Messages in this thread | | | Date | Sat, 05 Mar 2011 19:20:54 -0800 | From | "Justin P. Mattock" <> | Subject | Re: <IRQ> [<ffffffff810c9204>] ? __alloc_pages_nodemask+0x622/0x696 |
| |
On 03/05/2011 06:50 PM, David Rientjes wrote: > On Sat, 5 Mar 2011, Justin Mattock wrote: > >> [ 339.787008] SLUB: Unable to allocate memory on node -1 (gfp=0x20) >> [ 339.787008] cache: kmalloc-8192, object size: 8192, buffer size: >> 8264, default order: 3, min order: 2 >> [ 339.787008] kmalloc-8192 debugging increased min order, use >> slub_debug=O to disable. >> [ 339.787008] node 0: slabs: 181, objs: 543, free: 0 >> [ 339.790575] skbuff alloc of size 3904 failed >> [ 339.823775] swapper: page allocation failure. order:2, mode:0x4020 >> [ 339.823781] Pid: 0, comm: swapper Not tainted 2.6.38-rc7-00143-gd1b152c #7 >> [ 339.823783] Call Trace: >> [ 339.823786]<IRQ> [<ffffffff810c9204>] ? __alloc_pages_nodemask+0x622/0x696 >> [ 339.823811] [<ffffffffa015a077>] ? ath_rxbuf_alloc+0x29/0x9a [ath] >> [ 339.823815] [<ffffffff810f67cf>] ? new_slab+0x7b/0x1c6 >> [ 339.823819] [<ffffffffa015a077>] ? ath_rxbuf_alloc+0x29/0x9a [ath] >> [ 339.823823] [<ffffffff810f6dd6>] ? T.925+0x163/0x29c >> [ 339.823827] [<ffffffff810c4ecc>] ? mempool_destroy+0xf/0x17 >> [ 339.823832] [<ffffffffa015a077>] ? ath_rxbuf_alloc+0x29/0x9a [ath] >> [ 339.823836] [<ffffffff810f895a>] ? __kmalloc_track_caller+0xfe/0x190 >> > > Order-2 atomic allocations are always going to be difficult since we can't > do synchronous compaction or reclaim, but I'd try the suggestion shown > above: boot with slub_debug=O or disable slub debugging so that > kmalloc-8192 is actually order-1 like it should be. >
alright! I will leave that setting on in the boot options to see if this happens again.
Justin P. Mattock
| |