- 17 Nov, 2021 4 commits
-
-
Olga Kornievskaia authored
When the client receives ERR_NOSPC on reply to CREATE_SESSION it leads to a client hanging in nfs_wait_client_init_complete(). Instead, complete and fail the client initiation with an EIO error which allows for the mount command to fail instead of hanging. Signed-off-by:
Olga Kornievskaia <kolga@netapp.com> Signed-off-by:
Trond Myklebust <trond.myklebust@hammerspace.com>
-
Benjamin Coddington authored
The mechanism in use to allow the client to see the results of COPY/CLONE is to drop those pages from the pagecache. This forces the client to read those pages once more from the server. However, truncate_pagecache_range() zeros out partial pages instead of dropping them. Let us instead use invalidate_inode_pages2_range() with full-page offsets to ensure the client properly sees the results of COPY/CLONE operations. Cc: <stable@vger.kernel.org> # v4.7+ Fixes: 2e72448b ("NFS: Add COPY nfs operation") Signed-off-by:
Benjamin Coddington <bcodding@redhat.com> Signed-off-by:
-
Benjamin Coddington authored
This provides some insight into the client's invalidation behavior to show both when the client uses the helper, and the results of calling the helper which can vary depending on how the helper is called. Signed-off-by:
-
Trond Myklebust authored
The failure to retrieve post-op attributes has no bearing on whether or not the clone operation itself was successful. We must therefore ignore the return value of decode_getfattr() when looking at the success or failure of nfs4_xdr_dec_clone(). Fixes: 36022770 ("nfs42: add CLONE xdr functions") Signed-off-by:
-
- 07 Nov, 2021 1 commit
-
-
Trond Myklebust authored
Ensure that the values supplied by the server do not exceed the size of the largest allowed slot table. Reported-by: <rtm@csail.mit.edu> Signed-off-by:
-
- 05 Nov, 2021 21 commits
-
-
Anna Schumaker authored
Wa can check if the fattr has an allocated label when needed Signed-off-by:
Anna Schumaker <Anna.Schumaker@Netapp.com> Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
Pull the label from the fattr instead. Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
Signed-off-by:
-
Anna Schumaker authored
And usethe fattr's label field instead. I also adjust function calls to remove labels along the way. Signed-off-by:
-
Anna Schumaker authored
Again, use the fattr's label field instead. Signed-off-by:
-
Anna Schumaker authored
Instead, use the label embedded in the attached fattr. Signed-off-by:
-
Anna Schumaker authored
And instead allocate the fattr using nfs_alloc_fattr_with_label() Signed-off-by:
-
Anna Schumaker authored
For creating fattrs with the label field already allocated for us. I also update nfs_free_fattr() to free the label in the end. Signed-off-by:
-
Trond Myklebust authored
We're about to add a check in nfs_free_fattr() for whether or not the label is non-zero. Signed-off-by: -
Trond Myklebust authored
Signed-off-by: -
Trond Myklebust authored
The preferred behaviour is always to allocate struct nfs_fattr from the slab. Signed-off-by: -
Trond Myklebust authored
It is completely redundant to the server capability check. Signed-off-by: -
Trond Myklebust authored
The return value of xdr_inline_decode() is not being checked, leading to a potential Oops. Just replace the open coded array decode with the generic XDR version. Reported-by: <rtm@csail.mit.edu> Signed-off-by: -
Trond Myklebust authored
The server is supposed to return the same tag that the client sends in the outgoing RPC call, but we should still sanity check the length just in case. Reported-by: <rtm@csail.mit.edu> Signed-off-by: -
Trond Myklebust authored
If fhandle is NULL or fattr is NULL, then 'error' is uninitialised. Reported-by:
Nathan Chancellor <nathan@kernel.org> Signed-off-by:
-
- 04 Nov, 2021 10 commits
-
-
Trond Myklebust authored
If we have already set up the socket and are waiting for it to connect, then don't immediately close and retry. Signed-off-by: -
Trond Myklebust authored
Ensure that we bump the xprt->connect_cookie when we set the XPRT_CLOSE_WAIT flag so that another call to xprt_conditional_disconnect() won't race with the reconnection. Signed-off-by: -
Olga Kornievskaia authored
Add tracepoint to OFFLOAD_CANCEL operation. Signed-off-by:
-
Olga Kornievskaia authored
Add a tracepoint to COPY_NOTIFY operation. Signed-off-by:
-
Olga Kornievskaia authored
Add a tracepoint to the CB_OFFLOAD operation. Signed-off-by:
-
Olga Kornievskaia authored
Add a tracepoint to the CLONE operation. Signed-off-by:
-
Olga Kornievskaia authored
Add a tracepoint to the COPY operation. Signed-off-by:
-
Olga Kornievskaia authored
Add a tracepoint to the FALLOCATE/DEALLOCATE operations. Signed-off-by:
-
Olga Kornievskaia authored
Add a tracepoint to the SEEK operation. Signed-off-by:
-
Anna Schumaker authored
xprts don't immediately reconnect when changing the "dstaddr" property, instead this gets handled the next time an operation uses the transport. This could lead to NULL pointer dereferences when trying to read sysfs files between the disconnect and reconnect operations. Fix this by returning an error if the xprt is not connected. Signed-off-by:
-
- 03 Nov, 2021 2 commits
-
-
Jonathan Corbet authored
Commit 19fcae3d ("scsi: remove the SCSI OSD library") deleted the last file that included <linux/pnfs_osd_xdr.h> but left that file behind. It's unused, get rid of it now. Cc: Christoph Hellwig <hch@lst.de> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Anna Schumaker <anna.schumaker@netapp.com> Cc: linux-nfs@vger.kernel.org Signed-off-by:
Jonathan Corbet <corbet@lwn.net> Signed-off-by:
-
Scott Mayhew authored
During umount, the session slot tables are freed. If there are outstanding FREE_STATEID tasks, a use-after-free and slab corruption can occur when rpc_exit_task calls rpc_call_done -> nfs41_sequence_done -> nfs4_sequence_process/nfs41_sequence_free_slot. Prevent that from happening by taking a reference on the nfs_client in nfs41_free_stateid and putting it in nfs41_free_stateid_release. Signed-off-by:
Scott Mayhew <smayhew@redhat.com> Signed-off-by:
-
- 02 Nov, 2021 2 commits
-
-
Dave Wysochanski authored
Add the byte offset of the readahead request to the tracepoint output so we know where the read starts. Before this patch: cat-8104 [002] ..... 813.168775: nfs_aop_readahead: fileid=00:31:141 fhandle=0xe55807f6 version=1756509392533525500 nr_pages=256 cat-8104 [002] ..... 813.174973: nfs_aop_readahead_done: fileid=00:31:141 fhandle=0xe55807f6 version=1756509392533525500 nr_pages=256 ret=0 cat-8104 [002] ..... 813.175963: nfs_aop_readahead: fileid=00:31:141 fhandle=0xe55807f6 version=1756509392533525500 nr_pages=256 cat-8104 [002] ..... 813.183742: nfs_aop_readahead_done: fileid=00:31:141 fhandle=0xe55807f6 version=1756509392533525500 nr_pages=1 ret=0 After this patch: cat-6392 [001] ..... 73.107782: nfs_aop_readahead: fileid=00:31:141 fhandle=0xed22403f version=1756511950029502774 offset=5242880 nr_pages=256 cat-6392 [001] ..... 73.112466: nfs_aop_readahead_done: fileid=00:31:141 fhandle=0xed22403f version=1756511950029502774 nr_pages=256 ret=0 cat-6392 [001] ..... 73.115692: nfs_aop_readahead: fileid=00:31:141 fhandle=0xed22403f version=1756511950029502774 offset=6291456 nr_pages=256 cat-6392 [001] ..... 73.123283: nfs_aop_readahead_done: fileid=00:31:141 fhandle=0xed22403f version=1756511950029502774 nr_pages=256 ret=0 Signed-off-by:
Dave Wysochanski <dwysocha@redhat.com> Signed-off-by:
-
Benjamin Coddington authored
This minor fix-up keeps GCC from complaining that "last' may be used uninitialized", which breaks some build workflows that have been running with all warnings treated as errors. Signed-off-by:
-
