[PATCH net v5 0/2] net: core: improvements to device lookup by hardware address.

[PATCH net v5 0/2] net: core: improvements to device lookup by hardware address.

[Breno Leitao]

The first patch adds a new dev_getbyhwaddr() helper function for
finding devices by hardware address when the rtnl lock is held. This
prevents PROVE_LOCKING warnings that occurred when rtnl lock was held
but the RCU read lock wasn't. The common address comparison logic is
extracted into dev_comp_addr() to avoid code duplication.

The second coverts arp_req_set_public() to the new helper.

Signed-off-by: Breno Leitao <leitao@debian.org>
---
Changes in v5:
- Change the function name from dev_comp_addr() to dev_addr_cmp()
  (Jakub).
- Fix the kerneldoc function name.
- Link to v4: https://lore.kernel.org/r/20250213-arm_fix_selftest-v4-0-26714529a6cf@debian.org

Changes in v4:
- Split the patchset in two, and now targeting `net` instead of
  `net-next` (Kuniyuki Iwashima)
- Identended the kernel-doc in the new way. The other functions will
  come in a separate patchset. (Kuniyuki Iwashima)
- Link to v3: https://lore.kernel.org/r/20250212-arm_fix_selftest-v3-0-72596cb77e44@debian.org

Changes in v3:
- Fixed the cover letter (Kuniyuki Iwashima)
- Added a third patch converting arp_req_set_public() to the new helper
  (Kuniyuki Iwashima)
- Link to v2:
  https://lore.kernel.org/r/20250210-arm_fix_selftest-v2-0-ba84b5bc58c8@debian.org

Changes in v2:
- Fixed the documentation (Jakub)
- Renamed the function from dev_getbyhwaddr_rtnl() to dev_getbyhwaddr()
  (Jakub)
- Exported the function in the header (Jakub)
- Link to v1: https://lore.kernel.org/r/20250207-arm_fix_selftest-v1-1-487518d2fd1c@debian.org

---
Breno Leitao (2):
      net: Add non-RCU dev_getbyhwaddr() helper
      arp: switch to dev_getbyhwaddr() in arp_req_set_public()

 include/linux/netdevice.h |  2 ++
 net/core/dev.c            | 37 ++++++++++++++++++++++++++++++++++---
 net/ipv4/arp.c            |  2 +-
 3 files changed, 37 insertions(+), 4 deletions(-)
---
base-commit: 0469b410c888414c3505d8d2b5814eb372404638
change-id: 20250207-arm_fix_selftest-ee29dbc33a06

Best regards,
-- 
Breno Leitao <leitao@debian.org>

[PATCH net v5 1/2] net: Add non-RCU dev_getbyhwaddr() helper

[Breno Leitao]

Add dedicated helper for finding devices by hardware address when
holding rtnl_lock, similar to existing dev_getbyhwaddr_rcu(). This prevents
PROVE_LOCKING warnings when rtnl_lock is held but RCU read lock is not.

Extract common address comparison logic into dev_addr_cmp().

The context about this change could be found in the following
discussion:

Link: https://lore.kernel.org/all/20250206-scarlet-ermine-of-improvement-1fcac5@leitao/

Cc: kuniyu@amazon.com
Cc: ushankar@purestorage.com
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
---
 include/linux/netdevice.h |  2 ++
 net/core/dev.c            | 37 ++++++++++++++++++++++++++++++++++---
 2 files changed, 36 insertions(+), 3 deletions(-)

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 365f0e2098d13f40ce6d8865962678b052b39a16..ab550a89b9bfaa5682e65f1dcc7f5f99ce90eb94 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -3275,6 +3275,8 @@ static inline struct net_device *first_net_device_rcu(struct net *net)
 }
 
 int netdev_boot_setup_check(struct net_device *dev);
+struct net_device *dev_getbyhwaddr(struct net *net, unsigned short type,
+				   const char *hwaddr);
 struct net_device *dev_getbyhwaddr_rcu(struct net *net, unsigned short type,
 				       const char *hwaddr);
 struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type);
diff --git a/net/core/dev.c b/net/core/dev.c
index 55e356a68db667982e7e62d09d07feecc14deebe..ec92fd2465c23de35187b2a9eeee0f49f98f0757 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1121,6 +1121,12 @@ int netdev_get_name(struct net *net, char *name, int ifindex)
 	return ret;
 }
 
