Bug#740148: linux-image-3.12-1-amd64: kernel panic when entering sleep mode

[Jon Jahren <jj@linux.com>]

On 02/27/2014 12:30 AM, Ben Hutchings wrote:
> Control: tag -1 moreinfo
>
> On Wed, 2014-02-26 at 12:00 +0100, Jon Jahren wrote:
> > Package: src:linux
> > Version: 3.12.9-1
> > Severity: important
> >
> > Hello!
> >
> > When entering sleep mode on this computer it will generate a kernel panic. This
> > happens regardless of whether I use the menu or close the lid on the unit.
> > Attached is a log of the panic and I was asked to include these lines when
> > submitting the bug.
> >
> > linux-image-3.12-1-amd64/postinst/depmod-error-initrd-3.12-1-amd64: false
> > linux-image-3.12-1-amd64/prerm/removing-running-kernel-3.12-1-amd64: true
> [...]
> > <0>[   11.029360] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000009
> > <0>[   11.029360]
> > Panic#2 Part2
> > <4>[    5.939992] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> > <4>[    5.940000] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> > <4>[    5.940004] Process systemd (pid: 1, threadinfo ffff880069744000, task ffff880069748000)
> > <4>[    5.940004] Stack:
> > <4>[    5.940004]  ffffea0001b27180 ffff88006cc69090 ffff880069745d48 ffffffff8113abee
> > <4>[    5.940004]  ffffea0001b27180 ffffffffffffffff ffff880069745e28 ffffffff8113ad91
> > <4>[    5.940004]  ffff88003f1ba000 0000000000000050 0000000000000000 0000000000000000
> > <4>[    5.940004] Call Trace:
> > <4>[    5.940004]  [<ffffffff8113abee>] truncate_inode_page+0x4e/0xa0
> > <4>[    5.940004]  [<ffffffff8113ad91>] truncate_inode_pages_range+0x151/0x4b0
> > <4>[    5.940004]  [<ffffffff811b79bd>] ? __inode_wait_for_writeback+0x6d/0xc0
> > <4>[    5.940004]  [<ffffffff81080840>] ? autoremove_wake_function+0x50/0x50
> > <4>[    5.940004]  [<ffffffff8113b175>] truncate_inode_pages+0x15/0x20
> > Panic#2 Part3
> > <4>[    5.939898] CPU 0
> > <4>[    5.939904] Pid: 1, comm: systemd Not tainted 3.6.10-4.fc18.x86_64 #1 Apple Inc. MacBookAir3,2/Mac-942C5DF58193131B
> [...]
>
>
> Then, please test whether the bug is fixed in Linux 3.13 (currently in
> unstable).  If not, and you can't get a crash log, please try to find
> out how much of suspend/resume works by following the instructions at
> <https://www.kernel.org/doc/Documentation/power/basic-pm-debugging.txt>
> (section 2, but ignore the bit about s2ram).
>
> Ben.
Hey Ben!

Booting the kernel with init=/bin/bash results in an unresponsive 
keyboard, and I'm assuming this is because I lack some usb support 
compiled in the kernel as opposed to modules. I'll rebuild the kernels 
with usb compiled statically in the kernel for those things and see if I 
can get some more input.

Instead I just ran through the tests suggested in the document for 
hibernation, and I'm getting some seriously confusing results. Atleast 
from my perspective. I tested with a 3.14-rc4 vanilla kernel, the debian 
3.13-kernel, and the 3.12-kernel. On the 3.14-kernel it will fail only 
occasionally, and I can't seem to find a pattern in the failsures. At 
one point it failed while testing the 'core' of the system. While other 
times it fails on the 'devices' test. I tested this on three different 
kernels, 3.14, 3.13 and 3.12. The result is the same. None of the tests 
from the document can reliably make the system crash. However, on all 
three kernels the command pm-suspend and pm-hibernate will crash the 
system reliably every time.

Output from the /sys/debug/suspend_stats gives me this:

success: 0
fail: 0
failed_freeze: 0
failed_prepare: 0
failed_suspend: 0
failed_suspend_late: 0
failed_suspend_noirq: 0
failed_resume: 0
failed_resume_early: 0
failed_resume_noirq: 0
failures:
  last_failed_dev:

  last_failed_errno:    0
            0
  last_failed_step:


Which looks weird to me.