linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Sasha Levin <sashal@kernel.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: Alexander Aring <aahringo@redhat.com>,
	David Teigland <teigland@redhat.com>,
	Sasha Levin <sashal@kernel.org>,
	cluster-devel@redhat.com
Subject: [PATCH AUTOSEL 5.13 15/59] fs: dlm: fix srcu read lock usage
Date: Mon,  5 Jul 2021 11:27:31 -0400	[thread overview]
Message-ID: <20210705152815.1520546-15-sashal@kernel.org> (raw)
In-Reply-To: <20210705152815.1520546-1-sashal@kernel.org>

From: Alexander Aring <aahringo@redhat.com>

[ Upstream commit b38bc9c2b3171f4411d80015ecb876bc6f9bcd26 ]

This patch holds the srcu connection read lock in cases where we lookup
the connections and accessing it. We don't hold the srcu lock in workers
function where the scheduled worker is part of the connection itself.
The connection should not be freed if any worker is scheduled or
pending.

Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/dlm/lowcomms.c | 75 ++++++++++++++++++++++++++++++++---------------
 1 file changed, 52 insertions(+), 23 deletions(-)

diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 166e36fcf3e4..47bf99373f3e 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -113,6 +113,7 @@ struct writequeue_entry {
 	int len;
 	int end;
 	int users;
+	int idx; /* get()/commit() idx exchange */
 	struct connection *con;
 };
 
@@ -163,21 +164,14 @@ static inline int nodeid_hash(int nodeid)
 	return nodeid & (CONN_HASH_SIZE-1);
 }
 
-static struct connection *__find_con(int nodeid)
+static struct connection *__find_con(int nodeid, int r)
 {
-	int r, idx;
 	struct connection *con;
 
-	r = nodeid_hash(nodeid);
-
-	idx = srcu_read_lock(&connections_srcu);
 	hlist_for_each_entry_rcu(con, &connection_hash[r], list) {
-		if (con->nodeid == nodeid) {
-			srcu_read_unlock(&connections_srcu, idx);
+		if (con->nodeid == nodeid)
 			return con;
-		}
 	}
-	srcu_read_unlock(&connections_srcu, idx);
 
 	return NULL;
 }
@@ -216,7 +210,8 @@ static struct connection *nodeid2con(int nodeid, gfp_t alloc)
 	struct connection *con, *tmp;
 	int r, ret;
 
-	con = __find_con(nodeid);
+	r = nodeid_hash(nodeid);
+	con = __find_con(nodeid, r);
 	if (con || !alloc)
 		return con;
 
@@ -230,8 +225,6 @@ static struct connection *nodeid2con(int nodeid, gfp_t alloc)
 		return NULL;
 	}
 
-	r = nodeid_hash(nodeid);
-
 	spin_lock(&connections_lock);
 	/* Because multiple workqueues/threads calls this function it can
 	 * race on multiple cpu's. Instead of locking hot path __find_con()
@@ -239,7 +232,7 @@ static struct connection *nodeid2con(int nodeid, gfp_t alloc)
 	 * under protection of connections_lock. If this is the case we
 	 * abort our connection creation and return the existing connection.
 	 */
-	tmp = __find_con(nodeid);
+	tmp = __find_con(nodeid, r);
 	if (tmp) {
 		spin_unlock(&connections_lock);
 		kfree(con->rx_buf);
@@ -256,15 +249,13 @@ static struct connection *nodeid2con(int nodeid, gfp_t alloc)
 /* Loop round all connections */
 static void foreach_conn(void (*conn_func)(struct connection *c))
 {
-	int i, idx;
+	int i;
 	struct connection *con;
 
-	idx = srcu_read_lock(&connections_srcu);
 	for (i = 0; i < CONN_HASH_SIZE; i++) {
 		hlist_for_each_entry_rcu(con, &connection_hash[i], list)
 			conn_func(con);
 	}
-	srcu_read_unlock(&connections_srcu, idx);
 }
 
 static struct dlm_node_addr *find_node_addr(int nodeid)
@@ -518,14 +509,21 @@ static void lowcomms_state_change(struct sock *sk)
 int dlm_lowcomms_connect_node(int nodeid)
 {
 	struct connection *con;
+	int idx;
 
 	if (nodeid == dlm_our_nodeid())
 		return 0;
 
+	idx = srcu_read_lock(&connections_srcu);
 	con = nodeid2con(nodeid, GFP_NOFS);
-	if (!con)
+	if (!con) {
+		srcu_read_unlock(&connections_srcu, idx);
 		return -ENOMEM;
+	}
+
 	lowcomms_connect_sock(con);
+	srcu_read_unlock(&connections_srcu, idx);
+
 	return 0;
 }
 
