2.6.19-rc1: known regressions

2.6.19-rc1: known regressions

[Adrian Bunk]

Contrary to popular belief, there are people who test -rc kernels
and report bugs.

And there are even people who test -git kernels.

This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18.

If you find your name in the Cc header, you are either submitter of one
of the bugs, maintainer of an affectected subsystem or driver, a patch
of you was declared guilty for a breakage or I'm considering you in any
other way possibly involved with one or more of these issues.

Due to the huge amount of recipients, please trim the Cc when answering.


Subject    : CONFIG_M386=y rwsem compile error
References : http://lkml.org/lkml/2006/10/3/240
Submitter  : Klaus Knopper <knopper@knopper.net>
Guilty     : Andi Kleen <ak@suse.de>
             commit add659bf8aa92f8b3f01a8c0220557c959507fb1
Status     : unknown


Subject    : Lost all PCI devices
References : http://lkml.org/lkml/2006/9/30/128
Submitter  : Luca Tettamanti <kronos.it@gmail.com>
Guilty     : Andi Kleen <ak@suse.de>
             commit 5e544d618f0fb21011f36f28d5e3952b9dc109d2
Handled-By : Andi Kleen <ak@suse.de>
Status     : patch available (might not completely fix the problem?)


Subject    : SMP x86_64 boot problem
References : http://lkml.org/lkml/2006/9/28/330
Submitter  : art@usfltd.com
Status     : unknown


Subject    : sata-via doesn't detect anymore disks attached to VIA vt6421
References : http://bugzilla.kernel.org/show_bug.cgi?id=7255
Submitter  : Thierry Vignaud <tvignaud@mandriva.com>
Status     : unknown


Subject    : T60 stops triggering any ACPI events
References : http://lkml.org/lkml/2006/10/4/425
Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
Status     : unknown


Subject    : DVD drive lost DVD capabilities
References : http://lkml.org/lkml/2006/10/1/45
Submitter  : Olaf Hering <olaf@aepfle.de>
Guilty     : Jens Axboe <axboe@suse.de>
             commit 4aff5e2333c9a1609662f2091f55c3f6fffdad36
Handled-By : Jens Axboe <axboe@suse.de>
Status     : Jens is working on a fix

Re: 2.6.19-rc1: known regressions

[Nigel Cunningham]

Hi.

On Thu, 2006-10-05 at 06:28 +0200, Adrian Bunk wrote:
> Contrary to popular belief, there are people who test -rc kernels
> and report bugs.
> 
> And there are even people who test -git kernels.

Slightly off topic, but let me report a "metoo" as far as testing -git
goes (you can even find a suspend2-against-current-git tree on
kernel.org now!), and some positive progress: for the first time in
months, I'm reliably suspending and resuming my amd64 based laptop. I
don't have an exact cause, but would guess at Andi's multitude of
patches and/or any cpufreq fixes that might have slid in there too. It's
been really strange writing software that others can use reliably, and
not being able to reliably use it yourself! Kudos to all the authors of
suspend/resume bug fixes that have gone in!

Nigel

[patch] x86, fix rwsem build bug on CONFIG_M386=y

[Ingo Molnar]

* Adrian Bunk <bunk@stusta.de> wrote:

> This email lists some known regressions in 2.6.19-rc1 compared to 
> 2.6.18.

thanks Adrian - this list is extremely useful in my opinion! Please keep 
up the good work.

> Subject    : CONFIG_M386=y rwsem compile error
> References : http://lkml.org/lkml/2006/10/3/240
> Submitter  : Klaus Knopper <knopper@knopper.net>
> Guilty     : Andi Kleen <ak@suse.de>
>              commit add659bf8aa92f8b3f01a8c0220557c959507fb1
> Status     : unknown

find the fix below.

-------------------->
Subject: [patch] x86, fix rwsem build bug on CONFIG_M386=y
From: Ingo Molnar <mingo@elte.hu>

CONFIG_M386 turns on spinlock-based generic rwsems - which surprises the 
semaphore.S rwsem stubs. Tested both with and without CONFIG_M386.

Reported-by: Klaus Knopper <knopper@knopper.net>
Triaged-by: Adrian Bunk <bunk@stusta.de>

Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/i386/lib/semaphore.S |    3 +++
 1 file changed, 3 insertions(+)

Index: linux/arch/i386/lib/semaphore.S
===================================================================
--- linux.orig/arch/i386/lib/semaphore.S
+++ linux/arch/i386/lib/semaphore.S
@@ -152,6 +152,8 @@ ENTRY(__read_lock_failed)
 
 #endif
 
+#ifdef CONFIG_RWSEM_XCHGADD_ALGORITHM
+
 /* Fix up special calling conventions */
 ENTRY(call_rwsem_down_read_failed)
 	CFI_STARTPROC
@@ -214,3 +216,4 @@ ENTRY(call_rwsem_downgrade_wake)
 	CFI_ENDPROC
 	END(call_rwsem_downgrade_wake)
 
+#endif

Re: [discuss] [patch] x86, fix rwsem build bug on CONFIG_M386=y

[Andi Kleen]

> 
> -------------------->
> Subject: [patch] x86, fix rwsem build bug on CONFIG_M386=y
> From: Ingo Molnar <mingo@elte.hu>

added thanks

-Andi

2.6.19-rc1: known regressions (v2)

[Adrian Bunk]

This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
that are not yet fixed Linus' tree.

If you find your name in the Cc header, you are either submitter of one
of the bugs, maintainer of an affectected subsystem or driver, a patch
of you was declared guilty for a breakage or I'm considering you in any
other way possibly involved with one or more of these issues.

Due to the huge amount of recipients, please trim the Cc when answering.


Subject    : SMP x86_64 boot problem
References : http://lkml.org/lkml/2006/9/28/330
             http://lkml.org/lkml/2006/10/5/289
Submitter  : art@usfltd.com
Status     : submitter was asked to git bisect
             result of bisecting seems to be wrong


Subject    : snd-hda-intel <-> forcedeth MSI problem
References : http://lkml.org/lkml/2006/10/5/40
Submitter  : Prakash Punnoor <prakash@punnoor.de>
Status     : unknown


Subject    : oops in xfrm_register_mode
References : http://lkml.org/lkml/2006/10/4/170
Submitter  : Steve Fox <drfickle@us.ibm.com>
Status     : unknown


Subject    : T60 stops triggering any ACPI events
References : http://lkml.org/lkml/2006/10/4/425
Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
Status     : unknown


Subject    : thinkpad x60: brightness no longer adjustable in 2.6.18-git
References : http://lkml.org/lkml/2006/10/2/300
Submitter  : Pavel Machek <pavel@ucw.cz>
Status     : unknown, related to the issue above?


Subject    : monitor not active after boot
References : http://lkml.org/lkml/2006/10/5/338
Submitter  : Olaf Hering <olaf@aepfle.de>
Guilty     : Antonino Daplas <adaplas@pol.net>
             commit 346bc21026e7a92e1d7a4a1b3792c5e8b686133d
Status     : unknown


Subject    : sata-via doesn't detect anymore disks attached to VIA vt6421
References : http://bugzilla.kernel.org/show_bug.cgi?id=7255
Submitter  : Thierry Vignaud <tvignaud@mandriva.com>
Status     : unknown


Subject    : ueagle-atm Oops
References : http://lkml.org/lkml/2006/10/6/390
Submitter  : Ernst Herzberg <list-lkml@net4u.de>
Handled-By : Matthieu Castet <castet.matthieu@free.fr>
Status     : unknown


Subject    : DVD drive lost DVD capabilities
References : http://lkml.org/lkml/2006/10/1/45
Submitter  : Olaf Hering <olaf@aepfle.de>
Guilty     : Jens Axboe <axboe@oracle.com>
             commit 4aff5e2333c9a1609662f2091f55c3f6fffdad36
Handled-By : Jens Axboe <axboe@oracle.com>
Status     : Jens is working on a fix


Subject    : sleep/wakeup on powerbooks apparently busted
References : http://lkml.org/lkml/2006/10/5/13
Submitter  : Benjamin Herrenschmidt <benh@kernel.crashing.org>
Handled-By : Benjamin Herrenschmidt <benh@kernel.crashing.org>
Status     : Benjamin will investigate


Subject    : doesn't boot on iBook G4
References : http://lkml.org/lkml/2006/10/5/305
Submitter  : Andreas Schwab <schwab@suse.de>
Handled-By : Mel Gorman <mel@skynet.ie>
Patch      : http://lkml.org/lkml/2006/10/6/80
Status     : patch available


