Bug #2157
Run out of memory killed by oom-killer
0%
Description
I tested 3.9.952~gf161132~trusty for a few hours and found it killed by the oom-killer.
The dmesg showed as following:
[226256.256727] linuxdvb_fronte invoked oom-killer: gfp_mask=0x280da, order=0, oom_score_adj=0
[226256.256731] linuxdvb_fronte cpuset=/ mems_allowed=0
[226256.256733] CPU: 0 PID: 28937 Comm: linuxdvb_fronte Tainted: PF O 3.13.0-29-generic #53-Ubuntu
[226256.256734] Hardware name: ASUSTeK Computer INC. BM6835_BM6635_BP6335/BM6835_BM6635_BP6335, BIOS 1108 12/05/2013
[226256.256735] 0000000000000000 ffff8801175f1a80 ffffffff8171a214 ffff8800c652dfc0
[226256.256738] ffff8801175f1b08 ffffffff81714b4f 0000000000000000 0000000000000000
[226256.256740] 0000000000000000 0000000000000000 0000000000000000 0000000000000000
[226256.256741] Call Trace:
[226256.256747] [<ffffffff8171a214>] dump_stack+0x45/0x56
[226256.256749] [<ffffffff81714b4f>] dump_header+0x7f/0x1f1
[226256.256752] [<ffffffff81151a7e>] oom_kill_process+0x1ce/0x330
[226256.256755] [<ffffffff812d35e5>] ? security_capable_noaudit+0x15/0x20
[226256.256756] [<ffffffff811521b4>] out_of_memory+0x414/0x450
[226256.256758] [<ffffffff8115843c>] __alloc_pages_nodemask+0xa5c/0xb80
[226256.256761] [<ffffffff8119878a>] alloc_pages_vma+0x9a/0x140
[226256.256763] [<ffffffff811791eb>] handle_mm_fault+0xb2b/0xf10
[226256.256766] [<ffffffff81726164>] __do_page_fault+0x184/0x560
[226256.256769] [<ffffffff8111140c>] ? acct_account_cputime+0x1c/0x20
[226256.256771] [<ffffffff8109d77b>] ? account_user_time+0x8b/0xa0
[226256.256773] [<ffffffff8109dd94>] ? vtime_account_user+0x54/0x60
[226256.256774] [<ffffffff8172655a>] do_page_fault+0x1a/0x70
[226256.256776] [<ffffffff817229c8>] page_fault+0x28/0x30
[226256.256777] Mem-Info:
[226256.256778] Node 0 DMA per-cpu:
[226256.256780] CPU 0: hi: 0, btch: 1 usd: 0
[226256.256780] CPU 1: hi: 0, btch: 1 usd: 0
[226256.256781] CPU 2: hi: 0, btch: 1 usd: 0
[226256.256782] CPU 3: hi: 0, btch: 1 usd: 0
[226256.256782] CPU 4: hi: 0, btch: 1 usd: 0
History
Updated by Chuang Liang-Shing over 10 years ago
The memory leaks problem since have great improved after I pulled down some recently commits. It can survived one more night.