@@ -864,7 +862,7 @@ static int accept_from_sock(struct listen_connection *con)
 	int result;
 	struct sockaddr_storage peeraddr;
 	struct socket *newsock;
-	int len;
+	int len, idx;
 	int nodeid;
 	struct connection *newcon;
 	struct connection *addcon;
@@ -907,8 +905,10 @@ static int accept_from_sock(struct listen_connection *con)
 	 *  the same time and the connections cross on the wire.
 	 *  In this case we store the incoming one in "othercon"
 	 */
+	idx = srcu_read_lock(&connections_srcu);
 	newcon = nodeid2con(nodeid, GFP_NOFS);
 	if (!newcon) {
+		srcu_read_unlock(&connections_srcu, idx);
 		result = -ENOMEM;
 		goto accept_err;
 	}
@@ -924,6 +924,7 @@ static int accept_from_sock(struct listen_connection *con)
 			if (!othercon) {
 				log_print("failed to allocate incoming socket");
 				mutex_unlock(&newcon->sock_mutex);
+				srcu_read_unlock(&connections_srcu, idx);
 				result = -ENOMEM;
 				goto accept_err;
 			}
@@ -932,6 +933,7 @@ static int accept_from_sock(struct listen_connection *con)
 			if (result < 0) {
 				kfree(othercon);
 				mutex_unlock(&newcon->sock_mutex);
+				srcu_read_unlock(&connections_srcu, idx);
 				goto accept_err;
 			}
 
@@ -966,6 +968,8 @@ static int accept_from_sock(struct listen_connection *con)
 	if (!test_and_set_bit(CF_READ_PENDING, &addcon->flags))
 		queue_work(recv_workqueue, &addcon->rwork);
 
+	srcu_read_unlock(&connections_srcu, idx);
+
 	return 0;
 
 accept_err:
@@ -1403,7 +1407,9 @@ static struct writequeue_entry *new_wq_entry(struct connection *con, int len,
 
 void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc)
 {
+	struct writequeue_entry *e;
 	struct connection *con;
+	int idx;
 
 	if (len > DEFAULT_BUFFER_SIZE ||
 	    len < sizeof(struct dlm_header)) {
@@ -1413,11 +1419,23 @@ void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc)
 		return NULL;
 	}
 
+	idx = srcu_read_lock(&connections_srcu);
 	con = nodeid2con(nodeid, allocation);
-	if (!con)
+	if (!con) {
+		srcu_read_unlock(&connections_srcu, idx);
 		return NULL;
+	}
+
+	e = new_wq_entry(con, len, allocation, ppc);
+	if (!e) {
+		srcu_read_unlock(&connections_srcu, idx);
+		return NULL;
+	}
+
+	/* we assume if successful commit must called */
+	e->idx = idx;
 
-	return new_wq_entry(con, len, allocation, ppc);
+	return e;
 }
 
 void dlm_lowcomms_commit_buffer(void *mh)
@@ -1435,10 +1453,12 @@ void dlm_lowcomms_commit_buffer(void *mh)
 	spin_unlock(&con->writequeue_lock);
 
 	queue_work(send_workqueue, &con->swork);
+	srcu_read_unlock(&connections_srcu, e->idx);
 	return;
 
 out:
 	spin_unlock(&con->writequeue_lock);
+	srcu_read_unlock(&connections_srcu, e->idx);
 	return;
 }
 
@@ -1532,8 +1552,10 @@ int dlm_lowcomms_close(int nodeid)
 {
 	struct connection *con;
 	struct dlm_node_addr *na;
+	int idx;
 
 	log_print("closing connection to node %d", nodeid);
+	idx = srcu_read_lock(&connections_srcu);
 	con = nodeid2con(nodeid, 0);
 	if (con) {
 		set_bit(CF_CLOSE, &con->flags);
@@ -1542,6 +1564,7 @@ int dlm_lowcomms_close(int nodeid)
 		if (con->othercon)
 			clean_one_writequeue(con->othercon);
 	}
+	srcu_read_unlock(&connections_srcu, idx);
 
 	spin_lock(&dlm_node_addrs_spin);
 	na = find_node_addr(nodeid);
@@ -1621,6 +1644,8 @@ static void shutdown_conn(struct connection *con)
 
 void dlm_lowcomms_shutdown(void)
 {
+	int idx;
+
 	/* Set all the flags to prevent any
 	 * socket activity.
 	 */
@@ -1633,7 +1658,9 @@ void dlm_lowcomms_shutdown(void)
 
 	dlm_close_sock(&listen_con.sock);
 
+	idx = srcu_read_lock(&connections_srcu);
 	foreach_conn(shutdown_conn);
+	srcu_read_unlock(&connections_srcu, idx);
 }
 
 static void _stop_conn(struct connection *con, bool and_other)
@@ -1682,7 +1709,7 @@ static void free_conn(struct connection *con)
 
 static void work_flush(void)
 {
-	int ok, idx;
+	int ok;
 	int i;
 	struct connection *con;
 