Subject    : airo suspend fails
References : http://lkml.org/lkml/2006/10/6/3
Submitter  : Alex Romosan <romosan@sycorax.lbl.gov>
Guilty     : Sukadev Bhattiprolu <sukadev@us.ibm.com>  (?)
             commit 3b4c7d640376dbccfe80fc4f7b8772ecc7de28c5  (?)
Handled-By : Dave Kleikamp <shaggy@austin.ibm.com>
Patch      : http://lkml.org/lkml/2006/10/7/139
Status     : patch available


Subject    : NFSv4 fails to mount (timeout)
References : http://bugzilla.kernel.org/show_bug.cgi?id=7274
Submitter  : Torsten Kaiser <kernel@bardioc.dyndns.org>
Guilty     : Trond Myklebust <Trond.Myklebust@netapp.com>
             commit 51b6ded4d9a94a61035deba1d8f51a54e3a3dd86
Handled-By : Trond Myklebust <Trond.Myklebust@netapp.com>
Patch      : http://bugzilla.kernel.org/show_bug.cgi?id=7274
Status     : patch available


Subject    : kexec broken on x86_64
References : http://lkml.org/lkml/2006/10/5/86
Submitter  : Magnus Damm <magnus.damm@gmail.com>
Handled-By : Vivek Goyal <vgoyal@in.ibm.com>
Patch      : http://lkml.org/lkml/2006/10/5/199
Status     : patch available


Subject    : strange ieee1394 messages
References : http://lkml.org/lkml/2006/10/5/154
Submitter  : "Alistair John Strachan" <s0348365@sms.ed.ac.uk>
Guilty     : Stefan Richter <stefanr@s5r6.in-berlin.de>
             commit d2f119fe319528da8c76a1107459d6f478cbf28c
Handled-By : Stefan Richter <stefanr@s5r6.in-berlin.de>
Patch      : http://lkml.org/lkml/2006/10/6/235
Status     : harmless, patch available

x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Pavel Machek]

On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> that are not yet fixed Linus' tree.
> 
> If you find your name in the Cc header, you are either submitter of one
> of the bugs, maintainer of an affectected subsystem or driver, a patch
> of you was declared guilty for a breakage or I'm considering you in any
> other way possibly involved with one or more of these issues.
> 
> Due to the huge amount of recipients, please trim the Cc when answering.
...
> Subject    : T60 stops triggering any ACPI events
> References : http://lkml.org/lkml/2006/10/4/425
> Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> Status     : unknown
> 
> 
> Subject    : thinkpad x60: brightness no longer adjustable in 2.6.18-git
> References : http://lkml.org/lkml/2006/10/2/300
> Submitter  : Pavel Machek <pavel@ucw.cz>
> Status     : unknown, related to the issue above?

Strange, problem went away after reboot. I guess I'll write it off as
an acpi glitch... there's definitely something strange going on with
backlight around s2ram: during normal operation, backlight changes are
fast. After s2ram, backlight change from keyboard takes 300msec or so.

-- 
Thanks for all the (sleeping) penguins.

Re: 2.6.19-rc1: known regressions (v2)

[matthieu castet]

Hi,

Adrian Bunk wrote:
> This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> that are not yet fixed Linus' tree.
> 
> If you find your name in the Cc header, you are either submitter of one
> of the bugs, maintainer of an affectected subsystem or driver, a patch
> of you was declared guilty for a breakage or I'm considering you in any
> other way possibly involved with one or more of these issues.
> 
> Due to the huge amount of recipients, please trim the Cc when answering.
> 
> 

> 
> Subject    : ueagle-atm Oops
> References : http://lkml.org/lkml/2006/10/6/390
> Submitter  : Ernst Herzberg <list-lkml@net4u.de>
> Handled-By : Matthieu Castet <castet.matthieu@free.fr>
> Status     : unknown
> 
> 
Guilty     :
http://www.kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=e7ccdfec087f02930c5cdc81143d4a045ae8d361

Patch available at 
https://mail.gna.org/public/ueagleatm-dev/2006-10/msg00022.html , but 
some cleaning is need (I wasn't aware of "USB: fix __must_check warnings 
in drivers/usb/atm/" patch).

Matthieu

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Adrian Bunk]

On Sun, Oct 08, 2006 at 07:12:54AM +0000, Pavel Machek wrote:
> 
> On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> > This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> > that are not yet fixed Linus' tree.
> > 
> > If you find your name in the Cc header, you are either submitter of one
> > of the bugs, maintainer of an affectected subsystem or driver, a patch
> > of you was declared guilty for a breakage or I'm considering you in any
> > other way possibly involved with one or more of these issues.
> > 
> > Due to the huge amount of recipients, please trim the Cc when answering.
> ...
> > Subject    : T60 stops triggering any ACPI events
> > References : http://lkml.org/lkml/2006/10/4/425
> > Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> > Status     : unknown
> > 
> > 
> > Subject    : thinkpad x60: brightness no longer adjustable in 2.6.18-git
> > References : http://lkml.org/lkml/2006/10/2/300
> > Submitter  : Pavel Machek <pavel@ucw.cz>
> > Status     : unknown, related to the issue above?
> 
> Strange, problem went away after reboot. I guess I'll write it off as
> an acpi glitch... there's definitely something strange going on with
> backlight around s2ram: during normal operation, backlight changes are
> fast. After s2ram, backlight change from keyboard takes 300msec or so.

Is this a regression compared to 2.6.18 or an older issue?

If it's a regression can you bisect?

TIA
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Michael S. Tsirkin]

Quoting r. Adrian Bunk <bunk@stusta.de>:
> Subject: Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)
> 
> On Sun, Oct 08, 2006 at 07:12:54AM +0000, Pavel Machek wrote:
> > 
> > On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> > > This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> > > that are not yet fixed Linus' tree.
> > > 
> > ...
> > > Subject    : T60 stops triggering any ACPI events
> > > References : http://lkml.org/lkml/2006/10/4/425
> > > Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> > > Status     : unknown

This was on a pre -rc1 git tree.
I've been using -rc1 since it's out and does not happen to me anymore.
So we probably can write this off as a memory corruption
issue that got fixed in between.

-- 
MST

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Pavel Machek]

On Sun 08-10-06 19:38:09, Adrian Bunk wrote:
> On Sun, Oct 08, 2006 at 07:12:54AM +0000, Pavel Machek wrote:
> > 
> > On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> > > This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> > > that are not yet fixed Linus' tree.
> > > 
> > > If you find your name in the Cc header, you are either submitter of one
> > > of the bugs, maintainer of an affectected subsystem or driver, a patch
> > > of you was declared guilty for a breakage or I'm considering you in any
> > > other way possibly involved with one or more of these issues.
> > > 
> > > Due to the huge amount of recipients, please trim the Cc when answering.
> > ...
> > > Subject    : T60 stops triggering any ACPI events
> > > References : http://lkml.org/lkml/2006/10/4/425
> > > Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> > > Status     : unknown
> > > 
> > > 
> > > Subject    : thinkpad x60: brightness no longer adjustable in 2.6.18-git
> > > References : http://lkml.org/lkml/2006/10/2/300
> > > Submitter  : Pavel Machek <pavel@ucw.cz>
> > > Status     : unknown, related to the issue above?
> > 
> > Strange, problem went away after reboot. I guess I'll write it off as
> > an acpi glitch... there's definitely something strange going on with
> > backlight around s2ram: during normal operation, backlight changes are
> > fast. After s2ram, backlight change from keyboard takes 300msec or so.
> 
> Is this a regression compared to 2.6.18 or an older issue?

Slowness is not something I'd classify as serious problem, and
breakage is not reproducible. Ignore it for now.

-- 
Thanks for all the (sleeping) penguins.

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Adrian Bunk]

