Patchwork kvmvapic: align start address as well as size

login
register
mail settings
Submitter Avi Kivity
Date March 6, 2012, 3:50 p.m.
Message ID <1331049010-24312-1-git-send-email-avi@redhat.com>
Download mbox | patch
Permalink /patch/144956/
State New
Headers show

Comments

Avi Kivity - March 6, 2012, 3:50 p.m.
The kvmvapic code remaps a section of ROM as RAM to allow the guest to
maintain state there.  It is careful to align the section size to a page
boundary, to avoid creating subpages, but neglects to do the same for
the start address.  These leads to an assert later on when the memory
core tries to create a page which is half RAM and half ROM.

Fix by aligning the start address to a page boundary.

This can be triggered by running qemu-system-x86_64 -enable-kvm -vga none.

Signed-off-by: Avi Kivity <avi@redhat.com>
---
 hw/kvmvapic.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)
Jan Kiszka - March 6, 2012, 4:01 p.m.
On 2012-03-06 16:50, Avi Kivity wrote:
> The kvmvapic code remaps a section of ROM as RAM to allow the guest to
> maintain state there.  It is careful to align the section size to a page
> boundary, to avoid creating subpages, but neglects to do the same for
> the start address.  These leads to an assert later on when the memory
> core tries to create a page which is half RAM and half ROM.
> 
> Fix by aligning the start address to a page boundary.
> 
> This can be triggered by running qemu-system-x86_64 -enable-kvm -vga none.
> 
> Signed-off-by: Avi Kivity <avi@redhat.com>
> ---
>  hw/kvmvapic.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
> 
> diff --git a/hw/kvmvapic.c b/hw/kvmvapic.c
> index 36ccfbc..e8bfeec 100644
> --- a/hw/kvmvapic.c
> +++ b/hw/kvmvapic.c
> @@ -578,8 +578,10 @@ static void vapic_map_rom_writable(VAPICROMState *s)
>      rom_size = ram[rom_paddr + 2] * ROM_BLOCK_SIZE;
>      s->rom_size = rom_size;
>  
> -    /* We need to round up to avoid creating subpages
> +    /* We need to round to avoid creating subpages
>       * from which we cannot run code. */
> +    rom_size += rom_paddr & ~TARGET_PAGE_MASK;
> +    rom_paddr &= TARGET_PAGE_MASK;
>      rom_size = TARGET_PAGE_ALIGN(rom_size);
>  
>      memory_region_init_alias(&s->rom, "kvmvapic-rom", section.mr, rom_paddr,

Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>
Anthony Liguori - March 14, 2012, 8:33 p.m.
On 03/06/2012 09:50 AM, Avi Kivity wrote:
> The kvmvapic code remaps a section of ROM as RAM to allow the guest to
> maintain state there.  It is careful to align the section size to a page
> boundary, to avoid creating subpages, but neglects to do the same for
> the start address.  These leads to an assert later on when the memory
> core tries to create a page which is half RAM and half ROM.
>
> Fix by aligning the start address to a page boundary.
>
> This can be triggered by running qemu-system-x86_64 -enable-kvm -vga none.
>
> Signed-off-by: Avi Kivity<avi@redhat.com>

Tested-by: Anthony Liguori <aliguori@us.ibm.com>

Are you going to do a PULL request or do you want me to apply directly?

Regards,

Anthony Liguori

> ---
>   hw/kvmvapic.c |    4 +++-
>   1 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/hw/kvmvapic.c b/hw/kvmvapic.c
> index 36ccfbc..e8bfeec 100644
> --- a/hw/kvmvapic.c
> +++ b/hw/kvmvapic.c
> @@ -578,8 +578,10 @@ static void vapic_map_rom_writable(VAPICROMState *s)
>       rom_size = ram[rom_paddr + 2] * ROM_BLOCK_SIZE;
>       s->rom_size = rom_size;
>
> -    /* We need to round up to avoid creating subpages
> +    /* We need to round to avoid creating subpages
>        * from which we cannot run code. */
> +    rom_size += rom_paddr&  ~TARGET_PAGE_MASK;
> +    rom_paddr&= TARGET_PAGE_MASK;
>       rom_size = TARGET_PAGE_ALIGN(rom_size);
>
>       memory_region_init_alias(&s->rom, "kvmvapic-rom", section.mr, rom_paddr,
Anthony Liguori - March 14, 2012, 9:18 p.m.
On 03/06/2012 09:50 AM, Avi Kivity wrote:
> The kvmvapic code remaps a section of ROM as RAM to allow the guest to
> maintain state there.  It is careful to align the section size to a page
> boundary, to avoid creating subpages, but neglects to do the same for
> the start address.  These leads to an assert later on when the memory
> core tries to create a page which is half RAM and half ROM.
>
> Fix by aligning the start address to a page boundary.
>
> This can be triggered by running qemu-system-x86_64 -enable-kvm -vga none.
>
> Signed-off-by: Avi Kivity<avi@redhat.com>

Applied.  Thanks.

