Merge tag 'nfs-for-4.19-1' of git://git.linux-nfs.org/projects/anna/l…

…inux-nfs

Pull NFS client updates from Anna Schumaker:
 "These patches include adding async support for the v4.2 COPY
  operation. I think Bruce is planning to send the server patches for
  the next release, but I figured we could get the client side out of
  the way now since it's been in my tree for a while. This shouldn't
  cause any problems, since the server will still respond with
  synchronous copies even if the client requests async.

  Features:
   - Add support for asynchronous server-side COPY operations

  Stable bufixes:
   - Fix an off-by-one in bl_map_stripe() (v3.17+)
   - NFSv4 client live hangs after live data migration recovery (v4.9+)
   - xprtrdma: Fix disconnect regression (v4.18+)
   - Fix locking in pnfs_generic_recover_commit_reqs (v4.14+)
   - Fix a sleep in atomic context in nfs4_callback_sequence() (v4.9+)

  Other bugfixes and cleanups:
   - Optimizations and fixes involving NFS v4.1 / pNFS layout handling
   - Optimize lseek(fd, SEEK_CUR, 0) on directories to avoid locking
   - Immediately reschedule writeback when the server replies with an
     error
   - Fix excessive attribute revalidation in nfs_execute_ok()
   - Add error checking to nfs_idmap_prepare_message()
   - Use new vm_fault_t return type
   - Return a delegation when reclaiming one that the server has
     recalled
   - Referrals should inherit proto setting from parents
   - Make rpc_auth_create_args a const
   - Improvements to rpc_iostats tracking
   - Fix a potential reference leak when there is an error processing a
     callback
   - Fix rmdir / mkdir / rename nlink accounting
   - Fix updating inode change attribute
   - Fix error handling in nfsn4_sp4_select_mode()
   - Use an appropriate work queue for direct-write completion
   - Don't busy wait if NFSv4 session draining is interrupted"

* tag 'nfs-for-4.19-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (54 commits)
  pNFS: Remove unwanted optimisation of layoutget
  pNFS/flexfiles: ff_layout_pg_init_read should exit on error
  pNFS: Treat RECALLCONFLICT like DELAY...
  pNFS: When updating the stateid in layoutreturn, also update the recall range
  NFSv4: Fix a sleep in atomic context in nfs4_callback_sequence()
  NFSv4: Fix locking in pnfs_generic_recover_commit_reqs
  NFSv4: Fix a typo in nfs4_init_channel_attrs()
  NFSv4: Don't busy wait if NFSv4 session draining is interrupted
  NFS recover from destination server reboot for copies
  NFS add a simple sync nfs4_proc_commit after async COPY
  NFS handle COPY ERR_OFFLOAD_NO_REQS
  NFS send OFFLOAD_CANCEL when COPY killed
  NFS export nfs4_async_handle_error
  NFS handle COPY reply CB_OFFLOAD call race
  NFS add support for asynchronous COPY
  NFS COPY xdr handle async reply
  NFS OFFLOAD_CANCEL xdr
  NFS CB_OFFLOAD xdr
  NFS: Use an appropriate work queue for direct-write completion
  NFSv4: Fix error handling in nfs4_sp4_select_mode()
  ...

fs/nfs/blocklayout/blocklayout.c

@@ -753,6 +753,7 @@ bl_alloc_lseg(struct pnfs_layout_hdr *lo, struct nfs4_layoutget_res *lgr,
 	case -ENODEV:
 		/* Our extent block devices are unavailable */
 		set_bit(NFS_LSEG_UNAVAILABLE, &lseg->pls_flags);
+		/* Fall through */
 	case 0:
 		return lseg;
 	default:

fs/nfs/blocklayout/dev.c

@@ -204,7 +204,7 @@ static bool bl_map_stripe(struct pnfs_block_dev *dev, u64 offset,
 	chunk = div_u64(offset, dev->chunk_size);
 	div_u64_rem(chunk, dev->nr_children, &chunk_idx);
 
