[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <m13b12rqel.fsf@ebiederm.dsl.xmission.com>
Date: Fri, 08 Jun 2007 17:13:22 -0600
From: ebiederm@...ssion.com (Eric W. Biederman)
To: Jesse Barnes <jesse.barnes@...el.com>
Cc: Andi Kleen <andi@...stfloor.org>, linux-kernel@...r.kernel.org,
Justin Piszcz <jpiszcz@...idpixels.com>
Subject: Re: [PATCH] trim memory not covered by WB MTRRs
Jesse Barnes <jesse.barnes@...el.com> writes:
>> - Overlapping MTRRs.
>
> Overlapping should be ok, since that's usually intentional (e.g. one big
> wb range with a portion of uc space due to another mtrr).
I'm not say overlapping was a bug. I was saying that you don't handle
overlapping mtrrs in figuring the last cached addresses. Therefore
when a UC range overlaps a WB range we might thing the last page
in the WB range is cached when it is not.
>> - What happens if we have uncached memory lower down?
>
> Holes definitely aren't dealt with, but then we haven't seen any yet...
>
>> Except for performance problems I guess that case is relatively
>> harmless. - Is it possible and worth it to amend the e820 map, so it
>> shows the problem area as Reserved or otherwise not usable RAM?
>
> That would be useful, but only if we moved the check to a little
> earlier, prior to the addition of the active ranges from the e820.
> Might be a little nicer than adjusting end_pfn, but will ultimately
> achieve the same thing...
Yes, with perhaps a little more consistency.
Eric
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists