Re: [PATCH v2] Proportional Rate Reduction for TCP.

On Fri, Aug 19, 2011 at 3:26 AM, David Miller [off-list ref] wrote:

From: Nandita Dukkipati <redacted>
Date: Fri, 19 Aug 2011 00:33:32 -0700

quoted

@@ -2830,9 +2830,14 @@ static int tcp_try_undo_loss(struct sock *sk)

 static inline void tcp_complete_cwr(struct sock *sk)
 {
      struct tcp_sock *tp = tcp_sk(sk);
-     /* Do not moderate cwnd if it's already undone in cwr or recovery */
-     if (tp->undo_marker && tp->snd_cwnd > tp->snd_ssthresh) {
-             tp->snd_cwnd = tp->snd_ssthresh;
+
+     /* Do not moderate cwnd if it's already undone in cwr or recovery. */
+     if (tp->undo_marker) {
+
+             if (inet_csk(sk)->icsk_ca_state == TCP_CA_CWR)

Please get rid of that empty line before the TCP_CA_CWR case.

Done.

quoted

+             sndcnt = DIV_ROUND_UP((u64)(tp->prr_delivered *
+                                         tp->snd_ssthresh),
+                                   (u64)tp->prior_cwnd) - tp->prr_out;

This won't link on 32-bit unless __divdi3 libgcc routine is provided
by the architecture.  To portably do 64-bit division you need to use
do_div() or something based upon it.  Perhaps DIV_ROUND_UP_LL() will
work best in this case.

patch v3 uses div_u64 which is based on do_div() for 32-bit archs.

Thanks
Nandita