LP:1054732 Added To Kernel Team Hot List
On 24/09/12 17:04, Joseph Salisbury wrote:
I've added the following bug to the Kernel Team Hot List:
* This bug is related to suspend/resume error messages.
* Many duplicates spiked over the weekend.
* Still triaging.
I've given these a quick review:
LP# kernel machine apparent cause
1054732 3.5.0-15.22-generic Lenovo X220 Tablet resume failed
1054503 3.5.0-15.22-generic Lenovo Edge E320 resume failed
1054781 3.5.0-15.22-generic ASUSTek 1001PXD resume failed
1054838 3.5.0-15.22-generic Acer AOA150 resume failed
1055100 3.5.0-15.22-generic MacBookPro6,2 resume failed
1055101 3.5.0-15.22-generic MacBookPro6,2 resume failed
1055371 3.5.0-15.22-generic Dell Studio 1555 S4 failed 
1055462 3.5.0-15.22-generic Lenovo TP T61 resume 
1055472 3.5.0-15.22-generic Lenovo TP W530 resume 
1055603 3.5.0-15.22-generic Gigabyte GA-MA785GM-US2H resumed failed
1055784 3.5.0-15.22-generic Toshiba A100 see 
1055789 3.5.0-15.22-generic Toshiba A100 resume failed
 This is a resume from hibernate failure, so not a duplicate.
 SleepLog.txt states it was suspended and never resumed, user says
they just shut it down.
 Apparently the machine suspended because of low power, and probably
powered off hence it never resumed, false positive bug report.
 No SleepLog.txt, I don't think this is a duplicate, no idea why this
failed, doubt its a resume failure.
Unfortunately none of these have any oops messages, so it's hard to see
what is failing. We should see if these can be gathered for the resume
Best way to approach this:
1. Find a willing user who can easily reproduce this
2. Ask them to edit /etc/default/grub (using sudo and favourite editor)
3. Run "sudo update-grub"
5. Do a bunch of suspend/resumes, see if they catch any oopses (photo)
which we can then work on. If need be, maybe install fwts and run:
sudo fwts s3 --s3-multiple=50
to run 50 S3 cycles to see if we can trip the oops.
BTW, I've exercised this kernel with 50+ S3 cycles on an Lenovo X220 and
can't trip any issues and this H/W is very close to the one reported in
I'm still in the process of triaging this bug. I figured I'd add it to
the hot list since so many duplicates were reported in a short period of
kernel-team mailing list