LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* Opps in 2.6.21-rc4 nfsd
@ 2007-03-22 19:24 Phy Prabab
  2007-03-23 15:29 ` Adrian Bunk
  2007-03-23 17:47 ` Chuck Ebbert
  0 siblings, 2 replies; 7+ messages in thread
From: Phy Prabab @ 2007-03-22 19:24 UTC (permalink / raw)
  To: linux-kernel

Hello,

I have a recent oopps on a small file server I have running.  The
machine in question is a dual, dual core AMD Opteron 2.2GHz w/4G RAM,
3Ware 9650SE (14 drive RAID 10), using XFS, LVM2 (latest), exported
via knfsd.  Kernel is 2.6.21-rc4 64b.  Here is the bt:

Mar 22 12:09:01 rcfs05 kernel: Modules linked in:
Mar 22 12:09:01 rcfs05 kernel: Pid: 3102, comm: nfsd Not tainted
2.6.21-rc4-smp #3
Mar 22 12:09:01 rcfs05 kernel: RIP: 0010:[<ffffffff80350613>]
[<ffffffff80350613>] fh_compose+0x3e3/0x4d0
Mar 22 12:09:01 rcfs05 kernel: RSP: 0018:ffff8101197fdd70  EFLAGS: 00010297
Mar 22 12:09:01 rcfs05 kernel: RAX: 0000000000000006 RBX:
000000000ee00004 RCX: 0000000000000000
Mar 22 12:09:01 rcfs05 kernel: RDX: 0000000000000018 RSI:
ffff81011975a940 RDI: 0000000000000080
Mar 22 12:09:01 rcfs05 kernel: RBP: 0000000000000001 R08:
00000000ffffffff R09: 00002ba24f06ca50
Mar 22 12:09:01 rcfs05 kernel: R10: ffff81011975a808 R11:
ffffffff8022f500 R12: ffff81011975a938
Mar 22 12:09:01 rcfs05 kernel: R13: 0000000000000006 R14:
ffff81010eaab0c0 R15: ffff8100cc714940
Mar 22 12:09:01 rcfs05 kernel: FS:  00002b020c71b6d0(0000)
GS:ffff81011dcf56c0(0000) knlGS:0000000000000000
Mar 22 12:09:01 rcfs05 kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000 CR3:
00000001173d1000 CR4: 00000000000006e0
Mar 22 12:09:01 rcfs05 kernel: Process nfsd (pid: 3102, threadinfo
ffff8101197fc000, task ffff81011b7ef020)
Mar 22 12:09:01 rcfs05 kernel: Stack:  ffff81010fa7b868
ffffffff80239d8a 0000000000000286 ffff810111aee5e8
Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
ffff810105af2890 000000000eaab0c0
Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
ffff810105af2890 000000000000000a
Mar 22 12:09:01 rcfs05 kernel: Call Trace:
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80239d8a>] __lookup_hash+0x7a/0x160
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff803547a5>] nfsd_lookup+0x435/0x4d0
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8035aaa8>]
nfsd3_proc_lookup+0xe8/0x110
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034d69d>] nfsd_dispatch+0xfd/0x1f0
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80608094>] svc_process+0x3e4/0x730
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8026bbf2>] __down_read+0x12/0xa2
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034dd00>] nfsd+0x1a0/0x2d0
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f38>] child_rip+0xa/0x12
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034db60>] nfsd+0x0/0x2d0
Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f2e>] child_rip+0x0/0x12
Mar 22 12:09:01 rcfs05 kernel:
Mar 22 12:09:01 rcfs05 kernel:
Mar 22 12:09:01 rcfs05 kernel: Code: 48 8b 01 ba 10 00 00 00 bf 10 00
00 00 48 89 06 48 8b 41 08
Mar 22 12:09:01 rcfs05 kernel: RIP  [<ffffffff80350613>] fh_compose+0x3e3/0x4d0
Mar 22 12:09:01 rcfs05 kernel:  RSP <ffff8101197fdd70>
Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000

Once this oops happens, the machine is still usable, however NFS
crawls to a halt (in fact one of the machines with a mount point of
this machine, failed to respond anymore - I assume that thread is the
one that ooops'ed).

