[1/4] 2.6.23-rc6: known regressions

[1/4] 2.6.23-rc6: known regressions

[Michal Piotrowski]

Hi all,

Here is a list of some known regressions in 2.6.23-rc6.

Feel free to add new regressions/remove fixed etc.
http://kernelnewbies.org/known_regressions

List of Aces

Name                    Regressions fixed since 21-Jun-2007
Adrian Bunk                            10
Linus Torvalds                         6
Alan Stern                             5
Andi Kleen                             5
Hugh Dickins                           5
Trond Myklebust                        5
Andrew Morton                          4
Al Viro                                3
Alexey Starikovskiy                    3
Cornelia Huck                          3
David S. Miller                        3
Jens Axboe                             3
Stephen Hemminger                      3
Tejun Heo                              3



Unclassified

Subject         : x86_64 vdso patch is broken somehow?
References      : http://lkml.org/lkml/2007/8/29/136
Last known good : ?
Submitter       : Chuck Ebbert <cebbert@redhat.com>
Caused-By       : Andi Kleen <ak@suse.de>
                  commit 2aae950b21e4bc789d1fc6668faf67e8748300b7
Handled-By      : ?
Status          : unknown

Subject         : cpu hotplug support broken in 2.6.23-rc3/highres timers break cpu hotplug in 2.6.23-rc5
References      : http://lkml.org/lkml/2007/8/27/58
                  http://lkml.org/lkml/2007/9/3/65
Last known good : ?
Submitter       : Pavel Machek <pavel@ucw.cz>
Caused-By       : ?
Handled-By      : ?
Status          : problem is being debugged

Subject         : 2.6.23-rc3-git1 crash/stuck on VIA CN700 system
References      : http://lkml.org/lkml/2007/8/20/174
Last known good : ?
Submitter       : Stefan Becker <Stefan.Becker@nokia.com>
Caused-By       : Andi Kleen <ak@suse.de>
                  commit 19d36ccdc34f5ed444f8a6af0cbfdb6790eb1177
Handled-By      : Andi Kleen <ak@suse.de>
                  Dave Jones <davej@redhat.com>
Workaround      : http://lkml.org/lkml/2007/9/5/221
Status          : problem is being debugged

Subject         : console is messed up after resume from s2ram or switching to console from X
References      : http://lkml.org/lkml/2007/8/4/6
Last known good : ?
Submitter       : Jeff Chua <jeff.chua.linux@gmail.com>
Caused-By       : ?
Handled-By      : H. Peter Anvin <hpa@zytor.com>
                  Antonino A. Daplas <adaplas@gmail.com>
Workaround      : "s2ram --force --acpi_sleep 1 --vbe_mode"
Status          : problem is being debugged

Subject         : konqueror suddenly vanishing, "konqueror: Fatal IO error: client killed"
References      : http://lkml.org/lkml/2007/7/22/86
Last known good : ?
Submitter       : Markus <lists4me@web.de>
Caused-By       : ?
Handled-By      : Ingo Molnar <mingo@elte.hu>
Status          : problem is being debugged

Subject         : uml on x86_64 compile error
References      : http://lkml.org/lkml/2007/9/3/86
Last known good : ?
Submitter       : Adrian Bunk <bunk@kernel.org>
Caused-By       : Jeff Dike <jdike@linux.intel.com>
                  commit d1254b12c93e1e586137a2ffef71fd33cf273f35
Handled-By      : ?
Status          : unknown



Regards,
Michal

--
LOG
http://www.stardust.webpages.pl/log/

Re: [1/4] 2.6.23-rc6: known regressions

[Linus Torvalds]

On Wed, 12 Sep 2007, Michal Piotrowski wrote:
> 
> Subject         : x86_64 vdso patch is broken somehow?
> References      : http://lkml.org/lkml/2007/8/29/136
> Last known good : ?
> Submitter       : Chuck Ebbert <cebbert@redhat.com>
> Caused-By       : Andi Kleen <ak@suse.de>
>                   commit 2aae950b21e4bc789d1fc6668faf67e8748300b7
> Handled-By      : ?
> Status          : unknown

This should be fixed now. Just pushed out with commit 
95b08679963c78ce0d675224a6efdb5169f2bf75 ("x86_64: Add missing mask 
operation to vdso").

