[PATCH] Subject: [PATCH] x86, AMD: K8 Rev.A-E processors are subject to erratum 400

[PATCH] Subject: [PATCH] x86, AMD: K8 Rev.A-E processors are subject to erratum 400

[Boris Ostrovsky]

Signed-off-by: Boris Ostrovsky <Boris.Ostrovsky@amd.com>
Acked-by: Borislav Petkov <borislav.petkov@amd.com>
Reported-by: Joerg-Volker Peetz <jvpeetz@web.de>
Tested-by: Joerg-Volker Peetz <jvpeetz@web.de>

Older AMD K8 processors (Revisions A-E) are affected by erratum 400 (APIC timer
interrupts don't occur in C states greater than C1). This, for example, means that
X86_FEATURE_ARAT flag should not be set for these parts.

This addresses regression introduced by commit b87cf80af3ba4b4c008b4face3c68d604e1715c6
(x86, AMD: Set ARAT feature on AMD processors") so it is somewhat high priority.
---
 arch/x86/kernel/cpu/amd.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 3532d3b..71499b4 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -698,7 +698,7 @@ cpu_dev_register(amd_cpu_dev);
  */
 
 const int amd_erratum_400[] =
-	AMD_OSVW_ERRATUM(1, AMD_MODEL_RANGE(0xf, 0x41, 0x2, 0xff, 0xf),
+	AMD_OSVW_ERRATUM(1, AMD_MODEL_RANGE(0xf, 0x4, 0x2, 0xff, 0xf),
 			    AMD_MODEL_RANGE(0x10, 0x2, 0x1, 0xff, 0xf));
 EXPORT_SYMBOL_GPL(amd_erratum_400);
 
-- 
1.7.3.4

Re: [PATCH] Subject: [PATCH] x86, AMD: K8 Rev.A-E processors are subject to erratum 400

[Ingo Molnar]

* Boris Ostrovsky <ostr@amd64.org> wrote:

> Signed-off-by: Boris Ostrovsky <Boris.Ostrovsky@amd.com>
> Acked-by: Borislav Petkov <borislav.petkov@amd.com>
> Reported-by: Joerg-Volker Peetz <jvpeetz@web.de>
> Tested-by: Joerg-Volker Peetz <jvpeetz@web.de>
> 
> Older AMD K8 processors (Revisions A-E) are affected by erratum 400 (APIC 
> timer interrupts don't occur in C states greater than C1). This, for example, 
> means that X86_FEATURE_ARAT flag should not be set for these parts.
> 
> This addresses regression introduced by commit 
> b87cf80af3ba4b4c008b4face3c68d604e1715c6 (x86, AMD: Set ARAT feature on AMD 
> processors") so it is somewhat high priority.

Hm, what is the regression, how do users see it? Hangs? Delays? Crashes? 
Something more benign? The changelog should really contain such information.

Thanks,

	Ingo

Re: [PATCH] Subject: [PATCH] x86, AMD: K8 Rev.A-E processors are subject to erratum 400

[Boris Ostrovsky]

On 04/29/2011 03:55 PM, Ingo Molnar wrote:
>
> * Boris Ostrovsky<ostr@amd64.org>  wrote:
>
>> Signed-off-by: Boris Ostrovsky<Boris.Ostrovsky@amd.com>
>> Acked-by: Borislav Petkov<borislav.petkov@amd.com>
>> Reported-by: Joerg-Volker Peetz<jvpeetz@web.de>
>> Tested-by: Joerg-Volker Peetz<jvpeetz@web.de>
>>
>> Older AMD K8 processors (Revisions A-E) are affected by erratum 400 (APIC
>> timer interrupts don't occur in C states greater than C1). This, for example,
>> means that X86_FEATURE_ARAT flag should not be set for these parts.
>>
>> This addresses regression introduced by commit
>> b87cf80af3ba4b4c008b4face3c68d604e1715c6 (x86, AMD: Set ARAT feature on AMD
>> processors") so it is somewhat high priority.
>
> Hm, what is the regression, how do users see it? Hangs? Delays? Crashes?
> Something more benign? The changelog should really contain such information.


The system becomes unresponsive until an external interrupt (e.g. 
keyboard input) happens. The most obvious effect is that time is not 
reported correctly. We haven't seen anything worse but that's probably 
bad enough.

Do you want me to resubmit the patch with updated comments?

