net: bridge: avoid redundant strlen() in show_backup_port()

Started by Dmitry Antipov on Nov 17, 2026 · 4 messages from 4 authors · Last reply Nov 18, 2026

Subsystems (3)

Maintainers (7)

Show 3 more
WARM3d

open the first message on its own page

Message details and related content
In store_backup_port(), prefer strchrnul() over strchr() to
avoid redundant call to strlen().

Signed-off-by: Dmitry Antipov <redacted>
---
 net/bridge/br_sysfs_if.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/net/bridge/br_sysfs_if.c b/net/bridge/br_sysfs_if.c
index 621d62a063d9..03205b3f465f 100644
--- a/net/bridge/br_sysfs_if.c
+++ b/net/bridge/br_sysfs_if.c
@@ -231,12 +231,12 @@ static ssize_t show_backup_port(struct net_bridge_port *p, char *buf)
 static int store_backup_port(struct net_bridge_port *p, char *buf)
 {
 	struct net_device *backup_dev = NULL;
-	char *nl = strchr(buf, '\n');
+	char *nl = strchrnul(buf, '\n');
 
-	if (nl)
+	if (*nl != '\0')
 		*nl = '\0';
 
-	if (strlen(buf) > 0) {
+	if (nl - buf > 0) {
 		backup_dev = __dev_get_by_name(dev_net(p->dev), buf);
 		if (!backup_dev)
 			return -ENOENT;
-- 
2.56.1
Message details and related content
On Tue, Nov 17, 2026 at 07:51:06PM +0300, Dmitry Antipov wrote:
In store_backup_port(), prefer strchrnul() over strchr() to
avoid redundant call to strlen().
a couple things:

- i dont't think the change is worth it because it feels like unnecessary
  churn in a path that isn't a hot path. how many bytes are we really avoiding
  counting here?

- commit message mentions store_backup_port but subject line says
  show_backup_port, so commit would need to be respun

but imho i think this should be dropped
Message details and related content
On Tue, Nov 17, 2026 at 09:49:18AM -0800, Joe Damato wrote:
On Tue, Nov 17, 2026 at 07:51:06PM +0300, Dmitry Antipov wrote:
quoted
In store_backup_port(), prefer strchrnul() over strchr() to
avoid redundant call to strlen().
a couple things:

- i dont't think the change is worth it because it feels like unnecessary
  churn in a path that isn't a hot path. how many bytes are we really avoiding
  counting here?

- commit message mentions store_backup_port but subject line says
  show_backup_port, so commit would need to be respun

but imho i think this should be dropped
Yep.

pw-bot: rejected
Message details and related content
On Tue, 17 Nov 2026 19:51:06 +0300
Dmitry Antipov [off-list ref] wrote:
In store_backup_port(), prefer strchrnul() over strchr() to
avoid redundant call to strlen().
For most architectures I think the compiler optimises that strlen()
to a check of buf[0].
So you don't get the function call at all.

David
quoted hunk
Signed-off-by: Dmitry Antipov <redacted>
---
 net/bridge/br_sysfs_if.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/net/bridge/br_sysfs_if.c b/net/bridge/br_sysfs_if.c
index 621d62a063d9..03205b3f465f 100644
--- a/net/bridge/br_sysfs_if.c
+++ b/net/bridge/br_sysfs_if.c
@@ -231,12 +231,12 @@ static ssize_t show_backup_port(struct net_bridge_port *p, char *buf)
 static int store_backup_port(struct net_bridge_port *p, char *buf)
 {
 	struct net_device *backup_dev = NULL;
-	char *nl = strchr(buf, '\n');
+	char *nl = strchrnul(buf, '\n');
 
-	if (nl)
+	if (*nl != '\0')
 		*nl = '\0';
 
-	if (strlen(buf) > 0) {
+	if (nl - buf > 0) {
 		backup_dev = __dev_get_by_name(dev_net(p->dev), buf);
 		if (!backup_dev)
 			return -ENOENT;
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help