Re: [2.6.12-rc1-mm3] BUG: atomic counter underflow in smbfs

[Andrew Morton]

Benoit Boissinot <benoit.boissinot@xxxxxxxxxxxx> wrote:
>
> I had the following BUG with 2.6.12-rc1-mm3:
> 
> remote host is running 2.6.12-rc1-mm1 with samba 3.0.13.
> 
> [23156.357178] smb_lookup: find musique/Pink_Floyd-Dark_Side_of_the_Moon
> failed, error=-512
> [23157.057501] BUG: atomic counter underflow at:
> [23157.057508]  [<c0103c27>] dump_stack+0x17/0x20
> [23157.057516]  [<e0ed0f31>] smb_rput+0x51/0x60 [smbfs]
> [23157.057530]  [<e0ecd497>] smb_proc_query_cifsunix+0x77/0xa0 [smbfs]
> [23157.057538]  [<e0eca14c>] smb_newconn+0x2bc/0x310 [smbfs]
> [23157.057546]  [<e0ed05ac>] smb_ioctl+0xfc/0x100 [smbfs]
> [23157.057554]  [<c0162188>] do_ioctl+0x48/0x70
> [23157.057559]  [<c01622f9>] vfs_ioctl+0x59/0x1b0
> [23157.057563]  [<c0162489>] sys_ioctl+0x39/0x60
> [23157.057582]  [<c0102d8f>] sysenter_past_esp+0x54/0x75

Oh dear.  That warning is not necessarily telling us that there's a serious
problem - often it's fairly harmless.  Did the filesytem misbehave in any
other manner?

A problem we have here is that nobody really maintains smbfs any more, and
it has problems.  I was hoping that the stock answer to that would be "use
cifs", but for some reason that doesn't seem to be happening.  Have you
tried it?  (Last time I looked, cifs didn't work against win98 servers -
maybe that got fixed).


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/