+static bool dev_addr_cmp(struct net_device *dev, unsigned short type,
+			 const char *ha)
+{
+	return dev->type == type && !memcmp(dev->dev_addr, ha, dev->addr_len);
+}
+
 /**
  *	dev_getbyhwaddr_rcu - find a device by its hardware address
  *	@net: the applicable net namespace
@@ -1129,7 +1135,7 @@ int netdev_get_name(struct net *net, char *name, int ifindex)
  *
  *	Search for an interface by MAC address. Returns NULL if the device
  *	is not found or a pointer to the device.
- *	The caller must hold RCU or RTNL.
+ *	The caller must hold RCU.
  *	The returned device has not had its ref count increased
  *	and the caller must therefore be careful about locking
  *
@@ -1141,14 +1147,39 @@ struct net_device *dev_getbyhwaddr_rcu(struct net *net, unsigned short type,
 	struct net_device *dev;
 
 	for_each_netdev_rcu(net, dev)
-		if (dev->type == type &&
-		    !memcmp(dev->dev_addr, ha, dev->addr_len))
+		if (dev_addr_cmp(dev, type, ha))
 			return dev;
 
 	return NULL;
 }
 EXPORT_SYMBOL(dev_getbyhwaddr_rcu);
 
+/**
+ * dev_getbyhwaddr() - find a device by its hardware address
+ * @net: the applicable net namespace
+ * @type: media type of device
+ * @ha: hardware address
+ *
+ * Similar to dev_getbyhwaddr_rcu(), but the owner needs to hold
+ * rtnl_lock.
+ *
+ * Context: rtnl_lock() must be held.
+ * Return: pointer to the net_device, or NULL if not found
+ */
+struct net_device *dev_getbyhwaddr(struct net *net, unsigned short type,
+				   const char *ha)
+{
+	struct net_device *dev;
+
+	ASSERT_RTNL();
+	for_each_netdev(net, dev)
+		if (dev_addr_cmp(dev, type, ha))
+			return dev;
+
+	return NULL;
+}
+EXPORT_SYMBOL(dev_getbyhwaddr);
+
 struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type)
 {
 	struct net_device *dev, *ret = NULL;

-- 
2.43.5

[PATCH net v5 2/2] arp: switch to dev_getbyhwaddr() in arp_req_set_public()

[Breno Leitao]

The arp_req_set_public() function is called with the rtnl lock held,
which provides enough synchronization protection. This makes the RCU
variant of dev_getbyhwaddr() unnecessary. Switch to using the simpler
dev_getbyhwaddr() function since we already have the required rtnl
locking.

This change helps maintain consistency in the networking code by using
the appropriate helper function for the existing locking context.

Fixes: 941666c2e3e0 ("net: RCU conversion of dev_getbyhwaddr() and arp_ioctl()")
Suggested-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
---
 net/ipv4/arp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c
index f23a1ec6694cb2f1bd60f28faa357fcad83c165a..814300eee39de12b959caf225f65d9190594bba9 100644
--- a/net/ipv4/arp.c
+++ b/net/ipv4/arp.c
@@ -1077,7 +1077,7 @@ static int arp_req_set_public(struct net *net, struct arpreq *r,
 	__be32 mask = ((struct sockaddr_in *)&r->arp_netmask)->sin_addr.s_addr;
 
 	if (!dev && (r->arp_flags & ATF_COM)) {
-		dev = dev_getbyhwaddr_rcu(net, r->arp_ha.sa_family,
+		dev = dev_getbyhwaddr(net, r->arp_ha.sa_family,
 				      r->arp_ha.sa_data);
 		if (!dev)
 			return -ENODEV;

-- 
2.43.5

Re: [PATCH net v5 0/2] net: core: improvements to device lookup by hardware address.

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Tue, 18 Feb 2025 05:49:29 -0800 you wrote:
> The first patch adds a new dev_getbyhwaddr() helper function for
> finding devices by hardware address when the rtnl lock is held. This
> prevents PROVE_LOCKING warnings that occurred when rtnl lock was held
> but the RCU read lock wasn't. The common address comparison logic is
> extracted into dev_comp_addr() to avoid code duplication.
> 
> The second coverts arp_req_set_public() to the new helper.
> 
> [...]

Here is the summary with links:
  - [net,v5,1/2] net: Add non-RCU dev_getbyhwaddr() helper
    https://git.kernel.org/netdev/net/c/4b5a28b38c4a
  - [net,v5,2/2] arp: switch to dev_getbyhwaddr() in arp_req_set_public()
    https://git.kernel.org/netdev/net/c/4eae0ee0f1e6

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html