2.6.24-rc5-git7: Reported regressions from 2.6.23
[Note: From December 22 to December 29 inclusive I will be traveling with (most probably) limited Internet access, so the next report will likely be posted after December 30. I also may be unresponsive next week. ;-)]
This message contains a list of some regressions from 2.6.23 reported since 2.6.24-rc1 was released, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know.
If you know of any other unresolved regressions from 2.6.23, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid.
Listed regressions statistics
-----------------------------
Date Total Pending Unresolved
----------------------------------------
Today 118 21 13 2007-12-18 115 29 15 2007-12-12 106 31 17 2007-12-08 98 29 19 2007-12-01 85 29 18 2007-11-24 75 25 21 2007-11-19 68 26 21 2007-11-17 65 25 20
Unresolved regressions
----------------------
Subject : EHCI causes system to resume instantly from S4
Submitter : Maxim Levitsky <[email protected]>
Date : 2007-10-28 14:56
References : http://lkml.org/lkml/2007/10/27/66
http://bugzilla.kernel.org/show_bug.cgi?id=9258
David Brownell <[email protected]> Alan Stern <[email protected]>
Patch :
Workaround : http://bugzilla.kernel.org/show_bug.cgi?id=9258#c30
Subject : v2.6.24-rc2-409-g9418d5d: attempt to access beyond end of device
Submitter : Thomas Meyer <[email protected]>
Date : 2007-11-13 13:11
References : http://lkml.org/lkml/2007/11/13/250
http://bugzilla.kernel.org/show_bug.cgi?id=9370
Patch :
Subject : SError: { DevExch } occuring and causing disruption
Submitter : Avuton Olrich <[email protected]>
Date : 2007-11-15 22:39
References : http://bugzilla.kernel.org/show_bug.cgi?id=9393
Mark Lord <[email protected]>
Patch :
Subject : 20000+ wake-ups/second in 2.6.24
Submitter : Mark Lord <[email protected]>
Date : 2007-12-02 04:23
References : http://lkml.org/lkml/2007/12/1/141
http://bugzilla.kernel.org/show_bug.cgi?id=9489
Patch :
Subject : 2.6.24: false double-clicks from USB mouse
Submitter : Mark Lord <[email protected]>
Date : 2007-12-02 12:05
References : http://lkml.org/lkml/2007/12/2/86
http://bugzilla.kernel.org/show_bug.cgi?id=9492
Dmitry Torokhov <[email protected]>
Patch :
Subject : kobject ->k_name memory leak
Submitter : Alexey Dobriyan <[email protected]>
Date : 2007-12-03 13:42
References : http://lkml.org/lkml/2007/12/3/20
http://bugzilla.kernel.org/show_bug.cgi?id=9496
Patch :
Subject : 2.6.24-rc4 hwmon it87 probe fails
Submitter : Mike Houston <[email protected]>
Date : 2007-12-06 17:10
References : http://lkml.org/lkml/2007/12/4/466
http://bugzilla.kernel.org/show_bug.cgi?id=9514
Patch :
Subject : soft lockup - CPU#1 stuck for 15s! [swapper:0]
Submitter : "Parag Warudkar" <[email protected]>
Date : 2007-12-07 18:14
References : http://lkml.org/lkml/2007/12/7/299
http://bugzilla.kernel.org/show_bug.cgi?id=9525
Thomas Gleixner <[email protected]> Ingo Molnar <[email protected]>
Patch :
Subject : BUG: bad unlock balance detected!
Submitter : Krzysztof Oledzki <[email protected]>
Date : 2007-12-11 03:17
References : http://bugzilla.kernel.org/show_bug.cgi?id=9542
Herbert Xu <[email protected]>
Patch :
Subject : PATA_HPT37X embezzles two ports
Submitter : "Bjoern Olausson" <[email protected]>
Date : 2007-12-12 11:05
References : http://lkml.org/lkml/2007/12/12/161
http://bugzilla.kernel.org/show_bug.cgi?id=9551
Handled-By :
Patch :
Subject : Could not set non-blocking flag with 2.6.24-rc5
Submitter : Tino Keitel <[email protected]>
Date : 2007-12-13 16:27
References : http://lkml.org/lkml/2007/12/13/392
http://bugzilla.kernel.org/show_bug.cgi?id=9557
Handled-By :
Patch :
Subject : s2ram regression
Submitter : Jan Willies <[email protected]>
Date : 2007-12-16 12:02
References : http://bugzilla.kernel.org/show_bug.cgi?id=9585
Rafael J. Wysocki <[email protected]>
Patch :
Subject : swapping in 2.6.24-rc5-git3
Submitter : Lukas Hejtmanek <[email protected]>
Date : 2007-12-17 14:04
References : http://lkml.org/lkml/2007/12/17/98
http://bugzilla.kernel.org/show_bug.cgi?id=9592
Patch :
Regressions with patches
------------------------
Subject : snd_hda_intel 2.6.24-rc2 bug: interrupts don't always work on Lenovo X60s
Submitter : Roland Dreier <[email protected]>
Date : 2007-11-08 14:55
References : http://lkml.org/lkml/2007/11/8/255
http://bugzilla.kernel.org/show_bug.cgi?id=9332
Ingo Molnar <[email protected]>