Any more information or debugging, please let me know.

TIA!
Phy

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Opps in 2.6.21-rc4 nfsd
  2007-03-22 19:24 Opps in 2.6.21-rc4 nfsd Phy Prabab
@ 2007-03-23 15:29 ` Adrian Bunk
  2007-03-23 15:45   ` Phy Prabab
  2007-03-23 17:47 ` Chuck Ebbert
  1 sibling, 1 reply; 7+ messages in thread
From: Adrian Bunk @ 2007-03-23 15:29 UTC (permalink / raw)
  To: Phy Prabab, neilb; +Cc: linux-kernel, nfs

On Thu, Mar 22, 2007 at 12:24:12PM -0700, Phy Prabab wrote:
> Hello,
> 
> I have a recent oopps on a small file server I have running.  The
> machine in question is a dual, dual core AMD Opteron 2.2GHz w/4G RAM,
> 3Ware 9650SE (14 drive RAID 10), using XFS, LVM2 (latest), exported
> via knfsd.  Kernel is 2.6.21-rc4 64b.  Here is the bt:
> 
> Mar 22 12:09:01 rcfs05 kernel: Modules linked in:
> Mar 22 12:09:01 rcfs05 kernel: Pid: 3102, comm: nfsd Not tainted
> 2.6.21-rc4-smp #3
> Mar 22 12:09:01 rcfs05 kernel: RIP: 0010:[<ffffffff80350613>]
> [<ffffffff80350613>] fh_compose+0x3e3/0x4d0
> Mar 22 12:09:01 rcfs05 kernel: RSP: 0018:ffff8101197fdd70  EFLAGS: 00010297
> Mar 22 12:09:01 rcfs05 kernel: RAX: 0000000000000006 RBX:
> 000000000ee00004 RCX: 0000000000000000
> Mar 22 12:09:01 rcfs05 kernel: RDX: 0000000000000018 RSI:
> ffff81011975a940 RDI: 0000000000000080
> Mar 22 12:09:01 rcfs05 kernel: RBP: 0000000000000001 R08:
> 00000000ffffffff R09: 00002ba24f06ca50
> Mar 22 12:09:01 rcfs05 kernel: R10: ffff81011975a808 R11:
> ffffffff8022f500 R12: ffff81011975a938
> Mar 22 12:09:01 rcfs05 kernel: R13: 0000000000000006 R14:
> ffff81010eaab0c0 R15: ffff8100cc714940
> Mar 22 12:09:01 rcfs05 kernel: FS:  00002b020c71b6d0(0000)
> GS:ffff81011dcf56c0(0000) knlGS:0000000000000000
> Mar 22 12:09:01 rcfs05 kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 
> 000000008005003b
> Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000 CR3:
> 00000001173d1000 CR4: 00000000000006e0
> Mar 22 12:09:01 rcfs05 kernel: Process nfsd (pid: 3102, threadinfo
> ffff8101197fc000, task ffff81011b7ef020)
> Mar 22 12:09:01 rcfs05 kernel: Stack:  ffff81010fa7b868
> ffffffff80239d8a 0000000000000286 ffff810111aee5e8
> Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
> ffff810105af2890 000000000eaab0c0
> Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
> ffff810105af2890 000000000000000a
> Mar 22 12:09:01 rcfs05 kernel: Call Trace:
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80239d8a>] 
> __lookup_hash+0x7a/0x160
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff803547a5>] nfsd_lookup+0x435/0x4d0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8035aaa8>]
> nfsd3_proc_lookup+0xe8/0x110
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034d69d>] 
> nfsd_dispatch+0xfd/0x1f0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80608094>] svc_process+0x3e4/0x730
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8026bbf2>] __down_read+0x12/0xa2
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034dd00>] nfsd+0x1a0/0x2d0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f38>] child_rip+0xa/0x12
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034db60>] nfsd+0x0/0x2d0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f2e>] child_rip+0x0/0x12
> Mar 22 12:09:01 rcfs05 kernel:
> Mar 22 12:09:01 rcfs05 kernel:
> Mar 22 12:09:01 rcfs05 kernel: Code: 48 8b 01 ba 10 00 00 00 bf 10 00
> 00 00 48 89 06 48 8b 41 08
> Mar 22 12:09:01 rcfs05 kernel: RIP  [<ffffffff80350613>] 
> fh_compose+0x3e3/0x4d0
> Mar 22 12:09:01 rcfs05 kernel:  RSP <ffff8101197fdd70>
> Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000
> 
> Once this oops happens, the machine is still usable, however NFS
> crawls to a halt (in fact one of the machines with a mount point of
> this machine, failed to respond anymore - I assume that thread is the
> one that ooops'ed).
> 
> Any more information or debugging, please let me know.


