Re: Kernel 2.6.9 Multiple Page Allocation Failures

From: Marcelo Tosatti
Date: Tue Nov 09 2004 - 19:04:19 EST


On Tue, Nov 09, 2004 at 11:44:23PM +0100, Lukas Hejtmanek wrote:
> On Tue, Nov 09, 2004 at 02:46:07PM -0800, Andrew Morton wrote:
> > > > Can you please run your workload which cause 0-order page allocation
> > > > failures with the following patch, pretty please?
> > > >
> > > > We will have more information on the free areas state when the allocation
> > > > fails.
> > > >
> > > > Andrew, please apply it to the next -mm, will you?
> > >
> > > here is the trace:
> > > klogd: page allocation failure. order:0, mode: 0x20
> > > [__alloc_pages+441/862] __alloc_pages+0x1b9/0x363
> > > [__get_free_pages+42/63] __get_free_pages+0x25/0x3f
> > > [kmem_getpages+37/201] kmem_getpages+0x21/0xc9
> > > [cache_grow+175/333] cache_grow+0xab/0x14d
> > > [cache_alloc_refill+376/537] cache_alloc_refill+0x174/0x219
> > > [__kmalloc+137/140] __kmalloc+0x85/0x8c
> > > [alloc_skb+75/224] alloc_skb+0x47/0xe0
> > > [e1000_alloc_rx_buffers+72/227] e1000_alloc_rx_buffers+0x44/0xe3
> > > [e1000_clean_rx_irq+402/1095] e1000_clean_rx_irq+0x18e/0x447
> > > [e1000_clean+85/202] e1000_clean+0x51/0xca
> >
> > What kernel is in use here?
> >
> > There was a problem related to e1000 and TSO which was leading to these
> > over-aggressive atomic allocations. That was fixed (within ./net/)
> > post-2.6.9.
>
> I use vanilla 2.6.9.

Lukas,

So you can be hitting the e1000/TSO issue - care to retest with
2.6.10-rc1-mm3/4 please?

Thanks a lot for testing!
-
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/