[PATCH AUTOSEL 4.19 46/68] NFSv4: Fix a NFSv4 state manager deadlock

[PATCH AUTOSEL 4.19 46/68] NFSv4: Fix a NFSv4 state manager deadlock

[Sasha Levin]

From: Trond Myklebust <trond.myklebust@hammerspace.com>

[ Upstream commit aeabb3c96186a0f944fc2b1f25c84d5eb3a93fa9 ]

Fix a deadlock whereby the NFSv4 state manager can get stuck in the
delegation return code, waiting for a layout return to complete in
another thread. If the server reboots before that other thread
completes, then we need to be able to start a second state
manager thread in order to perform recovery.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/nfs/nfs4_fs.h   |  2 ++
 fs/nfs/nfs4state.c | 16 +++++++++++-----
 2 files changed, 13 insertions(+), 5 deletions(-)

diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index 3a6904173214..63287d911c08 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -41,6 +41,8 @@ enum nfs4_client_state {
 	NFS4CLNT_MOVED,
 	NFS4CLNT_LEASE_MOVED,
 	NFS4CLNT_DELEGATION_EXPIRED,
+	NFS4CLNT_RUN_MANAGER,
+	NFS4CLNT_DELEGRETURN_RUNNING,
 };
 
 #define NFS4_RENEW_TIMEOUT		0x01
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 18920152da14..d2f645d34eb1 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -1210,6 +1210,7 @@ void nfs4_schedule_state_manager(struct nfs_client *clp)
 	struct task_struct *task;
 	char buf[INET6_ADDRSTRLEN + sizeof("-manager") + 1];
 
+	set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
 	if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
 		return;
 	__module_get(THIS_MODULE);
@@ -2485,6 +2486,7 @@ static void nfs4_state_manager(struct nfs_client *clp)
 
 	/* Ensure exclusive access to NFSv4 state */
 	do {
+		clear_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
 		if (test_bit(NFS4CLNT_PURGE_STATE, &clp->cl_state)) {
 			section = "purge state";
 			status = nfs4_purge_lease(clp);
@@ -2575,14 +2577,18 @@ static void nfs4_state_manager(struct nfs_client *clp)
 		}
 
 		nfs4_end_drain_session(clp);
-		if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
-			nfs_client_return_marked_delegations(clp);
-			continue;
+		nfs4_clear_state_manager_bit(clp);
+
+		if (!test_and_set_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state)) {
+			if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
+				nfs_client_return_marked_delegations(clp);
+				set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
+			}
+			clear_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state);
 		}
 
-		nfs4_clear_state_manager_bit(clp);
 		/* Did we race with an attempt to give us more work? */
-		if (clp->cl_state == 0)
+		if (!test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state))
 			return;
 		if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
 			return;
-- 
2.17.1

[PATCH AUTOSEL 4.19 57/68] NFSv4.2 copy do not allocate memory under the lock

[Sasha Levin]

From: Olga Kornievskaia <kolga@netapp.com>

[ Upstream commit 99f2c55591fb5c1b536263970d98c2ebc2089906 ]

Bruce pointed out that we shouldn't allocate memory while holding
a lock in the nfs4_callback_offload() and handle_async_copy()
that deal with a racing CB_OFFLOAD and reply to COPY case.

Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/nfs/callback_proc.c | 22 +++++++++++-----------
 fs/nfs/nfs42proc.c     | 19 ++++++++++---------
 2 files changed, 21 insertions(+), 20 deletions(-)

diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c
index fa515d5ea5ba..48b2e9063b0a 100644
--- a/fs/nfs/callback_proc.c
+++ b/fs/nfs/callback_proc.c
@@ -686,20 +686,24 @@ __be32 nfs4_callback_offload(void *data, void *dummy,
 {
 	struct cb_offloadargs *args = data;
 	struct nfs_server *server;
-	struct nfs4_copy_state *copy;
+	struct nfs4_copy_state *copy, *tmp_copy;
 	bool found = false;
 
