Messages in this thread | | | Date | Thu, 6 Dec 2012 13:23:40 +0100 (CET) | From | Thomas Gleixner <> | Subject | Re: need help on a DEADLOCK problem related to function try_one_irq() |
| |
Ed,
On Thu, 6 Dec 2012, Edward Donovan wrote: > So, let me try to confirm some things now, so I can learn as I go. To > spell out what's in the new patch: > > The later line, > > action = desc->action; > > should have been this all along? > > action = action->next;
Yes. My bad :(
> And with that fixed, the other test can go. Which is good, now, > because it was creating locking problems for you. > > Ok, thanks for bearing with me. I'll repeat the test here, some > more, but they've generally behaved consistently and probably will > keep showing the same. > > I might as well add that I would happily ship one of these computers, > with a reliably spurious interrupt, to Thomas in Germany. if that > would help for testing.
I'm always happy to have machines which expose strange symptoms.
Thanks,
tglx
| |