On 07/27/2011 01:25 PM, Trond Myklebust wrote:
> On Wed, 2011-07-27 at 13:20 -0700, Boaz Harrosh wrote:
>> On 07/27/2011 12:53 PM, Trond Myklebust wrote:
>>>> diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
>>>> index bddd8b9..f271425 100644
>>>> --- a/fs/nfs/pnfs.h
>>>> +++ b/fs/nfs/pnfs.h
>>>> @@ -113,6 +113,9 @@ struct pnfs_layoutdriver_type {
>>>> struct xdr_stream *xdr,
>>>> const struct nfs4_layoutreturn_args *args);
>>>>
>>>> + void (*cleanup_layoutcommit) (struct pnfs_layout_hdr *layoutid,
>>>> + struct nfs4_layoutcommit_data *data);
>>>> +
>>>> void (*encode_layoutcommit) (struct pnfs_layout_hdr *layoutid,
>>>> struct xdr_stream *xdr,
>>>> const struct nfs4_layoutcommit_args *args);
>>>
>>> This really needs to go. We should have
>>>
>>> int (*layoutcommit)()...
>>>
>>> instead of 'encode' and 'cleanup' methods...
>>>
>>
>> Theoretically it is not possible because the blocks-layout protocol mandates
>> different handling depending on the "error" response from the Server which
>> will be received on RPC done.
>
> ???? If the blocks code is in charge of actually doing the RPC call, why
> would it not be able to perform its own error handling?
>
Is it? I thought it was the Generic layer that was Initiating the layoutcommit
(From the pnfs_layoutcommit_inode called from nfs_write_inode)
The LD only has a chance to encode the payload on rpc-setup and here the blocks
code needs cleanup depending on the return-status of rpc-done
[I do think that setup and done might be better names to reflect the rpc states)
Thanks
Boaz
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
