FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian Kernel

 
 
LinkBack Thread Tools
 
Old 10-01-2011, 02:31 AM
Jonathan Nieder
 
Default A few linux-2.6 bugs, September edition

# Hi,
#
# As you all know, the linux-2.6 has bugs. More stalled important or
# normal bugs at the end of September than at the start, so we're
# still losing the battle to recover all useful data from each report
# before the submitters get bored.
#
# Acting as bug receptionist is not very fun, so let's see if there
# are other ways to make progress. Here's a summary of the first few
# important and serious bugs not tagged "moreinfo" or "forwarded" in
# the bug list. Please feel free to follow up to them with
# information or hints to move them forward.
#
# Bug: http://bugs.debian.org/642988
# Submitter: Tshepang Lekhonkhobe <tshepang@gmail.com>
# Summary: [experimental] linux-headers depends on linux-kbuild-3.1.0
# which is not in the archive
# Workaround: build src:linux-tools from svn.debian.org. Or don't use
# out-of-tree modules.
#
# Bug: http://bugs.debian.org/643817
# Submitter: Ben Hutchings <ben@decadent.org.uk>
# Summary: [squeeze, sid] NULL pointer dereference in the new IPv6
# fragment ID generator
# Reference: http://thread.gmane.org/gmane.linux.kernel/1196272/focus=1196503
# Regression: yes
# Comment: will presumably revert CVE-2011-2699 fix for now
#
tags 643817 + upstream
#
# Bug: http://bugs.debian.org/593304
# Submitter: Ben Hutchings <ben@decadent.org.uk>
# Summary: [lenny] Removal and reinsertion of floppy or mtdblock device can
# cause oops
# Fixed-upstream: 2.6.28, 2.6.27.51
# Comment: requires physical access to exploit, but the fix seems low risk.
# I think we should take the fix.
#
# Bug: http://bugs.debian.org/598104
# Submitter: Mirosław Zalewski <miniopl@gmail.com>
# Summary: [squeeze] BIOS on Toshiba Satellite machines corrupts DSDT
# during bootup
# Fixed-upstream: 2.6.36
# Comment: can't reproduce the bug on my Satellite C650D, BIOS 1.60 09/02/2010
# Once we have confirmation the fix works, we can apply and pass it to gregkh
#
tags 598104 = upstream patch moreinfo
#
# Bug: http://bugs.debian.org/516785
# Submitter: Hermann Lauer <Hermann.Lauer@iwr.uni-heidelberg.de>
# Summary: [lenny, squeeze, sid] System Hardware FATAL RESET with cassini driver
# on sparc
# Reference: http://thread.gmane.org/gmane.linux.ports.sparc/10405/focus=11193
# Comments: DaveM does enough great work as it is. We shouldn't have to
# rely on him to carry the brunt of this work.
#
# The fundamental questions are: where is it panicking? And why?
# Maybe someone can make progress by the printk method.
#
# Some people are using the cassini driver with more or less success on
# similar (but presumably not identical) machines:
# http://thread.gmane.org/gmane.linux.ports.sparc/14032/focus=14044
# http://thread.gmane.org/gmane.linux.debian.ports.sparc/14019/focus=14025
#
tags 516785 = upstream moreinfo
forwarded 516785 http://thread.gmane.org/gmane.linux.ports.sparc/10405/focus=11193
#
# Bug: http://bugs.debian.org/517627
# Submitter: Thibaut VARENE <varenet@debian.org>
# Summary: [lenny, squeeze, sid] AceNIC driver requires unpackaged
# firmware
# Reference: http://wiki.debian.org/KernelFirmwareLicensing#acenic_.28drivers.2BAC8-net.29
# Comments: From <http://web.archive.org/web/20000711071330/sanjose.alteon.com/license-agree.shtml>:
#
# "Alteon grants you a limited, non-transferable, non-exclusive,
# perpetual license to use, execute and compile the Software solely
# for the purpose of creating, testing and providing software
# programs (herein, the "Authorized Drivers") for use with ALTEON
# ACEnic adapters and ALTEON network cards. You may copy the Software
# and may publicly display the source code of any Authorized Driver
# in educational journals and periodicals; provided that you
# reproduce all applicable copyright and other proprietary notices
# that are contained within the original copy of the Software. You
# may license the object code of an Authorized Driver, including the
# Software contained therein; provided that such Authorized Driver is
# restricted to use solely with ALTEON ACEnic adapters and ALTEON
# network cards."
#
# Debian is not an educational journal or periodical (well, we do
# periodically make releases ), as far as I know, but it looks like
# the compiled firmware should be distributable. Next step is
# presumably to clarify this in linux-firmware:WHENCE.
#
clone 517627 -1
reassign -1 src:firmware-nonfree 0.33