Does it work with kernel 2.6.20?


> TIA!
> Phy

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Opps in 2.6.21-rc4 nfsd
  2007-03-23 15:29 ` Adrian Bunk
@ 2007-03-23 15:45   ` Phy Prabab
  2007-03-26  0:50     ` Neil Brown
  0 siblings, 1 reply; 7+ messages in thread
From: Phy Prabab @ 2007-03-23 15:45 UTC (permalink / raw)
  To: Adrian Bunk; +Cc: neilb, linux-kernel, nfs

Adrian et al,

Under 2.6.20.2+3 I get a similar crash and more frequently.  At least
under 2.6.21rc4 I only get this about once every two hours under heavy
nfs testing.

Let me know if I can provide more debugging or more information.

Thanks,
Phy

On 3/23/07, Adrian Bunk <bunk@stusta.de> wrote:
> On Thu, Mar 22, 2007 at 12:24:12PM -0700, Phy Prabab wrote:
> > Hello,
> >
> > I have a recent oopps on a small file server I have running.  The
> > machine in question is a dual, dual core AMD Opteron 2.2GHz w/4G RAM,
> > 3Ware 9650SE (14 drive RAID 10), using XFS, LVM2 (latest), exported
> > via knfsd.  Kernel is 2.6.21-rc4 64b.  Here is the bt:
> >
> > Mar 22 12:09:01 rcfs05 kernel: Modules linked in:
> > Mar 22 12:09:01 rcfs05 kernel: Pid: 3102, comm: nfsd Not tainted
> > 2.6.21-rc4-smp #3
> > Mar 22 12:09:01 rcfs05 kernel: RIP: 0010:[<ffffffff80350613>]
> > [<ffffffff80350613>] fh_compose+0x3e3/0x4d0
> > Mar 22 12:09:01 rcfs05 kernel: RSP: 0018:ffff8101197fdd70  EFLAGS: 00010297
> > Mar 22 12:09:01 rcfs05 kernel: RAX: 0000000000000006 RBX:
> > 000000000ee00004 RCX: 0000000000000000
> > Mar 22 12:09:01 rcfs05 kernel: RDX: 0000000000000018 RSI:
> > ffff81011975a940 RDI: 0000000000000080
> > Mar 22 12:09:01 rcfs05 kernel: RBP: 0000000000000001 R08:
> > 00000000ffffffff R09: 00002ba24f06ca50
> > Mar 22 12:09:01 rcfs05 kernel: R10: ffff81011975a808 R11:
> > ffffffff8022f500 R12: ffff81011975a938
> > Mar 22 12:09:01 rcfs05 kernel: R13: 0000000000000006 R14:
> > ffff81010eaab0c0 R15: ffff8100cc714940
> > Mar 22 12:09:01 rcfs05 kernel: FS:  00002b020c71b6d0(0000)
> > GS:ffff81011dcf56c0(0000) knlGS:0000000000000000
> > Mar 22 12:09:01 rcfs05 kernel: CS:  0010 DS: 0000 ES: 0000 CR0:
> > 000000008005003b
> > Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000 CR3:
> > 00000001173d1000 CR4: 00000000000006e0
> > Mar 22 12:09:01 rcfs05 kernel: Process nfsd (pid: 3102, threadinfo
> > ffff8101197fc000, task ffff81011b7ef020)
> > Mar 22 12:09:01 rcfs05 kernel: Stack:  ffff81010fa7b868
> > ffffffff80239d8a 0000000000000286 ffff810111aee5e8
> > Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
> > ffff810105af2890 000000000eaab0c0
> > Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
> > ffff810105af2890 000000000000000a
> > Mar 22 12:09:01 rcfs05 kernel: Call Trace:
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80239d8a>]
> > __lookup_hash+0x7a/0x160
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff803547a5>] nfsd_lookup+0x435/0x4d0
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8035aaa8>]
> > nfsd3_proc_lookup+0xe8/0x110
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034d69d>]
> > nfsd_dispatch+0xfd/0x1f0
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80608094>] svc_process+0x3e4/0x730
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8026bbf2>] __down_read+0x12/0xa2
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034dd00>] nfsd+0x1a0/0x2d0
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f38>] child_rip+0xa/0x12
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034db60>] nfsd+0x0/0x2d0
> > Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f2e>] child_rip+0x0/0x12
> > Mar 22 12:09:01 rcfs05 kernel:
> > Mar 22 12:09:01 rcfs05 kernel:
> > Mar 22 12:09:01 rcfs05 kernel: Code: 48 8b 01 ba 10 00 00 00 bf 10 00
> > 00 00 48 89 06 48 8b 41 08
> > Mar 22 12:09:01 rcfs05 kernel: RIP  [<ffffffff80350613>]
> > fh_compose+0x3e3/0x4d0
> > Mar 22 12:09:01 rcfs05 kernel:  RSP <ffff8101197fdd70>
> > Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000
> >
> > Once this oops happens, the machine is still usable, however NFS
> > crawls to a halt (in fact one of the machines with a mount point of
> > this machine, failed to respond anymore - I assume that thread is the
> > one that ooops'ed).
> >
> > Any more information or debugging, please let me know.
>
>
> Does it work with kernel 2.6.20?
>
>
> > TIA!
> > Phy
>
> cu
> Adrian
>
> --
>
>        "Is there not promise of rain?" Ling Tan asked suddenly out
>         of the darkness. There had been need of rain for many days.
>        "Only a promise," Lao Er said.
>                                        Pearl S. Buck - Dragon Seed
>
>

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Opps in 2.6.21-rc4 nfsd
  2007-03-22 19:24 Opps in 2.6.21-rc4 nfsd Phy Prabab
  2007-03-23 15:29 ` Adrian Bunk
@ 2007-03-23 17:47 ` Chuck Ebbert
  1 sibling, 0 replies; 7+ messages in thread
From: Chuck Ebbert @ 2007-03-23 17:47 UTC (permalink / raw)
  To: Phy Prabab; +Cc: linux-kernel

