Skip to contentSkip to conversation

net: use the %px format to display sock

PATCH In review [email protected] 3 participants 4 messages May 5, 2022 – May 7, 2022

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net: use the %px format to display sock

From: Jason Xing <[email protected]>

I found that the current socket address, say 000000009842d952, cannot be searched in messages because %p format function hashes and converts it into an unique identifier which is currently useless for debugging.

---
 net/ipv4/af_inet.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/ipv4/af_inet.c +2 -2
index 72fde28..b17a4d4 100644
@@ -139,12 +139,12 @@ void inet_sock_destruct(struct sock *sk)
 	sk_mem_reclaim_final(sk);
 
 	if (sk->sk_type == SOCK_STREAM && sk->sk_state != TCP_CLOSE) {
-		pr_err("Attempt to release TCP socket in state %d %p\n",
+		pr_err("Attempt to release TCP socket in state %d %px\n",
 		       sk->sk_state, sk);
 		return;
 	}
 	if (!sock_flag(sk, SOCK_DEAD)) {
-		pr_err("Attempt to release alive inet socket %p\n", sk);
+		pr_err("Attempt to release alive inet socket %px\n", sk);
 		return;
 	}
 
Signature · 1 line
1.8.3.1
#1 lore raw

Hi Jason,

net/ipv4/af_inet.c void inet_sock_destruct(struct sock *sk) · line 142

Quoted 2 lines · On Thu, May 05, 2022 at 09:08:26PM +0800, [email protected] wrote:
net/ipv4/af_inet.c +1 -1
-		pr_err("Attempt to release TCP socket in state %d %p\n",
+		pr_err("Attempt to release TCP socket in state %d %px\n",

I think we cannot use %px here for security reasons? checkpatch is also warning about it:

WARNING: Using vsprintf specifier '%px' potentially exposes the kernel memory layout, if you don't really need the address please consider using '%p'. #21: FILE: net/ipv4/af_inet.c:142:
+ pr_err("Attempt to release TCP socket in state %d %px\n",
sk->sk_state, sk);

Thanks,
Peilin Ye

author
#2 lore raw
Quoted 10 lines · Peilin Ye

Hi Jason,

net/ipv4/af_inet.c void inet_sock_destruct(struct sock *sk) · line 142

Quoted 2 lines · On Thu, May 05, 2022 at 09:08:26PM +0800, [email protected] wrote:
net/ipv4/af_inet.c +1 -1
-             pr_err("Attempt to release TCP socket in state %d %p\n",
+             pr_err("Attempt to release TCP socket in state %d %px\n",

I think we cannot use %px here for security reasons? checkpatch is also warning about it:

I noticed this warning before submitting. Since the %p format doesn't print the real address, printing the address here will be helpless and we cannot trace what exactly the bad socket is.

What do you suggest?

Thanks,
Jason

Quoted 8 lines

WARNING: Using vsprintf specifier '%px' potentially exposes the kernel memory layout, if you don't really need the address please consider using '%p'. #21: FILE: net/ipv4/af_inet.c:142:
+ pr_err("Attempt to release TCP socket in state %d %px\n",

                       sk->sk_state, sk);

Thanks,
Peilin Ye

#3 lore raw
Quoted 12 lines · Jason Xing
Quoted 10 lines · Peilin Ye

Hi Jason,

net/ipv4/af_inet.c void inet_sock_destruct(struct sock *sk) · line 142

Quoted 2 lines · On Thu, May 05, 2022 at 09:08:26PM +0800, [email protected] wrote:
net/ipv4/af_inet.c +1 -1
-             pr_err("Attempt to release TCP socket in state %d %p\n",
+             pr_err("Attempt to release TCP socket in state %d %px\n",

I think we cannot use %px here for security reasons? checkpatch is also warning about it:

Quoted 5 lines

I noticed this warning before submitting. Since the %p format doesn't print the real address, printing the address here will be helpless and we cannot trace what exactly the bad socket is.

What do you suggest?

How is a socket identified in places like /proc/<PID>/net/tcp ? Could you print the local and remote port to identify the socket?

How does the address of the structure actually help you? Do you see this address somewhere else?

Andrew