Re: linux-next: build failure after merge of the nfsd tree

[Chuck Lever <chuck.lever@oracle.com>]

On Apr 29, 2013, at 12:21 PM, Trond Myklebust <trond.myklebust@fys.uio.no> wrote:

> On Mon, 2013-04-29 at 12:05 -0400, Chuck Lever wrote:
>> On Apr 29, 2013, at 11:45 AM, "J. Bruce Fields" <bfields@fieldses.org> wrote:
>> 
>>> On Mon, Apr 29, 2013 at 10:53:37AM -0400, Chuck Lever wrote:
>>>> 
>>>> On Apr 28, 2013, at 9:24 PM, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>>>> 
>>>>> Hi J.,
>>>>> 
>>>>> After merging the nfsd tree, today's linux-next build (powerpc
>>>>> ppc64_defconfig) failed like this:
>>>>> 
>>>>> net/sunrpc/auth_gss/svcauth_gss.c: In function 'gss_proxy_save_rsc':
>>>>> net/sunrpc/auth_gss/svcauth_gss.c:1182:3: error: implicit declaration of function 'gss_mech_get_by_OID' [-Werror=implicit-function-declaration]
>>>>> 
>>>>> Caused byc ommit 030d794bf498 ("SUNRPC: Use gssproxy upcall for server
>>>>> RPCGSS authentication").  gss_mech_get_by_OID() made static to
>>>>> net/sunrpc/auth_gss/gss_mech_switch.c by commit 9568c5e9a61d ("SUNRPC:
>>>>> Introduce rpcauth_get_pseudoflavor()") in the nfs tree (part of the nfs
>>>>> tree that you did not merge).
>>>>> 
>>>>> I don't know how to fix this, so I have used the nfsd tree from
>>>>> next-20130426 for today.
>>>> 
>>>> Bruce, it might make sense for me to submit the three server-side RPC GSS patches, and then you can rebase the gssproxy work on top of those.  Let me know how you would like to proceed.
>>> 
>>> I'm happy to take those patches whenever you consider them ready.  Would
>>> that fix the problem?
>> 
>> Someone would need to modify the gssproxy patches to use the new interfaces.
>> 
>>> Also: it looks like 030d794bf498 "SUNRPC: Introduce
>>> rpcauth_get_pseudoflavor()" is in Trond's linux-next, but not his
>>> nfs-for-next.  I'm not sure what that means--is it safe to rebase on top
>>> of *that*?
>> 
>> That doesn't seem right to me.
> 
> I've now pulled the rpcsec_gss changes into the nfs-for-next. The main
> reason why they were not pulled in earlier was due to uncertainty what
> to do about the increase in "AUTH_GSS upcall timed out." syslog
> warnings.

Trond's nfs-for-next now has the new rpcauth_get_gssinfo() and rpcauth_get_pseudoflavor() APIs, which are replacements for direct calls into the GSS mech switch.  These APIs are a little more generic, and more robust in the face of unloaded GSS kernel modules.

Instead of gss_mech_get_by_OID(), I suspect you want rpcauth_get_pseudoflavor(), but I haven't looked at the gssproxy code.

-- 
Chuck Lever
chuck[dot]lever[at]oracle[dot]com