From: f6bvp <bernard.f6bvp@gmail.com>
To: linux-hams@vger.kernel.org
Cc: netdev@vger.kernel.org, edumazet@google.com, pabeni@redhat.com,
f6bvp <bernard.f6bvp@gmail.com>
Subject: [PATCH net-next 1/3] rose: fix race between loopback timer and module removal
Date: Mon, 13 Apr 2026 19:42:36 +0200 [thread overview]
Message-ID: <20260413174238.112418-1-bernard.f6bvp@gmail.com> (raw)
In-Reply-To: <5a88b747-bb06-4ebd-99de-80ceb574cf22@free.fr>
rose_loopback_clear() used timer_delete() which returns immediately
without waiting for any running callback to complete. If the timer
fired concurrently with module removal, rose_loopback_timer() would
access rose_loopback_neigh after it was freed, causing a use-after-free.
Three changes fix the race:
1. Add a loopback_stopping atomic flag. rose_loopback_timer() checks
this at entry and mid-loop; when set it drains the queue and bails
out without re-arming the timer.
2. Switch rose_loopback_clear() to timer_delete_sync() so it blocks
until any in-flight callback has returned.
3. Wrap the timer body with rose_neigh_hold()/rose_neigh_put() so the
loopback neighbour cannot be freed while the callback is running.
Also fix a pre-existing bug: dev_put(dev) was only called on the
failure path of rose_rx_call_request(); it is now called unconditionally
so the device reference is always released.
Remove a dead check (!neigh->dev && !neigh->loopback) that can never
be true for the loopback neighbour, which always has loopback=1.
Signed-off-by: f6bvp <bernard.f6bvp@gmail.com>
---
net/rose/rose_loopback.c | 53 +++++++++++++++++++++++++++-------------
1 file changed, 36 insertions(+), 17 deletions(-)
diff --git a/net/rose/rose_loopback.c b/net/rose/rose_loopback.c
index b538e39b3df5..80d7879ef36a 100644
--- a/net/rose/rose_loopback.c
+++ b/net/rose/rose_loopback.c
@@ -12,13 +12,15 @@
#include <net/rose.h>
#include <linux/init.h>
-static struct sk_buff_head loopback_queue;
#define ROSE_LOOPBACK_LIMIT 1000
-static struct timer_list loopback_timer;
+static struct timer_list loopback_timer;
+static struct sk_buff_head loopback_queue;
static void rose_set_loopback_timer(void);
static void rose_loopback_timer(struct timer_list *unused);
+static atomic_t loopback_stopping = ATOMIC_INIT(0);
+
void rose_loopback_init(void)
{
skb_queue_head_init(&loopback_queue);
@@ -66,10 +68,25 @@ static void rose_loopback_timer(struct timer_list *unused)
unsigned int lci_i, lci_o;
int count;
+ if (atomic_read(&loopback_stopping))
+ return;
+
+ if (rose_loopback_neigh)
+ rose_neigh_hold(rose_loopback_neigh);
+ else
+ return;
+
for (count = 0; count < ROSE_LOOPBACK_LIMIT; count++) {
skb = skb_dequeue(&loopback_queue);
if (!skb)
- return;
+ goto out;
+
+ if (atomic_read(&loopback_stopping)) {
+ kfree_skb(skb);
+ skb_queue_purge(&loopback_queue);
+ goto out;
+ }
+
if (skb->len < ROSE_MIN_LEN) {
kfree_skb(skb);
continue;
@@ -96,27 +113,24 @@ static void rose_loopback_timer(struct timer_list *unused)
}
if (frametype == ROSE_CALL_REQUEST) {
- if (!rose_loopback_neigh->dev &&
- !rose_loopback_neigh->loopback) {
- kfree_skb(skb);
- continue;
- }
-
dev = rose_dev_get(dest);
if (!dev) {
kfree_skb(skb);
continue;
}
- if (rose_rx_call_request(skb, dev, rose_loopback_neigh, lci_o) == 0) {
- dev_put(dev);
+ if (rose_rx_call_request(skb, dev, rose_loopback_neigh, lci_o) == 0)
kfree_skb(skb);
- }
+ dev_put(dev);
} else {
kfree_skb(skb);
}
}
- if (!skb_queue_empty(&loopback_queue))
+
+out:
+ rose_neigh_put(rose_loopback_neigh);
+
+ if (!atomic_read(&loopback_stopping) && !skb_queue_empty(&loopback_queue))
mod_timer(&loopback_timer, jiffies + 1);
}
@@ -124,10 +138,15 @@ void __exit rose_loopback_clear(void)
{
struct sk_buff *skb;
- timer_delete(&loopback_timer);
+ atomic_set(&loopback_stopping, 1);
+ /* Pairs with atomic_read() in rose_loopback_timer(): ensure the
+ * stopping flag is visible before we cancel, so a concurrent
+ * callback aborts its loop early rather than re-arming the timer.
+ */
+ smp_mb();
- while ((skb = skb_dequeue(&loopback_queue)) != NULL) {
- skb->sk = NULL;
+ timer_delete_sync(&loopback_timer);
+
+ while ((skb = skb_dequeue(&loopback_queue)) != NULL)
kfree_skb(skb);
- }
}
--
2.51.0
next prev parent reply other threads:[~2026-04-13 17:42 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-01-17 6:39 [PATCH] rose: Fix use-after-free in rose_timer_expiry Deepanshu Kartikey
2026-01-19 20:19 ` F6BVP
2026-04-13 17:42 ` f6bvp [this message]
2026-04-13 17:42 ` [PATCH net-next 2/3] rose: clear neighbour pointer after rose_neigh_put() in state machines f6bvp
2026-04-13 20:53 ` Jakub Kicinski
2026-04-13 17:42 ` [PATCH net-next 3/3] rose: guard rose_neigh_put() against NULL in timer expiry f6bvp
2026-04-13 21:21 ` [PATCH net-next 1/3] rose: fix race between loopback timer and module removal Andrew Lunn
2026-04-13 21:34 ` Jakub Kicinski
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=20260413174238.112418-1-bernard.f6bvp@gmail.com \
--to=bernard.f6bvp@gmail.com \
--cc=edumazet@google.com \
--cc=linux-hams@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox