Re: Issue #5876 : assertion failure in rbd_img_obj_callback()

[Alex Elder <elder@ieee.org>]

On 04/25/2014 06:37 AM, Olivier Bonvalet wrote:
> Le vendredi 04 avril 2014 à 20:57 -0500, Alex Elder a écrit :
>> On 04/04/2014 08:16 PM, Olivier Bonvalet wrote:
>>> Le mardi 25 mars 2014 à 09:39 +0100, Olivier Bonvalet a écrit :
>>>> Hi,
>>>>
>>>> what can/should I do to help fix that problem ?
>>>>
>>>> for now, RBD kernel client hang on : 
>>>>         Assertion failure in rbd_img_obj_callback() at line 2131:
>>>>            rbd_assert(which >= img_request->next_completion);
>>>>
>>>> or on :
>>>>         Assertion failure in rbd_img_obj_callback() at line 2127:
>>>>             rbd_assert(img_request != NULL);
>>>>
>>>>
>>>> I have both case at least once per week, on latest 3.13.5 kernels.
>>>>
>>>> It seems that the problem occurs only on more loaded servers (I have 4
>>>> near same servers, and crash occurs on two of them. If I move the VM,
>>>> crash follows...).
>>>>
>>>> Olivier
>>>>
>>>> --
>>>
>>> Hi,
>>>
>>> so. After some days without any problems, RBD crashed toonight :
>>
>> Unfortunately this could be a symptom of the same sort of race.
>> When a object request is removed from its image request's list
>> the request count gets decremented.  To be honest, all of these
>> assertions in rbd_img_obj_callback() are probably unsafe, at
>> least until I get the patch that does proper reference counting
>> implemented:
>>
>>         rbd_assert(img_request != NULL);
>>         rbd_assert(img_request->obj_request_count > 0);
>>         rbd_assert(which != BAD_WHICH);
>>         rbd_assert(which < img_request->obj_request_count);
>>
>> Until then I think you can avoid this by commenting out those
>> assertions.  I'm afraid there will remain a (smaller) window
>> of opportunity for a problem to occur, but I believe commenting
>> those out will help for now.
>>
>> I'm very sorry you're hitting these.  I'll see if I can get
>> a comprehensive fix this weekend.
>>
>> 					-Alex
>>
> 
> Hi,
> 
> I suppose that I should add :
>     if (img_request == NULL) goto out;
> 
> Right ?

Sure, why not?

To be serious we need to get you a proper fix.  I have one
written (I think I've had it for two weeks) but have been
unable to test it at all.  And this is one I don't want to
just give to a customer to test, I want to make sure it works
before sending it out.

I was hoping we had made the window of vulnerability small
enough that the problem wouldn't occur.  Your new report
shows we're not that lucky.   I'll see what I can do.

