On Saturday, January 19, 2013 12:41:11 AM Sedat Dilek wrote: > On Sat, Jan 19, 2013 at 12:39 AM, Rafael J. Wysocki <rjw@xxxxxxx> wrote: > > On Saturday, January 19, 2013 12:28:55 AM Sedat Dilek wrote: > >> On Sat, Jan 19, 2013 at 12:25 AM, Rafael J. Wysocki <rjw@xxxxxxx> wrote: > >> > On Friday, January 18, 2013 11:56:53 PM Sedat Dilek wrote: > >> >> On Fri, Jan 18, 2013 at 11:35 PM, Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote: > >> >> > On Fri, Jan 18, 2013 at 11:20 PM, Rafael J. Wysocki <rjw@xxxxxxx> wrote: > >> >> >> On Friday, January 18, 2013 11:11:07 PM Sedat Dilek wrote: > >> >> >>> On Fri, Jan 18, 2013 at 5:37 AM, Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx> wrote: > >> >> >>> > Hi all, > >> >> >>> > > >> >> >>> > Changes since 20130117: > >> >> >>> > > >> >> >>> > Undropped tree: samung > >> >> >>> > > >> >> >>> > The powerpc tree still had a build failure. > >> >> >>> > > >> >> >>> > The driver-core tree gained a build failure for which I applied a merge > >> >> >>> > fix patch. > >> >> >>> > > >> >> >>> > The gpio-lw tree gained a build failure so I used the version from > >> >> >>> > next-20130117. > >> >> >>> > > >> >> >>> > The samsung tree lost the majority of its conflicts but gained more > >> >> >>> > against the arm-soc and slave-dma tree. > >> >> >>> > > >> >> >>> > ---------------------------------------------------------------------------- > >> >> >>> > > >> >> >>> > >> >> >>> From my dmesg diff-file: > >> >> >>> > >> >> >>> +[ 288.730849] PM: Syncing filesystems ... done. > >> >> >>> +[ 294.050498] Freezing user space processes ... (elapsed 0.04 seconds) done. > >> >> >>> +[ 294.097024] Freezing remaining freezable tasks ... > >> >> >>> +[ 314.098849] Freezing of tasks failed after 20.01 seconds (1 tasks > >> >> >>> refusing to freeze, wq_busy=0): > >> >> >>> +[ 314.098862] jbd2/loop0-8 D ffffffff8180d780 0 297 2 0x00000000 > >> >> >>> +[ 314.098865] ffff880117ec5b68 0000000000000046 ffff880117ec5b08 > >> >> >>> ffffffff81044c29 > >> >> >>> +[ 314.098868] ffff88011829dc80 ffff880117ec5fd8 ffff880117ec5fd8 > >> >> >>> ffff880117ec5fd8 > >> >> >>> +[ 314.098871] ffff880119b34560 ffff88011829dc80 ffff880117ec5b68 > >> >> >>> ffff88011fad4738 > >> >> >>> +[ 314.098873] Call Trace: > >> >> >>> +[ 314.098881] [<ffffffff81044c29>] ? default_spin_lock_flags+0x9/0x10 > >> >> >>> +[ 314.098885] [<ffffffff811c63e0>] ? __wait_on_buffer+0x30/0x30 > >> >> >>> +[ 314.098888] [<ffffffff816b4b59>] schedule+0x29/0x70 > >> >> >>> +[ 314.098890] [<ffffffff816b4c2f>] io_schedule+0x8f/0xd0 > >> >> >>> +[ 314.098892] [<ffffffff811c63ee>] sleep_on_buffer+0xe/0x20 > >> >> >>> +[ 314.098896] [<ffffffff816b342f>] __wait_on_bit+0x5f/0x90 > >> >> >>> +[ 314.098898] [<ffffffff811c5aa1>] ? submit_bh+0x121/0x1e0 > >> >> >>> +[ 314.098900] [<ffffffff811c63e0>] ? __wait_on_buffer+0x30/0x30 > >> >> >>> +[ 314.098903] [<ffffffff816b34dc>] out_of_line_wait_on_bit+0x7c/0x90 > >> >> >>> +[ 314.098906] [<ffffffff8107eb00>] ? autoremove_wake_function+0x40/0x40 > >> >> >>> +[ 314.098909] [<ffffffff811c63de>] __wait_on_buffer+0x2e/0x30 > >> >> >>> +[ 314.098913] [<ffffffff8128a6a1>] > >> >> >>> jbd2_journal_commit_transaction+0x1791/0x1960 > >> >> >>> +[ 314.098917] [<ffffffff8109269d>] ? sched_clock_cpu+0xbd/0x110 > >> >> >>> +[ 314.098920] [<ffffffff8107eac0>] ? add_wait_queue+0x60/0x60 > >> >> >>> +[ 314.098923] [<ffffffff81069fbf>] ? try_to_del_timer_sync+0x4f/0x70 > >> >> >>> +[ 314.098925] [<ffffffff8128e4e8>] kjournald2+0xb8/0x240 > >> >> >>> +[ 314.098927] [<ffffffff8107eac0>] ? add_wait_queue+0x60/0x60 > >> >> >>> +[ 314.098929] [<ffffffff8128e430>] ? commit_timeout+0x10/0x10 > >> >> >>> +[ 314.098931] [<ffffffff8107ded0>] kthread+0xc0/0xd0 > >> >> >>> +[ 314.098933] [<ffffffff8107de10>] ? flush_kthread_worker+0xb0/0xb0 > >> >> >>> +[ 314.098936] [<ffffffff816be52c>] ret_from_fork+0x7c/0xb0 > >> >> >>> +[ 314.098938] [<ffffffff8107de10>] ? flush_kthread_worker+0xb0/0xb0 > >> >> >>> +[ 314.098969] > >> >> >>> +[ 314.098970] Restarting kernel threads ... done. > >> >> >>> +[ 314.099052] Restarting tasks ... done. > >> >> >>> > >> >> >>> Please, have a lot at it. > >> >> >> > >> >> >> This is a freezer failure while freezing kernel threads, so I don't think it's > >> >> >> related to ACPI or PM directly. > >> >> >> > >> >> >> Does it happen on every suspend? > >> >> >> > >> >> > > >> >> > No, I only did one S/R. > >> >> > > >> >> > I have built a 2nd new kernel where I pulled-in latest pm.git#linux-next. > >> >> > With this kernel two S/Rs were fine - but that says not much. > >> >> > > >> >> > >> >> After several S/Rs on the "buggy" -1 kernel I know see in my syslogs: > >> >> > >> >> Jan 18 23:50:02 fambox kernel: [ 141.853828] Disabling non-boot CPUs ... > >> >> Jan 18 23:50:02 fambox kernel: [ 141.956943] smpboot: CPU 1 is now offline > >> >> Jan 18 23:50:02 fambox kernel: [ 141.957438] NOHZ: local_softirq_pending 02 > >> >> Jan 18 23:50:02 fambox kernel: [ 141.957454] NOHZ: local_softirq_pending 02 > >> >> Jan 18 23:50:02 fambox kernel: [ 142.060830] smpboot: CPU 2 is now offline > >> >> Jan 18 23:50:02 fambox kernel: [ 142.164639] smpboot: CPU 3 is now offline > >> > > >> > Are you worried about the "local_softirq_pending" messages? > >> > > >> > >> That's the only new messages I have seen after several S/Rs. > > > > They are kind of unusual. > > > > Anyway, they seem to be related to CPU hotplug (CPU offline), so you can try > > if you can trigger them through the sysfs CPU offline/online interface. > > > > Can you explain that a bit clearer or give some sample lines for testing? There is a sysfs file /sys/devices/system/cpu/cpuX/online (where X=0,1,2,3,...) for each CPU core in the system. The value read from it indicates whether or not the given core is online (1 means online). Writing 0 to it means that the given core should be put offline. Writing 1 means to put it back online. You can simply write first 0s and than 1s to those files for CPUs > 0 multiple times in a row and see if that triggers messages like the above. If it does, that may mean there's been a change in kernel/cpu.c, for example, that causes it to appear. The change may have been made somewhere in arch/x86 too, though. > >> If you have a testcase for me to reproduce it here, I would be happy. > > > > Do you mean the freezer-related issue? > > > > Any one as I am still stepping in the dark. > I checked my disc-space as I built a lot of software today and run > once out of space. > But 1.7GiB should be enough on / for testing. > I wanted to run the new LTP version I built the last days. > Let's see what I get... Stress-testing the freezer is rather easy and doesn't require disk space. All it takes is to echo "freezer" to /sys/power/pm_test and then do "echo mem > /sys/power/state && sleep 1" in a loop. This is described in Documentation/power/basic-pm-debugging.txt IIRC. Thanks, Rafael -- I speak only for myself. Rafael J. Wysocki, Intel Open Source Technology Center.