Phy Prabab wrote:
> Hello,
> 
> I have a recent oopps on a small file server I have running.  The
> machine in question is a dual, dual core AMD Opteron 2.2GHz w/4G RAM,
> 3Ware 9650SE (14 drive RAID 10), using XFS, LVM2 (latest), exported
> via knfsd.  Kernel is 2.6.21-rc4 64b.  Here is the bt:
> 
> Mar 22 12:09:01 rcfs05 kernel: Modules linked in:
> Mar 22 12:09:01 rcfs05 kernel: Pid: 3102, comm: nfsd Not tainted
> 2.6.21-rc4-smp #3
> Mar 22 12:09:01 rcfs05 kernel: RIP: 0010:[<ffffffff80350613>]
> [<ffffffff80350613>] fh_compose+0x3e3/0x4d0
> Mar 22 12:09:01 rcfs05 kernel: RSP: 0018:ffff8101197fdd70  EFLAGS: 00010297
> Mar 22 12:09:01 rcfs05 kernel: RAX: 0000000000000006 RBX:
> 000000000ee00004 RCX: 0000000000000000
> Mar 22 12:09:01 rcfs05 kernel: RDX: 0000000000000018 RSI:
> ffff81011975a940 RDI: 0000000000000080
> Mar 22 12:09:01 rcfs05 kernel: RBP: 0000000000000001 R08:
> 00000000ffffffff R09: 00002ba24f06ca50
> Mar 22 12:09:01 rcfs05 kernel: R10: ffff81011975a808 R11:
> ffffffff8022f500 R12: ffff81011975a938
> Mar 22 12:09:01 rcfs05 kernel: R13: 0000000000000006 R14:
> ffff81010eaab0c0 R15: ffff8100cc714940
> Mar 22 12:09:01 rcfs05 kernel: FS:  00002b020c71b6d0(0000)
> GS:ffff81011dcf56c0(0000) knlGS:0000000000000000
> Mar 22 12:09:01 rcfs05 kernel: CS:  0010 DS: 0000 ES: 0000 CR0:
> 000000008005003b
> Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000 CR3:
> 00000001173d1000 CR4: 00000000000006e0
> Mar 22 12:09:01 rcfs05 kernel: Process nfsd (pid: 3102, threadinfo
> ffff8101197fc000, task ffff81011b7ef020)
> Mar 22 12:09:01 rcfs05 kernel: Stack:  ffff81010fa7b868
> ffffffff80239d8a 0000000000000286 ffff810111aee5e8
> Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
> ffff810105af2890 000000000eaab0c0
> Mar 22 12:09:01 rcfs05 kernel:  ffff810105af2890 ffff81011975a808
> ffff810105af2890 000000000000000a
> Mar 22 12:09:01 rcfs05 kernel: Call Trace:
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80239d8a>]
> __lookup_hash+0x7a/0x160
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff803547a5>]
> nfsd_lookup+0x435/0x4d0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8035aaa8>]
> nfsd3_proc_lookup+0xe8/0x110
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034d69d>]
> nfsd_dispatch+0xfd/0x1f0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80608094>]
> svc_process+0x3e4/0x730
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8026bbf2>] __down_read+0x12/0xa2
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034dd00>] nfsd+0x1a0/0x2d0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f38>] child_rip+0xa/0x12
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff8034db60>] nfsd+0x0/0x2d0
> Mar 22 12:09:01 rcfs05 kernel:  [<ffffffff80264f2e>] child_rip+0x0/0x12
> Mar 22 12:09:01 rcfs05 kernel:
> Mar 22 12:09:01 rcfs05 kernel:
> Mar 22 12:09:01 rcfs05 kernel: Code: 48 8b 01 ba 10 00 00 00 bf 10 00
> 00 00 48 89 06 48 8b 41 08
> Mar 22 12:09:01 rcfs05 kernel: RIP  [<ffffffff80350613>]
> fh_compose+0x3e3/0x4d0
> Mar 22 12:09:01 rcfs05 kernel:  RSP <ffff8101197fdd70>
> Mar 22 12:09:01 rcfs05 kernel: CR2: 0000000000000000
> 
> Once this oops happens, the machine is still usable, however NFS
> crawls to a halt (in fact one of the machines with a mount point of
> this machine, failed to respond anymore - I assume that thread is the
> one that ooops'ed).
> 
> Any more information or debugging, please let me know.

The first few lines of the oops are missing.