retitle -1 firmware-nonfree: please package acenic firmware
# Once the firmware is packaged, it would be a lot easier to tell
# whether it's feasible to get the driver working using only free
# software.
block 517627 by -1
tags 517627 + upstream
found 517627 linux-2.6/2.6.26-13
#
# Bug: http://bugs.debian.org/524876
# Submitter: Dave Alitz <satabug@powersys.com>
# Summary: [lenny, (later versions?)] sata_mv: timeouts/hard resets on
# 4-port 5041, fixed by replacing disks
# Comments: "The drives in the server were standard Western Digital
# drives instead of their RAID drives. Apparently the older
# driver was slow enough that the drive latency didn't cause
# problems. Replacing the drives with RE3 drives fixed everything."
#
# Doesn't leave me happy. But the submitter is happy, so I don't
# see how we can make progress on it --- close the bug? Would be
# nice if someone knowledgeable about RAID could glance over this
# and tell whether it is a known problem or not.
#
# Bug: http://bugs.debian.org/532005
# Submitter: Matthew Wakeling <mnw21@cam.ac.uk>
# Submitter: Nick Daly <nick.m.daly@gmail.com>
# Summary: HDA digital beep turns a pleasant PC speaker beep into an
# unpleasant honk on some machines.
# Regression: yes
# Comments: I can reproduce this with my roommate's laptop. Will try
# to get time on that computer to look into it.
#
tags 532005 + upstream
#
# Bug: http://bugs.debian.org/533620
# Submitter: Rainer Typke
# Summary: [lenny] (unreproducible?) BUG at fs/ocfs2/dlm/dlmmaster.c:671!
# Comment: sounds like a race. I don't see any obvious bugs or
# patches in the area (not that I've looked hard).
#
tags 533620 + upstream unreproducible
#
# Bug: http://bugs.debian.org/534545
# Submitter: Szabolcs Gyuris <gyuris.szabolcs@integrity.hu>
# Summary: [lenny] BUG at fs/ocfs2/dlmglue.c:678 and hang
# Comment: the tripped assertion is
# BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY)) in
# ocfs2_generic_handle_convert_action. Would be useful to know how
# reproducible it is, and if it is reproducible whether it was fixed
# in newer kernels.
#
# Bug: http://bugs.debian.org/535460
# Submitter: Youssef Eldakar <youssef.eldakar@bibalex.org>
# Summary: [lenny] unbounded memory usage by nfs-kernel-server
# Comment: submitter moved on to OpenAFS but _might_ be able to debug
# anyway
#
tags 535460 + moreinfo
#
# Bug: http://bugs.debian.org/538158
# Submitter: Arcady Genkin <agenkin@cdf.toronto.edu>
# Submitter: Juan Miguel Corral Cano <cde_ruylopez@yahoo.es>
# Summary: [lenny] soft lockup in default_idle
# Comment: Juan reports the squeeze kernel working well on affected
# machines. No obviously relevant patches in the 2.6.27.y tree.
# Would be hard to bisect the fix because is intermittent, but might
# be worth a try.
#
tags 538158 + upstream
#
# Bug: http://bugs.debian.org/556433
# Submitter: Steve Karg <steve@kargs.net>
# Summary: 3c59x: timeout after high network activity
# Comment: Earlier kernels also required a reboot to get the network
# working again. That should probably be treated as a separate bug,
# but it does not affect squeeze and it is not obvious that it
# affects lenny, so meh.
#
# Next step is to test against 3.x and the latest 2.6.32.y and then
# actually pass it upstream this time.
#
found 556433 linux-2.6/2.6.30-8
found 556433 linux-2.6/2.6.32~rc8-1~experimental.1
#
# Bug: http://bugs.debian.org/558258
# Submitter: Peter Neal <doabackflip@gmail.com>
# Summary: BUG in isolate_lru_pages at mm/vmscan.c:725!
# Comment: Has attached oops, ripe for converting to text and
# disassembling with decodecode.
#
# Bug: http://bugs.debian.org/561203
# Submitter: dann frazier <dannf@debian.org>
# Summary: [hppa] cache corruption on VIPT-WB machines
# Reference: http://wiki.parisc-linux.org/TestCases
# Workaround: http://bugs.debian.org/590889
# Comment: long known upstream. Perhaps should be split at some
# point into two bugs, one for fixing this in UP kernels and
# one for SMP kernels.
#
tags 561203 + upstream experimental
forwarded 561203 http://thread.gmane.org/gmane.linux.ports.parisc/3161/focus=3291
#
# Bug: http://bugs.debian.org/568557
# Submitter: cteg <comsublant@gmx.net>
# Submitter: Rik Theys <Rik.Theys@esat.kuleuven.be>
# Summary: asus_atk0110: keyboard and mouse stop working after loading on
# ASUS P7P55D-E
# Workaround: blacklist asus_atk0110
# Workaround: pcie_aspm=off
# Comment: would be helpful to know what exactly "stop working" means
# and whether sid is affected
#
# Thanks for reading. Happy investigating!


