Re: DEBUG_PAGEALLOC + order-10 alloc/free_pages = lockup

From: Ingo Molnar
Date: Fri Dec 26 2008 - 04:49:33 EST



* Tony Battersby <tonyb@xxxxxxxxxxxxxxx> wrote:

> An order-10 alloc_pages followed by free_pages with DEBUG_PAGEALLOC
> causes a lockup during subsequent memory allocations. Order-9 and lower
> do not trigger the problem. This problem was introduced in 2.6.25-rc1
> and fixed in 2.6.28-rc1. Even though the bug is now fixed, I am
> reporting it because:
>
> 1) I am not sure that anyone ever realized that the bug existed.

Correct.

> 2) I want to make sure that the bug is really fixed and not just hidden
> from view.

i think it got hidden. Apparently splitting up a large kernel linear page
in IRQ context has a bug. I dont see it immediately what it could be -
Thomas, Suresh, do you have any ideas?

> 3) To see if anyone thinks that the fix should be included in
> 2.6.27 -stable.

if 0b8fdcbcd287a1fbe66817491e6149841ae25705 applies cleanly to .27 -stable
then i'd agree it should be added. If there's lots of dependencies then
maybe not.

Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/