Re: kernel BUG at mm/slab.c:2980 (was Re: [<c019c63f>] xfs_bmap_search_multi_extents+0x6f/0xe0)

[Christoph Lameter <clameter@xxxxxxx>]

On Tue, 4 Sep 2007, Marco Berizzi wrote:

> After a week uptime I got this error. I hope it
> will be useful for you.

Yes indeed but this is a different type of failure. Looks like a higher 
allocation failure in the networking code. Someone created objects that
required an order 2 allocations that failed.

> SLUB: Genslabs=22, HWalign=32, Order=0-1, MinObjects=4, CPUs=1, Nodes=1

Slub is restricted to order 0 and 1 allocs.

>   PREFETCH window: e5c00000-e7cfffff
> NET: Registered protocol family 2
> IP route cache hash table entries: 2048 (order: 1, 8192 bytes)
> TCP established hash table entries: 8192 (order: 4, 65536 bytes)
> TCP bind hash table entries: 8192 (order: 3, 32768 bytes)

Hmmmm.

> eth2:  setting full-duplex.
> swapper: page allocation failure. order:2, mode:0x4020
>  [<c013b21d>] __alloc_pages+0x1ed/0x2e0
>  [<c014ee3b>] allocate_slab+0x4b/0x90
>  [<c014eef2>] new_slab+0x32/0x150
>  [<c014f47b>] __slab_alloc+0xcb/0x120
>  [<c014f429>] __slab_alloc+0x79/0x120
>  [<c02b0183>] tcp_collapse+0x113/0x3b0

tcp_collapse? This is due to a network configuration that required an
order 2 kmalloc block. Jumbo frames?