Re: [nfs] WARNING: CPU: 1 PID: 1392 at kernel/sched/core.c:7300 __might_sleep+0xbd/0xd0()

From: Trond Myklebust
Date: Tue Feb 03 2015 - 13:02:24 EST

On Tue, Feb 3, 2015 at 12:40 PM, Josh Boyer <jwboyer@xxxxxxxxxxxxxxxxx> wrote:
> On Mon, Feb 2, 2015 at 8:43 AM, Trond Myklebust
> <trond.myklebust@xxxxxxxxxxxxxxx> wrote:
>> On Mon, Feb 2, 2015 at 2:33 AM, Christoph Hellwig <hch@xxxxxx> wrote:
>>>
>>> On Sat, Jan 31, 2015 at 07:19:17PM -0800, Fengguang Wu wrote:
>>> > Hi Christoph,
>>> >
>>> > FYI, this patch discloses an 100% reproducible boot warning.
>>> >
>>> > git://git.infradead.org/users/hch/pnfs.git flexfiles+pnfsd
>>> > commit 34c311faa8dcd323907c6075ab24b4d9e3c6dcb0 ("nfs: force version 4.1")
>>>
>>> The branch is just test branch for some new pnfs patches. But the fact
>>> that forcing the protocol version to 4.1 makes your boot fail still seems
>>> like an interesting observation.
>>>
>>> >
>>> > +------------------------------------------------------------------+------------+------------+
>>> > | | 457be31a00 | 34c311faa8 |
>>> > +------------------------------------------------------------------+------------+------------+
>>> > | boot_successes | 20 | 10 |
>>> > | early-boot-hang | 1 | |
>>> > | boot_failures | 0 | 12 |
>>> > | Kernel_panic-not_syncing:Out_of_memory_and_no_killable_processes | 0 | 2 |
>>> > | backtrace:vfs_write | 0 | 2 |
>>> > | backtrace:SyS_write | 0 | 2 |
>>> > | backtrace:populate_rootfs | 0 | 2 |
>>> > | backtrace:kernel_init_freeable | 0 | 2 |
>>> > | WARNING:at_kernel/sched/core.c:#__might_sleep() | 0 | 10 |
>>> > | backtrace:nfs41_callback_svc | 0 | 10 |
>>> > +------------------------------------------------------------------+------------+------------+
>>> >
>>> >
>>> > [ 12.520894] Key type id_resolver registered
>>> > [ 12.522364] Key type id_legacy registered
>>> > [ 12.530530] ------------[ cut here ]------------
>>> > [ 12.532061] WARNING: CPU: 1 PID: 1392 at kernel/sched/core.c:7300 __might_sleep+0xbd/0xd0()
>>> > [ 12.534114] do not call blocking ops when !TASK_RUNNING; state=1 set at [<ffffffff810b63af>] prepare_to_wait+0x2f/0x90
>>> > [ 12.536264] Modules linked in: rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver sg sr_mod cdrom ata_generic pata_acpi parport_pc floppy parport cirrus syscopyarea snd_pcm sysfillrect sysimgblt snd_timer ttm snd drm_kms_helper ata_piix soundcore libata drm pcspkr i2c_piix4
>>> > [ 12.542569] CPU: 1 PID: 1392 Comm: nfsv4.1-svc Not tainted 3.19.0-rc5-wl-ga224126 #1
>>> > [ 12.544509] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
>>> > [ 12.546252] ffffffff81b7a3c0 ffff88007f627bd8 ffffffff818a735f 0000000000003658
>>> > [ 12.548252] ffff88007f627c28 ffff88007f627c18 ffffffff810725ea ffff88007f627bf8
>>> > [ 12.550261] ffffffff81b90e59 00000000000004d9 0000000000000000 0000000000000001
>>> > [ 12.552359] Call Trace:
>>> > [ 12.553868] [<ffffffff818a735f>] dump_stack+0x4c/0x65
>>> > [ 12.555666] [<ffffffff810725ea>] warn_slowpath_common+0x8a/0xc0
>>> > [ 12.557512] [<ffffffff81072666>] warn_slowpath_fmt+0x46/0x50
>>> > [ 12.559338] [<ffffffff810a0104>] ? try_to_wake_up+0x1f4/0x380
>>> > [ 12.561154] [<ffffffff810b63af>] ? prepare_to_wait+0x2f/0x90
>>> > [ 12.562993] [<ffffffff810b63af>] ? prepare_to_wait+0x2f/0x90
>>> > [ 12.564793] [<ffffffff810986fd>] __might_sleep+0xbd/0xd0
>>> > [ 12.566553] [<ffffffff811c81e7>] kmem_cache_alloc_trace+0x1d7/0x250
>>> > [ 12.568383] [<ffffffff81095b0e>] ? groups_alloc+0x3e/0x130
>>> > [ 12.570159] [<ffffffff81095b0e>] groups_alloc+0x3e/0x130
>>> > [ 12.571878] [<ffffffff818781be>] svcauth_unix_accept+0x16e/0x290
>>> > [ 12.573677] [<ffffffff81876f11>] svc_authenticate+0xe1/0xf0
>>> > [ 12.575405] [<ffffffff818738f4>] svc_process_common+0x224/0x680
>>> > [ 12.577184] [<ffffffff818740d4>] bc_svc_process+0x1c4/0x260
>>> > [ 12.578904] [<ffffffffa01c3b64>] nfs41_callback_svc+0x104/0x1b0 [nfsv4]
>>> > [ 12.580752] [<ffffffff810b6790>] ? wait_woken+0xc0/0xc0
>>> > [ 12.582441] [<ffffffffa01c3a60>] ? nfs4_callback_svc+0x60/0x60 [nfsv4]
>>> > [ 12.584268] [<ffffffff81091fcf>] kthread+0xef/0x110
>>> > [ 12.585859] [<ffffffff81091ee0>] ? kthread_create_on_node+0x180/0x180
>>> > [ 12.587572] [<ffffffff818af9fc>] ret_from_fork+0x7c/0xb0
>>> > [ 12.589175] [<ffffffff81091ee0>] ? kthread_create_on_node+0x180/0x180
>>> > [ 12.590895] ---[ end trace 7b39108134f7677c ]---
>>> > RESULT_ROOT=/result/vm-vp-2G/boot/1/debian-x86_64-2015-01-13.cgz/x86_64-rhel/a224126be542547c3d3040d2b4c145c0c024cc04/0
>>>
>> <snip away the looong .config>
>>
>> That warning should hopefully be fixed by the following commit by Jeff:
>> http://git.linux-nfs.org/?p=trondmy/linux-nfs.git;a=commitdiff;h=6ffa30d3f734d4f6b478081dfc09592021028f90
>>
>> I've already pulled it into my linux-next branch.
>
> If that's marked for stable, why wouldn't it go to Linus to get into
> the final 3.19 release?

Even stable patches need soak time. This is something that came up as
a result of a new sleep test that was added to 3.19-rc, so we've lived
with the problem for a while. It is a real bug, so it does need to be
solved, however we can afford to give ourselves an extra week to make
sure that the timeout we're now introducing is not a problem.

--
Trond Myklebust
Linux NFS client maintainer, PrimaryData
trond.myklebust@xxxxxxxxxxxxxxx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/