Lots of disk activity on resume from s2ram

Lots of disk activity on resume from s2ram

[Johan Brannlund]

Hi. I've noticed that with recent kernels (starting somewhere between 
2.6.20 and 2.6.22) I sometimes get *lots* of disk activity on resume from 
suspend to ram. About 2/3 of the time, the system resumes normally but in 
the remaining 1/3 of the time, the hard drive light stays on almost solid 
and the machine is very, very slow to respond. The only way I've found to 
reliably recover from this is if I can get to a command prompt fast 
enough and do "shutdown -r now". There's not much cpu activity at all, it 
just seems to be disk io that's killing interactivity. This also happens 
if I resume just from an empty gnome desktop with no applications 
running, so I don't think it's due to swapping.

Some kernels affected: 2.6.22, 2.6.23+hrt patches, Ubuntu Gutsy kernel 
(2.6.22-14). These are all 64-bit kernels running on an HP nx6125 laptop 
- single-core Turion 64 processor, 1 gig of ram. To the best of my 
recollection, this problem did *not* appear with 2.6.20.

I put some dmesg output from the vm block dumping and some vmstat output 
at http://nullinfinity.org/tmp/s2ram/ . The dmesg logs are from the same 
resume, just a little while apart. The vmstat is from a different resume.

Any workarounds, patches, tips for further debugging etc would be 
appreciated.

- Johan

Re: Lots of disk activity on resume from s2ram

[Rafael J. Wysocki]

On Monday, 15 October 2007 23:53, Johan Brannlund wrote:
> Hi. I've noticed that with recent kernels (starting somewhere between 
> 2.6.20 and 2.6.22) I sometimes get *lots* of disk activity on resume from 
> suspend to ram. About 2/3 of the time, the system resumes normally but in 
> the remaining 1/3 of the time, the hard drive light stays on almost solid 
> and the machine is very, very slow to respond. The only way I've found to 
> reliably recover from this is if I can get to a command prompt fast 
> enough and do "shutdown -r now". There's not much cpu activity at all, it 
> just seems to be disk io that's killing interactivity. This also happens 
> if I resume just from an empty gnome desktop with no applications 
> running, so I don't think it's due to swapping.
> 
> Some kernels affected: 2.6.22, 2.6.23+hrt patches, Ubuntu Gutsy kernel 
> (2.6.22-14). These are all 64-bit kernels running on an HP nx6125 laptop 
> - single-core Turion 64 processor, 1 gig of ram. To the best of my 
> recollection, this problem did *not* appear with 2.6.20.
> 
> I put some dmesg output from the vm block dumping and some vmstat output 
> at http://nullinfinity.org/tmp/s2ram/ . The dmesg logs are from the same 
> resume, just a little while apart. The vmstat is from a different resume.
> 
> Any workarounds, patches, tips for further debugging etc would be 
> appreciated.

Please file a bug report at http://bugzilla.kernel.org , under
"Power management"->Hibernation/Suspend and add my address to the CC list.

Greetings,
Rafael