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20111001023108.GA2678@elie.sbx02827.chicail.waypor t.net">http://lists.debian.org/20111001023108.GA2678@elie.sbx02827.chicail.waypor t.net
 
Old 10-01-2011, 05:54 AM
Ben Hutchings
 
Default A few linux-2.6 bugs, September edition

On Fri, 2011-09-30 at 21:31 -0500, Jonathan Nieder wrote:
> # Hi,
> #
> # As you all know, the linux-2.6 has bugs. More stalled important or
> # normal bugs at the end of September than at the start, so we're
> # still losing the battle to recover all useful data from each report
> # before the submitters get bored.
> #
> # Acting as bug receptionist is not very fun, so let's see if there
> # are other ways to make progress. Here's a summary of the first few
> # important and serious bugs not tagged "moreinfo" or "forwarded" in
> # the bug list. Please feel free to follow up to them with
> # information or hints to move them forward.
> #
> # Bug: http://bugs.debian.org/642988
> # Submitter: Tshepang Lekhonkhobe <tshepang@gmail.com>
> # Summary: [experimental] linux-headers depends on linux-kbuild-3.1.0
> # which is not in the archive
> # Workaround: build src:linux-tools from svn.debian.org. Or don't use
> # out-of-tree modules.

This keeps happening, because we can't really be bothered to support OOT
crap in experimental. I'm tempted to resolve this by not building the
linux-headers packages in experimental.

> # Bug: http://bugs.debian.org/643817
> # Submitter: Ben Hutchings <ben@decadent.org.uk>
> # Summary: [squeeze, sid] NULL pointer dereference in the new IPv6
> # fragment ID generator
> # Reference: http://thread.gmane.org/gmane.linux.kernel/1196272/focus=1196503
> # Regression: yes
> # Comment: will presumably revert CVE-2011-2699 fix for now
> #
> tags 643817 + upstream

Not sure this is correct, as 3.1 has a very different fix for
CVE-2011-2699 that doesn't suffer this problem.

> # Bug: http://bugs.debian.org/593304
> # Submitter: Ben Hutchings <ben@decadent.org.uk>
> # Summary: [lenny] Removal and reinsertion of floppy or mtdblock device can
> # cause oops
> # Fixed-upstream: 2.6.28, 2.6.27.51
> # Comment: requires physical access to exploit, but the fix seems low risk.
> # I think we should take the fix.

At this point it's hardly worth it, but I leave it to dann to decide.

[...]
> # Bug: http://bugs.debian.org/517627
> # Submitter: Thibaut VARENE <varenet@debian.org>
> # Summary: [lenny, squeeze, sid] AceNIC driver requires unpackaged
> # firmware
> # Reference: http://wiki.debian.org/KernelFirmwareLicensing#acenic_.28drivers.2BAC8-net.29
> # Comments: From <http://web.archive.org/web/20000711071330/sanjose.alteon.com/license-agree.shtml>:
> #
> # "Alteon grants you a limited, non-transferable, non-exclusive,
> # perpetual license to use, execute and compile the Software solely
> # for the purpose of creating, testing and providing software
> # programs (herein, the "Authorized Drivers") for use with ALTEON
> # ACEnic adapters and ALTEON network cards. You may copy the Software

This means, presumably, the "SOFTWARE DRIVER DEVELOPMENT CODE,
INCLUDING ANY SAMPLE FIRMWARE ..." But does it also apply to copying
the firmware alone?

> # and may publicly display the source code of any Authorized Driver
> # in educational journals and periodicals; provided that you
> # reproduce all applicable copyright and other proprietary notices
> # that are contained within the original copy of the Software.

Do we have an original copy with those notices?

[...]
> # Bug: http://bugs.debian.org/524876
> # Submitter: Dave Alitz <satabug@powersys.com>
> # Summary: [lenny, (later versions?)] sata_mv: timeouts/hard resets on
> # 4-port 5041, fixed by replacing disks
> # Comments: "The drives in the server were standard Western Digital
> # drives instead of their RAID drives. Apparently the older
> # driver was slow enough that the drive latency didn't cause
> # problems. Replacing the drives with RE3 drives fixed everything."
> #
> # Doesn't leave me happy. But the submitter is happy, so I don't
> # see how we can make progress on it --- close the bug? Would be
> # nice if someone knowledgeable about RAID could glance over this
> # and tell whether it is a known problem or not.

Given that (1) the submitter was using software RAID (2) there was a
regression in the driver between etch and lenny, it seems unlikely (but
not impossible) that this was actually a defect in the drives or an
incompatibility between the controller and drives.

However, I don't see what we can do with this bug report.

[...]
> # Bug: http://bugs.debian.org/533620
> # Submitter: Rainer Typke
> # Summary: [lenny] (unreproducible?) BUG at fs/ocfs2/dlm/dlmmaster.c:671!
> # Comment: sounds like a race. I don't see any obvious bugs or
> # patches in the area (not that I've looked hard).
> #
> tags 533620 + upstream unreproducible
> #
> # Bug: http://bugs.debian.org/534545
> # Submitter: Szabolcs Gyuris <gyuris.szabolcs@integrity.hu>
> # Summary: [lenny] BUG at fs/ocfs2/dlmglue.c:678 and hang
> # Comment: the tripped assertion is
> # BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY)) in
> # ocfs2_generic_handle_convert_action. Would be useful to know how
> # reproducible it is, and if it is reproducible whether it was fixed
> # in newer kernels.

Also quite possibly OpenVZ brain-damage.

[...]
> # Bug: http://bugs.debian.org/538158
> # Submitter: Arcady Genkin <agenkin@cdf.toronto.edu>
> # Submitter: Juan Miguel Corral Cano <cde_ruylopez@yahoo.es>
> # Summary: [lenny] soft lockup in default_idle
> # Comment: Juan reports the squeeze kernel working well on affected
> # machines. No obviously relevant patches in the 2.6.27.y tree.
> # Would be hard to bisect the fix because is intermittent, but might
> # be worth a try.
> #
> tags 538158 + upstream

Not confirmed by the submitter, though.

[...]
> # Bug: http://bugs.debian.org/558258
> # Submitter: Peter Neal <doabackflip@gmail.com>
> # Summary: BUG in isolate_lru_pages at mm/vmscan.c:725!
> # Comment: Has attached oops, ripe for converting to text and
> # disassembling with decodecode.

Not sure there's much point, as we have the source line. We can tell
that one of the first two tests in __isolate_lru_page() must have
failed.

My inclination is to blame vmmemctl (VMware balloon driver). (I've read
some of their other OOT drivers; they're not good.)

> # Bug: http://bugs.debian.org/561203
> # Submitter: dann frazier <dannf@debian.org>
> # Summary: [hppa] cache corruption on VIPT-WB machines
> # Reference: http://wiki.parisc-linux.org/TestCases
> # Workaround: http://bugs.debian.org/590889
> # Comment: long known upstream. Perhaps should be split at some
> # point into two bugs, one for fixing this in UP kernels and
> # one for SMP kernels.
> #
> tags 561203 + upstream experimental
> forwarded 561203 http://thread.gmane.org/gmane.linux.ports.parisc/3161/focus=3291
[...]

RIP hppa.

Ben.

--
Ben Hutchings
Everything should be made as simple as possible, but not simpler.
- Albert Einstein
 

Thread Tools




All times are GMT. The time now is 05:28 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright ©2007 - 2008, www.linux-archive.org