From: Robert Hancock <hancockr@shaw.ca>
To: linux-kernel <linux-kernel@vger.kernel.org>
Cc: prakash@punnoor.de
Subject: Re: AMD64, 4GB, mttr questions
Date: Tue, 24 Jan 2006 18:13:49 -0600 [thread overview]
Message-ID: <43D6C2BD.50301@shaw.ca> (raw)
In-Reply-To: <5ywpz-49F-11@gated-at.bofh.it>
Prakash Punnoor wrote:
> I have a machine with 4GB RAM, an Athlon64 X2 and following mttr entries:
>
> reg00: base=0x00000000 ( 0MB), size=4096MB: write-back, count=1
> reg01: base=0x100000000 (4096MB), size=2048MB: write-back, count=1
> reg02: base=0x80000000 (2048MB), size=2048MB: uncachable, count=1
>
> First of all, why is there an uncachable region? Is it the upper half of
> memory? Or is this just a hole and the remaining 2GB are seated at
> 0x100000000 ?
Your e820 memory map shows the first 2GB of RAM is at 0-2GB and the
remaining 2GB is at 4-6GB, so yes there is a hole. This doesn't explain
why all of 0-4GB is set as write-back and then the top half of it is
also set as uncacheable. This will presumably have been set up by the
BIOS though, I don't think the kernel does this.
>
> I am also wondering why the kernel doesn't propery set up write-combining
> regions: (I noticed this on 32bit x86, as well, on various machines)
Just because there is MMIO there, does not mean it can be safely set as
write combining. For devices that support write combining, I believe
it's the responsibility of the driver (or in the case of a graphics
card, X) to set this up if it wants. Generally this is only done for
graphics cards due to the limited number of MTRR entries.
> X also comlains about this:
>
> mtrr: type mismatch for b0000000,2000000 old: write-back new: write-combining
Sounds like X is doing something reasonable, that range shouldn't have
been set as write-back in the first place. Presumably the fault of the
BIOS however.
--
Robert Hancock Saskatoon, SK, Canada
To email, remove "nospam" from hancockr@nospamshaw.ca
Home Page: http://www.roberthancock.com/
next parent reply other threads:[~2006-01-25 0:18 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <5ywpz-49F-11@gated-at.bofh.it>
2006-01-25 0:13 ` Robert Hancock [this message]
2006-01-26 7:37 ` AMD64, 4GB, mttr questions Prakash Punnoor
2006-01-24 13:33 Prakash Punnoor
2006-01-25 3:17 ` thockin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=43D6C2BD.50301@shaw.ca \
--to=hancockr@shaw.ca \
--cc=linux-kernel@vger.kernel.org \
--cc=prakash@punnoor.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox