Skip to contentSkip to conversation

Fixes for devlink rate objects API

PATCH Applied [email protected] 2 participants 5 messages Jun 23, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover Fixes for devlink rate objects API

From: Dmytro Linkin <[email protected]>

Patch #1 fixes not decreased refcount of parent node for destroyed leaf object.

Patch #2 fixes incorect eswitch mode check.

Patch #3 protects list traversing with a lock.

Dmytro Linkin (3):
devlink: Decrease refcnt of parent rate object on leaf destroy
devlink: Remove eswitch mode check for mode set call
devlink: Protect rate list with lock while switching modes

 net/core/devlink.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)
Signature · 1 line
1.8.3.1
#1 lore raw

PATCH 1/3 devlink: Decrease refcnt of parent rate object on leaf destroy

From: Dmytro Linkin <[email protected]>

Port functions, like SFs, can be deleted by the user when its leaf rate object has parent node. In such case node refcnt won't be decreased which blocks the node from deletion later.
Do simple refcnt decrease, since driver in cleanup stage. This:
1) assumes that driver took proper internal parent unset action;
2) allows to avoid nested callbacks call and deadlock.

Fixes d75559845078 ("devlink: Allow setting parent node of rate objects")
---
 net/core/devlink.c | 2 ++
 1 file changed, 2 insertions(+)
net/core/devlink.c +2 -0
index 566ddd1..ba27395 100644
@@ -9275,6 +9275,8 @@ void devlink_rate_leaf_destroy(struct devlink_port *devlink_port)
 
 	mutex_lock(&devlink->lock);
 	devlink_rate_notify(devlink_rate, DEVLINK_CMD_RATE_DEL);
+	if (devlink_rate->parent)
+		refcount_dec(&devlink_rate->parent->refcnt);
 	list_del(&devlink_rate->list);
 	devlink_port->devlink_rate = NULL;
 	mutex_unlock(&devlink->lock);
Signature · 1 line
1.8.3.1
#2 lore raw

PATCH 2/3 devlink: Remove eswitch mode check for mode set call

In reply to [email protected]

From: Dmytro Linkin <[email protected]>

When eswitch is disabled, querying its current mode results in error. Due to this when trying to set the eswitch mode for mlx5 devices, it fails to set the eswitch switchdev mode.
Hence remove such check.

Fixes a8ecb93ef03d ("devlink: Introduce rate nodes")
---
 net/core/devlink.c | 11 -----------
 1 file changed, 11 deletions(-)
net/core/devlink.c +0 -11
index ba27395..153d432 100644
@@ -2709,17 +2709,6 @@ static int devlink_rate_nodes_check(struct devlink *devlink, u16 mode,
 				    struct netlink_ext_ack *extack)
 {
 	struct devlink_rate *devlink_rate;
-	u16 old_mode;
-	int err;
-
-	if (!devlink->ops->eswitch_mode_get)
-		return -EOPNOTSUPP;
-	err = devlink->ops->eswitch_mode_get(devlink, &old_mode);
-	if (err)
-		return err;
-
-	if (old_mode == mode)
-		return 0;
 
 	list_for_each_entry(devlink_rate, &devlink->rate_list, list)
 		if (devlink_rate_is_node(devlink_rate)) {
Signature · 1 line
1.8.3.1
#3 lore raw

PATCH 3/3 devlink: Protect rate list with lock while switching modes

In reply to [email protected]

From: Dmytro Linkin <[email protected]>

Devlink eswitch set command doesn't hold devlink->lock, which makes possible race condition between rate list traversing and others devlink rate KAPI calls, like devlink_rate_nodes_destroy().
Hold devlink lock while traversing the list.

Fixes a8ecb93ef03d ("devlink: Introduce rate nodes")
---
 net/core/devlink.c | 4 ++++
 1 file changed, 4 insertions(+)
net/core/devlink.c +4 -0
index 153d432..8fdd04f 100644
@@ -2710,11 +2710,15 @@ static int devlink_rate_nodes_check(struct devlink *devlink, u16 mode,
 {
 	struct devlink_rate *devlink_rate;
 
+	/* Take the lock to sync with devlink_rate_nodes_destroy() */
+	mutex_lock(&devlink->lock);
 	list_for_each_entry(devlink_rate, &devlink->rate_list, list)
 		if (devlink_rate_is_node(devlink_rate)) {
+			mutex_unlock(&devlink->lock);
 			NL_SET_ERR_MSG_MOD(extack, "Rate node(s) exists.");
 			return -EBUSY;
 		}
+	mutex_unlock(&devlink->lock);
 	return 0;
 }
 
Signature · 1 line
1.8.3.1
#4 lore raw
In reply to [email protected]
This series was applied to netdev/net-next.git (refs/heads/master):

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

Quoted 8 lines · On Wed, 23 Jun 2021 16:43:12 +0300 you wrote:

From: Dmytro Linkin <[email protected]>

Patch #1 fixes not decreased refcount of parent node for destroyed leaf object.

Patch #2 fixes incorect eswitch mode check.

[...]

Here is the summary with links:
- [net-next,1/3] devlink: Decrease refcnt of parent rate object on leaf destroy

    https://git.kernel.org/netdev/net-next/c/1321ed5e7648
  - [net-next,2/3] devlink: Remove eswitch mode check for mode set call
    https://git.kernel.org/netdev/net-next/c/ff99324ded01
  - [net-next,3/3] devlink: Protect rate list with lock while switching modes
    https://git.kernel.org/netdev/net-next/c/a3e5e5797faa

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