Bug#666021: [powerpc] Kernel reports page allocation failure: order:1, mode:0x20

[Jonathan Nieder <jrnieder@gmail.com>]

Hi Petr,

Petr Tichy wrote:

> [958021.754593] afpd: page allocation failure: order:1, mode:0x20
> [958021.759018] Call Trace:
> [958021.763346] [c00000000fffee10] [c0000000000134f4] .show_stack+0x80/0x130 (unreliable)
> [958021.772169] [c00000000fffeec0] [c000000000118ea0] .warn_alloc_failed+0xf0/0x108
> [958021.781514] [c00000000fffef80] [c00000000011c170] .__alloc_pages_nodemask+0x700/0x7c4
> [958021.791170] [c00000000ffff110] [c000000000158754] .kmem_getpages+0x5c/0x140
> [958021.796193] [c00000000ffff1b0] [c000000000158a78] .fallback_alloc+0x174/0x200
> [958021.805902] [c00000000ffff290] [c00000000015a474] .kmem_cache_alloc+0x104/0x1f8
> [958021.815720] [c00000000ffff350] [c0000000003c2970] .sk_prot_alloc+0x38/0x1c4
> [958021.820730] [c00000000ffff400] [c0000000003c3a0c] .sk_clone+0x20/0x2cc
> [958021.825597] [c00000000ffff4a0] [c000000000412bcc] .inet_csk_clone+0x1c/0x94
> [958021.830363] [c00000000ffff530] [c00000000042bb14] .tcp_create_openreq_child+0x24/0x3e8
> [958021.839509] [c00000000ffff5e0] [c00000000042a044] .tcp_v4_syn_recv_sock+0x3c/0x31c
[...]

Interesting.  What was happening when this happened?  (Were you using
the machine?  Was it idle or busy?  What is the usual workload like?
Had you just upgraded the kernel recently?)  Is this reproducible?

Thanks for writing,
Jonathan