xen guest crashes if it runs out of memory

Binary package hint: linux-source-2.6.22

I set up a new guest system with 128 MB RAM and 128 MB swap with xen-tools, followed the instructions to work around bug #144631 booted it and did and aptitude full-upgrade. This used up all the memory of the domU and it chrashed:

[71370.652372] BUG: unable to handle kernel NULL pointer dereference at virtual address 00000000
[71370.652386] printing eip:
[71370.652394] 07bf0000 -> *pde = 00000000:4d76d001
[71370.652398] 01d8a000 -> *pme = 00000000:00000000
[71370.652403] Oops: 0000 [#1]
[71370.652405] SMP
[71370.652411] Modules linked in: ipv6 evdev xfs dm_mirror dm_snapshot dm_mod fuse apparmor commoncap
[71370.652432] CPU: 0
[71370.652433] EIP: 0061:[<c019d5a6>] Not tainted VLI
[71370.652435] EFLAGS: 00010282 (2.6.22-14-xen #1)
[71370.652445] EIP is at block_prepare_write+0x16/0x40
[71370.652449] eax: 00000000 ebx: 00000000 ecx: 00000181 edx: c9178d70
[71370.652454] esi: c16a2600 edi: c9184120 ebp: c16a2600 esp: c394dd38
[71370.652457] ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0069
[71370.652462] Process dpkg (pid: 3125, ti=c394c000 task=c2b24a60 task.ti=c394c000)
[71370.652466] Stack: 00000000 00000001 00000000 00000181 c01539a8 c9178d70 00000181 00000200
[71370.652481] 00000001 c394ded8 00000001 00000181 c14fc0c0 c1a400c0 c9184120 c1a4001c
[71370.652497] c16d0b20 00001000 08316350 00000000 000040e9 c1594180 c15a4980 c394df60
[71370.652512] Call Trace:
[71370.652517] [<c01539a8>] generic_file_buffered_write+0x208/0x690
[71370.652526] [<c9178d70>] xfs_get_blocks+0x0/0x30 [xfs]
[71370.652557] [<c011c34f>] try_to_wake_up+0x3f/0x400
[71370.652564] [<c0198a7c>] __mark_inode_dirty+0x5c/0x190
[71370.652571] [<c0126df1>] current_fs_time+0x41/0x50
[71370.652578] [<c918141b>] xfs_write+0x73b/0xc60 [xfs]
[71370.652606] [<c018f7da>] touch_atime+0x7a/0x120
[71370.652613] [<c917cdbe>] xfs_file_aio_write+0x6e/0x80 [xfs]
[71370.652638] [<c017a6d5>] do_sync_write+0xd5/0x120
[71370.652645] [<c01365d0>] autoremove_wake_function+0x0/0x50
[71370.652653] [<c017afbe>] vfs_write+0xbe/0x170
[71370.652659] [<c017a600>] do_sync_write+0x0/0x120
[71370.652665] [<c017b6d1>] sys_write+0x41/0x70
[71370.652671] [<c01055a2>] syscall_call+0x7/0xb
[71370.652678] =======================
[71370.652680] Code: ff e8 6f 92 f6 ff 8b 5c 24 54 e9 00 fe ff ff 8d b6 00 00 00 00 83 ec 10 89 5c 24 08 89 d3 8b 54 24 14 89 74 24 0c 89 c6 8b 40 10 <8b> 00 89 54 24 04 89 f2 89 0c 24 89 d9 e8 08 fb ff ff 85 c0 74
[71370.652762] EIP: [<c019d5a6>] block_prepare_write+0x16/0x40 SS:ESP 0069:c394dd38
[71408.109600] BUG: unable to handle kernel paging request at virtual address 00100100
[71408.109614] printing eip:
[71408.109617] c0151f53
[71408.109624] 07494000 -> *pde = 00000000:5133b001
[71408.109628] 010bc000 -> *pme = 00000000:00000000
[71408.109633] Oops: 0000 [#2]
[71408.109635] SMP
[71408.109642] Modules linked in: ipv6 evdev xfs dm_mirror dm_snapshot dm_mod fuse apparmor commoncap
[71408.109664] CPU: 0
[71408.109665] EIP: 0061:[<c0151f53>] Not tainted VLI
[71408.109666] EFLAGS: 00010006 (2.6.22-14-xen #1)
[71408.109679] EIP is at find_get_pages_tag+0x43/0xa0
[71408.109683] eax: 00000000 ebx: 00000008 ecx: 00000002 edx: 00100100
[71408.109687] esi: c150fe88 edi: c1a400c0 ebp: c150fec0 esp: c150fe24
[71408.109692] ds: 007b es: 007b fs: 00d8 gs: 0000 ss: 0069
[71408.109697] Process pdflush (pid: 90, ti=c150e000 task=c1508a60 task.ti=c150e000)
[71408.109700] Stack: 0000000e 00000000 c150fec0 c150fe80 c150ff70 00000000 c015966a 0000000e
[71408.109716] c150fe88 00000000 00000000 c0157dd6 00000000 0000000e c0157a30 c1a400c0
[71408.109732] 00000000 12950786 00000000 c6d5c0b4 00000000 ffffffff 00000000 00000000
[71408.109746] Call Trace:
[71408.109750] [<c015966a>] pagevec_lookup_tag+0x2a/0x40
[71408.109758] [<c0157dd6>] write_cache_pages+0xf6/0x310
[71408.109764] [<c0157a30>] __writepage+0x0/0x30
[71408.109771] [<c0106e20>] do_IRQ+0x40/0x70
[71408.109778] [<c0158010>] generic_writepages+0x20/0x30
[71408.109784] [<c015804b>] do_writepages+0x2b/0x50
[71408.109789] [<c01980c3>] __writeback_single_inode+0x93/0x3c0
[71408.109796] [<c01057a6>] hypervisor_callback+0x46/0x4e
[71408.109803] [<c019877e>] sync_sb_inodes+0x17e/0x240
[71408.109810] [<c0198c49>] writeback_inodes+0x99/0xd0
[71408.109816] [<c0158715>] wb_kupdate+0x85/0xf0
[71408.109822] [<c0158ab0>] pdflush+0x0/0x260
[71408.109827] [<c0158bf8>] pdflush+0x148/0x260
[71408.109833] [<c0158690>] wb_kupdate+0x0/0xf0
[71408.109839] [<c0136312>] kthread+0x42/0x70
[71408.109845] [<c01362d0>] kthread+0x0/0x70
[71408.109851] [<c0105927>] kernel_thread_helper+0x7/0x10
[71408.109857] =======================
[71408.109860] Code: 00 8d 47 04 8b 54 24 1c 89 5c 24 04 89 14 24 8b 4d 00 89 f2 e8 8f 61 0a 00 85 c0 89 c3 74 32 31 c9 8d b4 26 00 00 00 00 8b 14 8e <8b> 02 25 00 40 02 00 3d 00 40 02 00 74 3e f0 ff 42 04 83 c1 01
[71408.109943] EIP: [<c0151f53>] find_get_pages_tag+0x43/0xa0 SS:ESP 0069:c150fe24

I tried this serveral times and one time it crashed with:

BUG: unable to handle kernel paging request at virtual address 00100100

Adding more memory helped for me.