Re: Regression in suspend to ram in 2.6.31-rc kernels

[Zdenek Kabelac <zdenek.kabelac@gmail.com>]

2009/9/1 Zdenek Kabelac <zdenek.kabelac@gmail.com>:
> 2009/8/31 Rafael J. Wysocki <rjw@sisk.pl>:
>> On Monday 31 August 2009, Zdenek Kabelac wrote:
>>> Hi
>>>
>>> I've noticed that my machine freezes while doing s2r and having
>>> mounted filesystem from SD card.
>>>
>>> My system: Lenovo T61, 4GB, C2D, kernel 2.6.31-rc8
>>>
>>> Here is the trace of  suspend when SD card is inserted, follows resume
>>> and again suspend and this time with mounted filesystem from SD card.
>>>
>>> System locks with: PM: Removing info for mmc:mmc0:b368
>>> Also btusb_bulk_complete: errors are interesting - I've noticed few
>>> bugzillas in the google.
>>> I could try bisect to find the patch - but it will take time....
>>> So hopefully this trace will help.
>>
>> Not really. :-(
>>
>> Is this a new regression in 2.6.31-rc, or does 2.6.30 also fail?
>>
>
> Well - all I know for now is this:  v2.6.30 goes to suspend (i.e.
> machine turns to sleep), but resume fails - I do not get a single line
> of output on serial console either - so it's hard to tell whether it
> works or not - before suspend there are some INFO traces about lockdep
> problems in cpu frequency.   v2.6.31-rc1 doesn't boot on my machine.
> v2.6.31-rc2 seems to lock on suspend so this version is already
> definitely broken. I may try to do some bisecting between these
> kernels - but it might probably easily lead to some dead-ends probably
> :(...
> Also I'm using  mmc debugs for compilation of these kernels - but
> nothing seems to printed from mmc subsystem at the end of suspend log
> and the end is still the same - last printed lines:
>
>  mmc0: card b368 removed
>  PM: Removing info for mmc:mmc0:b368
>

Ok - another bisect game played - and unexpected winner is:

(fat: add ->sync_fs)

f83d6d46e7adf241a064a4a425e5cd8a8fd8925f

Reverting this commit with current -rc8 kernel makes the system happy
during the suspend/resume cycle. Obviously it has it price :) so just
plain revert is probably not a good solution so the problem looks
'more serious'  (fat is not the only fs with this patch) thus adding
original author to this thread.

Zdenek