linux-kernel - Re: [PATCH][net-next] tcp: remove the now redundant non-null check on tskb

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <1510688414.2849.179.camel@edumazet-glaptop3.roam.corp.google.com>
Date:   Tue, 14 Nov 2017 11:40:14 -0800
From:   Eric Dumazet <eric.dumazet@...il.com>
To:     Colin King <colin.king@...onical.com>
Cc:     "David S . Miller" <davem@...emloft.net>,
        Alexey Kuznetsov <kuznet@....inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        netdev@...r.kernel.org, kernel-janitors@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH][net-next] tcp: remove the now redundant non-null check
 on tskb

On Tue, 2017-11-14 at 18:41 +0000, Colin King wrote:
> From: Colin Ian King <colin.king@...onical.com>
> 
> The non-null check on tskb is redundant as it is in an else
> section of a check on tskb where tskb is always null. Remove
> the redundant if statement and the label coalesce.
> 
> Detected by CoverityScan, CID#1457751 ("Logically dead code")
> 
> Fixes: 75c119afe14f ("tcp: implement rb-tree based retransmit queue")
> Signed-off-by: Colin Ian King <colin.king@...onical.com>
> ---
>  net/ipv4/tcp_output.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> index 071bdd34f8eb..b58c986b2b27 100644
> --- a/net/ipv4/tcp_output.c
> +++ b/net/ipv4/tcp_output.c
> @@ -3053,7 +3053,6 @@ void tcp_send_fin(struct sock *sk)
>  		tskb = skb_rb_last(&sk->tcp_rtx_queue);
>  
>  	if (tskb) {
> -coalesce:
>  		TCP_SKB_CB(tskb)->tcp_flags |= TCPHDR_FIN;
>  		TCP_SKB_CB(tskb)->end_seq++;
>  		tp->write_seq++;
> @@ -3069,11 +3068,8 @@ void tcp_send_fin(struct sock *sk)
>  		}
>  	} else {
>  		skb = alloc_skb_fclone(MAX_TCP_HEADER, sk->sk_allocation);
> -		if (unlikely(!skb)) {
> -			if (tskb)
> -				goto coalesce;
> +		if (unlikely(!skb))
>  			return;
> -		}
>  		INIT_LIST_HEAD(&skb->tcp_tsorted_anchor);
>  		skb_reserve(skb, MAX_TCP_HEADER);
>  		sk_forced_mem_schedule(sk, skb->truesize);

Hmm... I would rather try to use skb_rb_last(), because
alloc_skb_fclone() might fail even if tcp_under_memory_pressure() is
false.

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 76dbe884f2469660028684a46fc19afa000a1353..eea017b8a8918815226fd1412c0a7b8e484aeca8 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -3070,6 +3070,7 @@ void tcp_send_fin(struct sock *sk)
 	} else {
 		skb = alloc_skb_fclone(MAX_TCP_HEADER, sk->sk_allocation);
 		if (unlikely(!skb)) {
+			tskb = skb_rb_last(&sk->tcp_rtx_queue);
 			if (tskb)
 				goto coalesce;
 			return;