Re: linux-3.14 nfsd regression

[Mark Lord <mlord@pobox.com>]

On 14-04-03 05:28 PM, J. Bruce Fields wrote:
> On Thu, Apr 03, 2014 at 04:48:11PM -0400, Mark Lord wrote:
>> On 14-04-03 03:30 PM, J. Bruce Fields wrote:
>>> On Thu, Apr 03, 2014 at 01:51:06PM -0400, Mark Lord wrote:
>>>> On 14-04-03 01:16 PM, J. Bruce Fields wrote:
>>>>> On Thu, Apr 03, 2014 at 12:33:55PM -0400, Mark Lord wrote:
>>>>>> This commit from linux-3.14 breaks our NFS-root clients here:
>>>>>>
>>>>>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=6e14b46b91fee8a049b0940333ce13a820beaaa5
>>>>>>
>>>>>>
>>>>>> - *p++ = htonl((u32) stat->mode);
>>>>>> + *p++ = htonl((u32) (stat->mode & S_IALLUGO));
>>>>>>
>>>>>>
>>>>>> Reverting the one-liner above (on the server) fixes it for us,
>>>>>> as does reverting back to linux-3.13.8 on the server.
>>>>>>
>>>>>> The NFS-root clients are on PowerPC (big-endian) architecture,
>>>>>> running linux-3.12.16. The NFS server is on an Intel PC running linux-3.14.
>>>>>>
>>>>>> ACL is completely disabled on server and client,
>>>>>> and we're using NFSv2/v3.  No support for v4.
>>>>>>
>>>>>> I instrumented the function to see what other bits were being cleared
>>>>>> by the (stat->mode & S_IALLUGO) masking.  The results are attached.
>>>>>
>>>>> Hm, it sounds like a bug in the client if it's depending on those high
>>>>> bits.
>>>>
>>>> But only for mounting / starting up from the nfsroot, it seems.
>>>> I wonder if there's an unusual code path for that in there?
>>>> The regular stuff looks mostly fine:
>>>>
>>>>         p = xdr_decode_ftype3(p, &fmode);
>>>>         fattr->mode = (be32_to_cpup(p++) & ~S_IFMT) | fmode;
>>>
>>> Hm, but that's in nfs3xdr.c; in nfs2xdr.c we have just
>>>
>>> 	fattr->mode = be32_to_cpup(p+);
>>>
>>> and NFSv2 is the default for nfsroot.  Do you have some reason to
>>> believe you're not using NFSv2?
>>
>> Oh, the client here was using NFS2, absolutely.
>> I just don't know my way around the code very well yet.  :)
>>
>> But that mask in nfs3xdr.c (client) doesn't match what the server side is using.
> 
> Not sure there's anything to see there.
> 
> Looking at include/uapi/linux/stat.h and include/linux/stat.h, if I'm
> doing this right...
> 
> S_ISFMT is   0170000
> S_IALLUGO is 0007777
> 
> So they're 16-bit complements.  And we're storing the result in a short?

Oh, is it a short?  I was just going by the be32_to_cpup() macro (not a short).
But if the target field is, then okay for now, until someone expands it.

Cheers
-- 
Mark Lord
Real-Time Remedies Inc.
mlord@pobox.com