					-Alex

> When commenting the asserts I obtain a NULL pointer dereference :
> 
> Apr 25 13:03:15 murmillia kernel: [124049.097927] BUG: unable to handle kernel NULL pointer dereference at 000000000000003c
> Apr 25 13:03:15 murmillia kernel: [124049.098008] IP: [<ffffffff8105d922>] do_raw_spin_lock+0x5/0x22
> Apr 25 13:03:15 murmillia kernel: [124049.098056] PGD 0 
> Apr 25 13:03:15 murmillia kernel: [124049.098091] Oops: 0002 [#1] SMP 
> Apr 25 13:03:15 murmillia kernel: [124049.098133] Modules linked in: cbc rbd libceph xen_gntdev ip6table_mangle ip6t_REJECT ip6table_filter ip6_tables xt_DSCP iptable_mangle xt_LOG xt_physdev ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables xfs libcrc32c bridge loop iTCO_wdt gpio_ich iTCO_vendor_support serio_raw sb_edac edac_core i2c_i801 evdev lpc_ich mfd_core ioatdma shpchp ipmi_si ipmi_msghandler wmi ac button dm_mod hid_generic usbhid hid sg sd_mod crc_t10dif crct10dif_common megaraid_sas ahci libahci isci ehci_pci ehci_hcd libsas usbcore libata igb ixgbe scsi_transport_sas i2c_algo_bit i2c_core usb_common scsi_mod dca ptp pps_core mdio
> Apr 25 13:03:15 murmillia kernel: [124049.098695] CPU: 0 PID: 31669 Comm: kworker/0:0 Not tainted 3.13-dae-dom0 #1
> Apr 25 13:03:15 murmillia kernel: [124049.098739] Hardware name: Supermicro X9DRW-7TPF+/X9DRW-7TPF+, BIOS 2.0a 03/11/2013
> Apr 25 13:03:15 murmillia kernel: [124049.098809] Workqueue: ceph-msgr con_work [libceph]
> Apr 25 13:03:15 murmillia kernel: [124049.098851] task: ffff8802458b38a0 ti: ffff88023cfcc000 task.ti: ffff88023cfcc000
> Apr 25 13:03:15 murmillia kernel: [124049.098916] RIP: e030:[<ffffffff8105d922>]  [<ffffffff8105d922>] do_raw_spin_lock+0x5/0x22
> Apr 25 13:03:15 murmillia kernel: [124049.098987] RSP: e02b:ffff88023cfcdce0  EFLAGS: 00010002
> Apr 25 13:03:15 murmillia kernel: [124049.099026] RAX: 0000000000010000 RBX: ffff88025749a3c8 RCX: 0000000000002201
> Apr 25 13:03:15 murmillia kernel: [124049.099091] RDX: 000000000000003c RSI: ffff88025749a3e0 RDI: 000000000000003c
> Apr 25 13:03:15 murmillia kernel: [124049.099154] RBP: 0000000000000000 R08: 0000000000000001 R09: 0000000000000001
> Apr 25 13:03:15 murmillia kernel: [124049.099218] R10: ffff88024749d07d R11: ffff8802476929f8 R12: ffff880269f6b701
> Apr 25 13:03:15 murmillia kernel: [124049.099281] R13: 00000000ffffffff R14: ffff8802476927c0 R15: 0000000000000000
> Apr 25 13:03:15 murmillia kernel: [124049.099349] FS:  00007f01384088e0(0000) GS:ffff88027fc00000(0000) knlGS:0000000000000000
> Apr 25 13:03:15 murmillia kernel: [124049.099415] CS:  e033 DS: 0000 ES: 0000 CR0: 0000000080050033
> Apr 25 13:03:15 murmillia kernel: [124049.099455] CR2: 000000000000003c CR3: 0000000243dec000 CR4: 0000000000042660
> Apr 25 13:03:15 murmillia kernel: [124049.099519] Stack:
> Apr 25 13:03:15 murmillia kernel: [124049.099549]  ffffffffa032caad 000000000000003c ffff8802476929f8 0000000000002201
> Apr 25 13:03:15 murmillia kernel: [124049.099629]  ffff8802411ea218 ffff8802476927b8 ffff880269f6b718 0000000000000000
> Apr 25 13:03:15 murmillia kernel: [124049.099708]  ffff8802476927c0 0000000000000000 ffffffffa030b69b 0000000000000025
> Apr 25 13:03:15 murmillia kernel: [124049.099786] Call Trace:
> Apr 25 13:03:15 murmillia kernel: [124049.099823]  [<ffffffffa032caad>] ? rbd_img_obj_callback+0x56/0x308 [rbd]
> Apr 25 13:03:15 murmillia kernel: [124049.099871]  [<ffffffffa030b69b>] ? dispatch+0x3e4/0x55e [libceph]
> Apr 25 13:03:15 murmillia kernel: [124049.099915]  [<ffffffffa03060fc>] ? con_work+0xf6e/0x1a65 [libceph]
> Apr 25 13:03:15 murmillia kernel: [124049.099959]  [<ffffffff8100122a>] ? xen_hypercall_xen_version+0xa/0x20
> Apr 25 13:03:15 murmillia kernel: [124049.100004]  [<ffffffff81005959>] ? xen_force_evtchn_callback+0x9/0xa
> Apr 25 13:03:15 murmillia kernel: [124049.100048]  [<ffffffff810484e8>] ? process_one_work+0x15a/0x214
> Apr 25 13:03:15 murmillia kernel: [124049.100100]  [<ffffffff8104896c>] ? worker_thread+0x139/0x1de
> Apr 25 13:03:15 murmillia kernel: [124049.100141]  [<ffffffff81048833>] ? rescuer_thread+0x26e/0x26e
> Apr 25 13:03:15 murmillia kernel: [124049.100183]  [<ffffffff8104d007>] ? kthread+0x9e/0xa6
> Apr 25 13:03:15 murmillia kernel: [124049.100223]  [<ffffffff8104cf69>] ? __kthread_parkme+0x55/0x55
> Apr 25 13:03:15 murmillia kernel: [124049.100268]  [<ffffffff81372a0c>] ? ret_from_fork+0x7c/0xb0
> Apr 25 13:03:15 murmillia kernel: [124049.100309]  [<ffffffff8104cf69>] ? __kthread_parkme+0x55/0x55
> Apr 25 13:03:15 murmillia kernel: [124049.100349] Code: d0 f0 0f b1 0f 39 d0 0f 94 c0 0f b6 c0 c3 31 c0 48 81 ff e8 db 36 81 72 0c 31 c0 48 81 ff af df 36 81 0f 92 c0 c3 b8 00 00 01 00 <f0> 0f c1 07 89 c2 c1 ea 10 66 39 c2 89 d1 74 0c 66 8b 07 66 39 
> Apr 25 13:03:15 murmillia kernel: [124049.100727] RIP  [<ffffffff8105d922>] do_raw_spin_lock+0x5/0x22
> Apr 25 13:03:15 murmillia kernel: [124049.100773]  RSP <ffff88023cfcdce0>
> Apr 25 13:03:15 murmillia kernel: [124049.100807] CR2: 000000000000003c
> Apr 25 13:03:15 murmillia kernel: [124049.101120] ---[ end trace 7f81ace5e0aed716 ]---
> 
> 
> 
> 

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html