From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from aserp1040.oracle.com ([141.146.126.69]:43927 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755901AbbCCLFu (ORCPT ); Tue, 3 Mar 2015 06:05:50 -0500 Date: Tue, 3 Mar 2015 19:05:43 +0800 From: Liu Bo To: Marcel Ritter Cc: linux-btrfs Subject: Re: Regression: kernel 4.0.0-rc1 - soft lockups Message-ID: <20150303110542.GC18292@localhost.localdomain> Reply-To: bo.li.liu@oracle.com References: <20150303063725.GC30093@localhost.localdomain> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Tue, Mar 03, 2015 at 08:31:10AM +0100, Marcel Ritter wrote: > Hi, > > yes it is reproducible. > > Just creating a new btrfs filesystem (14 disks, data/mdata raid6, > latest git btrfs-progs) > and mounting this filesystems causes the system to hang (I think I once even got > it mounted, but it did hang shortly after when dd started writing to it). > > I just ran some quick tests and (at least at first sight) it looks > like the raid5/6 > code may cause the trouble: > > I created different btrfs filesystem types, mounted them and (if possible) > did a big "dd" on the filesystem: > > mkfs.btrfs /dev/cciss/c1d* -m raid0 -d raid0 -f -> no problem (only short test) > mkfs.btrfs /dev/cciss/c1d* -m raid1 -d raid1 -f -> no problem (only short test) > mkfs.btrfs /dev/cciss/c1d* -m raid5 -d raid5 -f -> (almost) instant hang > mkfs.btrfs /dev/cciss/c1d* -m raid6 -d raid6 -f -> (almost) instant > hang (standard test) > > Once the machine is up again I'll do some more testing (variing the combination > of data and mdata raid levels) Hmm, just FYI, raid5&6 works good on my box with 4.0.0 rc1. Thanks, -liubo > > Bye, > Marcel > > > 2015-03-03 7:37 GMT+01:00 Liu Bo : > > On Tue, Mar 03, 2015 at 07:02:01AM +0100, Marcel Ritter wrote: > >> Hi, > >> > >> yesterday I did a kernel update on my btrfs test system (Ubuntu > >> 14.04.2) from custom-build kernel 3.19-rc6 to 4.0.0-rc1. > >> > >> Almost instantly after starting my test script, the system got stuck > >> with soft lockups (the machine was running the very same test for > >> weeks on the old kernel without problems, > >> basically doing massive streaming i/o on a raid6 btrfs volume): > >> > >> I found 2 types of messages in the logs: > >> > >> one btrfs related: > >> > >> [34165.540004] INFO: rcu_sched detected stalls on CPUs/tasks: { 3 7} > >> (detected by 6, t=6990777 jiffies, g=67455, c=67454, q=0) > >> [34165.540004] Task dump for CPU 3: > >> [34165.540004] mount D ffff8803ed266000 0 15156 15110 0x00000000 > >> [34165.540004] 0000000000000158 0000000000000014 ffff8803ecc13718 > >> ffff8803ecc136d8 > >> [34165.540004] ffffffff8106075a 0000000000000000 0000000000000002 > >> 0000000000000000 > >> [34165.540004] 00000000ecc13728 ffff8803eb603128 0000000000000000 > >> 0000000000000000 > >> [34165.540004] Call Trace: > >> [34165.540004] [] ? __do_page_fault+0x2fa/0x440 > >> [34165.540004] [] ? do_page_fault+0x31/0x70 > >> [34165.540004] [] ? page_fault+0x28/0x30 > >> [34165.540004] [] ? pick_next_task_fair+0x53e/0x880 > >> [34165.540004] [] ? pick_next_task_fair+0x53e/0x880 > >> [34165.540004] [] ? dequeue_task+0x5c/0x80 > >> [34165.540004] [] ? __schedule+0xf3/0x960 > >> [34165.540004] [] ? schedule+0x37/0x90 > >> [34165.540004] [] ? > >> btrfs_start_ordered_extent+0xd5/0x110 [btrfs] > >> [34165.540004] [] ? prepare_to_wait_event+0x110/0x110 > >> [34165.540004] [] ? > >> btrfs_wait_ordered_range+0xc4/0x120 [btrfs] > >> [34165.540004] [] ? > >> __btrfs_write_out_cache+0x378/0x470 [btrfs] > >> [34165.540004] [] ? btrfs_write_out_cache+0x9a/0x100 [btrfs] > >> [34165.540004] [] ? > >> btrfs_write_dirty_block_groups+0x159/0x560 [btrfs] > >> [34165.540004] [] ? commit_cowonly_roots+0x18d/0x2a4 [btrfs] > >> [34165.540004] [] ? > >> btrfs_commit_transaction+0x521/0xa50 [btrfs] > >> [34165.540004] [] ? btrfs_create_uuid_tree+0x5e/0x110 [btrfs] > >> [34165.540004] [] ? open_ctree+0x1dff/0x2200 [btrfs] > >> [34165.540004] [] ? btrfs_mount+0x75e/0x8f0 [btrfs] > >> [34165.540004] [] ? mount_fs+0x39/0x180 > >> [34165.540004] [] ? __alloc_percpu+0x15/0x20 > >> [34165.540004] [] ? vfs_kern_mount+0x6b/0x120 > >> [34165.540004] [] ? do_mount+0x204/0xb30 > >> [34165.540004] [] ? SyS_mount+0x8b/0xe0 > >> [34165.540004] [] ? system_call_fastpath+0x16/0x1b > >> [34165.540004] Task dump for CPU 7: > >> [34165.540004] kworker/u16:1 R running task 0 14518 2 0x00000008 > >> [34165.540004] Workqueue: btrfs-freespace-write > >> btrfs_freespace_write_helper [btrfs] > >> [34165.540004] 0000000000000200 ffff8803eac6fdf8 ffffffffa08ac242 > >> ffff8803eac6fe48 > >> [34165.540004] ffffffff8108b64f 00000000f1091400 0000000000000000 > >> ffff8803eca58000 > >> [34165.540004] ffff8803ea9ed3c0 ffff8803f1091418 ffff8803f1091400 > >> ffff8803eca58000 > >> [34165.540004] Call Trace: > >> [34165.540004] [] ? > >> btrfs_freespace_write_helper+0x12/0x20 [btrfs] > >> [34165.540004] [] ? process_one_work+0x14f/0x420 > >> [34165.540004] [] ? worker_thread+0x118/0x510 > >> [34165.540004] [] ? rescuer_thread+0x3d0/0x3d0 > >> [34165.540004] [] ? kthread+0xd2/0xf0 > >> [34165.540004] [] ? kthread_create_on_node+0x180/0x180 > >> [34165.540004] [] ? ret_from_fork+0x7c/0xb0 > >> [34165.540004] [] ? kthread_create_on_node+0x180/0x180 > >> > >> > >> and one general (related to "native_flush_tlb_other": > >> > >> [34152.604004] NMI watchdog: BUG: soft lockup - CPU#6 stuck for 23s! > >> [rs:main Q:Reg:490] > >> [34152.604004] Modules linked in: btrfs(E) xor(E) radeon(E) ttm(E) > >> drm_kms_helper(E) kvm(E) drm(E) raid6_pq(E) i2c_algo_bit(E) ipmi_si > >> (E) amd64_edac_mod(E) serio_raw(E) hpilo(E) hpwdt(E) edac_core(E) > >> shpchp(E) k8temp(E) mac_hid(E) edac_mce_amd(E) nfsd(E) auth_rpcgss(E > >> ) nfs_acl(E) nfs(E) lockd(E) grace(E) sunrpc(E) fscache(E) lp(E) > >> parport(E) hpsa(E) pata_acpi(E) hid_generic(E) psmouse(E) usbhid(E) b > >> nx2(E) cciss(E) hid(E) pata_amd(E) > >> [34152.604004] CPU: 6 PID: 490 Comm: rs:main Q:Reg Tainted: G D W > >> EL 4.0.0-rc1-custom #1 > >> [34152.604004] Hardware name: HP ProLiant DL585 G2 , BIOS A07 05/02/2011 > >> [34152.604004] task: ffff8803eecd9910 ti: ffff8803ecb30000 task.ti: > >> ffff8803ecb30000 > >> [34152.604004] RIP: 0010:[] [] > >> smp_call_function_many+0x20a/0x270 > >> [34152.604004] RSP: 0018:ffff8803ecb33cf8 EFLAGS: 00000202 > >> [34152.604004] RAX: 0000000000000000 RBX: ffffffff81cdd140 RCX: ffff8803ffc19700 > >> [34152.604004] RDX: 0000000000000000 RSI: 0000000000000100 RDI: 0000000000000000 > >> [34152.604004] RBP: ffff8803ecb33d38 R08: ffff8803ffd961c8 R09: 0000000000000004 > >> [34152.604004] R10: 0000000000000004 R11: 0000000000000246 R12: 0000000000000000 > >> [34152.604004] R13: ffff880300000040 R14: ffff8803ecb33ca0 R15: ffff8803ecb33ca8 > >> [34152.604004] FS: 00007f9cf6dae700(0000) GS:ffff8803ffd80000(0000) > >> knlGS:0000000000000000 > >> [34152.672920] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > >> [34152.672920] CR2: 00007f9ce80091a8 CR3: 00000003e50fe000 CR4: 00000000000007e0 > >> [34152.672920] Stack: > >> [34152.672920] 00000001ecb33de8 0000000000016180 00007f9cf0024fff > >> ffff8803eb726900 > >> [34152.672920] ffff8803eb726bd0 00007f9cf0025000 00007f9cf0021000 > >> 0000000000000004 > >> [34152.672920] ffff8803ecb33d68 ffffffff8106722e ffff8803ecb33d68 > >> ffff8803eb726900 > >> [34152.672920] Call Trace: > >> [34152.672920] [] native_flush_tlb_others+0x2e/0x30 > >> [34152.672920] [] flush_tlb_mm_range+0x64/0x170 > >> [34152.672920] [] tlb_flush_mmu_tlbonly+0x7e/0xe0 > >> [34152.672920] [] tlb_finish_mmu+0x14/0x50 > >> [34152.672920] [] zap_page_range+0xca/0x100 > >> [34152.672920] [] SyS_madvise+0x363/0x790 > >> [34152.672920] [] system_call_fastpath+0x16/0x1b > >> [34152.672920] Code: 9d 5c 2b 00 3b 05 7b b2 c2 00 89 c2 0f 8d 83 fe > >> ff ff 48 98 49 8b 4d 00 48 03 0c c5 40 b1 d1 81 f6 41 18 01 74 cb > >> 0f 1f 00 f3 90 41 18 01 75 f8 eb be 0f b6 4d c4 4c 89 fa 4c 89 f6 44 89 ef > >> > >> So I'm not totally sure if this is a btrfs problem, or something else > >> got broken in 4.0.0-rc1. > >> > >> Maybe someone can have a look. > >> > >> If you need more information just let me know. > > > > Is it reproducible? > > > > From the stacks about btrfs, it's been stopped at mount stage, so it's likely to > > be unrelated to btrfs. > > > > Thanks, > > > > -liubo > > > >> > >> Bye, > >> Marcel > >> -- > >> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in > >> the body of a message to majordomo@vger.kernel.org > >> More majordomo info at http://vger.kernel.org/majordomo-info.html