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Opps in 2.6.21-rc4 nfsd
  2007-03-23 15:45   ` Phy Prabab
@ 2007-03-26  0:50     ` Neil Brown
  2007-03-26  1:08       ` Neil Brown
  0 siblings, 1 reply; 7+ messages in thread
From: Neil Brown @ 2007-03-26  0:50 UTC (permalink / raw)
  To: Phy Prabab; +Cc: Adrian Bunk, linux-kernel, nfs

On Friday March 23, phyprabab@gmail.com wrote:
> Adrian et al,
> 
> Under 2.6.20.2+3 I get a similar crash and more frequently.  At least
> under 2.6.21rc4 I only get this about once every two hours under heavy
> nfs testing.

That is very odd.  The code where it is crashing in the trace you
provide does not exist in 2.6.20 at all.

Do you have an Oops trace from when it crashed under 2.6.20.x ??

I'll try to have a patch for you to try on 2.6.21rc4 shortly.

What version of nfs-utils are you using?

Thanks,
NeilBrown

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Opps in 2.6.21-rc4 nfsd
  2007-03-26  0:50     ` Neil Brown
@ 2007-03-26  1:08       ` Neil Brown
  2007-03-26  7:07         ` Phy Prabab
  0 siblings, 1 reply; 7+ messages in thread
From: Neil Brown @ 2007-03-26  1:08 UTC (permalink / raw)
  To: Phy Prabab, Adrian Bunk, linux-kernel, nfs

On Monday March 26, neilb@suse.de wrote:
> 
> I'll try to have a patch for you to try on 2.6.21rc4 shortly.
> 

This patch should remove the Oops, but I have a suspicion that
something else is wrong, so I'd be very interested if you could run
with this for a few days and report the results.

NeilBrown

Signed-off-by: Neil Brown <neilb@suse.de>

### Diffstat output
 ./fs/nfsd/nfsfh.c |   37 +++++++++++++++++++++++++++++++++++--
 1 file changed, 35 insertions(+), 2 deletions(-)

diff .prev/fs/nfsd/nfsfh.c ./fs/nfsd/nfsfh.c
--- .prev/fs/nfsd/nfsfh.c	2007-03-26 10:51:01.000000000 +1000
+++ ./fs/nfsd/nfsfh.c	2007-03-26 11:04:20.000000000 +1000
@@ -324,7 +324,7 @@ fh_compose(struct svc_fh *fhp, struct sv
 	 *
 	 */
 
-	u8 version = 1;
+	u8 version;
 	u8 fsid_type = 0;
 	struct inode * inode = dentry->d_inode;
 	struct dentry *parent = dentry->d_parent;
@@ -342,12 +342,45 @@ fh_compose(struct svc_fh *fhp, struct sv
 	 * the reference filehandle (if it is in the same export)
 	 * or the export options.
 	 */
+ retry:
+	version = 1;
 	if (ref_fh && ref_fh->fh_export == exp) {
 		version = ref_fh->fh_handle.fh_version;
 		if (version == 0xca)
 			fsid_type = FSID_DEV;
-		else
+		else {
 			fsid_type = ref_fh->fh_handle.fh_fsid_type;
+			/* need to check that this type works for this
+			 * export point.  As the lookup was done with
+			 * help from user-space, there is no guarantee
+			 */
+			if (ref_fh == fhp)
+				fh_put(ref_fh);
+			ref_fh = NULL;
+			switch(fsid_type) {
+			case FSID_DEV:
+				if (!old_valid_dev(ex_dev))
+					goto retry;
+				break;
+			case FSID_NUM:
+				if (! (exp->ex_flags & NFSEXP_FSID))
+					goto retry;
+				break;
+			case FSID_MAJOR_MINOR:
+			case FSID_ENCODE_DEV:
+				break;
+			case FSID_UUID8:
+			case FSID_UUID16:
+				if (!root_export)
+					goto retry;
+				/* fall through */
+			case FSID_UUID4_INUM:
+			case FSID_UUID16_INUM:
+				if (exp->ex_uuid == NULL)
+					goto retry;
+				break;
+			}
+		}
 		/* We know this version/type works for this export
 		 * so there is no need for further checks.
 		 */

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: Opps in 2.6.21-rc4 nfsd
  2007-03-26  1:08       ` Neil Brown
@ 2007-03-26  7:07         ` Phy Prabab
  0 siblings, 0 replies; 7+ messages in thread
From: Phy Prabab @ 2007-03-26  7:07 UTC (permalink / raw)
  To: Neil Brown; +Cc: Adrian Bunk, linux-kernel, nfs

Opps, forgot, this is for 2.6.20.x, guess I'll wait for the updated patch.

