[openib-general] Re: oops with user level verbs

[Michael S. Tsirkin]

Quoting r. Michael S. Tsirkin <mst at mellanox.co.il>:
> Subject: oops with user level verbs
> 
> Hi, Roland!
> On one of my machines, I am often getting this oops when running
> the pingpong example.
> 
> Feb 27 15:18:40 swlab156 kernel: Unable to handle kernel paging request at 000030616368746d RIP: 
> Feb 27 15:18:40 swlab156 kernel: [<000030616368746d>]
> Feb 27 15:18:40 swlab156 kernel: PML4 0 
> Feb 27 15:18:40 swlab156 kernel: Oops: 0000 [1] SMP 
> Feb 27 15:18:40 swlab156 kernel: CPU 1 
> Feb 27 15:18:40 swlab156 kernel: Modules linked in: ib_uverbs ib_umad autofs usbserial parport_pc lp parport joydev st sr_mod floppy thermal processor fan button battery ipv6 ac ib_mthca ib_mad ib_core e1000 i2c_i801 i2c_core hw_random ehci_hcd uhci_hcd usbcore evdev dm_mod ext3 jbd aic79xx ata_piix libata sd_mod scsi_mod
> Feb 27 15:18:40 swlab156 kernel: Pid: 12588, comm: ibv_pingpong Not tainted 2.6.10-openib
> Feb 27 15:18:40 swlab156 kernel: RIP: 0010:[<000030616368746d>] [<000030616368746d>]
> Feb 27 15:18:40 swlab156 kernel: RSP: 0018:0000010170adde90  EFLAGS: 00010212
> Feb 27 15:18:40 swlab156 kernel: RAX: 000001017479b800 RBX: 0000010175c04618 RCX: 0000000000000000
> Feb 27 15:18:40 swlab156 kernel: RDX: 00000101736d8880 RSI: 0000010175c04618 RDI: 0000010170985480
> Feb 27 15:18:40 swlab156 kernel: RBP: 000001017f705cc0 R08: 0000010170addf08 R09: 0000000000000000
> Feb 27 15:18:40 swlab156 kernel: R10: 0000010175c04618 R11: 00000000000000b0 R12: 0000000000001000
> Feb 27 15:18:40 swlab156 kernel: R13: 0000002a95dba000 R14: 00000000ffffffea R15: 00000100bffaf6c0
> Feb 27 15:18:41 swlab156 kernel: FS:  0000002a95cb50a0(0000) GS:ffffffff8045bc00(0000) knlGS:0000000000000000
> Feb 27 15:18:41 swlab156 kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> Feb 27 15:18:41 swlab156 kernel: CR2: 000030616368746d CR3: 000000000b14a000 CR4: 00000000000006e0
> Feb 27 15:18:41 swlab156 kernel: Process ibv_pingpong (pid: 12588, threadinfo 0000010170adc000, task 0000010174a94a40)
> Feb 27 15:18:41 swlab156 kernel: Stack: ffffffffa01db022 0000000000000212 ffffffff80164cb5 0000000000000000 
> Feb 27 15:18:41 swlab156 kernel:        0000000000000018 0000007fbfffee20 00000101736d8880 0000000000000000 
> Feb 27 15:18:41 swlab156 kernel:        00000001bfffee20 001000fa00000000 
> Feb 27 15:18:41 swlab156 kernel: Call Trace:<ffffffffa01db022>{:ib_uverbs:ib_uverbs_mmap+32} <ffffffff80164cb5>{do_mmap_pgoff+1330} 
> Feb 27 15:18:41 swlab156 kernel:        <ffffffff80114584>{sys_mmap+142} <ffffffff8010e226>{system_call+126} 
> Feb 27 15:18:41 swlab156 kernel:        
> Feb 27 15:18:41 swlab156 kernel: 
> Feb 27 15:18:41 swlab156 kernel: Code:  Bad RIP value.
> Feb 27 15:18:41 swlab156 kernel: RIP [<000030616368746d>] RSP <0000010170adde90>
> Feb 27 15:18:41 swlab156 kernel: CR2: 000030616368746d
> 

Happens now every time for me, but somewhat surprisingly not on
another, identical box.

-- 
MST - Michael S. Tsirkin