Thread (4 messages) flat view 4 messages, 3 authors, 2024-10-29

Re: [PATCH net] netlink: Fix off-by-one error in netlink_proto_init()

From: Eric Dumazet <edumazet@google.com>
Date: 2024-10-28 15:20:36
Also in: lkml

On Mon, Oct 28, 2024 at 9:05 AM Jinjie Ruan [off-list ref] wrote:
quoted hunk ↗ jump to hunk
In the error path of netlink_proto_init(), frees the already allocated
bucket table for new hash tables in a loop, but the loop condition
terminates when the index reaches zero, which fails to free the first
bucket table at index zero.

Check for >= 0 so that nl_table[0].hash is freed as well.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Jinjie Ruan <redacted>
---
 net/netlink/af_netlink.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index 0a9287fadb47..9601b85dda95 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -2936,7 +2936,7 @@ static int __init netlink_proto_init(void)
        for (i = 0; i < MAX_LINKS; i++) {
                if (rhashtable_init(&nl_table[i].hash,
                                    &netlink_rhashtable_params) < 0) {
-                       while (--i > 0)
+                       while (--i >= 0)
                                rhashtable_destroy(&nl_table[i].hash);
                        kfree(nl_table);
                        goto panic;
--
Note that the host is going to panic, many other pieces of memory are
left behind.

A Fixes: tag seems unnecessary.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help