+	copy = kzalloc(sizeof(struct nfs4_copy_state), GFP_NOFS);
+	if (!copy)
+		return htonl(NFS4ERR_SERVERFAULT);
+
 	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) {
+		list_for_each_entry(tmp_copy, &server->ss_copies, copies) {
 			if (memcmp(args->coa_stateid.other,
-					copy->stateid.other,
+					tmp_copy->stateid.other,
 					sizeof(args->coa_stateid.other)))
 				continue;
-			nfs4_copy_cb_args(copy, args);
-			complete(&copy->completion);
+			nfs4_copy_cb_args(tmp_copy, args);
+			complete(&tmp_copy->completion);
 			found = true;
 			goto out;
 		}
@@ -707,15 +711,11 @@ __be32 nfs4_callback_offload(void *data, void *dummy,
 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);
-	}
+	} else
+		kfree(copy);
 	spin_unlock(&cps->clp->cl_lock);
 
 	return 0;
diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c
index ac5b784a1de0..fed06fd9998d 100644
--- a/fs/nfs/nfs42proc.c
+++ b/fs/nfs/nfs42proc.c
@@ -137,31 +137,32 @@ static int handle_async_copy(struct nfs42_copy_res *res,
 			     struct file *dst,
 			     nfs4_stateid *src_stateid)
 {
-	struct nfs4_copy_state *copy;
+	struct nfs4_copy_state *copy, *tmp_copy;
 	int status = NFS4_OK;
 	bool found_pending = false;
 	struct nfs_open_context *ctx = nfs_file_open_context(dst);
 
+	copy = kzalloc(sizeof(struct nfs4_copy_state), GFP_NOFS);
+	if (!copy)
+		return -ENOMEM;
+
 	spin_lock(&server->nfs_client->cl_lock);
-	list_for_each_entry(copy, &server->nfs_client->pending_cb_stateids,
+	list_for_each_entry(tmp_copy, &server->nfs_client->pending_cb_stateids,
 				copies) {
-		if (memcmp(&res->write_res.stateid, &copy->stateid,
+		if (memcmp(&res->write_res.stateid, &tmp_copy->stateid,
 				NFS4_STATEID_SIZE))
 			continue;
 		found_pending = true;
-		list_del(&copy->copies);
+		list_del(&tmp_copy->copies);
 		break;
 	}
 	if (found_pending) {
 		spin_unlock(&server->nfs_client->cl_lock);
+		kfree(copy);
+		copy = tmp_copy;
 		goto out;
 	}
 
-	copy = kzalloc(sizeof(struct nfs4_copy_state), GFP_NOFS);
-	if (!copy) {
-		spin_unlock(&server->nfs_client->cl_lock);
-		return -ENOMEM;
-	}
 	memcpy(&copy->stateid, &res->write_res.stateid, NFS4_STATEID_SIZE);
 	init_completion(&copy->completion);
 	copy->parent_state = ctx->state;
-- 
2.17.1

[PATCH AUTOSEL 4.19 58/68] flexfiles: use per-mirror specified stateid for IO

[Sasha Levin]

From: Tigran Mkrtchyan <tigran.mkrtchyan@desy.de>

[ Upstream commit bb21ce0ad227b69ec0f83279297ee44232105d96 ]

rfc8435 says:

  For tight coupling, ffds_stateid provides the stateid to be used by
  the client to access the file.

However current implementation replaces per-mirror provided stateid with
by open or lock stateid.

Ensure that per-mirror stateid is used by ff_layout_write_prepare_v4 and
nfs4_ff_layout_prepare_ds.