Patch : http://lkml.org/lkml/2007/11/16/66
Subject : jiffies counter leaps in 2.6.24-rc3
Submitter : Stefano Brivio <[email protected]>
Date : 2007-11-29 08:36
References : http://lkml.org/lkml/2007/11/24/53
http://bugzilla.kernel.org/show_bug.cgi?id=9475
Patch : http://lkml.org/lkml/2007/12/7/132
Subject : Battery shows up twice in kpowersave
Submitter : Rolf Eike Beer <[email protected]>
Date : 2007-12-03 12:06
References : http://bugzilla.kernel.org/show_bug.cgi?id=9494
Patch : http://bugzilla.kernel.org/attachment.cgi?id=14137&action=view
Subject : PS3: trouble with SPARSEMEM_VMEMMAP and kexec
Submitter : Geoff Levand <[email protected]>
Date : 2007-12-03 14:12
References : http://lkml.org/lkml/2007/12/3/137
http://bugzilla.kernel.org/show_bug.cgi?id=9499
Geert Uytterhoeven <[email protected]> Yasunori Goto <[email protected]>
Patch : http://lkml.org/lkml/2007/12/5/73
Subject : RTNL: assertion failed at net/ipv6/addrconf.c (2164)/RTNL: assertion failed at net/ipv4/devinet.c (1055)
Submitter : Krzysztof Oledzki <[email protected]>
Date : 2007-12-11 03:20
References : http://bugzilla.kernel.org/show_bug.cgi?id=9543
Herbert Xu <[email protected]> Jay Vosburgh <[email protected]>
Patch : http://bugzilla.kernel.org/show_bug.cgi?id=9543#c6
Subject : Important regression with XFS update for 2.6.24-rc6
Submitter : Damien Wyart <[email protected]>
Date : 2007-12-19 15:35
References : http://lkml.org/lkml/2007/12/18/126
http://bugzilla.kernel.org/show_bug.cgi?id=9607
Patch : http://lkml.org/lkml/2007/12/19/62
Subject : xfs mknod regression
Submitter : "Bret Towe" <[email protected]>
Date : 2007-12-19 15:42
References : http://lkml.org/lkml/2007/12/17/526
http://bugzilla.kernel.org/show_bug.cgi?id=9608
Patch : http://lkml.org/lkml/2007/12/18/200
Subject : Linux 2.6.24-rc5 x86 architecture no longer Oopses...
Submitter : Trond Myklebust <[email protected]>
Date : 2007-12-20 16:32
References : http://lkml.org/lkml/2007/12/20/552
http://bugzilla.kernel.org/show_bug.cgi?id=9610
Ingo Molnar <[email protected]>
Patch : http://lkml.org/lkml/2007/12/20/596
For details, please follow the links given in references.
As you can see, there is a Bugzilla entry for each of the listed regressions. There also is a Bugzilla entry used for tracking the regressions from 2.6.23, unresolved as well as resolved, at:
http://bugzilla.kernel.org/show_bug.cgi?id=9243
Please let me know if there are any Bugzilla entries that should be added to the list in there.
Thanks,
Rafael
Hi,
Quoted 7 lines · Rafael J. Wysocki
This message contains a list of some regressions from 2.6.23 reported since 2.6.24-rc1 was released, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know.
If you know of any other unresolved regressions from 2.6.23, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid.
I still have a bug with cpufreq when using ondemand governor as default.
The performance governor, which has been the essential default until 1c2562459faedc35927546cfa5273ec6c2884cce, was initialized with fs_initcall() instead of module_init() to make sure the driver is up and running when the bootcode (speedstep_init in my case) calls into it.
Since the above mentioned commit, other governors can also be chosen to be the default but they are not correctly initialized before first use then.
I poked Dave (added to CC) but no response since a few days. I am getting edgy now because this really breaks stuff and the attached patch should get into .24 or at least the problem should be taken more seriously. I have not yet seen a discussion about this particular problem, so if I missed something, please tell me and consider this mail obsolete.
Hannes
From: Johannes Weiner <[email protected]>
Date: Tue, 18 Dec 2007 13:03:49 +0100
Subject: [PATCH] cpufreq: Initialise default governor before use
When the cpufreq driver starts up at boot time, it calls into the default governor which might not be initialised yet. This hurts when the governor's worker function relies on memory that is not yet set up by its init function.
This migrates all governors from module_init() to fs_initcall() when being the default, as was already done in cpufreq_performance when it was the only possible choice.
Fixes at least one actual oops where ondemand is the default governor and cpufreq_governor_dbs() uses the uninitialised kondemand_wq work-queue during boot-time.