Thanks!
Phy

On 3/25/07, Neil Brown <neilb@suse.de> wrote:
> On Monday March 26, neilb@suse.de wrote:
> >
> > I'll try to have a patch for you to try on 2.6.21rc4 shortly.
> >
>
> This patch should remove the Oops, but I have a suspicion that
> something else is wrong, so I'd be very interested if you could run
> with this for a few days and report the results.
>
> NeilBrown
>
> Signed-off-by: Neil Brown <neilb@suse.de>
>
> ### Diffstat output
>  ./fs/nfsd/nfsfh.c |   37 +++++++++++++++++++++++++++++++++++--
>  1 file changed, 35 insertions(+), 2 deletions(-)
>
> diff .prev/fs/nfsd/nfsfh.c ./fs/nfsd/nfsfh.c
> --- .prev/fs/nfsd/nfsfh.c       2007-03-26 10:51:01.000000000 +1000
> +++ ./fs/nfsd/nfsfh.c   2007-03-26 11:04:20.000000000 +1000
> @@ -324,7 +324,7 @@ fh_compose(struct svc_fh *fhp, struct sv
>          *
>          */
>
> -       u8 version = 1;
> +       u8 version;
>         u8 fsid_type = 0;
>         struct inode * inode = dentry->d_inode;
>         struct dentry *parent = dentry->d_parent;
> @@ -342,12 +342,45 @@ fh_compose(struct svc_fh *fhp, struct sv
>          * the reference filehandle (if it is in the same export)
>          * or the export options.
>          */
> + retry:
> +       version = 1;
>         if (ref_fh && ref_fh->fh_export == exp) {
>                 version = ref_fh->fh_handle.fh_version;
>                 if (version == 0xca)
>                         fsid_type = FSID_DEV;
> -               else
> +               else {
>                         fsid_type = ref_fh->fh_handle.fh_fsid_type;
> +                       /* need to check that this type works for this
> +                        * export point.  As the lookup was done with
> +                        * help from user-space, there is no guarantee
> +                        */
> +                       if (ref_fh == fhp)
> +                               fh_put(ref_fh);
> +                       ref_fh = NULL;
> +                       switch(fsid_type) {
> +                       case FSID_DEV:
> +                               if (!old_valid_dev(ex_dev))
> +                                       goto retry;
> +                               break;
> +                       case FSID_NUM:
> +                               if (! (exp->ex_flags & NFSEXP_FSID))
> +                                       goto retry;
> +                               break;
> +                       case FSID_MAJOR_MINOR:
> +                       case FSID_ENCODE_DEV:
> +                               break;
> +                       case FSID_UUID8:
> +                       case FSID_UUID16:
> +                               if (!root_export)
> +                                       goto retry;
> +                               /* fall through */
> +                       case FSID_UUID4_INUM:
> +                       case FSID_UUID16_INUM:
> +                               if (exp->ex_uuid == NULL)
> +                                       goto retry;
> +                               break;
> +                       }
> +               }
>                 /* We know this version/type works for this export
>                  * so there is no need for further checks.
>                  */
>

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2007-03-26  7:07 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-03-22 19:24 Opps in 2.6.21-rc4 nfsd Phy Prabab
2007-03-23 15:29 ` Adrian Bunk
2007-03-23 15:45   ` Phy Prabab
2007-03-26  0:50     ` Neil Brown
2007-03-26  1:08       ` Neil Brown
2007-03-26  7:07         ` Phy Prabab
2007-03-23 17:47 ` Chuck Ebbert

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).