Dave Chinner wrote:
On Wed, Sep 24, 2008 at 05:42:38PM +1000, Lachlan McIlroy wrote:
Looks good Pete.
No, it is not yet good. Pete cannot explain the underlying problem
and we need to understand if this is fixing the problem or just
changing the timing so it doesn't show up....
Pete clearly demonstrated to me that the sv_wait is being prematurely
woken up when it should not be. There may be an underlying problem or
this could just be another peculiarity of the Linux kernel. Either way
Pete will continue to look into this and whatever the explanation turns
out to be, we deperately need a way to prevent our customers from
running into this deadlock. Even if there is another problem lurking
this code change is fine and if nothing else will serve as a performance
improvement.
|