lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <49E6D8FE.80908@xenontk.org>
Date:	Thu, 16 Apr 2009 12:36:38 +0530
From:	David John <davidjon@...ontk.org>
To:	linux-kernel@...r.kernel.org
CC:	venkatesh.pallipadi@...el.com, Ingo Molnar <mingo@...e.hu>,
	Yinghai Lu <yhlu.kernel@...il.com>,
	Jesse Barnes <jbarnes@...tuousgeek.org>
Subject: X freezes intermittently with 2.6.29.1

Hi All,

With 2.6.29.1 (stock kernel), on my Dell Inspiron 1545, X freezes for
15-20 seconds at a time randomly every 10 seconds or so making it
unusable. Currently, I use the 2.6.29.1-54.fc11.x86_64 kernel that comes
with Fedora 11 Beta which has the same problem but only when X is left
idle for a period of time. The kernel log is overrun by messages of the
type:

X:2802 freeing invalid memtype e4b42000-e4b43000
X:2802 freeing invalid memtype e4b43000-e4b44000
X:2802 freeing invalid memtype e4b44000-e4b45000

This I believe is related to this:

mtrr: type mismatch for e0000000,10000000 old: write-back new:
write-combining
[drm] MTRR allocation failed.  Graphics performance may suffer.

The MTRR setup is below:
reg00: base=0x000000000 (    0MB), size=32768MB, count=1: write-back
reg01: base=0x0e0000000 ( 3584MB), size=  512MB, count=1: uncachable
reg02: base=0x07dc00000 ( 2012MB), size=    4MB, count=1: uncachable
reg03: base=0x07e000000 ( 2016MB), size=   32MB, count=1: uncachable

I think the first entry is bogus. Both kernels come with the
MTRR sanitize option enabled but the code is unable to find an
optimal value. Is there any way to correct this? What values for
mtrr_gran_size and mtrr_chunk_size would be appropriate?

While the stock kernel is not usable, the Fedora kernel is okay and X
performance is not slow as well (500fps with glxgears). I've attached
the dmesg log from the Fedora kernel (stock kernel buffer is overrun
in a few seconds) and the configs of both. Let me know if you need
anything else.

Regards,
David.

View attachment "config-2.6.29.1-54.fc11.x86_64" of type "text/plain" (92300 bytes)

View attachment "2.6.29.1-54.fc11.x86_64.log" of type "text/plain" (50996 bytes)

Download attachment "config-2.6.29.1" of type "application/x-troff-man" (82274 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