2.6.11 issues with S3

2.6.11 issues with S3

[Ranjan Parthasarathy]

[-- Attachment #1: Type: text/plain, Size: 892 bytes --]

I am trying to get S3 working on my system running a 2.6.11 (upgraded from
2.6.4) kernel. On boot up I can put the system to S3 and then wol using magic
packets. Subsequent attempts to put the system to S3 fail with the error
"stopping tasks failed (208 tasks remaining)". I tried changing the TIMEOUT
from the default 6 HZ in kernel/power/process.c to a larger number and the
system is now able to suspend but takes about 20 - 30 seconds for it to go to
suspend. Why is it taking that long? Also shouldn't the timeout be dependent
on the number of processes that we are trying to put to suspend vs a constant
value?
 
On my system, the error is easily reproducible with a simple shell script
that spawns a lot of processes. As the number of processes increase, after
around 70 odd processes in the system I start seeing failures while trying to
suspend.
 
Thanks,
 
Ranjan

[-- Attachment #2: Type: text/html, Size: 4553 bytes --]

Re: 2.6.11 issues with S3

[Pavel Machek]

Hi!

> I am trying to get S3 working on my system running a 2.6.11 (upgraded from
> 2.6.4) kernel. On boot up I can put the system to S3 and then wol using magic
> packets. Subsequent attempts to put the system to S3 fail with the error

Try with recent kernel.

> "stopping tasks failed (208 tasks remaining)". I tried changing the TIMEOUT
> from the default 6 HZ in kernel/power/process.c to a larger number and the
> system is now able to suspend but takes about 20 - 30 seconds for it to go to
> suspend. Why is it taking that long? Also shouldn't the timeout be dependent
> on the number of processes that we are trying to put to suspend vs a constant
> value?
>  
> On my system, the error is easily reproducible with a simple shell script
> that spawns a lot of processes. As the number of processes increase, after
> around 70 odd processes in the system I start seeing failures while trying to
> suspend.

It was not really design to work on stressed system, but feel free to fix it.
Sometimes even make -j 10 on kernel breaks it.
-- 
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl

RE: 2.6.11 issues with S3

[Ranjan Parthasarathy]

Thanks,

Ranjan

-----Original Message-----
From: Pavel Machek [mailto:pavel-+ZI9xUNit7I@public.gmane.org] 
Sent: Tuesday, September 27, 2005 5:44 AM
To: Ranjan Parthasarathy
Cc: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
Subject: Re: [ACPI] 2.6.11 issues with S3

Hi!

> I am trying to get S3 working on my system running a 2.6.11 (upgraded from
> 2.6.4) kernel. On boot up I can put the system to S3 and then wol using
magic
> packets. Subsequent attempts to put the system to S3 fail with the error

Try with recent kernel.

> "stopping tasks failed (208 tasks remaining)". I tried changing the TIMEOUT
> from the default 6 HZ in kernel/power/process.c to a larger number and the
> system is now able to suspend but takes about 20 - 30 seconds for it to go
to
> suspend. Why is it taking that long? Also shouldn't the timeout be
dependent
> on the number of processes that we are trying to put to suspend vs a
constant
> value?
>  
> On my system, the error is easily reproducible with a simple shell script
> that spawns a lot of processes. As the number of processes increase, after
> around 70 odd processes in the system I start seeing failures while trying
to
> suspend.

It was not really design to work on stressed system, but feel free to fix it.
Sometimes even make -j 10 on kernel breaks it.
-- 
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl