I can confirm this bug in Kubuntu - Lucid with kernel 2.6.32-11-generic #15-Ubuntu. System was unusable for 17 minutes and consumed the entire swap. Then the X window system rebooted itself (Jan 22 09:25:32 sweetie kernel: [59679.495033] Out of memory: kill process 1928 (kdeinit4) score 5312589 or a child). I was connected to the network via UTP cable. I think, this problem occurs whenever there is no leisure physical memory.

Important lines from the attached kern.log:
from: Jan 22 09:13:14 sweetie kernel: [58951.841921]
to: Jan 22 09:32:11 sweetie kernel: [60089.640345]

1#
Call Trace:
<IRQ>  [<ffffffff810f771e>] __alloc_pages_slowpath+0x56e/0x580
[<ffffffff810f788e>] __alloc_pages_nodemask+0x15e/0x1a0
[<ffffffff8112f432>] kmalloc_large_node+0x62/0xb0
[<ffffffff811338b5>] __kmalloc_node_track_caller+0x145/0x190
[<ffffffffa0347d47>] ? iwl_rx_allocate+0x197/0x2f0 [iwlcore]
[<ffffffff81471780>] __alloc_skb+0x80/0x190
[<ffffffffa0347d47>] iwl_rx_allocate+0x197/0x2f0 [iwlcore]
[<ffffffffa03491ab>] iwl_rx_replenish_now+0x1b/0x30 [iwlcore]
[<ffffffffa03eb258>] iwl_rx_handle+0x288/0x2f0 [iwlagn]
[<ffffffff8105a043>] ? enqueue_task_fair+0x43/0x90
[<ffffffffa03eba70>] iwl_irq_tasklet+0x140/0x4f0 [iwlagn]
[<ffffffff81059a74>] ? try_to_wake_up+0x284/0x380
[<ffffffff8106ac05>] tasklet_action+0xd5/0xe0
[<ffffffff8106c627>] __do_softirq+0xb7/0x1e0
[<ffffffff810c2fa0>] ? handle_IRQ_event+0x60/0x170
[<ffffffff810142ec>] call_softirq+0x1c/0x30
[<ffffffff81015c85>] do_softirq+0x65/0xa0
[<ffffffff8106c4c5>] irq_exit+0x85/0x90
[<ffffffff81564e55>] do_IRQ+0x75/0xf0
[<ffffffff81013b13>] ret_from_intr+0x0/0x11

2# 
Call Trace:
[<ffffffff810b17ad>] ? cpuset_print_task_mems_allowed+0x9d/0xb0
[<ffffffff810f3eb4>] oom_kill_process+0xd4/0x2c0
[<ffffffff810f4440>] ? select_bad_process+0xd0/0x110
[<ffffffff810f44d8>] __out_of_memory+0x58/0xc0
[<ffffffff810f466e>] out_of_memory+0x12e/0x1a0
[<ffffffff8155fc8e>] ? _spin_lock+0xe/0x20
[<ffffffff810f76c1>] __alloc_pages_slowpath+0x511/0x580
[<ffffffff810f788e>] __alloc_pages_nodemask+0x15e/0x1a0
[<ffffffff8112f432>] kmalloc_large_node+0x62/0xb0
[<ffffffff811338b5>] __kmalloc_node_track_caller+0x145/0x190
[<ffffffffa0347d47>] ? iwl_rx_allocate+0x197/0x2f0 [iwlcore]
[<ffffffff81471780>] __alloc_skb+0x80/0x190
[<ffffffffa0347d47>] iwl_rx_allocate+0x197/0x2f0 [iwlcore]
[<ffffffff81019cd0>] ? nommu_map_page+0x0/0xc0
[<ffffffffa03e9f20>] ? iwl_bg_rx_replenish+0x0/0x50 [iwlagn]
[<ffffffffa03491ed>] iwl_rx_replenish+0x2d/0x60 [iwlcore]
[<ffffffffa03e9f20>] ? iwl_bg_rx_replenish+0x0/0x50 [iwlagn]
[<ffffffffa03e9f5d>] iwl_bg_rx_replenish+0x3d/0x50 [iwlagn]
[<ffffffff8107e977>] run_workqueue+0xc7/0x1a0
[<ffffffff8107eaf3>] worker_thread+0xa3/0x110
[<ffffffff81083520>] ? autoremove_wake_function+0x0/0x40
[<ffffffff8107ea50>] ? worker_thread+0x0/0x110
[<ffffffff810831a6>] kthread+0x96/0xa0
[<ffffffff810141ea>] child_rip+0xa/0x20
[<ffffffff81083110>] ? kthread+0x0/0xa0
[<ffffffff810141e0>] ? child_rip+0x0/0x20

3# 
process_name: page allocation failure. order:2, mode:0x4020
(mostly swapper: page allocation failure. order:2, mode:0x4020)

4#
iwlagn 0000:03:00.0: Failed to allocate SKB buffer with GFP_ATOMIC. Only X free buffers remaining.