Re: kernel panic with 2.6.37-rc1 +pnfs

[Benny Halevy <bhalevy@panasas.com>]

On 2010-11-09 16:41, Tigran Mkrtchyan wrote:
> 
> 
> pnfs-all-2.6.37-rc1-2010-11-03 from Banny's tree (git 
> 6a1df873544d146fcdc493034b170879985909e8)
> I am not sure that this is NFS code issue, but anyway. With NFS IO  
> minutes machine becomes unresponcive.
> This prevents us to update our cluster with latest code base.
> 
> Tigran.
> 
> [  293.145013] general protection fault: 0000 [#1] SMP
> [  293.145017] last sysfs file: 
> /sys/devices/pci0000:00/0000:00:0d.0/resource
> [  293.145019] CPU 0
> [  293.145021] Modules linked in: ipv6 af_packet binfmt_misc dm_mirror 
> dm_multipath scsi_dh video output thermal sbs sbshc pci_slot fan 
> container battery lp sg option usb_wwan ac usbserial button thermal_sys 
> parport_pc serio_raw parport tpm_tis tpm tpm_bios e1000 i2c_piix4 
> pata_mpiix dm_region_hash dm_log dm_mod [last unloaded: mperf]
> [  293.145042]
> [  293.145045] Pid: 1953, comm: sadc Not tainted 2.6.37-rc1.pnfs.1 #1 
> /VirtualBox
> [  293.145047] RIP: 0010:[<ffffffff811a2cfa>]  [<ffffffff811a2cfa>] 
> strnlen+0xa/0x30
> [  293.145054] RSP: 0018:ffff88002d4cfcd0  EFLAGS: 00010297
> [  293.145056] RAX: ffffffff814d5cfa RBX: ffffffffffffffff RCX: 
> ffffffffffffffff
> [  293.145058] RDX: fffffffffffffffe RSI: ffffffffffffffff RDI: 
> 9a81ffffffff814d
> [  293.145060] RBP: 9a81ffffffff814d R08: 0000000000000000 R09: 
> ffffffff8165c4a8
> [  293.145062] R10: 00007f73dd7243fb R11: 0000000000000246 R12: 
> 0000000000000000
> [  293.145064] R13: ffff88002d47d000 R14: ffff88002d47c000 R15: 
> 00000000ffffffff
> [  293.145071] FS:  00007f73dd70b6e0(0000) GS:ffff88003fc00000(0000) 
> knlGS:0000000000000000
> [  293.145074] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [  293.145076] CR2: 00007f73dd724000 CR3: 000000002d4a6000 CR4: 
> 00000000000006f0
> [  293.145083] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 
> 0000000000000000
> [  293.145085] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 
> 0000000000000400
> [  293.145088] Process sadc (pid: 1953, threadinfo ffff88002d4ce000, 
> task ffff88003d5a51f0)
> [  293.145091] Stack:
> [  293.145093]  ffffffff811a46bc ffff88002d41e840 ffffffff814d98f9 
> 0000000000000002
> [  293.145096]  ffff88002d47c000 ffff88002d4cfd78 ffff88002d47d000 
> ffffffff814d98fb
> [  293.145100]  ffffffff811a5488 0000000000001000 ffff88002d47c000 
> ffffffffff0a0004
> [  293.145104] Call Trace:
> [  293.145107]  [<ffffffff811a46bc>] ? string+0x4c/0x100
> [  293.145110]  [<ffffffff811a5488>] ? vsnprintf+0x218/0x560
> [  293.145114]  [<ffffffff810efac7>] ? seq_printf+0x67/0xa0
> [  293.145118]  [<ffffffff81043d32>] ? get_online_cpus+0x22/0x50
> [  293.145121]  [<ffffffff81043d82>] ? put_online_cpus+0x22/0x70
> [  293.145125]  [<ffffffff810ac71f>] ? vmstat_start+0x7f/0xa0
> [  293.145128]  [<ffffffff810abeb3>] ? vmstat_show+0x23/0x30
> [  293.145130]  [<ffffffff810eff2f>] ? seq_read+0xaf/0x3a0
> [  293.145134]  [<ffffffff8111e223>] ? proc_reg_read+0x73/0xb0
> [  293.145138]  [<ffffffff810d43ed>] ? vfs_read+0xcd/0x170
> [  293.145141]  [<ffffffff810d4633>] ? sys_read+0x53/0x90
> [  293.145144]  [<ffffffff81002c6b>] ? system_call_fastpath+0x16/0x1b
> [  293.145146] Code: 01 e8 38 10 74 0b 48 83 e8 01 48 39 c5 76 f3 31 c0 
> 5b 5d c3 66 0f 1f 44 00 00 0f 1f 80 00 00 00 00 48 8d 56 ff 48 83 fa ff 
> 74 21 <80> 3f 00 74 1c 48 89 f8 eb 05 80 38 00 74 0e 48 83 ea 01 48 83
> [  293.145171] RIP  [<ffffffff811a2cfa>] strnlen+0xa/0x30
> [  293.145174]  RSP <ffff88002d4cfcd0>
> [  293.145177] ---[ end trace 63b42edd98a8dea9 ]---
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

I guess that ff8b16d vmstat: fix offset calculation on void*
could fix this problem.  It's currently in Linus's master branch
and he hasn't released 2.6.37-rc2 yet.
Can you please try to reproduce the problem you saw with this patch?

Benny