diff mbox series

[net,v2] mctp: Fix incorrect netdev unref for extended addr

Message ID 20220218062908.1994506-1-matt@codeconstruct.com.au (mailing list archive)
State Changes Requested
Delegated to: Netdev Maintainers
Headers show
Series [net,v2] mctp: Fix incorrect netdev unref for extended addr | expand

Checks

Context Check Description
netdev/tree_selection success Clearly marked for net
netdev/fixes_present success Fixes tag present in non-next series
netdev/subject_prefix success Link
netdev/cover_letter success Single patches do not need cover letters
netdev/patch_count success Link
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 4 this patch: 4
netdev/cc_maintainers success CCed 5 of 5 maintainers
netdev/build_clang success Errors and warnings before: 18 this patch: 18
netdev/module_param success Was 0 now: 0
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/verify_fixes success Fixes tag looks correct
netdev/build_allmodconfig_warn success Errors and warnings before: 9 this patch: 9
netdev/checkpatch success total: 0 errors, 0 warnings, 0 checks, 46 lines checked
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0

Commit Message

Matt Johnston Feb. 18, 2022, 6:29 a.m. UTC 
In the extended addressing local route output codepath
dev_get_by_index_rcu() doesn't take a dev_hold() so we shouldn't
dev_put(). Instead we need to hold/put the mctp_dev.

Fixes: 99ce45d5e7db ("mctp: Implement extended addressing")
Signed-off-by: Matt Johnston <matt@codeconstruct.com.au>
---
v2: Fix incorrect "Fixes:" ID
---
 net/mctp/route.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

Comments

Jakub Kicinski Feb. 19, 2022, 5:42 a.m. UTC | #1 
On Fri, 18 Feb 2022 14:29:08 +0800 Matt Johnston wrote:
>  		rt->dev = __mctp_dev_get(dev);
> +		if (rt->dev)
> +			mctp_dev_hold(rt->dev);

Is it safe to have the ref like that? mctp_dev_hold() just does a
refcount_inc(), does something prevent the refcount from dropping
to zero between __mctp_dev_get() and mctp_dev_hold()?
diff mbox series

Patch

diff --git a/net/mctp/route.c b/net/mctp/route.c
index e52cef750500..6ef79ab4a076 100644
--- a/net/mctp/route.c
+++ b/net/mctp/route.c
@@ -786,9 +786,8 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 {
 	struct mctp_sock *msk = container_of(sk, struct mctp_sock, sk);
 	struct mctp_skb_cb *cb = mctp_cb(skb);
-	struct mctp_route tmp_rt;
+	struct mctp_route tmp_rt = {0};
 	struct mctp_sk_key *key;
-	struct net_device *dev;
 	struct mctp_hdr *hdr;
 	unsigned long flags;
 	unsigned int mtu;
@@ -801,12 +800,12 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 
 	if (rt) {
 		ext_rt = false;
-		dev = NULL;
-
 		if (WARN_ON(!rt->dev))
 			goto out_release;
 
 	} else if (cb->ifindex) {
+		struct net_device *dev;
+
 		ext_rt = true;
 		rt = &tmp_rt;
 
@@ -816,8 +815,9 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 			rcu_read_unlock();
 			return rc;
 		}
-
 		rt->dev = __mctp_dev_get(dev);
+		if (rt->dev)
+			mctp_dev_hold(rt->dev);
 		rcu_read_unlock();
 
 		if (!rt->dev)
@@ -891,10 +891,10 @@  int mctp_local_output(struct sock *sk, struct mctp_route *rt,
 	if (!ext_rt)
 		mctp_route_release(rt);
 
-	dev_put(dev);
+	if (tmp_rt.dev)
+		mctp_dev_put(tmp_rt.dev);
 
 	return rc;
-
 }
 
 /* route management */