Re: nfs4.1 locking broken

["J. Bruce Fields" <bfields@fieldses.org>]

And, yuck, I think this is why I let such a bad screwup get by in the
server lock code: my usual automated testing depends on booting the same
kernel to client and server, and when they saw the client failure they
gave up and as a result never hit the bad server case.  Doh.  Something
else to fix in my testing.

--b.

On Fri, Mar 25, 2011 at 10:34:03PM -0400, J. Bruce Fields wrote:
> Commit 114f64b5f24abac33a42f4f1856eb3a9766d497e "NFSv4: remove duplicate
> clientid in struct nfs_client" breaks locking over 4.1; from a cthon
> run:
> 
> On Fri, Mar 25, 2011 at 07:28:31PM -0400, J. Bruce Fields wrote:
> > Starting LOCKING tests: test directory /mnt/TMP (arg: -f)
> > 
> > Testing native post-LFS locking
> > 
> > Creating parent/child synchronization pipes.
> > 
> > Test #1 - Test regions of an unlocked file.
> > 	Parent: 1.1  - F_TEST  [               0,               1] FAILED!
> > 	Parent: **** Expected success, returned EINVAL...
> > 	Parent: **** Probably implementation error.
> > 
> > ** PARENT pass 1 results: 0/0 pass, 0/0 warn, 1/1 fail (pass/total).
> > 
> > **  CHILD pass 1 results: 0/0 pass, 0/0 warn, 0/0 fail (pass/total).
> > lock tests failed
> 
> I haven't figured out why.
> 
> Presumably the lock code assumed the clientid would be unset in the 4.1
> case?
> 
> --b.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html