Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
From:   Oleksandr Natalenko <oleksandr@natalenko.name>
To:     Neal Cardwell <ncardwell@google.com>
Cc:     Yuchung Cheng <ycheng@google.com>,
        Yousuk Seung <ysseung@google.com>,
        Soheil Hassas Yeganeh <soheil@google.com>,
        Adithya Abraham Philip <abrahamphilip@google.com>,
        Eric Dumazet <edumazet@google.com>,
        "David S. Miller" <davem@davemloft.net>,
        Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
        David Ahern <dsahern@kernel.org>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Konstantin Demin <rockdrilla@gmail.com>
Subject: [RFC] tcp_bbr2: use correct 64-bit division
Date:   Mon, 23 May 2022 00:29:37 +0200
Message-ID: <4740526.31r3eYUQgx@natalenko.name>
MIME-Version: 1.0
Content-Transfer-Encoding: 7Bit
Content-Type: text/plain; charset="us-ascii"
Precedence: bulk

Hello Neal.

It was reported to me [1] by Konstantin (in Cc) that BBRv2 code suffers from integer division issue on 32 bit systems.

Konstantin suggested a solution available in the same linked merge request and copy-pasted by me below for your convenience:

```
diff --git a/net/ipv4/tcp_bbr.c b/net/ipv4/tcp_bbr.c
index 664c9e119787..fd3f89e3a8a6 100644
--- a/net/ipv4/tcp_bbr.c
+++ b/net/ipv4/tcp_bbr.c
@@ -312,7 +312,7 @@ static u32 bbr_tso_segs_generic(struct sock *sk, unsigned int mss_now,
 	bytes = sk->sk_pacing_rate >> sk->sk_pacing_shift;
 
 	bytes = min_t(u32, bytes, gso_max_size - 1 - MAX_TCP_HEADER);
-	segs = max_t(u32, bytes / mss_now, bbr_min_tso_segs(sk));
+	segs = max_t(u32, div_u64(bytes, mss_now), bbr_min_tso_segs(sk));
 	return segs;
 }
 
diff --git a/net/ipv4/tcp_bbr2.c b/net/ipv4/tcp_bbr2.c
index fa49e17c47ca..488429f0f3d0 100644
--- a/net/ipv4/tcp_bbr2.c
+++ b/net/ipv4/tcp_bbr2.c
@@ -588,7 +588,7 @@ static void bbr_debug(struct sock *sk, u32 acked,
 		 bbr_rate_kbps(sk, bbr_max_bw(sk)), /* bw: max bw */
 		 0ULL,				    /* lb: [obsolete] */
 		 0ULL,				    /* ib: [obsolete] */
-		 (u64)sk->sk_pacing_rate * 8 / 1000,
+		 div_u64((u64)sk->sk_pacing_rate * 8, 1000),
 		 acked,
 		 tcp_packets_in_flight(tp),
 		 rs->is_ack_delayed ? 'd' : '.',
@@ -698,7 +698,7 @@ static u32 bbr_tso_segs_generic(struct sock *sk, unsigned int mss_now,
 	}
 
 	bytes = min_t(u32, bytes, gso_max_size - 1 - MAX_TCP_HEADER);
-	segs = max_t(u32, bytes / mss_now, bbr_min_tso_segs(sk));
+	segs = max_t(u32, div_u64(bytes, mss_now), bbr_min_tso_segs(sk));
 	return segs;
 }
```

Could you please evaluate this report and check whether it is correct, and also check whether the suggested patch is acceptable?

Thanks.

[1] https://gitlab.com/post-factum/pf-kernel/-/merge_requests/6

-- 
Oleksandr Natalenko (post-factum)