pmdisk performance problem

pmdisk performance problem

[Narayan Desai]

When using 2.6.0-test7 on an ibm thinkpad t21, pmdisk works properly,
though it takes quite a while to write out pages to disk. On my last
suspend to disk, it took on the order of 8-10 minutes. After this
completed, i was able to successfully resume, fairly speedily, however
my hardware clock was 8-10 minutes behind. Does anyone have any ideas
why this is happening? thanks...
 -nld

Re: pmdisk performance problem

[Patrick Mochel]

> When using 2.6.0-test7 on an ibm thinkpad t21, pmdisk works properly,
> though it takes quite a while to write out pages to disk. On my last
> suspend to disk, it took on the order of 8-10 minutes. After this
> completed, i was able to successfully resume, fairly speedily, however
> my hardware clock was 8-10 minutes behind. Does anyone have any ideas
> why this is happening? thanks...

Do you know if this was just for the write operation, or also included 
stopping all processes? 

Could you increase the log-level before you suspend and let me know -- by 
looking at the debug messages -- what seems to take the most time? 

Thanks,


	Pat

Re: pmdisk performance problem

[Patrick Mochel]

>   Pat> Do you know if this was just for the write operation, or also
>   Pat> included stopping all processes?
> 
> This was during the write of ~4600 pages to disk.

Hm, that's definitely bad. 

>   Pat> Could you increase the log-level before you suspend and let me
>   Pat> know -- by looking at the debug messages -- what seems to take
>   Pat> the most time?
> 
> sorry, but how do i do this?

Sorry, 

	dmesg -n 8 

should do it.


	Pat

Re: pmdisk performance problem

[Nigel Cunningham]

For the clock issue, you need to run /sbin/hwclock --hctosys post resume
(this is not a bug in Software Suspend).

Regards,

Nigel

On Fri, 2003-10-10 at 08:11, Narayan Desai wrote:
> When using 2.6.0-test7 on an ibm thinkpad t21, pmdisk works properly,
> though it takes quite a while to write out pages to disk. On my last
> suspend to disk, it took on the order of 8-10 minutes. After this
> completed, i was able to successfully resume, fairly speedily, however
> my hardware clock was 8-10 minutes behind. Does anyone have any ideas
> why this is happening? thanks...
>  -nld
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
-- 
Nigel Cunningham
495 St Georges Road South, Hastings 4201, New Zealand

Evolution (n): A process whereby infinitely improbable events occur with
alarming frequency, order arises from chaos, and no one is given credit.

Re: pmdisk performance problem

[Narayan Desai]

>>>>> "Pat" == Patrick Mochel <mochel@osdl.org> writes:

  >> When using 2.6.0-test7 on an ibm thinkpad t21, pmdisk works
  >> properly, though it takes quite a while to write out pages to
  >> disk. On my last suspend to disk, it took on the order of 8-10
  >> minutes. After this completed, i was able to successfully resume,
  >> fairly speedily, however my hardware clock was 8-10 minutes
  >> behind. Does anyone have any ideas why this is happening?
  >> thanks...

  Pat> Do you know if this was just for the write operation, or also
  Pat> included stopping all processes?

This was during the write of ~4600 pages to disk. 

  Pat> Could you increase the log-level before you suspend and let me
  Pat> know -- by looking at the debug messages -- what seems to take
  Pat> the most time?

sorry, but how do i do this?
thanks..
 -nld

Re: pmdisk performance problem

[Narayan Desai]

>>>>> "Pat" == Patrick Mochel <mochel@osdl.org> writes:

  Pat> Do you know if this was just for the write operation, or also
  Pat> included stopping all processes?
  >>
  >> This was during the write of ~4600 pages to disk.

  Pat> Hm, that's definitely bad.

  Pat> Could you increase the log-level before you suspend and let me
  Pat> know -- by looking at the debug messages -- what seems to take
  Pat> the most time?
  >>
  >> sorry, but how do i do this?

the particular line that takes the most time is the writing 4700 pages
to disk followed by the dots. each dot takes about 10 seconds to
appear on screen.

I also notice that both hda and hdc are both suspended and resumed
before this point. Is this intentional?
 -nld

Re: pmdisk performance problem

[Pavel Machek]

Hi!

> When using 2.6.0-test7 on an ibm thinkpad t21, pmdisk works properly,
> though it takes quite a while to write out pages to disk. On my last
> suspend to disk, it took on the order of 8-10 minutes. After this
> completed, i was able to successfully resume, fairly speedily, however
> my hardware clock was 8-10 minutes behind. Does anyone have any ideas
> why this is happening? thanks...

I have seen this too, with swsusp on x86-64 machine. Suspend took
awfully long, in "writing data" phase.

								Pavel

-- 
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]