> I installed the first patch to enable_bh yesterday a few minutes
> after I received it, it ran fine for about an hour. Then I got the
> email from Andrea with a conceptually safer patch, which I also installed.
> It ran fine all afternoon, no problems. I let it run all night
> and this morning I found my kernel trace triggered on a similar
> problem (maybe the same one). It appears that irq must have
> been disabled when enable_bh was called because it seems that the timer_bh
> did not run when it should have:
have you also fixed start_atomic_bh()/end_atomic_bh() in a similar manner?
> I'll try to reproduce and put in a printk to verify. Is it *always*
> unsafe to do_bottom_half in enable_bh when irq disabled or does that
> only apply to the SMP case?
if you are using ktrace it's probably simpler to use a 'dummy' empty
function which you call only if the bad condition happens. Thus it will
show up in the ktrace.
-- mingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/