-	if (chunk_idx > dev->nr_children) {
+	if (chunk_idx >= dev->nr_children) {
 		dprintk("%s: invalid chunk idx %d (%lld/%lld)\n",
 			__func__, chunk_idx, offset, dev->chunk_size);
 		/* error, should not happen */

fs/nfs/callback.h

@@ -184,6 +184,18 @@ struct cb_notify_lock_args {
 extern __be32 nfs4_callback_notify_lock(void *argp, void *resp,
 					 struct cb_process_state *cps);
 #endif /* CONFIG_NFS_V4_1 */
+#ifdef CONFIG_NFS_V4_2
+struct cb_offloadargs {
+	struct nfs_fh		coa_fh;
+	nfs4_stateid		coa_stateid;
+	uint32_t		error;
+	uint64_t		wr_count;
+	struct nfs_writeverf	wr_writeverf;
+};
+
+extern __be32 nfs4_callback_offload(void *args, void *dummy,
+				    struct cb_process_state *cps);
+#endif /* CONFIG_NFS_V4_2 */
 extern int check_gss_callback_principal(struct nfs_client *, struct svc_rqst *);
 extern __be32 nfs4_callback_getattr(void *argp, void *resp,
 				    struct cb_process_state *cps);

fs/nfs/callback_proc.c

@@ -215,9 +215,9 @@ static u32 pnfs_check_callback_stateid(struct pnfs_layout_hdr *lo,
 {
 	u32 oldseq, newseq;
 
-	/* Is the stateid still not initialised? */
+	/* Is the stateid not initialised? */
 	if (!pnfs_layout_is_valid(lo))
-		return NFS4ERR_DELAY;
+		return NFS4ERR_NOMATCHING_LAYOUT;
 
 	/* Mismatched stateid? */
 	if (!nfs4_stateid_match_other(&lo->plh_stateid, new))
@@ -273,7 +273,6 @@ static u32 initiate_file_draining(struct nfs_client *clp,
 	rv = pnfs_check_callback_stateid(lo, &args->cbl_stateid);
 	if (rv != NFS_OK)
 		goto unlock;
-	pnfs_set_layout_stateid(lo, &args->cbl_stateid, true);
 
 	/*
 	 * Enforce RFC5661 Section 12.5.5.2.1.5 (Bulk Recall and Return)
@@ -283,19 +282,23 @@ static u32 initiate_file_draining(struct nfs_client *clp,
 		goto unlock;
 	}
 
-	if (pnfs_mark_matching_lsegs_return(lo, &free_me_list,
+	pnfs_set_layout_stateid(lo, &args->cbl_stateid, true);
+	switch (pnfs_mark_matching_lsegs_return(lo, &free_me_list,
 				&args->cbl_range,
 				be32_to_cpu(args->cbl_stateid.seqid))) {
+	case 0:
+	case -EBUSY:
+		/* There are layout segments that need to be returned */
 		rv = NFS4_OK;
-		goto unlock;
-	}
-
-	/* Embrace your forgetfulness! */
-	rv = NFS4ERR_NOMATCHING_LAYOUT;
+		break;
+	case -ENOENT:
+		/* Embrace your forgetfulness! */
+		rv = NFS4ERR_NOMATCHING_LAYOUT;
 
-	if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) {
-		NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo,
-			&args->cbl_range);
+		if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) {
+			NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo,
+				&args->cbl_range);
+		}
 	}
 unlock:
 	spin_unlock(&ino->i_lock);
@@ -328,8 +331,6 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 static u32 do_callback_layoutrecall(struct nfs_client *clp,
 				    struct cb_layoutrecallargs *args)
 {
-	write_seqcount_begin(&clp->cl_callback_count);
-	write_seqcount_end(&clp->cl_callback_count);
 	if (args->cbl_recall_type == RETURN_FILE)
 		return initiate_file_draining(clp, args);
 	return initiate_bulk_draining(clp, args);
@@ -441,11 +442,14 @@ validate_seqid(const struct nfs4_slot_table *tbl, const struct nfs4_slot *slot,
  * a match.  If the slot is in use and the sequence numbers match, the
  * client is still waiting for a response to the original request.
  */
-static bool referring_call_exists(struct nfs_client *clp,
+static int referring_call_exists(struct nfs_client *clp,
 				  uint32_t nrclists,
-				  struct referring_call_list *rclists)
+				  struct referring_call_list *rclists,
+				  spinlock_t *lock)
+	__releases(lock)
+	__acquires(lock)
 {
-	bool status = false;
+	int status = 0;
 	int i, j;
 	struct nfs4_session *session;
 	struct nfs4_slot_table *tbl;
@@ -468,8 +472,10 @@ static bool referring_call_exists(struct nfs_client *clp,
 