--- drivers/cpufreq/cpufreq_conservative.c | 4 ++++ drivers/cpufreq/cpufreq_ondemand.c | 5 ++++- drivers/cpufreq/cpufreq_performance.c | 4 ++++ drivers/cpufreq/cpufreq_userspace.c | 4 ++++ 4 files changed, 16 insertions(+), 1 deletions(-)
drivers/cpufreq/cpufreq_conservative.c +4 -0
@@ -603,5 +603,9 @@ MODULE_DESCRIPTION ("'cpufreq_conservative' - A dynamic cpufreq governor for " "optimised for use in a battery environment");MODULE_LICENSE ("GPL" ); +#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE +fs_initcall (cpufreq_gov_dbs_init); +#else module_init (cpufreq_gov_dbs_init); +#endif module_exit (cpufreq_gov_dbs_exit);
drivers/cpufreq/cpufreq_ondemand.c +4 -1
@@ -610,6 +610,9 @@ MODULE_DESCRIPTION("'cpufreq_ondemand' - A dynamic cpufreq governor for " "Low Latency Frequency Transition capable processors");MODULE_LICENSE ("GPL" ); +#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_ONDEMAND +fs_initcall (cpufreq_gov_dbs_init); +#else module_init (cpufreq_gov_dbs_init); +#endif module_exit (cpufreq_gov_dbs_exit); -
drivers/cpufreq/cpufreq_performance.c +4 -0
@@ -60,5 +60,9 @@ MODULE_AUTHOR("Dominik Brodowski <[email protected]>");MODULE_DESCRIPTION ("CPUfreq policy governor 'performance'" );MODULE_LICENSE ("GPL" ); +#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE fs_initcall (cpufreq_gov_performance_init); +#else +module_init (cpufreq_gov_performance_init); +#endif module_exit (cpufreq_gov_performance_exit);
drivers/cpufreq/cpufreq_userspace.c +4 -0
@@ -231,5 +231,9 @@ MODULE_AUTHOR ("Dominik Brodowski <[email protected]>, Russell King <[email protected].MODULE_DESCRIPTION ("CPUfreq policy governor 'userspace'" );MODULE_LICENSE ("GPL" ); +#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE fs_initcall (cpufreq_gov_userspace_init); +#else +module_init (cpufreq_gov_userspace_init); +#endif module_exit (cpufreq_gov_userspace_exit);
Signature · 1 line
1.5.3.7
Johannes Weiner wrote:
[]
Quoted 10 lines
I still have a bug with cpufreq when using ondemand governor as default.
The performance governor, which has been the essential default until
1c2562459faedc35927546cfa5273ec6c2884cce, was initialized withfs_initcall()instead ofmodule_init()to make sure the driver is up and running when the bootcode (speedstep_initin my case) calls into it.Since the above mentioned commit, other governors can also be chosen to be the default but they are not correctly initialized before first use then.
By the way, is there any real need to specify default governor at a compile time in the first place? Performance governor (which was the only default so far) is a very simple one (not large to consider its size effects for embedded systems for example), and switching governors at run time is trivial as well. What's the motivation behind this new config option?
[]
Quoted 3 lines
This migrates all governors from
module_init()tofs_initcall()when being the default, as was already done incpufreq_performancewhen it was the only possible choice.
Oh well. Which leads to more surprises in the future, I think...
Thanks.
/mjt
Hi,
Quoted 6 lines · Michael Tokarev
By the way, is there any real need to specify default governor at a compile time in the first place? Performance governor (which was the only default so far) is a very simple one (not large to consider its size effects for embedded systems for example), and switching governors at run time is trivial as well. What's the motivation behind this new config option?
I think it is just convenient. If you never use the performance governor, there is no need to compile it. I have no need for an init script that changes the governor on runtime, too. I just say, use this and nothing else, ever. Don't know if this convenience is worth the trouble, though ;)
Quoted 5 lines
Quoted 3 lines
This migrates all governors from
module_init()tofs_initcall()when being the default, as was already done incpufreq_performancewhen it was the only possible choice.Oh well. Which leads to more surprises in the future, I think...
It appears a bit hackish. I would be interested in a cleaner way to force an earlier call to a module's init function.
Hannes
btw, here is the oops as I pencil-copied it:
NULL pointer deref EIP: wq_per_cpu queue_delayed_work_on() do_dbs_timer() cpufreq_governor_dbs() __cpufreq_governor() __cpufreq_set_policy() speedstep_get() cpufreq_add_dev() handle_update() sysdev_driver_register() cpufreq_register_driver() speedstep_init() kernel_init() ...
HTH
Hannes
Quoted 8 lines · Johannes Weiner
Hi,
Quoted 4 lines · Rafael J. Wysocki
This message contains a list of some regressions from 2.6.23 reported since 2.6.24-rc1 was released, for which there are no fixes in the mainline I know of. If any of them have been fixed already, please let me know.
Quoted 5 lines
Quoted 3 lines
If you know of any other unresolved regressions from 2.6.23, please let me know either and I'll add them to the list. Also, please let me know if any of the entries below are invalid.
I still have a bug with cpufreq when using ondemand governor as default.
Added as http://bugzilla.kernel.org/show_bug.cgi?id=9615 .
Thanks,
Rafael