> Subject         : 2.6.23-rc3-git1 crash/stuck on VIA CN700 system
> References      : http://lkml.org/lkml/2007/8/20/174
> Last known good : ?
> Submitter       : Stefan Becker <Stefan.Becker@nokia.com>
> Caused-By       : Andi Kleen <ak@suse.de>
>                   commit 19d36ccdc34f5ed444f8a6af0cbfdb6790eb1177
> Handled-By      : Andi Kleen <ak@suse.de>
>                   Dave Jones <davej@redhat.com>
> Workaround      : http://lkml.org/lkml/2007/9/5/221
> Status          : problem is being debugged

This *seems* to be a VIA bug. Regardless, it should be fixed by commit 
a534b679180025aa324ebd63c05516e478551cfd ("x86_64: Remove CLFLUSH in 
text_poke()").

> Subject         : console is messed up after resume from s2ram or switching to console from X
> References      : http://lkml.org/lkml/2007/8/4/6
> Last known good : ?
> Submitter       : Jeff Chua <jeff.chua.linux@gmail.com>
> Caused-By       : ?
> Handled-By      : H. Peter Anvin <hpa@zytor.com>
>                   Antonino A. Daplas <adaplas@gmail.com>
> Workaround      : "s2ram --force --acpi_sleep 1 --vbe_mode"
> Status          : problem is being debugged

This needs to be re-tested (was it the BIOS register overwriting, 
perhaps?).

> Subject         : konqueror suddenly vanishing, "konqueror: Fatal IO error: client killed"
> References      : http://lkml.org/lkml/2007/7/22/86
> Last known good : ?
> Submitter       : Markus <lists4me@web.de>
> Caused-By       : ?
> Handled-By      : Ingo Molnar <mingo@elte.hu>
> Status          : problem is being debugged

This should probably be dropped. I don't think we have any new reports on 
it, and nobody else ever saw it.

> Subject         : uml on x86_64 compile error
> References      : http://lkml.org/lkml/2007/9/3/86
> Last known good : ?
> Submitter       : Adrian Bunk <bunk@kernel.org>
> Caused-By       : Jeff Dike <jdike@linux.intel.com>
>                   commit d1254b12c93e1e586137a2ffef71fd33cf273f35
> Handled-By      : ?
> Status          : unknown

This should be fixed by commit 980abe215b861891c39aba0936817c46f372143b 
("UML: Fix ELF_CORE_COPY_REGS build botch").

			Linus

Re: [1/4] 2.6.23-rc6: known regressions

[Thomas Gleixner]

On Wed, 2007-09-12 at 18:58 +0200, Michal Piotrowski wrote: 
> Subject         : x86_64 vdso patch is broken somehow?
> References      : http://lkml.org/lkml/2007/8/29/136
> Last known good : ?
> Submitter       : Chuck Ebbert <cebbert@redhat.com>
> Caused-By       : Andi Kleen <ak@suse.de>
>                   commit 2aae950b21e4bc789d1fc6668faf67e8748300b7
> Handled-By      : ?
> Status          : unknown

Patch available:
http://lkml.org/lkml/2007/9/11/161

> Subject         : cpu hotplug support broken in 2.6.23-rc3/highres timers break cpu hotplug in 2.6.23-rc5
> References      : http://lkml.org/lkml/2007/8/27/58
>                   http://lkml.org/lkml/2007/9/3/65
> Last known good : ?
> Submitter       : Pavel Machek <pavel@ucw.cz>
> Caused-By       : ?
> Handled-By      : ?
> Status          : problem is being debugged

This might be the same problem as the jinxed VAIO suspend/resume thingy,
which is a regression as well.

Please assign the Handled-By to me.

