Re: [patch] O(1) scheduler, -H5

[Arjan van de Ven]

Andi Kleen wrote:
> 
> Russell King <[e-mail]> writes:
> >
> > The serial driver (old or new) open/close functions are one of the worst
> > offenders of the global-cli-and-hold-kernel-lock-and-schedule problem.
> > I'm currently working on fixing this in the new serial driver.
> 
> When they hold the kernel lock in addition to the global cli() before
> schedule() it should be ok. Only the behaviour of code not holding
> kernel lock but global cli and calling schedule() has changed.

well the biggest serial.c offender is block_til_ready of course...
oh and there's quite some dusty old code that does

save_flags();
cli();

while (some_condition)
	sleep_on(&queue);

eg not re-disabling interrupts after the sleep_on().....
to the point where just about every use of
sleep_on/interruptible_sleep_on is buggy
except in serial.c ;(