Re: [PATCH][Regression] x86, 32-bit: trim memory not covered by wb mtrrs - FIX

[Yinghai Lu]

On Feb 7, 2008 12:21 AM, Balaji Rao <balajirrao@xxxxxxxxx> wrote:
> On Thursday 07 February 2008 01:32:45 pm Ingo Molnar wrote:
> > * Balaji Rao <balajirrao@xxxxxxxxx> wrote:
> > > Hello,
> > >
> > > The following commit caused my X server to stop working.
> > >
> > > commit 99fc8d424bc5d803fe92cad56c068fe64e73747a
> > > Author: Jesse Barnes <jesse.barnes@xxxxxxxxx>
> > > Date:   Wed Jan 30 13:33:18 2008 +0100
> > >
> > >     x86, 32-bit: trim memory not covered by wb mtrrs
> > >
> > > This patch fixes the improper handling of addresses > 4G by
> > > mtrr_trim_uncached_memory. This, now brings up X on my system.
> >
> > thanks. Incidentally this same bug was reported and fixed yesterday, and
> > that fix is upstream already. Could you please compare your solution to
> > Yinghai Lu's fix below, and send us a patch for any further improvements
> > (or cleanups) you might notice in that code? It seems to be almost the
> > same fix as yours. (and hopefully it fixes your X problem too)
>
> Cool! Yes, Yinghai Lu's patch indeed is the same as mine and its really
> surprising that we've used the same variable name too! :)

minor difference
+               trim_start = highest_pfn << PAGE_SHIFT;
+               trim_size = end_pfn << PAGE_SHIFT;

could cause some problem with 32 bit kernel when mem > 4g.
becase highest_pfn and end_pfn is unsigned long aka 32 bit ...could overflow.

so need to assign thtem to trim_start/trim_end at first
or
+               trim_start = (u64)highest_pfn << PAGE_SHIFT;
+               trim_size = (u64)end_pfn << PAGE_SHIFT;

YH
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/