- 05 Oct, 2022 2 commits
-
-
Jeff Layton authored
nfsd_file is RCU-freed, so we need to hold the rcu_read_lock long enough to get a reference after finding it in the hash. Take the rcu_read_lock() and call rhashtable_lookup directly. Switch to using rhashtable_lookup_insert_key as well, and use the usual retry mechanism if we hit an -EEXIST. Rename the "retry" bool to open_retry, and eliminiate the insert_err goto target. Signed-off-by:
Jeff Layton <jlayton@kernel.org> Signed-off-by:
Chuck Lever <chuck.lever@oracle.com>
-
Jeff Layton authored
nfsd_file_unhash_and_dispose() is called for two reasons: We're either shutting down and purging the filecache, or we've gotten a notification about a file delete, so we want to go ahead and unhash it so that it'll get cleaned up when we close. We're either walking the hashtable or doing a lookup in it and we don't take a reference in either case. What we want to do in both cases is to try and unhash the object and put it on the dispose list if that was successful. If it's no longer hashed, then we don't want to touch it, with the assumption being that something else is already cleaning up the sentinel reference. Instead of trying to selectively decrement the refcount in this function, just unhash it, and if that was successful, move it to the dispose list. Then, the disposal routine will just clean that up as usual. Also, just make this a void function, drop the WARN_ON_ONCE, and the comments about deadlocking since the nature of the purported deadlock is no longer clear. Signed-off-by:
-
- 26 Sep, 2022 38 commits
-
-
Jeff Layton authored
We've had some reports of problems in the refcounting for delegation stateids that we've yet to track down. Add some extra checks to ensure that we've removed the object from various lists before freeing it. Link: https://bugzilla.redhat.com/show_bug.cgi?id=2127067Signed-off-by:
-
Jeff Layton authored
queue_work can return false and not queue anything, if the work is already queued. If that happens in the case of a CB_RECALL, we'll have taken an extra reference to the stid that will never be put. Ensure we throw a warning in that case. Signed-off-by:
-
Jeff Layton authored
Signed-off-by:
-
Jeff Layton authored
In the case of a revoked delegation, we still fill out the pointer even when returning an error, which is bad form. Only overwrite the pointer on success. Signed-off-by:
-
Dai Ngo authored
Use-after-free occurred when the laundromat tried to free expired cpntf_state entry on the s2s_cp_stateids list after inter-server copy completed. The sc_cp_list that the expired copy state was inserted on was already freed. When COPY completes, the Linux client normally sends LOCKU(lock_state x), FREE_STATEID(lock_state x) and CLOSE(open_state y) to the source server. The nfs4_put_stid call from nfsd4_free_stateid cleans up the copy state from the s2s_cp_stateids list before freeing the lock state's stid. However, sometimes the CLOSE was sent before the FREE_STATEID request. When this happens, the nfsd4_close_open_stateid call from nfsd4_close frees all lock states on its st_locks list without cleaning up the copy state on the sc_cp_list list. When the time the FREE_STATEID arrives the server returns BAD_STATEID since the lock state was freed. This causes the use-after-free error to occur when the laundromat tries to free the expired cpntf_state. This patch adds a call to nfs4_free_cpntf_statelist in nfsd4_close_open_stateid to clean up the copy state before calling free_ol_stateid_reaplist to free the lock state's stid on the reaplist. Signed-off-by:
Dai Ngo <dai.ngo@oracle.com> Signed-off-by:
-
Chuck Lever authored
Since before the git era, NFSD has conserved the number of pages held by each nfsd thread by combining the RPC receive and send buffers into a single array of pages. This works because there are no cases where an operation needs a large RPC Call message and a large RPC Reply at the same time. Once an RPC Call has been received, svc_process() updates svc_rqst::rq_res to describe the part of rq_pages that can be used for constructing the Reply. This means that the send buffer (rq_res) shrinks when the received RPC record containing the RPC Call is large. Add an NFSv4 helper that computes the size of the send buffer. It replaces svc_max_payload() in spots where svc_max_payload() returns a value that might be larger than the remaining send buffer space. Callers who need to know the transport's actual maximum payload size will continue to use svc_max_payload(). Signed-off-by: -
Chuck Lever authored
