nfs: handle request add failure properly
authorPeng Tao <tao.peng@primarydata.com>
Sat, 5 Dec 2015 07:57:31 +0000 (15:57 +0800)
committerTrond Myklebust <trond.myklebust@primarydata.com>
Mon, 28 Dec 2015 19:32:37 +0000 (14:32 -0500)
commit0bcbf039f6b2bcefe4f5dada76079080edf9ecd0
tree8edf8f581dcd12258d7ab84898261a3a3d0ab2ef
parent2bff2288579f1e4af2f05a7f7443c85b7766d5ac
nfs: handle request add failure properly

When we fail to queue a read page to IO descriptor,
we need to clean it up otherwise it is hanging around
preventing nfs module from being removed.

When we fail to queue a write page to IO descriptor,
we need to clean it up and also save the failure status
to open context. Then at file close, we can try to write
pages back again and drop the page if it fails to writeback
in .launder_page, which will be done in the next patch.

Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
fs/nfs/inode.c
fs/nfs/internal.h
fs/nfs/pnfs.c
fs/nfs/read.c
fs/nfs/write.c