[PATCH] Fix potential small memory leak

[Michele Baldessari]

In __reg_val() in case the 'if (per_port && tdport)' condition is
true, we leak the previously allocated char *subpath.
Instead jump to 'errorout' which frees it, just like the other error
paths do in this function.

Signed-off-by: Michele Baldessari 
---
 teamd/teamd_state.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/teamd/teamd_state.c b/teamd/teamd_state.c
index 6191ea123954..ddfd85505e52 100644
--- a/teamd/teamd_state.c
+++ b/teamd/teamd_state.c
@@ -98,8 +98,10 @@ int __reg_val(struct teamd_context *ctx, const struct teamd_state_val *val,
    }
    if (val->per_port)
    	per_port = true;
-	if (per_port && tdport)
-		return -EINVAL;
+	if (per_port && tdport) {
+		err = -EINVAL;
+		goto errout;
+	}

if (val->type == TEAMD_STATE_ITEM_TYPE_NODE) {
    	int i;

[Jiri Pirko]

Wed, Mar 19, 2014 at 05:51:46PM CET, [e-mail] wrote:
> In __reg_val() in case the 'if (per_port && tdport)' condition is
> true, we leak the previously allocated char *subpath.
> Instead jump to 'errorout' which frees it, just like the other error
> paths do in this function.
>
> Signed-off-by: Michele Baldessari

applied, thanks!