suspend on 'idle'

Mikus Grinbergs mikus at bga.com
Thu Aug 7 14:05:32 EDT 2008


> ... but the idleness threshold obviously isn't working for you.
> If I allowed any CPU use at all (ie. the minimal amount you're using
> to write to the NAND from the network) to inhibit suspend, suspend would
> never happen; kernel threads use CPU in the background all the time.
> We look for CPU use over 20% for a few seconds of time when deciding
> whether to suspend.

Originally, I distrusted 'suspend' because of reports of it 
corrupting the partition tables of removable storage devices.  Even 
though that problem is supposed to have been fixed, I still turn off 
suspend because it interferes with other things I do.


I believe I now remember -- I recently needed 'inhibit-idle-suspend' 
to prevent my XO from suspending during long ethernet downloads.

But what got me started on this topic was that on one XO I run a 
background (nice 19) task 24/7.  If I don't "inhibit" it, I've seen 
that XO 'suspend' - even though its CPU was 100% utilized!  That is 
a Linux application for which I only have the binary -- it would be 
out of the question for me to "patch" the application so as to give 
"hints" to ohm.


Actually, I posted in order to "let others know" what I've 
experienced.  Now that I've seen your thoughts, my opinion is that I 
do  NOT  wish for any changes to your code.  As long as the Control 
Panel makes it easy, I'm willing to manually specify "inhibit 
suspend" for when I'm executing long-duration tasks on my XO, and 
willing to manually specify "allow suspend" for when I'm using the 
XO as an ebook reader.  ["Please, mom - I'd rather do it myself".]


mikus




More information about the Devel mailing list