Code maintenance: The name of the copy_stateid_t::sc_count field collides with the sc_count field in struct nfs4_stid, making the latter difficult to grep for when auditing stateid reference counting. No behavior change expected. Signed-off-by: -
ChenXiaoSong authored
Use DEFINE_SHOW_ATTRIBUTE helper macro to simplify the code. Signed-off-by:
ChenXiaoSong <chenxiaosong2@huawei.com> Signed-off-by:
-
ChenXiaoSong authored
Use DEFINE_SHOW_ATTRIBUTE helper macro to simplify the code. nfsd_net is converted from seq_file->file instead of seq_file->private in nfsd_reply_cache_stats_show(). Signed-off-by:
-
ChenXiaoSong authored
Use DEFINE_SHOW_ATTRIBUTE helper macro to simplify the code. inode is converted from seq_file->file instead of seq_file->private in client_info_show(). Signed-off-by:
-
ChenXiaoSong authored
Use DEFINE_SHOW_ATTRIBUTE helper macro to simplify the code. Signed-off-by:
-
ChenXiaoSong authored
Use DEFINE_PROC_SHOW_ATTRIBUTE helper macro to simplify the code. Signed-off-by:
-
Chuck Lever authored
Remove a couple of 4-byte holes on platforms with 64-bit pointers. Signed-off-by: -
Chuck Lever authored
This field was added by commit 1091006c ("nfsd: turn on reply cache for NFSv4") but was never put to use. Signed-off-by:
-
Chuck Lever authored
These helpers are always invoked indirectly, so the compiler can't inline these anyway. While we're updating the synopses of these helpers, defensively convert their parameters to const pointers. Signed-off-by: -
Chuck Lever authored
In today's Linux NFS server implementation, the NFS dispatcher initializes each XDR result stream, and the NFSv4 .pc_func and .pc_encode methods all use xdr_stream-based encoding. This keeps rq_res.len automatically updated. There is no longer a need for the WARN_ON_ONCE() check in nfs4svc_encode_compoundres(). Reviewed-by:
-
Chuck Lever authored
Fix a typo. Reviewed-by:
-
Chuck Lever authored
xdr_stream_subsegment() already returns a boolean value. Reviewed-by:
-
Chuck Lever authored
Replace the check for buffer over/underflow with a helper that is commonly used for this purpose. The helper also sets xdr->nwords correctly after successfully linearizing the symlink argument into the stream's scratch buffer. Reviewed-by:
-
Chuck Lever authored
The dust has settled a bit and it's become obvious what code is totally common between nfsd_init_dirlist_pages() and nfsd3_init_dirlist_pages(). Move that common code to SUNRPC. The new helper brackets the existing xdr_init_decode_pages() API. Signed-off-by: -
Chuck Lever authored
Note the function returns a per-transport value, not a per-request value (eg, one that is related to the size of the available send or receive buffer space). Signed-off-by: -
Chuck Lever authored
Have SunRPC clear everything except for the iops array. Then have each NFSv4 XDR decoder clear it's own argument before decoding. Now individual operations may have a large argument struct while not penalizing the vast majority of operations with a small struct. And, clearing the argument structure occurs as the argument fields are initialized, enabling the CPU to do write combining on that memory. In some cases, clearing is not even necessary because all of the fields in the argument structure are initialized by the decoder. Signed-off-by: -
Chuck Lever authored
Currently, SUNRPC clears the whole of .pc_argsize before processing each incoming RPC transaction. Add an extra parameter to struct svc_procedure to enable upper layers to reduce the amount of each operation's argument structure that is zeroed by SUNRPC. The size of struct nfsd4_compoundargs, in particular, is a lot to clear on each incoming RPC Call. A subsequent patch will cut this down to something closer to what NFSv2 and NFSv3 uses. This patch should cause no behavior changes. Signed-off-by: -
Chuck Lever authored
Move exception handling code out of the hot path, and avoid the need for a bswap of a non-constant. Signed-off-by: -
Dai Ngo authored
Add courtesy_client_reaper to react to low memory condition triggered by the system memory shrinker. The delayed_work for the courtesy_client_reaper is scheduled on the shrinker's count callback using the laundry_wq. The shrinker's scan callback is not used for expiring the courtesy clients due to potential deadlocks. Signed-off-by:
-
Dai Ngo authored
Add counter nfs4_courtesy_client_count to nfsd_net to keep track of the number of courtesy clients in the system. Signed-off-by:
-
Anna Schumaker authored
This was discussed with Chuck as part of this patch set. Returning nfserr_resource was decided to not be the best error message here, and he suggested changing to nfserr_serverfault instead. Signed-off-by:
Anna Schumaker <Anna.Schumaker@Netapp.com> Link: https://lore.kernel.org/linux-nfs/20220907195259.926736-1-anna@kernel.org/T/#tSigned-off-by:
-
Chuck Lever authored
nfsd_unlink() can kick off a CB_RECALL (via vfs_unlink() -> leases_conflict()) if a delegation is present. Before returning NFS4ERR_DELAY, give the client holding that delegation a chance to return it and then retry the nfsd_unlink() again, once. Link: https://bugzilla.linux-nfs.org/show_bug.cgi?id=354Tested-by:
Igor Mammedov <imammedo@redhat.com> Signed-off-by:
-
Chuck Lever authored
nfsd_rename() can kick off a CB_RECALL (via vfs_rename() -> leases_conflict()) if a delegation is present. Before returning NFS4ERR_DELAY, give the client holding that delegation a chance to return it and then retry the nfsd_rename() again, once. This version of the patch handles renaming an existing file, but does not deal with renaming onto an existing file. That case will still always trigger an NFS4ERR_DELAY. Link: https://bugzilla.linux-nfs.org/show_bug.cgi?id=354Tested-by:
-
Chuck Lever authored
nfsd_setattr() can kick off a CB_RECALL (via notify_change() -> break_lease()) if a delegation is present. Before returning NFS4ERR_DELAY, give the client holding that delegation a chance to return it and then retry the nfsd_setattr() again, once. Link: https://bugzilla.linux-nfs.org/show_bug.cgi?id=354Tested-by:
-
Chuck Lever authored
Move code that will be retried (in a subsequent patch) into a helper function. Signed-off-by:
-
Chuck Lever authored
Subsequent patches will use this mechanism to wake up an operation that is waiting for a client to return a delegation. The new tracepoint records whether the wait timed out or was properly awoken by the expected DELEGRETURN: nfsd-1155 [002] 83799.493199: nfsd_delegret_wakeup: xid=0x14b7d6ef fh_hash=0xf6826792 (timed out) Suggested-by: -
Chuck Lever authored
Wireshark has always been lousy about dissecting NFSv4 callbacks, especially NFSv4.0 backchannel requests. Add tracepoints so we can surgically capture these events in the trace log. Tracepoints are time-stamped and ordered so that we can now observe the timing relationship between a CB_RECALL Reply and the client's DELEGRETURN Call. Example: nfsd-1153 [002] 211.986391: nfsd_cb_recall: addr=192.168.1.67:45767 client 62ea82e4:fee7492a stateid 00000003:00000001 nfsd-1153 [002] 212.095634: nfsd_compound: xid=0x0000002c opcnt=2 nfsd-1153 [002] 212.095647: nfsd_compound_status: op=1/2 OP_PUTFH status=0 nfsd-1153 [002] 212.095658: nfsd_file_put: hash=0xf72 inode=0xffff9291148c7410 ref=3 flags=HASHED|REFERENCED may=READ file=0xffff929103b3ea00 nfsd-1153 [002] 212.095661: nfsd_compound_status: op=2/2 OP_DELEGRETURN status=0 kworker/u25:8-148 [002] 212.096713: nfsd_cb_recall_done: client 62ea82e4:fee7492a stateid 00000003:00000001 status=0 Signed-off-by: -
Chuck Lever authored
The Linux NFSv4 client implementation does not use COMPOUND tags, but the Solaris and MacOS implementations do, and so does pynfs. Record these eye-catchers in the server's trace buffer to annotate client requests while troubleshooting. Signed-off-by:
-
Chuck Lever authored
Record permission errors in the trace log. Note that the new trace event is conditional, so it will only record non-zero return values from nfsd_permission(). Signed-off-by:
-
Gaosheng Cui authored
nfsd4_prepare_cb_recall() has been removed since commit 0162ac2b ("nfsd: introduce nfsd4_callback_ops"), so remove it. Signed-off-by:
Gaosheng Cui <cuigaosheng1@huawei.com> Signed-off-by:
-
Jeff Layton authored
We only need the inode number for this, not a full rack of attributes. Rename this function make it take a pointer to a u64 instead of struct kstat, and change it to just request STATX_INO. Signed-off-by:
-
Chuck Lever authored
If an NFS server returns NFS4ERR_RESOURCE on the first operation in an NFSv4 COMPOUND, there's no way for a client to know where the problem is and then simplify the compound to make forward progress. So instead, make NFSD process as many operations in an oversized COMPOUND as it can and then return NFS4ERR_RESOURCE on the first operation it did not process. pynfs NFSv4.0 COMP6 exercises this case, but checks only for the COMPOUND status code, not whether the server has processed any of the operations. pynfs NFSv4.1 SEQ6 and SEQ7 exercise the NFSv4.1 case, which detects too many operations per COMPOUND by checking against the limits negotiated when the session was created. Suggested-by:
Bruce Fields <bfields@fieldses.org> Fixes: 0078117c ("nfsd: return RESOURCE not GARBAGE_ARGS on too many ops") Signed-off-by:
-
