On Wed, 23 Nov 2011 at 11:18, Benjamin Herrenschmidt wrote:
> I'd say only this one liner for now, just don't do slabinfo :-) I just
> want to see whether your network + heavy IO load problem goes away with
> that one patch.
OK, with Christoph's patch applied, 3.2.0-rc2-00274-g6fe4c6d-dirty survives
on this machine, with the disk & cpu workload that caused the machine to
panic w/o the patch. Load was at 4-5 this time, which is expected for this
box. I'll run a few more tests later on, but it seems ok for now.
I couldn't resist and ran "slabinfo" anyway (after the workload!) - the
box survived, nothing was printed in syslog either. Output attached.
Thanks!
Christian.
--- linux-2.6.orig/mm/slub.c 2011-11-21 21:15:41.575673204 -0600
+++ linux-2.6/mm/slub.c 2011-11-21 21:16:33.442336849 -0600
@@ -1969,7 +1969,7 @@
page->pobjects = pobjects;
page->next = oldpage;
- } while (this_cpu_cmpxchg(s->cpu_slab->partial, oldpage, page) != oldpage);
+ } while (irqsafe_cpu_cmpxchg(s->cpu_slab->partial, oldpage, page) != oldpage);
stat(s, CPU_PARTIAL_FREE);
return pobjects;
}
--
BOFH excuse #6:
global warming
Name Objects Objsize Space Slabs/Part/Cpu O/S O %Fr %Ef Flg
:at-0000040 102 36 4.0K 0/0/1 102 0 0 89 *a
:at-0000064 41288 64 4.0M 964/565/15 64 0 57 65 *a
:t-0000008 2048 8 16.3K 4294967078/0/222 512 0 0 100 *
:t-0000016 2624 16 94.2K 4294967069/15/250 256 0 65 44 *
:t-0000024 3279 24 98.3K 19/17/5 170 0 70 80 *
:t-0000032 750 32 28.6K 4294967175/4/128 128 0 57 83 *
:t-0000056 771 56 69.6K 17/7/0 73 0 41 62 *
:t-0000064 5226 64 610.3K 137/101/12 64 0 67 54 *
:t-0000072 112 72 8.1K 4294967253/0/45 56 0 0 98 *
:t-0000088 4013 88 360.4K 69/3/19 46 0 3 97 *
:t-0000096 9184 96 1.0M 239/72/23 42 0 27 82 *
:t-0000104 26 104 4.0K 1/1/0 39 0 100 66 *
:t-0000128 2127 128 512.0K 106/91/19 32 0 72 53 *
:t-0000144 56 144 8.1K 4294967269/0/29 28 0 0 98 *
:t-0000152 430 152 147.4K 11/29/25 26 0 80 44 *
:t-0000160 7000 160 1.1M 268/0/12 25 0 0 97 *A
:t-0000192 2959 192 638.9K 147/24/9 21 0 15 88 *
:t-0000320 9482 320 3.6M 879/167/0 12 0 18 84 *A
:t-0000352 31 352 24.5K 4/4/2 11 0 66 44 *A
:t-0000480 36 480 20.4K 5/1/0 8 0 20 84 *A
:t-0000736 2 728 8.1K 1/1/0 11 1 100 17 *A
:t-0000768 268 768 253.9K 25/9/6 10 1 29 81 *A
:t-0002048 256 2048 573.4K 31/6/4 8 2 17 91 *
:t-0004096 168 4096 720.8K 20/1/2 8 3 4 95 *
anon_vma 2117 88 217.0K 51/4/2 42 0 7 85
arp_cache 6 380 4.0K 1/1/0 10 0 100 55 A
bdev_cache 25 744 24.5K 3/1/0 10 1 33 75 Aa
biovec-128 20 1536 32.7K 4294967287/0/11 10 2 0 93 A
biovec-256 98 3072 360.4K 11/2/0 10 3 18 83 A
blkdev_queue 15 1240 32.7K 2/1/0 13 2 50 56
blkdev_requests 62 216 24.5K 4294967273/3/29 18 0 50 54
dentry 219259 160 36.3M 8842/224/21 25 0 2 96 a
eventpoll_pwq 92 36 20.4K 5/5/0 102 0 100 16
ext4_allocation_context 0 108 4.0K 1/1/0 36 0 100 0 a
ext4_groupinfo_4k 7498 148 1.1M 289/1/0 26 0 0 93 a
ext4_inode_cache 206222 936 199.6M 12177/159/8 17 2 1 10 a
ext4_io_end 13 592 8.1K 0/0/1 13 1 0 93 a
ext4_io_page 512 8 4.0K 0/0/1 512 0 0 100 a
ext4_prealloc_space 17 96 4.0K 1/1/0 42 0 100 39 a
files_cache 90 208 20.4K 4294967288/0/13 18 0 0 91 A
inode_cache 3735 512 2.0M 248/0/1 15 1 0 93 a
jbd2_revoke_record 0 24 4.0K 1/1/0 128 0 100 0 Aa
jbd2_revoke_table 4 12 4.0K 1/1/0 256 0 100 1 a
kmalloc-1024 310 1024 352.2K 36/7/7 8 1 16 90
kmalloc-256 110 256 65.5K 4/13/12 16 0 81 42
kmalloc-512 532 512 282.6K 58/5/11 8 0 7 96
kmalloc-8192 7 8192 98.3K 3/2/0 4 3 66 58
kmem_cache 28 116 4.0K 1/1/0 32 0 100 79 *A
kmem_cache_node 100 60 8.1K 2/1/0 64 0 50 73 *A
mm_struct 70 540 49.1K 4/4/2 15 1 66 76 A
mqueue_inode_cache 1 744 8.1K 1/1/0 10 1 100 9 A
ndisc_cache 2 392 8.1K 1/1/0 19 1 100 9 A
posix_timers_cache 0 136 4.0K 1/1/0 30 0 100 0
proc_inode_cache 1194 544 753.6K 79/24/13 14 1 26 86 a
radix_tree_node 17934 296 6.8M 1669/962/14 13 0 57 77 a
RAWv6 5 896 8.1K 1/1/0 9 1 100 54 A
rpc_inode_cache 12 680 16.3K 2/1/0 11 1 50 49 Aa
sgpool-128 2 2560 32.7K 1/1/0 12 3 100 15 A
sgpool-32 2 640 8.1K 1/1/0 12 1 100 15 A
sgpool-64 2 1280 16.3K 1/1/0 12 2 100 15 A
shmem_inode_cache 700 576 409.6K 42/0/8 14 1 0 98
sighand_cache 141 1356 229.3K 3/3/4 23 3 42 83 A
sock_inode_cache 225 544 155.6K 1/7/18 14 1 36 78 Aa
task_delay_info 159 112 20.4K 4294967295/1/6 36 0 20 86
task_struct 143 2992 524.2K 11/5/5 10 3 31 81
taskstats 1 328 4.0K 1/1/0 12 0 100 8
TCP 59 1480 262.1K 2/6/6 21 3 75 33 A
TCPv6 6 1616 16.3K 1/1/0 10 2 100 59 A
tw_sock_TCP 10 104 4.0K 1/1/0 32 0 100 25 A
UDP 30 744 24.5K 2/0/1 10 1 0 90 A
UDPv6 9 880 8.1K 0/0/1 9 1 0 96 A
xfs_da_state 0 392 4.0K 1/1/0 10 0 100 0
xfs_inode 192270 984 200.9M 24533/1199/0 8 1 4 8 Aa
xfs_log_ticket 1 208 4.0K 1/1/0 19 0 100 5