arm64: vdso: Fix ABI regression in compat vdso
diff mbox series

Message ID 20190710130452.44111-1-vincenzo.frascino@arm.com
State New
Headers show
Series
  • arm64: vdso: Fix ABI regression in compat vdso
Related show

Commit Message

Vincenzo Frascino July 10, 2019, 1:04 p.m. UTC
Prior to the introduction of Unified vDSO support and compat layer for
vDSO on arm64, AT_SYSINFO_EHDR was not defined for compat tasks.
In the current implementation, AT_SYSINFO_EHDR is defined even if the
compat vdso layer is not built and this causes a regression in the
expected behavior of the ABI.

Restore the ABI behavior making sure that AT_SYSINFO_EHDR for compat
tasks is defined only when CONFIG_GENERIC_COMPAT_VDSO and
CONFIG_COMPAT_VDSO are enabled.

Reported-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
---
 arch/arm64/include/asm/elf.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Will Deacon July 10, 2019, 1:25 p.m. UTC | #1
On Wed, Jul 10, 2019 at 02:04:52PM +0100, Vincenzo Frascino wrote:
> Prior to the introduction of Unified vDSO support and compat layer for
> vDSO on arm64, AT_SYSINFO_EHDR was not defined for compat tasks.
> In the current implementation, AT_SYSINFO_EHDR is defined even if the
> compat vdso layer is not built and this causes a regression in the
> expected behavior of the ABI.
> 
> Restore the ABI behavior making sure that AT_SYSINFO_EHDR for compat
> tasks is defined only when CONFIG_GENERIC_COMPAT_VDSO and
> CONFIG_COMPAT_VDSO are enabled.

I think you could do a better job in the changelog of explaining what's
actually going on here. The problem seems to be that you're advertising
the presence of a non-existent vDSO to userspace.

> Reported-by: John Stultz <john.stultz@linaro.org>
> Signed-off-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
> ---
>  arch/arm64/include/asm/elf.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
> index 3c7037c6ba9b..b7992bb9d414 100644
> --- a/arch/arm64/include/asm/elf.h
> +++ b/arch/arm64/include/asm/elf.h
> @@ -202,7 +202,7 @@ typedef compat_elf_greg_t		compat_elf_gregset_t[COMPAT_ELF_NGREG];
>  ({									\
>  	set_thread_flag(TIF_32BIT);					\
>   })
> -#ifdef CONFIG_GENERIC_COMPAT_VDSO
> +#if defined(CONFIG_COMPAT_VDSO) && defined(CONFIG_GENERIC_COMPAT_VDSO)

Can't this just be #ifdef CONFIG_COMPAT_VDSO ?

John -- can you give this a whirl, please?

Cheers,

Will
Vincenzo Frascino July 10, 2019, 1:42 p.m. UTC | #2
On 10/07/2019 14:25, Will Deacon wrote:
> On Wed, Jul 10, 2019 at 02:04:52PM +0100, Vincenzo Frascino wrote:
>> Prior to the introduction of Unified vDSO support and compat layer for
>> vDSO on arm64, AT_SYSINFO_EHDR was not defined for compat tasks.
>> In the current implementation, AT_SYSINFO_EHDR is defined even if the
>> compat vdso layer is not built and this causes a regression in the
>> expected behavior of the ABI.
>>
>> Restore the ABI behavior making sure that AT_SYSINFO_EHDR for compat
>> tasks is defined only when CONFIG_GENERIC_COMPAT_VDSO and
>> CONFIG_COMPAT_VDSO are enabled.
> 
> I think you could do a better job in the changelog of explaining what's
> actually going on here. The problem seems to be that you're advertising
> the presence of a non-existent vDSO to userspace.
> 
>> Reported-by: John Stultz <john.stultz@linaro.org>
>> Signed-off-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
>> ---
>>  arch/arm64/include/asm/elf.h | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
>> index 3c7037c6ba9b..b7992bb9d414 100644
>> --- a/arch/arm64/include/asm/elf.h
>> +++ b/arch/arm64/include/asm/elf.h
>> @@ -202,7 +202,7 @@ typedef compat_elf_greg_t		compat_elf_gregset_t[COMPAT_ELF_NGREG];
>>  ({									\
>>  	set_thread_flag(TIF_32BIT);					\
>>   })
>> -#ifdef CONFIG_GENERIC_COMPAT_VDSO
>> +#if defined(CONFIG_COMPAT_VDSO) && defined(CONFIG_GENERIC_COMPAT_VDSO)
> 
> Can't this just be #ifdef CONFIG_COMPAT_VDSO ?
>

Yes, I realized it after I pushed the patch that CONFIG_GENERIC_COMPAT_VDSO can
be removed. Posting v2 shortly.

> John -- can you give this a whirl, please?
> 
> Cheers,
> 
> Will
>

Patch
diff mbox series

diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
index 3c7037c6ba9b..b7992bb9d414 100644
--- a/arch/arm64/include/asm/elf.h
+++ b/arch/arm64/include/asm/elf.h
@@ -202,7 +202,7 @@  typedef compat_elf_greg_t		compat_elf_gregset_t[COMPAT_ELF_NGREG];
 ({									\
 	set_thread_flag(TIF_32BIT);					\
  })
-#ifdef CONFIG_GENERIC_COMPAT_VDSO
+#if defined(CONFIG_COMPAT_VDSO) && defined(CONFIG_GENERIC_COMPAT_VDSO)
 #define COMPAT_ARCH_DLINFO						\
 do {									\
 	/*								\