@@ -1693,7 +1720,6 @@ static void work_flush(void)
 			flush_workqueue(recv_workqueue);
 		if (send_workqueue)
 			flush_workqueue(send_workqueue);
-		idx = srcu_read_lock(&connections_srcu);
 		for (i = 0; i < CONN_HASH_SIZE && ok; i++) {
 			hlist_for_each_entry_rcu(con, &connection_hash[i],
 						 list) {
@@ -1707,14 +1733,17 @@ static void work_flush(void)
 				}
 			}
 		}
-		srcu_read_unlock(&connections_srcu, idx);
 	} while (!ok);
 }
 
 void dlm_lowcomms_stop(void)
 {
+	int idx;
+
+	idx = srcu_read_lock(&connections_srcu);
 	work_flush();
 	foreach_conn(free_conn);
+	srcu_read_unlock(&connections_srcu, idx);
 	work_stop();
 	deinit_local();
 }
-- 
2.30.2


  parent reply	other threads:[~2021-07-05 15:28 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-05 15:27 [PATCH AUTOSEL 5.13 01/59] HID: do not use down_interruptible() when unbinding devices Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 02/59] EDAC/ti: Add missing MODULE_DEVICE_TABLE Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 03/59] ACPI: PM: s2idle: Add missing LPS0 functions for AMD Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 04/59] ACPI: scan: Rearrange dep_unmet initialization Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 05/59] ACPI: processor idle: Fix up C-state latency if not ordered Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 06/59] hv_utils: Fix passing zero to 'PTR_ERR' warning Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 07/59] lib: vsprintf: Fix handling of number field widths in vsscanf Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 08/59] Input: goodix - platform/x86: touchscreen_dmi - Move upside down quirks to touchscreen_dmi.c Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 09/59] platform/x86: touchscreen_dmi: Add an extra entry for the upside down Goodix touchscreen on Teclast X89 tablets Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 10/59] platform/x86: touchscreen_dmi: Add info for the Goodix GT912 panel of TM800A550L tablets Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 11/59] ACPI: EC: Make more Asus laptops use ECDT _GPE Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 12/59] block_dump: remove block_dump feature in mark_inode_dirty() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 13/59] blk-mq: grab rq->refcount before calling ->fn in blk_mq_tagset_busy_iter Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 14/59] blk-mq: clear stale request in tags->rq[] before freeing one request pool Sasha Levin
2021-07-05 15:27 ` Sasha Levin [this message]
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 16/59] fs: dlm: reconnect if socket error report occurs Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 17/59] fs: dlm: cancel work sync othercon Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 18/59] fs: dlm: fix connection tcp EOF handling Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 19/59] random32: Fix implicit truncation warning in prandom_seed_state() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 20/59] open: don't silently ignore unknown O-flags in openat2() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 21/59] drivers: hv: Fix missing error code in vmbus_connect() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 22/59] fs: dlm: fix lowcomms_start error case Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 23/59] fs: dlm: fix memory leak when fenced Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 24/59] ACPICA: Fix memory leak caused by _CID repair function Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 25/59] ACPI: bus: Call kobject_put() in acpi_init() error path Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 26/59] ACPI: resources: Add checks for ACPI IRQ override Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 27/59] HID: hid-input: add Surface Go battery quirk Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 28/59] HID: sony: fix freeze when inserting ghlive ps3/wii dongles Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 29/59] block: fix race between adding/removing rq qos and normal IO Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 30/59] platform/x86: asus-nb-wmi: Revert "Drop duplicate DMI quirk structures" Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 31/59] platform/x86: asus-nb-wmi: Revert "add support for ASUS ROG Zephyrus G14 and G15" Sasha Levin
2021-07-05 17:08   ` Hans de Goede
2021-07-05 17:09     ` Hans de Goede
2021-07-09 22:50       ` Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 32/59] platform/x86: toshiba_acpi: Fix missing error code in toshiba_acpi_setup_keyboard() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 33/59] nvme-pci: fix var. type for increasing cq_head Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 34/59] nvmet-fc: do not check for invalid target port in nvmet_fc_handle_fcp_rqst() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 35/59] EDAC/Intel: Do not load EDAC driver when running as a guest Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 36/59] tools/power/x86/intel-speed-select: Fix uncore memory frequency display Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 37/59] PCI: hv: Add check for hyperv_initialized in init_hv_pci_drv() Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 38/59] cifs: improve fallocate emulation Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 39/59] cifs: fix check of dfs interlinks Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 40/59] cifs: retry lookup and readdir when EAGAIN is returned Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 41/59] smb3: fix uninitialized value for port in witness protocol move Sasha Levin
2021-07-05 15:27 ` [PATCH AUTOSEL 5.13 42/59] cifs: fix SMB1 error path in cifs_get_file_info_unix Sasha Levin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210705152815.1520546-15-sashal@kernel.org \
    --to=sashal@kernel.org \
    --cc=aahringo@redhat.com \
    --cc=cluster-devel@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=teigland@redhat.com \
    --subject='Re: [PATCH AUTOSEL 5.13 15/59] fs: dlm: fix srcu read lock usage' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).