Hit this one again while setting up to bisect another issue. I've verified
that the fix *is* someplace in 2.6.33-rc6-mmotm0201 (as that one lives at least
all the way to single-user mode), so I'm basically asking that somebody find
the commit that fixes it and make sure it gets into 2.6.33-final. I'd offer
the commit number for the fix, but I'm hoping somebody recognizes it and pushes
it so I don't have bisect to find the fixing commit.
Thomas? Did you have a fix for this that hasn't been pulled by Linus yet?
Refresher: The bad commit was:
commit 46aeb7430f79cb4d03e17fedd6399884ab3aa697
Author: Thomas Renninger <[email protected]>
Date: Mon Dec 14 11:44:11 2009 +0100
[CPUFREQ] Fix race in cpufreq_update_policy()
[ 4.950955] kernel BUG at drivers/cpufreq/cpufreq.c:88!
[ 4.951072] invalid opcode: 0000 [#1] PREEMPT SMP
[ 4.951452] last sysfs file:
[ 4.951560] CPU 1
[ 4.951754] Pid: 1, comm: swapper Not tainted 2.6.32-08667-g46aeb74 #15 0X564R/Latitude E6500
[ 4.951825] RIP: 0010:[<ffffffff813942f4>] [<ffffffff813942f4>] lock_policy_rwsem_write+0x43/0xad
[ 4.951825] RSP: 0000:ffff88011f87fd40 EFLAGS: 00010202
[ 4.951825] RAX: 00000000000104b0 RBX: 0000000000000001 RCX: 000000000000003e
[ 4.951825] RDX: 0000000000000000 RSI: 0000000000000001 RDI: ffffffff81ae1f88
[ 4.951825] RBP: ffff88011f87fd60 R08: 0000000000000000 R09: ffffffff81b04810
[ 4.951825] R10: ffffffff8103b7fa R11: ffffffff8184f00d R12: 00000000ffffffff
[ 4.951825] R13: 0000000000000000 R14: 0000000000000002 R15: 0000000000000000
[ 4.951825] FS: 0000000000000000(0000) GS:ffff880028300000(0000) knlGS:0000000000000000
[ 4.951825] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 4.951825] CR2: 00000034478a8440 CR3: 0000000001a1c000 CR4: 00000000000406e0
[ 4.951825] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 4.951825] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 4.951825] Process swapper (pid: 1, threadinfo ffff88011f87e000, task ffff88011f87c040)
[ 4.951825] Stack:
[ 4.951825] ffff88011f87c040 0000000000000000 0000000000000000 00000000ffffffea
[ 4.951825] <0> ffff88011f87feb0 ffffffff8139558c ffff88011f87fd90 ffffffff810641a8
[ 4.951825] <0> 0000000000000000 0000000000000000 ffff88011f87fe50 ffffffff81527c4c
[ 4.951825] Call Trace:
[ 4.951825] [<ffffffff8139558c>] cpufreq_update_policy+0x20/0xfc
[ 4.951825] [<ffffffff810641a8>] ? debug_mutex_free_waiter+0x2e/0x6a
[ 4.951825] [<ffffffff81527c4c>] ? __mutex_lock_common+0x58b/0x5aa
[ 4.951825] [<ffffffff8103b7fa>] ? cpu_maps_update_begin+0x12/0x14
[ 4.951825] [<ffffffff815291b0>] ? _raw_spin_unlock_irqrestore+0x72/0x80
[ 4.951825] [<ffffffff8103b7fa>] ? cpu_maps_update_begin+0x12/0x14
[ 4.951825] [<ffffffff81527478>] ? __mutex_unlock_slowpath+0x136/0x17b
[ 4.951825] [<ffffffff81066494>] ? trace_hardirqs_on_caller+0x118/0x13c
[ 4.951825] [<ffffffff810664c5>] ? trace_hardirqs_on+0xd/0xf
[ 4.951825] [<ffffffff815274ac>] ? __mutex_unlock_slowpath+0x16a/0x17b
[ 4.951825] [<ffffffff815274c6>] ? mutex_unlock+0x9/0xb
[ 4.951825] [<ffffffff8103b80c>] ? cpu_maps_update_done+0x10/0x12