Thanks.
-boris

Re: [PATCH] Subject: [PATCH] x86, AMD: K8 Rev.A-E processors are subject to erratum 400

[Borislav Petkov]

On Fri, Apr 29, 2011 at 04:11:26PM -0400, Boris Ostrovsky wrote:
> On 04/29/2011 03:55 PM, Ingo Molnar wrote:
> >
> > * Boris Ostrovsky<ostr@amd64.org>  wrote:
> >
> >> Signed-off-by: Boris Ostrovsky<Boris.Ostrovsky@amd.com>
> >> Acked-by: Borislav Petkov<borislav.petkov@amd.com>
> >> Reported-by: Joerg-Volker Peetz<jvpeetz@web.de>
> >> Tested-by: Joerg-Volker Peetz<jvpeetz@web.de>
> >>
> >> Older AMD K8 processors (Revisions A-E) are affected by erratum 400 (APIC
> >> timer interrupts don't occur in C states greater than C1). This, for example,
> >> means that X86_FEATURE_ARAT flag should not be set for these parts.
> >>
> >> This addresses regression introduced by commit
> >> b87cf80af3ba4b4c008b4face3c68d604e1715c6 (x86, AMD: Set ARAT feature on AMD
> >> processors") so it is somewhat high priority.
> >
> > Hm, what is the regression, how do users see it? Hangs? Delays? Crashes?
> > Something more benign? The changelog should really contain such information.
> 
> 
> The system becomes unresponsive until an external interrupt (e.g. 
> keyboard input) happens. The most obvious effect is that time is not 
> reported correctly. We haven't seen anything worse but that's probably 
> bad enough.
> 
> Do you want me to resubmit the patch with updated comments?

Yes please. Being too verbose is much better than being laconic with the
commit message.

Also, please put those lines

Signed-off-by: Boris Ostrovsky <Boris.Ostrovsky@amd.com>
Acked-by: Borislav Petkov <borislav.petkov@amd.com>
Reported-by: Joerg-Volker Peetz <jvpeetz@web.de>
Tested-by: Joerg-Volker Peetz <jvpeetz@web.de>

_after_ the commit message.

Thanks.

-- 
Regards/Gruss,
Boris.

Operating Systems Research Center
Advanced Micro Devices, Inc.

Re: [PATCH] Subject: [PATCH] x86, AMD: K8 Rev.A-E processors are subject to erratum 400

[Jiri Slaby]

On 04/29/2011 10:23 PM, Borislav Petkov wrote:
> On Fri, Apr 29, 2011 at 04:11:26PM -0400, Boris Ostrovsky wrote:
>> On 04/29/2011 03:55 PM, Ingo Molnar wrote:
>>>
>>> * Boris Ostrovsky<ostr@amd64.org>  wrote:
>>>
>>>> Signed-off-by: Boris Ostrovsky<Boris.Ostrovsky@amd.com>
>>>> Acked-by: Borislav Petkov<borislav.petkov@amd.com>
>>>> Reported-by: Joerg-Volker Peetz<jvpeetz@web.de>
>>>> Tested-by: Joerg-Volker Peetz<jvpeetz@web.de>
>>>>
>>>> Older AMD K8 processors (Revisions A-E) are affected by erratum 400 (APIC
>>>> timer interrupts don't occur in C states greater than C1). This, for example,
>>>> means that X86_FEATURE_ARAT flag should not be set for these parts.
>>>>
>>>> This addresses regression introduced by commit
>>>> b87cf80af3ba4b4c008b4face3c68d604e1715c6 (x86, AMD: Set ARAT feature on AMD
>>>> processors") so it is somewhat high priority.
>>>
>>> Hm, what is the regression, how do users see it? Hangs? Delays? Crashes?
>>> Something more benign? The changelog should really contain such information.
>>
>>
>> The system becomes unresponsive until an external interrupt (e.g. 
>> keyboard input) happens. The most obvious effect is that time is not 
>> reported correctly. We haven't seen anything worse but that's probably 
>> bad enough.
>>
>> Do you want me to resubmit the patch with updated comments?
> 
> Yes please. Being too verbose is much better than being laconic with the
> commit message.
> 
> Also, please put those lines
> 
> Signed-off-by: Boris Ostrovsky <Boris.Ostrovsky@amd.com>
> Acked-by: Borislav Petkov <borislav.petkov@amd.com>
> Reported-by: Joerg-Volker Peetz <jvpeetz@web.de>
> Tested-by: Joerg-Volker Peetz <jvpeetz@web.de>
> 
> _after_ the commit message.

And also please add
Cc: stable@kernel.org
as the original commit (b87cf80a) was merged into stable trees.

thanks,
-- 
js