Signed-off-by: Tigran Mkrtchyan <tigran.mkrtchyan@desy.de>
Signed-off-by: Rick Macklem <rmacklem@uoguelph.ca>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/nfs/flexfilelayout/flexfilelayout.c    | 21 +++++++++------------
 fs/nfs/flexfilelayout/flexfilelayout.h    |  4 ++++
 fs/nfs/flexfilelayout/flexfilelayoutdev.c | 19 +++++++++++++++++++
 3 files changed, 32 insertions(+), 12 deletions(-)

diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c
index cae43333ef16..86ac2c5b93fe 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.c
+++ b/fs/nfs/flexfilelayout/flexfilelayout.c
@@ -1361,12 +1361,7 @@ static void ff_layout_read_prepare_v4(struct rpc_task *task, void *data)
 				task))
 		return;
 
-	if (ff_layout_read_prepare_common(task, hdr))
-		return;
-
-	if (nfs4_set_rw_stateid(&hdr->args.stateid, hdr->args.context,
-			hdr->args.lock_context, FMODE_READ) == -EIO)
-		rpc_exit(task, -EIO); /* lost lock, terminate I/O */
+	ff_layout_read_prepare_common(task, hdr);
 }
 
 static void ff_layout_read_call_done(struct rpc_task *task, void *data)
@@ -1542,12 +1537,7 @@ static void ff_layout_write_prepare_v4(struct rpc_task *task, void *data)
 				task))
 		return;
 
-	if (ff_layout_write_prepare_common(task, hdr))
-		return;
-
-	if (nfs4_set_rw_stateid(&hdr->args.stateid, hdr->args.context,
-			hdr->args.lock_context, FMODE_WRITE) == -EIO)
-		rpc_exit(task, -EIO); /* lost lock, terminate I/O */
+	ff_layout_write_prepare_common(task, hdr);
 }
 
 static void ff_layout_write_call_done(struct rpc_task *task, void *data)
@@ -1742,6 +1732,10 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr)
 	fh = nfs4_ff_layout_select_ds_fh(lseg, idx);
 	if (fh)
 		hdr->args.fh = fh;
+
+	if (!nfs4_ff_layout_select_ds_stateid(lseg, idx, &hdr->args.stateid))
+		goto out_failed;
+
 	/*
 	 * Note that if we ever decide to split across DSes,
 	 * then we may need to handle dense-like offsets.
@@ -1804,6 +1798,9 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
 	if (fh)
 		hdr->args.fh = fh;
 
+	if (!nfs4_ff_layout_select_ds_stateid(lseg, idx, &hdr->args.stateid))
+		goto out_failed;
+
 	/*
 	 * Note that if we ever decide to split across DSes,
 	 * then we may need to handle dense-like offsets.
diff --git a/fs/nfs/flexfilelayout/flexfilelayout.h b/fs/nfs/flexfilelayout/flexfilelayout.h
index 411798346e48..de50a342d5a5 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.h
+++ b/fs/nfs/flexfilelayout/flexfilelayout.h
@@ -215,6 +215,10 @@ unsigned int ff_layout_fetch_ds_ioerr(struct pnfs_layout_hdr *lo,
 		unsigned int maxnum);
 struct nfs_fh *
 nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx);
+int
+nfs4_ff_layout_select_ds_stateid(struct pnfs_layout_segment *lseg,
+				u32 mirror_idx,
+				nfs4_stateid *stateid);
 
 struct nfs4_pnfs_ds *
 nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c
index 59aa04976331..a8df2f496898 100644
--- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c
+++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c
@@ -370,6 +370,25 @@ nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx)
 	return fh;
 }
 
+int
+nfs4_ff_layout_select_ds_stateid(struct pnfs_layout_segment *lseg,
+				u32 mirror_idx,
+				nfs4_stateid *stateid)
+{
+	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, mirror_idx);
+
+	if (!ff_layout_mirror_valid(lseg, mirror, false)) {
+		pr_err_ratelimited("NFS: %s: No data server for mirror offset index %d\n",
+			__func__, mirror_idx);
+		goto out;
+	}
+
+	nfs4_stateid_copy(stateid, &mirror->stateid);
+	return 1;
+out:
+	return 0;
+}
+
 /**
  * nfs4_ff_layout_prepare_ds - prepare a DS connection for an RPC call
  * @lseg: the layout segment we're operating on
-- 
2.17.1