[ 4.951825] [<ffffffff815113c7>] ? register_cpu_notifier+0x28/0x32
[ 4.951825] [<ffffffff81b48568>] cpufreq_stats_init+0x85/0xad
[ 4.951825] [<ffffffff81b484e3>] ? cpufreq_stats_init+0x0/0xad
[ 4.951825] [<ffffffff810001ef>] do_one_initcall+0x59/0x14e
[ 4.951825] [<ffffffff81b21667>] kernel_init+0x15f/0x1b5
[ 4.951825] [<ffffffff810031d4>] kernel_thread_helper+0x4/0x10
[ 4.951825] [<ffffffff81529ac0>] ? restore_args+0x0/0x30
[ 4.951825] [<ffffffff81b21508>] ? kernel_init+0x0/0x1b5
[ 4.951825] [<ffffffff810031d0>] ? kernel_thread_helper+0x0/0x10
[ 4.951825] Code: 88 1f ae 81 53 31 db 48 83 ec 08 48 8b 14 d5 60 46 b0 81 44 8b 24 10 41 83 fc ff 0f 94 c3 31 d2 89 de e8 f4 42 d0 ff 85 db 74 04 <0f> 0b eb fe 4d 63 e4 48 c7 c3 c0 04 01 00 48 89 df 4a 03 3c e5
[ 4.951825] RIP [<ffffffff813942f4>] lock_policy_rwsem_write+0x43/0xad
[ 4.951825] RSP <ffff88011f87fd40>
[ 4.971105] ---[ end trace 656ae57b6f9afe82 ]---
[ 4.971228] Kernel panic - not syncing: Attempted to kill init!
[ 4.971358] Pid: 1, comm: swapper Tainted: G D 2.6.32-08667-g46aeb74 #15
[ 4.971515] Call Trace:
[ 4.971646] [<ffffffff81526031>] panic+0x7f/0x13a
[ 4.971763] [<ffffffff81066392>] ? trace_hardirqs_on_caller+0x16/0x13c
[ 4.971878] [<ffffffff8103db1a>] do_exit+0xd7/0x8d6
[ 4.971991] [<ffffffff8103a328>] ? spin_unlock_irqrestore+0x9/0xb
[ 4.972111] [<ffffffff8103af22>] ? kmsg_dump+0x136/0x150
[ 4.972224] [<ffffffff8152a904>] oops_end+0x89/0x8e
[ 4.972351] [<ffffffff81005dd1>] die+0x55/0x5e
[ 4.972463] [<ffffffff8152a324>] do_trap+0x11c/0x12b
[ 4.972578] [<ffffffff81003ac9>] do_invalid_op+0x8f/0x98
[ 4.972691] [<ffffffff813942f4>] ? lock_policy_rwsem_write+0x43/0xad
[ 4.972804] [<ffffffff8152899f>] ? trace_hardirqs_off_thunk+0x3a/0x3c
[ 4.972924] [<ffffffff8103b7fa>] ? cpu_maps_update_begin+0x12/0x14
[ 4.973048] [<ffffffff81529af0>] ? irq_return+0x0/0x2
[ 4.973168] [<ffffffff81003055>] invalid_op+0x15/0x20
[ 4.973300] [<ffffffff8103b7fa>] ? cpu_maps_update_begin+0x12/0x14
[ 4.973421] [<ffffffff813942f4>] ? lock_policy_rwsem_write+0x43/0xad
[ 4.973536] [<ffffffff8139558c>] cpufreq_update_policy+0x20/0xfc
[ 4.973650] [<ffffffff810641a8>] ? debug_mutex_free_waiter+0x2e/0x6a
[ 4.973775] [<ffffffff81527c4c>] ? __mutex_lock_common+0x58b/0x5aa
[ 4.973887] [<ffffffff8103b7fa>] ? cpu_maps_update_begin+0x12/0x14
[ 4.974001] [<ffffffff815291b0>] ? _raw_spin_unlock_irqrestore+0x72/0x80
[ 4.974125] [<ffffffff8103b7fa>] ? cpu_maps_update_begin+0x12/0x14
[ 4.974239] [<ffffffff81527478>] ? __mutex_unlock_slowpath+0x136/0x17b
[ 4.974376] [<ffffffff81066494>] ? trace_hardirqs_on_caller+0x118/0x13c
[ 4.974490] [<ffffffff810664c5>] ? trace_hardirqs_on+0xd/0xf
[ 4.974609] [<ffffffff815274ac>] ? __mutex_unlock_slowpath+0x16a/0x17b
[ 4.974732] [<ffffffff815274c6>] ? mutex_unlock+0x9/0xb
[ 4.974844] [<ffffffff8103b80c>] ? cpu_maps_update_done+0x10/0x12
[ 4.974963] [<ffffffff815113c7>] ? register_cpu_notifier+0x28/0x32
[ 4.975087] [<ffffffff81b48568>] cpufreq_stats_init+0x85/0xad
[ 4.975210] [<ffffffff81b484e3>] ? cpufreq_stats_init+0x0/0xad
[ 4.975334] [<ffffffff810001ef>] do_one_initcall+0x59/0x14e
[ 4.975453] [<ffffffff81b21667>] kernel_init+0x15f/0x1b5
[ 4.975566] [<ffffffff810031d4>] kernel_thread_helper+0x4/0x10
[ 4.975678] [<ffffffff81529ac0>] ? restore_args+0x0/0x30
[ 4.975794] [<ffffffff81b21508>] ? kernel_init+0x0/0x1b5
[ 4.975905] [<ffffffff810031d0>] ? kernel_thread_helper+0x0/0x10
On Thursday 04 February 2010 17:00:02 [email protected] wrote:
> Hit this one again while setting up to bisect another issue. I've verified
> that the fix *is* someplace in 2.6.33-rc6-mmotm0201 (as that one lives at least
> all the way to single-user mode), so I'm basically asking that somebody find
> the commit that fixes it and make sure it gets into 2.6.33-final. I'd offer
> the commit number for the fix, but I'm hoping somebody recognizes it and pushes
> it so I don't have bisect to find the fixing commit.
>
> Thomas? Did you have a fix for this that hasn't been pulled by Linus yet?
The fix is to remove the patch.
> Refresher: The bad commit was:
>
> commit 46aeb7430f79cb4d03e17fedd6399884ab3aa697
> Author: Thomas Renninger <[email protected]>
> Date: Mon Dec 14 11:44:11 2009 +0100
>
> [CPUFREQ] Fix race in cpufreq_update_policy()
This one never hit Linus tree?
It was in -mm, but got reverted.
Did you bisect on another tree?
Thomas
PS: I checked by:
> cat .git/FETCH_HEAD
fc76be434d90bcd57a0ea6b93a2e66a3fec4b664 branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6
> git pull
Already up-to-date.
> git branch
* master
> git show 46aeb7430f79cb4d03e17fedd6399884ab3aa697
fatal: bad object 46aeb7430f79cb4d03e17fedd6399884ab3aa697
On Fri, 05 Feb 2010 10:26:32 +0100, Thomas Renninger said:
> On Thursday 04 February 2010 17:00:02 [email protected] wrote:
> > Hit this one again while setting up to bisect another issue. I've verified
> > that the fix *is* someplace in 2.6.33-rc6-mmotm0201 (as that one lives at least
> > all the way to single-user mode), so I'm basically asking that somebody find
> > the commit that fixes it and make sure it gets into 2.6.33-final. I'd offer
> > the commit number for the fix, but I'm hoping somebody recognizes it and pushes
> > it so I don't have bisect to find the fixing commit.
> >
> > Thomas? Did you have a fix for this that hasn't been pulled by Linus yet?
> The fix is to remove the patch.
>
> > Refresher: The bad commit was:
> >
> > commit 46aeb7430f79cb4d03e17fedd6399884ab3aa697
> > Author: Thomas Renninger <[email protected]>
> > Date: Mon Dec 14 11:44:11 2009 +0100
> >
> > [CPUFREQ] Fix race in cpufreq_update_policy()
> This one never hit Linus tree?
> It was in -mm, but got reverted.
>
> Did you bisect on another tree?
Good catch, Thomas.
Looks like a miscue on my part - although the timestamp on the vmlinuz in /boot
matched the one in arch/x86/boot/bzImage, the file size matched the one left
at the end of that bisect from last time. Looks like I managed to cd into the
wrong /usr/src/linux-* when I copied the bzImage to /boot.
"D'Oh!" - H. Simpson.
Sorry for the inconvenience.