On Sun, Oct 08, 2006 at 06:36:02PM +0000, Pavel Machek wrote:
> On Sun 08-10-06 19:38:09, Adrian Bunk wrote:
> > On Sun, Oct 08, 2006 at 07:12:54AM +0000, Pavel Machek wrote:
> > > 
> > > On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> > > > This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> > > > that are not yet fixed Linus' tree.
> > > > 
> > > > If you find your name in the Cc header, you are either submitter of one
> > > > of the bugs, maintainer of an affectected subsystem or driver, a patch
> > > > of you was declared guilty for a breakage or I'm considering you in any
> > > > other way possibly involved with one or more of these issues.
> > > > 
> > > > Due to the huge amount of recipients, please trim the Cc when answering.
> > > ...
> > > > Subject    : T60 stops triggering any ACPI events
> > > > References : http://lkml.org/lkml/2006/10/4/425
> > > > Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> > > > Status     : unknown
> > > > 
> > > > 
> > > > Subject    : thinkpad x60: brightness no longer adjustable in 2.6.18-git
> > > > References : http://lkml.org/lkml/2006/10/2/300
> > > > Submitter  : Pavel Machek <pavel@ucw.cz>
> > > > Status     : unknown, related to the issue above?
> > > 
> > > Strange, problem went away after reboot. I guess I'll write it off as
> > > an acpi glitch... there's definitely something strange going on with
> > > backlight around s2ram: during normal operation, backlight changes are
> > > fast. After s2ram, backlight change from keyboard takes 300msec or so.
> > 
> > Is this a regression compared to 2.6.18 or an older issue?
> 
> Slowness is not something I'd classify as serious problem, and
> breakage is not reproducible. Ignore it for now.

OK, removed.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Adrian Bunk]

On Sun, Oct 08, 2006 at 07:59:58PM +0200, Michael S. Tsirkin wrote:
> Quoting r. Adrian Bunk <bunk@stusta.de>:
> > Subject: Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)
> > 
> > On Sun, Oct 08, 2006 at 07:12:54AM +0000, Pavel Machek wrote:
> > > 
> > > On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> > > > This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> > > > that are not yet fixed Linus' tree.
> > > > 
> > > ...
> > > > Subject    : T60 stops triggering any ACPI events
> > > > References : http://lkml.org/lkml/2006/10/4/425
> > > > Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> > > > Status     : unknown
> 
> This was on a pre -rc1 git tree.
> I've been using -rc1 since it's out and does not happen to me anymore.
> So we probably can write this off as a memory corruption
> issue that got fixed in between.

Thanks for the information, I've removed it from the list.

> MST

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Jeremy Fitzhardinge]

Pavel Machek wrote:
> Strange, problem went away after reboot. I guess I'll write it off as
> an acpi glitch... there's definitely something strange going on with
> backlight around s2ram: during normal operation, backlight changes are
> fast. After s2ram, backlight change from keyboard takes 300msec or so.
>   

Yes, I'm seeing that too.  It's not just keyboard brightness changes; 
gnome does some fadeup effect, which takes ~300ms/step as well.

    J

Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)

[Michael S. Tsirkin]

Quoting r. Adrian Bunk <bunk@stusta.de>:
> Subject: Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)
> 
> On Sun, Oct 08, 2006 at 07:59:58PM +0200, Michael S. Tsirkin wrote:
> > Quoting r. Adrian Bunk <bunk@stusta.de>:
> > > Subject: Re: x60 backlight Re: [discuss] 2.6.19-rc1: known regressions (v2)
> > > 
> > > On Sun, Oct 08, 2006 at 07:12:54AM +0000, Pavel Machek wrote:
> > > > 
> > > > On Sat 07-10-06 23:46:21, Adrian Bunk wrote:
> > > > > This email lists some known regressions in 2.6.19-rc1 compared to 2.6.18
> > > > > that are not yet fixed Linus' tree.
> > > > > 
> > > > ...
> > > > > Subject    : T60 stops triggering any ACPI events
> > > > > References : http://lkml.org/lkml/2006/10/4/425
> > > > > Submitter  : "Michael S. Tsirkin" <mst@mellanox.co.il>
> > > > > Status     : unknown
> > 
> > This was on a pre -rc1 git tree.
> > I've been using -rc1 since it's out and does not happen to me anymore.
> > So we probably can write this off as a memory corruption
> > issue that got fixed in between.
> 
> Thanks for the information, I've removed it from the list.
 

Unfortunately, this came back after using 2.6.19-rc1 for a couple of days
without reboot. And I just re-tested and was able to re-produce the problem on
2.6.19-rc2, as well. So the problem is still there, with the same symptoms:
after usig machine for a while, or under stress (e.g. if I do a full kernel
compile), I stop getting any ACPI events.

tail -f /var/log/acpid
does not show anything, even on Fn/F4 which is supposed ot be always enabled.
Restarting the acpid doesn't do anything either - ACPI starts working
again, for a while, only after reboot.

Worked fine under 2.6.18 (I was running vanilla + this patch
http://lkml.org/lkml/2006/7/20/56 which made it into upstream since then).

Any ideas? How to debug this?

-- 
MST