Re: [2.6.25-git18 => 2.6.26-rc1-git1] Xorg crash with xf86MapVidMem error

["H. Peter Anvin" <hpa@zytor.com>]

Venki Pallipadi wrote:
> On Thu, May 08, 2008 at 01:08:01PM -0700, Rufus & Azrael wrote:
>>    Venki Pallipadi a ecrit :
>>    >
>>    > Use UC_MINUS in reserve_memtype call with -1, when MTRR lookup fails for
>>    any
>>    > reason.
>>    >
>>    > Change the logic in mtrr_type_lookup to just get the type from the start
>>    > address. Using start and end adddress is not right/complete as start and
>>    > end can be covered by different mtrr (where old code will fail) or
>>    > start and end can be in same mtrr, but still have some different
>>    > memory type region in between. Using only start is less restrictive and
>>    > deterministic.
>>    >
>>    > Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
>>    >
>>    > ---
>>    >  arch/x86/kernel/cpu/mtrr/generic.c |    7 ++-----
>>    >  arch/x86/mm/pat.c                  |    8 +-------
>>    >  2 files changed, 3 insertions(+), 12 deletions(-)
>>    >
>>
>>    Hi Venki,
>>
>>    Patch applied on 2.6.26-rc1-git5 and all works fine in X server with 
>>    CONFIG_X86_PAT=y.
>>
>>    See dmesg and mtrr files attached.
>>
> 
> Thanks for reporting the problem and testing the fix.
> ingo/thomas/hpa: Can you please pick up this patch. Fixes the PAT regression
> reported on this thread.

Hm...

I have *serious* concerns about this; this might paper over this 
particular problem, but it just isn't *correct*.

The fact that the range check is implemented incorrectly is not an 
excuse to just dump it and ignore the problem; it should be fixed instead.

	-hpa