Thanks,

	tglx

Re: [1/4] 2.6.23-rc6: known regressions

[Michal Piotrowski]

Thomas Gleixner pisze:
> On Wed, 2007-09-12 at 18:58 +0200, Michal Piotrowski wrote: 
>> Subject         : x86_64 vdso patch is broken somehow?
>> References      : http://lkml.org/lkml/2007/8/29/136
>> Last known good : ?
>> Submitter       : Chuck Ebbert <cebbert@redhat.com>
>> Caused-By       : Andi Kleen <ak@suse.de>
>>                   commit 2aae950b21e4bc789d1fc6668faf67e8748300b7
>> Handled-By      : ?
>> Status          : unknown
> 
> Patch available:
> http://lkml.org/lkml/2007/9/11/161

It is already fixed ;)

> 
>> Subject         : cpu hotplug support broken in 2.6.23-rc3/highres timers break cpu hotplug in 2.6.23-rc5
>> References      : http://lkml.org/lkml/2007/8/27/58
>>                   http://lkml.org/lkml/2007/9/3/65
>> Last known good : ?
>> Submitter       : Pavel Machek <pavel@ucw.cz>
>> Caused-By       : ?
>> Handled-By      : ?
>> Status          : problem is being debugged
> 
> This might be the same problem as the jinxed VAIO suspend/resume thingy,
> which is a regression as well.
> 
> Please assign the Handled-By to me.

Ok.

> 
> Thanks,
> 
> 	tglx
> 

Regards,
Michal

-- 
LOG
http://www.stardust.webpages.pl/log/

Re: [1/4] 2.6.23-rc6: known regressions

[Mark Lord]

Michal Piotrowski wrote:
> Hi all,
> 
> Here is a list of some known regressions in 2.6.23-rc6.
...

Missing from the list:

USB "autosuspend" feature (new in 2.6.23) breaks *lots* of devices.
Many have since been blacklisted in one-at-a-time discovery patches,
but that's really just the tip of the iceberg.

This "feature" breaks a TON of user-visible things,
mostly USB storage devices (USB drives/pens, digicams, etc..).

The functionality is broken for just too much stuff,
and needs to either be reverted or defaulted to "off" rather than "on".

Cheers

Re: [1/4] 2.6.23-rc6: known regressions

[Greg KH]

On Wed, Sep 12, 2007 at 10:50:26PM -0400, Mark Lord wrote:
> Michal Piotrowski wrote:
>> Hi all,
>> Here is a list of some known regressions in 2.6.23-rc6.
> ...
>
> Missing from the list:
>
> USB "autosuspend" feature (new in 2.6.23) breaks *lots* of devices.
> Many have since been blacklisted in one-at-a-time discovery patches,
> but that's really just the tip of the iceberg.
>
> This "feature" breaks a TON of user-visible things,
> mostly USB storage devices (USB drives/pens, digicams, etc..).
>
> The functionality is broken for just too much stuff,
> and needs to either be reverted or defaulted to "off" rather than "on".

Ok, after conferring with Oliver this morning, I think the best solution
here is to both revert the usb-storage suspend patch as it has some
basic problems that can cause data loss, _and_ add the change to the
kernel that by default disables autosuspend on all USB devices except
hubs.

If anyone has any objections to this, please let me know.

thanks,

greg k-h

Re: [1/4] 2.6.23-rc6: known regressions

[Mark Lord]

Greg KH wrote:
> On Wed, Sep 12, 2007 at 10:50:26PM -0400, Mark Lord wrote:
>> Michal Piotrowski wrote:
>>> Hi all,
>>> Here is a list of some known regressions in 2.6.23-rc6.
>> ...
>>
>> Missing from the list:
>>
>> USB "autosuspend" feature (new in 2.6.23) breaks *lots* of devices.
>> Many have since been blacklisted in one-at-a-time discovery patches,
>> but that's really just the tip of the iceberg.
>>
>> This "feature" breaks a TON of user-visible things,
>> mostly USB storage devices (USB drives/pens, digicams, etc..).
>>
>> The functionality is broken for just too much stuff,
>> and needs to either be reverted or defaulted to "off" rather than "on".
> 
> Ok, after conferring with Oliver this morning, I think the best solution
> here is to both revert the usb-storage suspend patch as it has some
> basic problems that can cause data loss, _and_ add the change to the
> kernel that by default disables autosuspend on all USB devices except
> hubs.

Sounds like a sane approach.

Hopefully the kinks can be worked out in time for a reattempt in 2.6.24.
I understand how important this stuff is for battery powered devices,
but it just doesn't appear to be ready yet.

Cheers