xen.c
index 117f5b8..d7b5109 100644
--- a/arch/x86/pci/xen.c
+++ b/arch/x86/pci/xen.c
@@ -147,8 +147,10 @@ static int xen_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
irq = xen_allocate_pirq(v[i], 0, /* not sharable */
(type == PCI_CAP_ID_MSIX) ?
"pcifront-msi-x" : "pcifront-msi");
- if (irq < 0)
- return -1;
+ if (irq < 0) {
+ ret = -1;
+ goto free;
+ }
ret = set_irq_msi(irq, msidesc);
if (ret)
@@ -164,7 +166,7 @@ error:
if (ret == -ENODEV)
dev_err(&dev->dev, "Xen PCI frontend has not registered" \
" MSI/MSI-X support!\n");
-
+free:
kfree(v);
return ret;
}
--
1.7.3.2
From c8ac3902fb7a98c45ed54d98ad6f1c8168f47021 Mon Sep 17 00:00:00 2001
From: Jesper Juhl <jj@chaosbits.net>
Date: Sat, 30 Oct 2010 14:51:30 +0200
Subject: [PATCH 3/5] xen-pcifront: Remove duplicate inclusion of headers.
In drivers/pci/xen-pcifront.c the xen/xenbus.h header is included twice -
once is enough.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
---
drivers/pci/xen-pcifront.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
index a87c498..0579273 100644
--- a/drivers/pci/xen-pcifront.c
+++ b/drivers/pci/xen-pcifront.c
@@ -13,7 +13,6 @@
#include <linux/spinlock.h>
#include <linux/pci.h>
#include <linux/msi.h>
-#include <xen/xenbus.h>
#include <xen/interface/io/pciif.h>
#include <asm/xen/pci.h>
#include <linux/interrupt.h>
--
1.7.3.2
From 2a63dd7275b2278bd7e9203f74b9aa4f07e82a7a Mon Sep 17 00:00:00 2001
From: Jiri Slaby <jslaby@suse.cz>
Date: Thu, 4 Nov 2010 15:31:30 +0100
Subject: [PATCH 4/5] xen-pcifront: fix PCI reference leak
Stanse found that when pdev is found and has no driver a reference is
leaked in pcifront_common_process. So add pci_dev_put there. For the
pdev == NULL case, pci_dev_put(NULL) is fine.
[v2: Updated to not dereference pcidev->dev per Milton's observation]
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Milton Miller <miltonm@bga.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
---
drivers/pci/xen-pcifront.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
index 0579273..3a5a6fc 100644
--- a/drivers/pci/xen-pcifront.c
+++ b/drivers/pci/xen-pcifront.c
@@ -575,8 +575,9 @@ static pci_ers_result_t pcifront_common_process(int cmd,
pcidev = pci_get_bus_and_slot(bus, devfn);
if (!pcidev || !pcidev->driver) {
- dev_err(&pcidev->dev,
- "device or driver is NULL\n");
+ dev_err(&pdev->xdev->dev, "device or AER driver is NULL\n");
+ if (pcidev)
+ pci_dev_put(pcidev);
return result;
}
pdrv = pcidev->driver;
--
1.7.3.2
From b74831e6437c0cbbd310dc587579390a146dc7a0 Mon Sep 17 00:00:00 2001
From: Joe Perches <joe@perches.com>
Date: Wed, 10 Nov 2010 09:47:51 -0800
Subject: [PATCH 5/5] MAINTAINERS: Mark XEN lists as moderated
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
---
MAINTAINERS | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/MAINTAINERS b/MAINTAINERS
index acf13f2..6a16f21 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -6598,14 +6598,14 @@ F: drivers/platform/x86
XEN PCI SUBSYSTEM
M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
-L: xen-devel@lists.xensource.com
+L: xen-devel@lists.xensource.com (moderated for non-subscribers)
S: Supported
F: arch/x86/pci/*xen*
F: drivers/pci/*xen*
