[pnfs] [PATCH 5/8] nfsd: tabulate nfs4 xdr encoding functions

[Benny Halevy]

On Jul. 03, 2008, 1:58 +0300, "J. Bruce Fields" <bfields at fieldses.org> wrote:
> Whoops, I should have seen this when I read the patch:
> 
> On Mon, Jun 30, 2008 at 09:07:05PM +0300, Benny Halevy wrote:
>> -	switch (op->opnum) {
>> -	case OP_ACCESS:
>> -		nfsd4_encode_access(resp, op->status, &op->u.access);
>> -		break;
>> -	case OP_CLOSE:
>> -		nfsd4_encode_close(resp, op->status, &op->u.close);
>> -		break;
>> -	case OP_COMMIT:
>> -		nfsd4_encode_commit(resp, op->status, &op->u.commit);
>> -		break;
>> -	case OP_CREATE:
>> -		nfsd4_encode_create(resp, op->status, &op->u.create);
>> -		break;
>> -	case OP_DELEGRETURN:
>> -		break;
> 
> Most of the encode operations don't return a status, but for those where
> the real work's done in the encode:
> 
>> -	case OP_GETATTR:
>> -		op->status = nfsd4_encode_getattr(resp, op->status, &op->u.getattr);
>> -		break;
> 
> we're now discarding that return:
> 
>> +	nfsd4_enc_ops[op->opnum](resp, op->status, &op->u);
> 
> I'd be inclined to standardize on having all operations return a status

Agreed.

> (always zero, in most cases).

hmm, I'd say it's always the input nfserr, in most cases...
(which hopefully is zero)

> 
> Anyway for now I'm just dropping this patch for now, and applying the
> others.

Thanks!
Will fix.

Benny

> 
> --b.