Re: [-mm patch] BUG at net/sunrpc/svc.c:128 (was Re: 2.6.20-rc6-mm2)

[Andrew Morton <akpm@osdl.org>]

On Mon, 29 Jan 2007 11:21:41 +0000
Frederik Deweerdt <deweerdt@free.fr> wrote:

> On Mon, Jan 29, 2007 at 12:12:07AM -0800, Andrew Morton wrote:
> > 	ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.20-rc6/2.6.20-rc6-mm2/
> > 
> 
> Hi,
> 
> The svc_pool_map_init_percpu() should get maxpool from the number of
> online cpus, not the number of nodes.  The following BUG is triggered
> when we try to check if the cpu index is smaller than the number of nodes.
> (The system is multi cpu, single node).
> 
> [  133.196276] ------------[ cut here ]------------
> [  133.196334] kernel BUG at net/sunrpc/svc.c:128!
> [  133.196391] invalid opcode: 0000 [#1]
> [  133.196444] PREEMPT SMP
> [  133.196571] last sysfs file: /devices/pci0000:00/0000:00:00.0/class
> [  133.196630] Modules linked in: nfsd exportfs lockd sunrpc ipv6 i2c_dev i2c_core video thermal sony_acpi processor fan button battery asus_acpi ac uhci_hcd ehci_hcd rng_core
> [  133.197473] CPU:    0
> [  133.197474] EIP:    0060:[<f8a5973c>]    Not tainted VLI
> [  133.197475] EFLAGS: 00010202   (2.6.20-rc6-mm2 #1)
> [  133.197656] EIP is at svc_pool_map_init_percpu+0x54/0x5a [sunrpc]
> [  133.197715] eax: 00000002   ebx: f8a762e4   ecx: 00000002   edx: 00000008
> [  133.197776] esi: 00000002   edi: f8aa71e0   ebp: c2fd3edc   esp: c2fd3ed4
> [  133.197835] ds: 007b   es: 007b   fs: 00d8  gs: 0033  ss: 0068
> [  133.197894] Process rpc.nfsd (pid: 3595, ti=c2fd2000 task=f7e22030 task.ti=c2fd2000)
> [  133.197954] Stack: f8a7c081 00040000 c2fd3ee4 f8a597de c2fd3efc f8a59a55 c2fd3f10 00000000
> [  133.198360]        00000008 00000801 c2fd3f54 f8a7c1ac f8a7c4e9 00000001 f8aa9480 fffdd1fe
> [  133.198769]        00000292 f8aa9140 00015f90 0007dbcc 00078804 00000000 0000076e 00000008
> [  133.199175] Call Trace:
> [  133.199275]  [<c010390d>] show_trace_log_lvl+0x1a/0x30
> [  133.199373]  [<c01039c4>] show_stack_log_lvl+0x8d/0xb1
> [  133.199467]  [<c0103bf1>] show_registers+0x1c0/0x336
> [  133.199561]  [<c0103efd>] die+0x11f/0x213
> [  133.199654]  [<c010406a>] do_trap+0x79/0xa7
> [  133.199748]  [<c010433b>] do_invalid_op+0xa3/0xad
> [  133.199842]  [<c03ee3a4>] error_code+0x7c/0x84
> [  133.199936]  [<f8a597de>] svc_pool_map_init+0x4c/0x60 [sunrpc]
> [  133.200047]  [<f8a59a55>] svc_create_pooled+0x14/0x44 [sunrpc]
> [  133.200156]  [<f8a7c1ac>] nfsd_create_serv+0x5a/0xd4 [nfsd]
> [  133.200262]  [<f8a7c498>] nfsd_svc+0x50/0xa1 [nfsd]
> [  133.200364]  [<f8a7ca1f>] write_svc+0x1a/0x20 [nfsd]
> [  133.200466]  [<f8a7c98a>] nfsctl_transaction_write+0x4a/0x6d [nfsd]
> [  133.200571]  [<c019f29e>] sys_nfsservctl+0xa3/0xe5
> [  133.200666]  [<c0102a94>] sysenter_past_esp+0x5d/0x99
> [  133.200762]  =======================
> [  133.200816] Code: c1 77 22 83 fe 01 77 23 8b 43 0c ba 10 3f 54 c0 89 34 88 8b 43 08 89 0c b0 89 c8 83 c6 01 e8 4e 8f 7b c7 eb d7 5b 89 f0 5e 5d c3 <0f> 0b 89 f6 eb fc 55 ba 01 00 00 00 89 e5 56 31 f6 53 89 c3 e8
> [  133.203330] EIP: [<f8a5973c>] svc_pool_map_init_percpu+0x54/0x5a [sunrpc] SS:ESP 0068:c2fd3ed4
> 
> The following patch fixes the problem for me.
> 
> Regards,
> Frederik
> 
> 
> Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
> 
> diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
> index c8c7c55..a553f9f 100644
> --- a/net/sunrpc/svc.c
> +++ b/net/sunrpc/svc.c
> @@ -115,7 +115,7 @@ fail:
>  static int
>  svc_pool_map_init_percpu(struct svc_pool_map *m)
>  {
> -	unsigned int maxpools = nr_node_ids;
> +	unsigned int maxpools = num_online_cpus();
>  	unsigned int pidx = 0;
>  	unsigned int cpu;
>  	int err;

Thanks.

Christoph, can you pleeeeeze be more careful?  A few seconds inattention
and a dopey copy-n-paste bug leads to large amounts of wasted time for
other people.