[PATCH] devlink: add NULL check in devlink_dpipe_entry_clear

[PATCH] devlink: add NULL check in devlink_dpipe_entry_clear

[Zhang Chujun]

This patch adds a NULL check for the pointer 'entry' at the entry
point of the function devlink_pipe_entry_clear. If the passed
'entry' is NULL, the function returns immediately, avoiding subsequent
dereferencing of a NULL pointer and thus preventing potential kernel
crahses or segmentation faults.

Signed-off-by: Zhang Chujun <zhangchujun@cmss.chinamobile.com>

diff --git a/net/devlink/dpipe.c b/net/devlink/dpipe.c
index e55701b007f0..4517d5112311 100644
--- a/net/devlink/dpipe.c
+++ b/net/devlink/dpipe.c
@@ -514,6 +514,9 @@ void devlink_dpipe_entry_clear(struct devlink_dpipe_entry *entry)
 	unsigned int value_count, value_index;
 	struct devlink_dpipe_value *value;
 
+	if (!entry)
+		return;
+
 	value = entry->action_values;
 	value_count = entry->action_values_count;
 	for (value_index = 0; value_index < value_count; value_index++) {
-- 
2.50.1.windows.1

Re: [PATCH] devlink: add NULL check in devlink_dpipe_entry_clear

[Simon Horman]

On Wed, Oct 29, 2025 at 03:02:56PM +0800, Zhang Chujun wrote:
> This patch adds a NULL check for the pointer 'entry' at the entry
> point of the function devlink_pipe_entry_clear. If the passed
> 'entry' is NULL, the function returns immediately, avoiding subsequent
> dereferencing of a NULL pointer and thus preventing potential kernel
> crahses or segmentation faults.
> 
> Signed-off-by: Zhang Chujun <zhangchujun@cmss.chinamobile.com>

Hi,

Do you have a case where entry can be NULL?
If not, this feels like defensive programming,
which, AFAIK, is not the preferred approach for Kernel code.

...