[PATCH 0/2] [Maveric] SRU: Fix CPU feature detection to stop stack dump

[Tim Gardner]

On 10/06/2010 03:46 PM, Colin King wrote:
> From: Colin Ian King<colin.king at canonical.com>
>
> BugLink: http://bugs.launchpad.net/bugs/615153 and
> https://bugzilla.kernel.org/show_bug.cgi?id=16322
>
> SRU Justification:
>
> Impact: Booting and suspend/resuming causes an kernel warning and a
> subsequent kernel stack dump on some machines with CPU features
> such as ARAT.
>
> Fix: Some CPU features such as ARAT are needed in early boot so
> that x86_init function pointers can be set up properly. Also the
> CPU feature detection code has to be re-run after uncapping the CPUID
> level.
>
> Testcase: Without the patches, booting and suspend/resuming causes the
> error: WARNING: at /build/buildd/linux-maverick-2.6.35/arch/x86/include \
> /asm/processor.h:1005 read_measured_perf_ctrs+0x6c/0x80(), and a kernel
> stack dump.  With the patches, boot and resume don't get produce the warning
> or stack dump.
>
> This is a cherrypick of upstream commits:
>    d900329e20f4476db6461752accebcf7935a8055 and
>    1dedefd1a066a795a87afca9c0236e1a94de9bf6
>
> H. Peter Anvin (1):
>    x86, cpu: After uncapping CPUID, re-run CPU feature detection
>
> Jacob Pan (1):
>    x86: detect scattered cpuid features earlier
>
>   arch/x86/kernel/cpu/common.c |    4 ++--
>   arch/x86/kernel/cpu/cpu.h    |    1 +
>   arch/x86/kernel/cpu/intel.c  |    1 +
>   3 files changed, 4 insertions(+), 2 deletions(-)
>
>

Both of these are targeted for stable. Is the severity of the warning 
sufficient to warrant a pre-stable patch?

-- 
Tim Gardner tim.gardner at canonical.com