 		for (j = 0; j < rclist->rcl_nrefcalls; j++) {
 			ref = &rclist->rcl_refcalls[j];
+			spin_unlock(lock);
 			status = nfs4_slot_wait_on_seqid(tbl, ref->rc_slotid,
 					ref->rc_sequenceid, HZ >> 1) < 0;
+			spin_lock(lock);
 			if (status)
 				goto out;
 		}
@@ -546,7 +552,8 @@ __be32 nfs4_callback_sequence(void *argp, void *resp,
 	 * related callback was received before the response to the original
 	 * call.
 	 */
-	if (referring_call_exists(clp, args->csa_nrclists, args->csa_rclists)) {
+	if (referring_call_exists(clp, args->csa_nrclists, args->csa_rclists,
+				&tbl->slot_tbl_lock) < 0) {
 		status = htonl(NFS4ERR_DELAY);
 		goto out_unlock;
 	}
@@ -660,3 +667,57 @@ __be32 nfs4_callback_notify_lock(void *argp, void *resp,
 	return htonl(NFS4_OK);
 }
 #endif /* CONFIG_NFS_V4_1 */
+#ifdef CONFIG_NFS_V4_2
+static void nfs4_copy_cb_args(struct nfs4_copy_state *cp_state,
+				struct cb_offloadargs *args)
+{
+	cp_state->count = args->wr_count;
+	cp_state->error = args->error;
+	if (!args->error) {
+		cp_state->verf.committed = args->wr_writeverf.committed;
+		memcpy(&cp_state->verf.verifier.data[0],
+			&args->wr_writeverf.verifier.data[0],
+			NFS4_VERIFIER_SIZE);
+	}
+}
+
+__be32 nfs4_callback_offload(void *data, void *dummy,
+			     struct cb_process_state *cps)
+{
+	struct cb_offloadargs *args = data;
+	struct nfs_server *server;
+	struct nfs4_copy_state *copy;
+	bool found = false;
+
+	spin_lock(&cps->clp->cl_lock);
+	rcu_read_lock();
+	list_for_each_entry_rcu(server, &cps->clp->cl_superblocks,
+				client_link) {
+		list_for_each_entry(copy, &server->ss_copies, copies) {
+			if (memcmp(args->coa_stateid.other,
+					copy->stateid.other,
+					sizeof(args->coa_stateid.other)))
+				continue;
+			nfs4_copy_cb_args(copy, args);
+			complete(&copy->completion);
+			found = true;
+			goto out;
+		}
+	}
+out:
+	rcu_read_unlock();
+	if (!found) {
+		copy = kzalloc(sizeof(struct nfs4_copy_state), GFP_NOFS);
+		if (!copy) {
+			spin_unlock(&cps->clp->cl_lock);
+			return htonl(NFS4ERR_SERVERFAULT);
+		}
+		memcpy(&copy->stateid, &args->coa_stateid, NFS4_STATEID_SIZE);
+		nfs4_copy_cb_args(copy, args);
+		list_add_tail(&copy->copies, &cps->clp->pending_cb_stateids);
+	}
+	spin_unlock(&cps->clp->cl_lock);
+
+	return 0;
+}
+#endif /* CONFIG_NFS_V4_2 */

fs/nfs/callback_xdr.c

@@ -38,6 +38,9 @@
 #define CB_OP_RECALLSLOT_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ)
 #define CB_OP_NOTIFY_LOCK_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ)
 #endif /* CONFIG_NFS_V4_1 */
+#ifdef CONFIG_NFS_V4_2
+#define CB_OP_OFFLOAD_RES_MAXSZ		(CB_OP_HDR_RES_MAXSZ)
+#endif /* CONFIG_NFS_V4_2 */
 
 #define NFSDBG_FACILITY NFSDBG_CALLBACK
 
@@ -527,7 +530,72 @@ static __be32 decode_notify_lock_args(struct svc_rqst *rqstp,
 }
 
 #endif /* CONFIG_NFS_V4_1 */
+#ifdef CONFIG_NFS_V4_2
+static __be32 decode_write_response(struct xdr_stream *xdr,
+					struct cb_offloadargs *args)
+{
+	__be32 *p;
+
+	/* skip the always zero field */
+	p = read_buf(xdr, 4);
+	if (unlikely(!p))
+		goto out;
+	p++;
+
+	/* decode count, stable_how, verifier */
+	p = xdr_inline_decode(xdr, 8 + 4);
+	if (unlikely(!p))
+		goto out;
+	p = xdr_decode_hyper(p, &args->wr_count);
+	args->wr_writeverf.committed = be32_to_cpup(p);
+	p = xdr_inline_decode(xdr, NFS4_VERIFIER_SIZE);
+	if (likely(p)) {
+		memcpy(&args->wr_writeverf.verifier.data[0], p,
+			NFS4_VERIFIER_SIZE);
+		return 0;
+	}
+out:
+	return htonl(NFS4ERR_RESOURCE);
+}
+
+static __be32 decode_offload_args(struct svc_rqst *rqstp,
+					struct xdr_stream *xdr,
+					void *data)
+{
+	struct cb_offloadargs *args = data;
+	__be32 *p;
+	__be32 status;
+
+	/* decode fh */
+	status = decode_fh(xdr, &args->coa_fh);
+	if (unlikely(status != 0))
+		return status;
 
+	/* decode stateid */
+	status = decode_stateid(xdr, &args->coa_stateid);
+	if (unlikely(status != 0))
+		return status;
+
+	/* decode status */
+	p = read_buf(xdr, 4);
+	if (unlikely(!p))
+		goto out;
+	args->error = ntohl(*p++);
+	if (!args->error) {
+		status = decode_write_response(xdr, args);
+		if (unlikely(status != 0))
+			return status;
+	} else {
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out;
+		p = xdr_decode_hyper(p, &args->wr_count);
+	}
+	return 0;
+out:
+	return htonl(NFS4ERR_RESOURCE);
+}
+#endif /* CONFIG_NFS_V4_2 */
 static __be32 encode_string(struct xdr_stream *xdr, unsigned int len, const char *str)
 {
 	if (unlikely(xdr_stream_encode_opaque(xdr, str, len) < 0))
@@ -773,7 +841,10 @@ preprocess_nfs42_op(int nop, unsigned int op_nr, struct callback_op **op)
 	if (status != htonl(NFS4ERR_OP_ILLEGAL))
 		return status;
 
-	if (op_nr == OP_CB_OFFLOAD)
+	if (op_nr == OP_CB_OFFLOAD) {
+		*op = &callback_ops[op_nr];
+		return htonl(NFS_OK);
+	} else
 		return htonl(NFS4ERR_NOTSUPP);
 	return htonl(NFS4ERR_OP_ILLEGAL);
 }
@@ -883,16 +954,21 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp)
 
 	if (hdr_arg.minorversion == 0) {
 		cps.clp = nfs4_find_client_ident(SVC_NET(rqstp), hdr_arg.cb_ident);
-		if (!cps.clp || !check_gss_callback_principal(cps.clp, rqstp))
+		if (!cps.clp || !check_gss_callback_principal(cps.clp, rqstp)) {
+			if (cps.clp)
+				nfs_put_client(cps.clp);
 			goto out_invalidcred;
+		}
 	}
 