XEN SWIOTLB SUBSYSTEM
M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
-L: xen-devel@lists.xensource.com
+L: xen-devel@lists.xensource.com (moderated for non-subscribers)
S: Supported
F: arch/x86/xen/*swiotlb*
F: drivers/xen/*swiotlb*
@@ -6613,7 +6613,7 @@ F: drivers/xen/*swiotlb*
XEN HYPERVISOR INTERFACE
M: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
-L: xen-devel@lists.xensource.com
+L: xen-devel@lists.xensource.com (moderated for non-subscribers)
L: virtualization@lists.osdl.org
S: Supported
F: arch/x86/xen/
--
1.7.3.2
[-- Attachment #4: Type: text/plain, Size: 138 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
next reply other threads:[~2010-11-16 19:20 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-16 19:20 Boris Derzhavets [this message]
-- strict thread matches above, loose matches on Subject: below --
2010-11-18 10:34 Re: 2.6.37-rc1 mainline domU - BUG: unable to handle kernel paging request Boris Derzhavets
2010-11-18 16:40 ` Bruce Edge
2010-11-18 17:05 ` Boris Derzhavets
2010-11-18 20:05 ` Bruce Edge
2010-11-19 7:12 ` Boris Derzhavets
2010-11-19 18:16 ` Bruce Edge
2010-11-19 18:52 ` Boris Derzhavets
[not found] ` <697216.62238.qm@web56104.mail.re3.yahoo.com>
2010-12-01 21:32 ` Bruce Edge
2010-12-02 6:33 ` Jeremy Fitzhardinge
2010-12-02 8:33 ` Boris Derzhavets
2010-12-02 14:41 ` Bruce Edge
2010-12-02 18:05 ` Bruce Edge
2010-12-02 18:28 ` Jeremy Fitzhardinge
2010-12-02 18:38 ` Bruce Edge
2010-12-02 18:50 ` Jeremy Fitzhardinge
2010-12-02 19:34 ` Bruce Edge
2010-11-19 14:32 ` Boris Derzhavets
2010-11-14 17:47 Boris Derzhavets
2010-11-14 17:52 ` Sander Eikelenboom
2010-11-10 22:15 Bruce Edge
2010-11-10 22:30 ` Bruce Edge
2010-11-10 23:03 ` Bruce Edge
2010-11-11 15:56 ` Konrad Rzeszutek Wilk
2010-11-11 16:32 ` Boris Derzhavets
2010-11-11 12:01 ` Boris Derzhavets
2010-11-11 12:08 ` Boris Derzhavets
2010-11-11 16:09 ` Konrad Rzeszutek Wilk
2010-11-11 16:29 ` Boris Derzhavets
2010-11-11 16:46 ` Konrad Rzeszutek Wilk
2010-11-11 16:53 ` Boris Derzhavets
2010-11-12 14:40 ` Bruce Edge
2010-11-12 16:06 ` Boris Derzhavets
2010-11-13 8:37 ` Boris Derzhavets
2010-11-12 16:27 ` Sander Eikelenboom
2010-11-12 17:01 ` Konrad Rzeszutek Wilk
2010-11-14 16:37 ` Boris Derzhavets
2010-11-14 16:56 ` Sander Eikelenboom
2010-11-14 17:09 ` Boris Derzhavets
2010-11-14 17:19 ` Sander Eikelenboom
2010-11-14 21:35 ` Bruce Edge
2010-11-15 8:06 ` Boris Derzhavets
2010-11-15 11:05 ` Boris Derzhavets
2010-11-15 20:21 ` Bruce Edge
2010-11-15 20:32 ` Sander Eikelenboom
2010-11-16 18:43 ` Boris Derzhavets
2010-11-16 19:00 ` Konrad Rzeszutek Wilk
2010-11-16 20:43 ` Boris Derzhavets
2010-11-16 21:15 ` Konrad Rzeszutek Wilk
2010-11-16 21:42 ` Boris Derzhavets
2010-11-16 21:49 ` Boris Derzhavets
2010-11-17 21:28 ` Bruce Edge
2010-11-16 20:50 ` Boris Derzhavets
2010-11-11 16:46 ` Boris Derzhavets
2010-11-11 12:26 ` Boris Derzhavets
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=535480.36725.qm@web56104.mail.re3.yahoo.com \
--to=bderzhavets@yahoo.com \
--cc=bruce.edge@gmail.com \
--cc=jeremy@goop.org \
--cc=konrad.wilk@oracle.com \
--cc=xen-devel@lists.xensource.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.