[lkml]   [2002]   [Aug]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH] misc. kernel preemption bits
Robert Love wrote:
> On Thu, 2002-08-29 at 14:24, Andrew Morton wrote:
> > Robert Love wrote:
> > >
> > > ...
> > > - we have a debug check in preempt_schedule that, even
> > > on detecting a schedule with irqs disabled, still goes
> > > ahead and reschedules. We should return. (me)
> > >
> >
> > OK, but that warning will still come out of the mess in mm/slab.c.
> Yah I saw your previous post. There is also a report or two re oops on
> poisoned memory with SLAB_DEBUG and preempt enabled - which should be
> fixed by the above return but one report says it does not.
> Anyhow, your new methods are fine... if they work, they work.
> I am starting to get concerned over the number of routines we are having
> to define. It is getting complicated (i.e. the
> inc_preempt_non_preempt() stuff).
> Maybe keep these local to mm/slab.c ?

Robert, I think what is needed is the changes I sent you for
spinlock.h and system.h. Could you convert to 2.5 and post?

All, I defined:

The preempt versions are to be used in the spin_unlock_*
macros. The ck_preempt versions in open code where we do
need to check preemption. I also redefined:
to call _local_irq_* and then test_preempt(x).
test_preempt() then, if CONFIG_ enables, prints warning
message the first time a test AFTER the irq restore finds a
preemptable condition AND if restore, the interrupt system
is now on.

The net result is that all the code is the same, but we can
now turn on a test to see if there are some preemptions we
are missing. Usually this would come from turning off
interrupts while wake-up is called.
George Anzinger
Preemption patch:
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:28    [W:0.088 / U:1.180 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site