Regards,

Anthony Liguori

> ---
>   hw/kvmvapic.c |    4 +++-
>   1 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/hw/kvmvapic.c b/hw/kvmvapic.c
> index 36ccfbc..e8bfeec 100644
> --- a/hw/kvmvapic.c
> +++ b/hw/kvmvapic.c
> @@ -578,8 +578,10 @@ static void vapic_map_rom_writable(VAPICROMState *s)
>       rom_size = ram[rom_paddr + 2] * ROM_BLOCK_SIZE;
>       s->rom_size = rom_size;
>
> -    /* We need to round up to avoid creating subpages
> +    /* We need to round to avoid creating subpages
>        * from which we cannot run code. */
> +    rom_size += rom_paddr&  ~TARGET_PAGE_MASK;
> +    rom_paddr&= TARGET_PAGE_MASK;
>       rom_size = TARGET_PAGE_ALIGN(rom_size);
>
>       memory_region_init_alias(&s->rom, "kvmvapic-rom", section.mr, rom_paddr,
Avi Kivity - March 15, 2012, 9:55 a.m.
On 03/14/2012 10:33 PM, Anthony Liguori wrote:
> On 03/06/2012 09:50 AM, Avi Kivity wrote:
>> The kvmvapic code remaps a section of ROM as RAM to allow the guest to
>> maintain state there.  It is careful to align the section size to a page
>> boundary, to avoid creating subpages, but neglects to do the same for
>> the start address.  These leads to an assert later on when the memory
>> core tries to create a page which is half RAM and half ROM.
>>
>> Fix by aligning the start address to a page boundary.
>>
>> This can be triggered by running qemu-system-x86_64 -enable-kvm -vga
>> none.
>>
>> Signed-off-by: Avi Kivity<avi@redhat.com>
>
> Tested-by: Anthony Liguori <aliguori@us.ibm.com>
>
> Are you going to do a PULL request or do you want me to apply directly?

I forgot all about it, sorry.
Anthony Liguori - March 15, 2012, 1:16 p.m.
On 03/15/2012 04:55 AM, Avi Kivity wrote:
> On 03/14/2012 10:33 PM, Anthony Liguori wrote:
>> On 03/06/2012 09:50 AM, Avi Kivity wrote:
>>> The kvmvapic code remaps a section of ROM as RAM to allow the guest to
>>> maintain state there.  It is careful to align the section size to a page
>>> boundary, to avoid creating subpages, but neglects to do the same for
>>> the start address.  These leads to an assert later on when the memory
>>> core tries to create a page which is half RAM and half ROM.
>>>
>>> Fix by aligning the start address to a page boundary.
>>>
>>> This can be triggered by running qemu-system-x86_64 -enable-kvm -vga
>>> none.
>>>
>>> Signed-off-by: Avi Kivity<avi@redhat.com>
>>
>> Tested-by: Anthony Liguori<aliguori@us.ibm.com>
>>
>> Are you going to do a PULL request or do you want me to apply directly?
>
> I forgot all about it, sorry.

No worries.  I actually caught the bug in the original pull request but do to a 
bug in qemu-test, when I tried to bisect, things got weird and I mistakenly 
thought the real problem had been introduced earlier.

Regards,

Anthony Liguori

>

Patch

diff --git a/hw/kvmvapic.c b/hw/kvmvapic.c
index 36ccfbc..e8bfeec 100644
--- a/hw/kvmvapic.c
+++ b/hw/kvmvapic.c
@@ -578,8 +578,10 @@  static void vapic_map_rom_writable(VAPICROMState *s)
     rom_size = ram[rom_paddr + 2] * ROM_BLOCK_SIZE;
     s->rom_size = rom_size;
 
-    /* We need to round up to avoid creating subpages
+    /* We need to round to avoid creating subpages
      * from which we cannot run code. */
+    rom_size += rom_paddr & ~TARGET_PAGE_MASK;
+    rom_paddr &= TARGET_PAGE_MASK;
     rom_size = TARGET_PAGE_ALIGN(rom_size);
 
     memory_region_init_alias(&s->rom, "kvmvapic-rom", section.mr, rom_paddr,