On Tue, 2010-05-11 at 15:47 +0200, maximilian attems wrote: > On Tue, May 11, 2010 at 02:09:06PM +0530, Ritesh Raj Sarraf wrote: > > On Tuesday 11 May 2010 12:32:45 Ritesh Raj Sarraf wrote: > > > Thanks Max. Am currently testing it. Will report results in a day or two. > > > > The bug got triggered in the very first run. Does not look fixed. dmesg > > attached. > > yes zut, but was worth a try, I can trigger it also: > > [45542.004293] swapper: page allocation failure. order:2, mode:0x4020 > [45542.004305] Pid: 0, comm: swapper Not tainted 2.6.32-5-amd64 #1 > [45542.004311] Call Trace: > [45542.004315] <IRQ> [<ffffffff810b9bcf>] ? > __alloc_pages_nodemask+0x55b/0x5d0 > [45542.004359] [<ffffffffa03aa666>] ? iwl_rx_allocate+0x9f/0x2cb > [iwlcore] > > cat /proc/version > Linux version 2.6.32-5-amd64 (Debian 2.6.32-13) (f...@debian.org) (gcc > version 4.3.2 (Debian 4.3.2-1.1) ) #1 SMP Sat May 8 22:47:58 CEST 2010 > > Ben other suggestion?
I think you've done something wrong in building that kernel. This patch removes padding of the RX buffers so that they are order-1 rather than order-2 (or on the 3945, order-0 rather than order-1). I don't see any way you could get the above error message with this patch properly applied. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse.
signature.asc
Description: This is a digitally signed message part
