linux-kernel - Re: 2.6.39-rc5-git2 boot crashs

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>] [day] [month] [year] [list]

Message-ID: <web-517410535@zbackend1.aha.ru>
Date:	Wed, 04 May 2011 00:13:23 -0400
From:	"werner" <w.landgraf@...ru>
To:	Linus Torvalds <torvalds@...ux-foundation.org>,
	jaxboe@...ionio.com, tj@...nel.org, linux-kernel@...r.kernel.org,
	Steven Rostedt <rostedt@...dmis.org>,
	"H. Peter Anvin" <hpa@...or.com>, werner <w.landgraf@...ru>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...e.hu>
Subject: Re: 2.6.39-rc5-git2 boot crashs

The result of the test below is:

The big-file-zip / unzip / very-big-file-moving crashs 
returned

The sync error didn't return

after switching on CONFIG-X86_EXTENDED_PLATFORMS inclusive 
CONFIG_ELAN, and also let ON the RDC321X_WDT,
ELAN-CPUFREQ, CONFIG-SC500-CPUFREQ (BTREE switched off 
anyway).

Thus, the first problem (zip..) cames from switching on 
EXTENDED PLATFORMS etc , but the second (sync) comes from 
MISC-FILESYSTEMS of which a part incl logfs continued 
unchanged switched OFF, like also the (not-crashing) error 
message came from this.

As said, before was used a config (with EXTENDED_PLATFORMS 
etc OFF) which yesterday runned stable almost the whole 
day and didnt show any of the reported problems.  But 
after this mail, for better security, I'll return to it 
and try to provoke a crash (unzipping bigger files etc) to 
reaffirm that its really stable.

Annexed are:

/ config used
/ The difference of that config to the previous stable 
one.  The zip error should be caused by these differences
/ The difference of the previous stable config, to the 
initial everythingyes config. If this stable config was 
really stable, then all problems reported by me since 
2.6.35-rc3 should be caused by these differences.  (this 
file I sent already, it's for comparison, and as a 
summary)

W.Landgraf

=====================
I started already the compilation suggested before, ie let
CONFIG_MISC_FILESYSTEMS unchanged like it was at the last 
stable config, with logfs OFF,
but switching on CONFIG-X86_EXTENDED_PLATFORMS, and also 
let ON the RDC321X_WDT,
ELAN-CPUFREQ, CONFIG-SC500-CPUFREQ (BTREE switched off 
anyway).
Thus, with exception of the items switched off in 
MISC-FILESYSTEMS (incl. logfs),
it's the same like on my first everyyesconfig.

If the problems don't return, then the reason is anywhere 
in MISC-FILESYSTEMS, one
of the few features which I switched off now, inclusive 
logfs.

If the problems return, then they are because of the 
EXTENDED-PLATFORMS (incl. ELAN) or in one of the 3 items 
quoted above. However, then still the reported boot error
message can be because logfs.

After this is ready, I report about the results, and I do 
what you write below.

wl
---
Professional hosting for everyone - http://www.host.ru

Download attachment "diff.zipproblem" of type "application/octet-stream" (1530 bytes)

Download attachment "config.zipproblem" of type "application/octet-stream" (137480 bytes)

View attachment "diff.config.txt" of type "text/plain" (2669 bytes)