commit 6c75dc0d498caa402fb17b1bf769835a9db875c8
author Fred Isaman <iisaman@netapp.com> Fri Apr 20 14:47:47 2012 -0400
committer Trond Myklebust <Trond.Myklebust@netapp.com> Fri Apr 27 14:10:37 2012 -0400
tree b367bb2428c011f20b8fd47a6dd0b8603ee136ba
parent 4db6e0b74c0f6dfc2f9c0690e8df512e3b635983

NFS: merge _full and _partial write rpc_ops

Decouple nfs_pgio_header and nfs_write_data, and have (possibly
multiple) nfs_write_datas each take a refcount on nfs_pgio_header.

For the moment keeps nfs_write_header as a way to preallocate a single
nfs_write_data with the nfs_pgio_header.  The code doesn't need this,
and would be prettier without, but given the amount of churn I am
already introducing I didn't want to play with tuning new mempools.

This also fixes bug in pnfs_ld_handle_write_error.  In the case of
desc->pg_bsize < PAGE_CACHE_SIZE, the pages list was empty, causing
replay attempt to do nothing.

Signed-off-by: Fred Isaman <iisaman@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

fs/nfs/direct.c

diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 90b00ce..22a40c4 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -768,11 +768,17 @@
 		bytes = min(wsize,count);
 
 		result = -ENOMEM;
-		whdr = nfs_writehdr_alloc(nfs_page_array_len(pgbase, bytes));
+		whdr = nfs_writehdr_alloc();
 		if (unlikely(!whdr))
 			break;
 
-		data = &whdr->rpc_data;
+		data = nfs_writedata_alloc(&whdr->header, nfs_page_array_len(pgbase, bytes));
+		if (!data) {
+			nfs_writehdr_free(&whdr->header);
+			break;
+		}
+		data->header = &whdr->header;
+		atomic_inc(&data->header->refcnt);
 		pages = &data->pages;
 
 		down_read(&current->mm->mmap_sem);