 	cps.minorversion = hdr_arg.minorversion;
 	hdr_res.taglen = hdr_arg.taglen;
 	hdr_res.tag = hdr_arg.tag;
-	if (encode_compound_hdr_res(&xdr_out, &hdr_res) != 0)
+	if (encode_compound_hdr_res(&xdr_out, &hdr_res) != 0) {
+		if (cps.clp)
+			nfs_put_client(cps.clp);
 		return rpc_system_err;
-
+	}
 	while (status == 0 && nops != hdr_arg.nops) {
 		status = process_op(nops, rqstp, &xdr_in,
 				    rqstp->rq_argp, &xdr_out, rqstp->rq_resp,
@@ -969,6 +1045,13 @@ static struct callback_op callback_ops[] = {
 		.res_maxsize = CB_OP_NOTIFY_LOCK_RES_MAXSZ,
 	},
 #endif /* CONFIG_NFS_V4_1 */
+#ifdef CONFIG_NFS_V4_2
+	[OP_CB_OFFLOAD] = {
+		.process_op = nfs4_callback_offload,
+		.decode_args = decode_offload_args,
+		.res_maxsize = CB_OP_OFFLOAD_RES_MAXSZ,
+	},
+#endif /* CONFIG_NFS_V4_2 */
 };
 
 /*

fs/nfs/client.c

@@ -886,6 +886,7 @@ struct nfs_server *nfs_alloc_server(void)
 	INIT_LIST_HEAD(&server->delegations);
 	INIT_LIST_HEAD(&server->layouts);
 	INIT_LIST_HEAD(&server->state_owners_lru);
+	INIT_LIST_HEAD(&server->ss_copies);
 
 	atomic_set(&server->active, 0);
 

fs/nfs/dir.c

@@ -904,23 +904,29 @@ static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence)
 	dfprintk(FILE, "NFS: llseek dir(%pD2, %lld, %d)\n",
 			filp, offset, whence);
 
-	inode_lock(inode);
 	switch (whence) {
-		case 1:
-			offset += filp->f_pos;
-		case 0:
-			if (offset >= 0)
-				break;
-		default:
-			offset = -EINVAL;
-			goto out;
+	default:
+		return -EINVAL;
+	case SEEK_SET:
+		if (offset < 0)
+			return -EINVAL;
+		inode_lock(inode);
+		break;
+	case SEEK_CUR:
+		if (offset == 0)
+			return filp->f_pos;
+		inode_lock(inode);
+		offset += filp->f_pos;
+		if (offset < 0) {
+			inode_unlock(inode);
+			return -EINVAL;
+		}
 	}
 	if (offset != filp->f_pos) {
 		filp->f_pos = offset;
 		dir_ctx->dir_cookie = 0;
 		dir_ctx->duped = 0;
 	}
-out:
 	inode_unlock(inode);
 	return offset;
 }
@@ -1032,7 +1038,7 @@ int nfs_lookup_verify_inode(struct inode *inode, unsigned int flags)
 	if (flags & LOOKUP_REVAL)
 		goto out_force;
 out:
-	return (inode->i_nlink == 0) ? -ENOENT : 0;
+	return (inode->i_nlink == 0) ? -ESTALE : 0;
 out_force:
 	if (flags & LOOKUP_RCU)
 		return -ECHILD;
@@ -2499,7 +2505,9 @@ static int nfs_execute_ok(struct inode *inode, int mask)
 	struct nfs_server *server = NFS_SERVER(inode);
 	int ret = 0;
 
-	if (nfs_check_cache_invalid(inode, NFS_INO_INVALID_ACCESS)) {
+	if (S_ISDIR(inode->i_mode))
+		return 0;
+	if (nfs_check_cache_invalid(inode, NFS_INO_INVALID_OTHER)) {
 		if (mask & MAY_NOT_BLOCK)
 			return -ECHILD;
 		ret = __nfs_revalidate_inode(server, inode);