xfs
[Top] [All Lists]

RE: xfs and nfs

To: "'Seref Tufan Sen'" <tufan@xxxxxxxxxx>, "'MOGUILNY Genevieve'" <moguilny@xxxxxxxxxxxxxxx>, linux-xfs@xxxxxxxxxxx
Subject: RE: xfs and nfs
From: "Christian, Chip" <chip_christian@xxxxxx>
Date: Wed, 7 Nov 2001 10:05:20 -0500
Sender: owner-linux-xfs@xxxxxxxxxxx
Since you mention NFS, and I see dput in the syslog output,
I'm gonna guess it's the problem I reported over the summer.

The patch got put in the mainline kernel, but if you want to
patch 2.4.3 yourself the problem I encountered was that 
nfsd_findparent() didn't find '..', then went and called
dput(tdentry) twice.  My patch added a return(), Neil's
dropped a dput():

struct dentry *nfsd_findparent(struct dentry *child)
{
        struct dentry *tdentry, *pdentry;
        tdentry = d_alloc(child, &(const struct qstr) {"..", 2, 0});
        if (!tdentry)
                return ERR_PTR(-ENOMEM);

        /* I'm going to assume that if the returned dentry is different, then
         * it is well connected.  But nobody returns different dentrys do they?
         */
        pdentry = child->d_inode->i_op->lookup(child->d_inode, tdentry);
        d_drop(tdentry); /* we never want ".." hashed */
        if (!pdentry && tdentry->d_inode == NULL) {
                /* File system cannot find ".." ... sad but possible */
Neil Dropped->  dput(tdentry);
                pdentry = ERR_PTR(-EINVAL);
I added------>  return pdentry;
        }
        if (!pdentry) {
...
        }
        dput(tdentry); /* it is not hashed, it will be discarded */
        return pdentry;
}

-----Original Message-----
From: Seref Tufan Sen [mailto:tufan@xxxxxxxxxx]
Sent: Wednesday, November 07, 2001 6:36
To: 'MOGUILNY Genevieve'; linux-xfs@xxxxxxxxxxx
Subject: RE: xfs and nfs


I had the same problems with NFS (RH 7.1 + MegaRAID Controler too) . I was
using RPMS from SGI (2.4.3) , then I compiled 2.4.9 with XFS patch. And the
problem has gone.

> -----Original Message-----
> From: owner-linux-xfs@xxxxxxxxxxx
> [mailto:owner-linux-xfs@xxxxxxxxxxx]On
> Behalf Of MOGUILNY Genevieve
> Sent: Wednesday, November 07, 2001 12:48 PM
> To: linux-xfs@xxxxxxxxxxx
> Subject: xfs and nfs
>
>
> Hello,
>
> We have installed XFS 1.0 on a RedHat 7.1 system
> with MegaRAID controller. On this machine,
> there are a hundred of homes, samba server...
> As we had problems with Remote/Local accesses
> on xfs partitions, we installed XFS 1.0.1.
> Since then, nfs stopiped working several times a day.
> In the var log, we have messages like :
>
> Nov  5 17:39:00 hera kernel: invalid operand: 0000
> Nov  5 17:39:00 hera kernel: CPU:    0
> Nov  5 17:39:00 hera kernel: EIP:    0010:[dput+20/324]
> Nov  5 17:39:00 hera kernel: EIP:    0010:[<c01436e4>]
> Nov  5 17:39:00 hera kernel: EFLAGS: 00010246
> Nov  5 17:39:00 hera kernel: eax: 00000000   ebx: d7878d60
> ecx: c458de44   edx: 00178780
> Nov  5 17:39:00 hera kernel: esi: d7878d60   edi: d7878ae0
> ebp: d7878ae0   esp: c458de4c
> Nov  5 17:39:00 hera kernel: ds: 0018   es: 0018   ss: 0018
> Nov  5 17:39:00 hera kernel: Process nfsd (pid: 1707,
> stackpage=c458d000)
> Nov  5 17:39:00 hera kernel: Stack: ffffffea d7878d60
> d8903ee1 d7878d60 00000000 10482e8e
> d890
> 4258 d7878ae0
> Nov  5 17:39:00 hera kernel:        00000000 c54ebe14
> 11270000 c54ebe04 4c1b9d86 c60779f8
> 0000
> 0000 ffffff8c
> Nov  5 17:39:00 hera kernel:        00000000 d89045d4
> c6077800 10482e8e 00000000 00000000
> 0000
> 0001 c54ebe04
> Nov  5 17:39:00 hera kernel: Call Trace: [<d8903ee1>]
> [<d8904258>] [<d89045d4>]
> [<d8904bde>] [
> schedule+647/976]
> Nov  5 17:39:00 hera kernel: Call Trace: [<d8903ee1>]
> [<d8904258>] [<d89045d4>]
> [<d8904bde>] [
> <c011357f>]
> Nov  5 17:39:00 hera kernel:    [<d890a84e>] [<d8911da0>]
> [<d890c51c>] [<d8911da0>]
> [<d8902553
> >] [<d8911da0>]
> Nov  5 17:39:00 hera kernel:    [<d88c9e98>] [<d8911ce0>]
> [<d89117d8>] [<d8902311>]
> [<d89117c0
> >] [kernel_thread+35/48]
> Nov  5 17:39:00 hera kernel:    [<d88c9e98>] [<d8911ce0>]
> [<d89117d8>] [<d8902311>]
> [<d89117c0
> >] [<c0105537>]
> Nov  5 17:39:00 hera kernel:
> Nov  5 17:39:00 hera kernel: Code: 0f 0b ff 0b 0f 94 c0 84 c0
> 0f 84 1e 01 00 00 8d 73 20 39
> 73
>
>
> And after that message, nobody could access their home space
> from a remote host.
>
> Do you have any idea of what could  cause  the problem ?
> Thank you in advance,
>
>
> Genevieve Moguilny
> DMPN - IPGP
> Case 89 - Tour 24
> 4 place Jussieu
> 75252 Paris cedex 05
> Tel. 01 44 27 24 15
> Fax  01 44 27 38 94
> moguilny@xxxxxxxxxxxxxxx


<Prev in Thread] Current Thread [Next in Thread>