Re: mutexes and preemption

[Sebastian Seifert <lathander-Mmb7MZpHnFY@public.gmane.org>]

On Wed, Mar 03, 2004 at 05:08:53PM -0800, Moore, Robert wrote:
> 
> Looking at the timestamps in the debug output, I suspect that the real
> issue here is this line:
> 
> Mar  2 23:03:10 tenshi kernel:      osl-0909 [2049] os_wait_semaphore
> : Failed to acquire semaphore[dffd8680|1|36863], AE_TIME
> 
> Because this wait should have been for 37 seconds (36863 milliseconds),
> yet not one second has gone by on the wall clock.
> 
> A close look at the Linux implementation of os_wait_semaphore may again
> be in order.

You are right, thank you. I believe there is a  bug in
os_wait_semaphore: the return value of down_trylock isn't checked
correctly, leading to the immediate failure of an Acquire with timeout
in the case the semaphore cannot be acquired immediately . I've sent a
patch to this list. I hope someone applies it.

Sebastian

-- 
A bad analogy is like leaky screwdriver.



-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click