> With respect to the lockup problem: it is due to a non-atomic __set_bit()
> in the new buffer_attached() implementation.
>
> Sure, we don't need atomic semantics for the BH_Attached bit because
> it is always read and modified under a global spinlock. But *other*
> users of buffer_head.b_state do not run under that lock so the nonatomic
> RMW will stomp on their changes. 2.4.20 does not have this bug.
>
> Here is a patch:
Hi Andrew,
The patch works (been running the unixbench subset for an hour now).
Your time and effort and very clear explanations are much appreciated.
Thanks,
Joe
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Jan 07 2003 - 22:00:27 EST