netfront regression in 3.16-rc1

netfront regression in 3.16-rc1

[Boris Ostrovsky]

We are having a couple of regressions in netfront and netback. I'll post 
netback failure in the next message since they appear independent.

Netfront dies on a resume:


[   85.142814] BUG: unable to handle kernel paging request at 
0000000000001540
[   85.143252] IP: [<ffffffff816f0485>] _raw_spin_lock_bh+0x15/0x50
[   85.143680] PGD 2f8a2067 PUD 301cd067 PMD 0
[   85.143788] Oops: 0002 [#1] SMP
[   85.143788] Modules linked in: sg sd_mod crct10dif_generic crc_t10dif 
crct10dif_common dm_multipath dm_mod xen_evtchn iscsi_boot_sysfs 
iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi scsi_mod libcrc32c 
crc32c_generic fbcon tileblit radeon font bitblit softcursor ttm 
drm_kms_helper x86_pkg_temp_thermal crc32c_intel xen_blkfront 
xen_netfront xen_fbfront fb_sys_fops sysimgblt sysfillrect syscopyarea 
xen_kbdfront xenfs xen_privcmd
[   85.143788] CPU: 0 PID: 31 Comm: xenwatch Not tainted 
3.16.0-rc1upstream-00003-g7074232-dirty #1
[   85.143788] task: ffff88003ce8f550 ti: ffff88003ce9c000 task.ti: 
ffff88003ce9c000
[   85.143788] RIP: e030:[<ffffffff816f0485>] [<ffffffff816f0485>] 
_raw_spin_lock_bh+0x15/0x50
[   85.143788] RSP: e02b:ffff88003ce9fc98  EFLAGS: 00010282
[   85.143788] RAX: 0000000000020000 RBX: 0000000000000000 RCX: 
0000000000000000
[   85.143788] RDX: 0000000000000000 RSI: ffffffff811d39a0 RDI: 
0000000000001540
[   85.143788] RBP: ffff88003ce9fc98 R08: ffff88002f942890 R09: 
ffff88002f942890
[   85.143788] R10: 0000000000000000 R11: 0000000000000001 R12: 
0000000000000001
[   85.143788] R13: 0000000000000000 R14: ffffffffa003c038 R15: 
ffffffff814049e0
[   85.143788] FS:  00007f89ba258700(0000) GS:ffff88003f600000(0000) 
knlGS:0000000000000000
[   85.143788] CS:  e033 DS: 0000 ES: 0000 CR0: 0000000080050033
[   85.143788] CR2: 0000000000001540 CR3: 00000000302f4000 CR4: 
0000000000042660
[   85.143788] Stack:
[   85.143788]  ffff88003ce9fcc8 ffffffffa0036c03 ffff88003ce9fcb8 
ffffffff814071e7
[   85.143788]  ffff88003cd68c00 ffff88003cd68c40 ffff88003ce9fcd8 
ffffffffa0036d80
[   85.143788]  ffff88003ce9fd08 ffffffff81404a38 ffff88003cd68c40 
0000000000000000
[   85.143788] Call Trace:
[   85.143788]  [<ffffffffa0036c03>] 
xennet_disconnect_backend+0x83/0x1f0 [xen_netfront]
[   85.143788]  [<ffffffff814071e7>] ? read_backend_details+0x17/0x20
[   85.143788]  [<ffffffffa0036d80>] netfront_resume+0x10/0x20 
[xen_netfront]
[   85.143788]  [<ffffffff81404a38>] xenbus_dev_resume+0x58/0x120
[   85.143788]  [<ffffffff81486cec>] dpm_run_callback+0x3c/0x160
[   85.143788]  [<ffffffff816f052b>] ? _raw_spin_unlock_irqrestore+0x1b/0x70
[   85.143788]  [<ffffffff8148794d>] device_resume+0xfd/0x1f0
[   85.143788]  [<ffffffff814888e9>] dpm_resume+0x129/0x330
[   85.143788]  [<ffffffff814034c0>] ? xenbus_transaction_start+0x60/0x60
[   85.143788]  [<ffffffff814891d0>] dpm_resume_end+0x10/0x20
[   85.143788]  [<ffffffff813fe268>] do_suspend+0xa8/0x120
[   85.143788]  [<ffffffff813fe402>] shutdown_handler+0x122/0x160
[   85.143788]  [<ffffffff810cd7d7>] ? __might_sleep+0xd7/0x100
[   85.143788]  [<ffffffff81403562>] xenwatch_thread+0xa2/0x160
[   85.143788]  [<ffffffff810e39e0>] ? bit_waitqueue+0xd0/0xd0
[   85.143788]  [<ffffffff814034c0>] ? xenbus_transaction_start+0x60/0x60
[   85.143788]  [<ffffffff810c3fde>] kthread+0xce/0xf0
[   85.143788]  [<ffffffff8103c359>] ? xen_end_context_switch+0x19/0x20
[   85.143788]  [<ffffffff810c3f10>] ? 
kthread_freezable_should_stop+0x80/0x80
[   85.143788]  [<ffffffff816f0abc>] ret_from_fork+0x7c/0xb0
[   85.143788]  [<ffffffff810c3f10>] ? 
kthread_freezable_should_stop+0x80/0x80
[   85.143788] Code: f1 ff 14 25 e0 e9 c2 81 eb e3 66 66 66 2e 0f 1f 84 
00 00 00 00 00 55 65 81 04 25 d4 c7 00 00 01 02 00 00 48 89 e5 b8 00 00 
02 00 <f0> 0f c1 07 89 c2 c1 ea 10 66 39 c2 75 02 c9 c3 83 e2 fe 0f b7
[   85.143788] RIP  [<ffffffff816f0485>] _raw_spin_lock_bh+0x15/0x50
[   85.143788]  RSP <ffff88003ce9fc98>
[   85.143788] CR2: 0000000000001540
[   85.143788] ---[ end trace ad465b8d7766ef51 ]---
[   85.143788] Kernel panic - not syncing: Fatal exception in interrupt
[   85.143788] Kernel Offset: 0x0 from 0xffffffff81000000 (relocation 
range: 0xffffffff80000000-0xffffffff9fffffff)


xennet_disconnect_backend() doesn't look right.

-boris

Re: netfront regression in 3.16-rc1

[David Vrabel]

On 18/06/14 02:57, Boris Ostrovsky wrote:
> We are having a couple of regressions in netfront and netback. I'll post
> netback failure in the next message since they appear independent.

I've posted a series fixing this and other netfront resume problems.
Can you give it a test?

David

Re: netfront regression in 3.16-rc1

[Boris Ostrovsky]

On 06/18/2014 05:51 AM, David Vrabel wrote:
> On 18/06/14 02:57, Boris Ostrovsky wrote:
>> We are having a couple of regressions in netfront and netback. I'll post
>> netback failure in the next message since they appear independent.
> I've posted a series fixing this and other netfront resume problems.
> Can you give it a test?

Netfront works fine (I somehow thought that these were also fixing the 
netback issue but that's apparently not the case).


-boris