[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [kpreempt-tech] Re: Oops in kernel 2.4.19-pre10-ac2-preempt



On Thu, 2002-06-20 at 10:22, William Lee Irwin III wrote:

> That'd be great, the two places are pte_chain_lock() and pte_chain_unlock().
> They're basically a spin_lock_bit() and spin_unlock_bit(), so they need
> the same kind of disable preempt before the spinloop and re-enable it after
> dropping the lock treatment as spinlocks.

Here is the patch... I will put out an updated -ac patch shortly with
this and some other bits.

Correct?

	Robert Love

diff -urN linux-2.4.19-pre10-ac2/include/linux/mm.h linux/include/linux/mm.h
--- linux-2.4.19-pre10-ac2/include/linux/mm.h	Thu Jun  6 11:16:03 2002
+++ linux/include/linux/mm.h	Thu Jun 20 10:57:01 2002
@@ -340,6 +340,7 @@
 	 * busywait with less bus contention for a good time to
 	 * attempt to acquire the lock bit.
 	 */
+	preempt_disable();
 	while (test_and_set_bit(PG_chainlock, &page->flags)) {
 		while (test_bit(PG_chainlock, &page->flags))
 			cpu_relax();
@@ -349,6 +350,7 @@
 static inline void pte_chain_unlock(struct page *page)
 {
 	clear_bit(PG_chainlock, &page->flags);
+	preempt_enable();
 }
 
 /*


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/