speed caps: not based on average speeds anymore

Speed limits (from CURLOPT_MAX_RECV_SPEED_LARGE &
CURLOPT_MAX_SEND_SPEED_LARGE) were applied simply by comparing limits
with the cumulative average speed of the entire transfer; While this
might work at times with good/constant connections, in other cases it
can result to the limits simply being "ignored" for more than "short
bursts" (as told in man page).

Consider a download that goes on much slower than the limit for some
time (because bandwidth is used elsewhere, server is slow, whatever the
reason), then once things get better, curl would simply ignore the limit
up until the average speed (since the beginning of the transfer) reached
the limit.  This could prove the limit useless to effectively avoid
using the entire bandwidth (at least for quite some time).

So instead, we now use a "moving starting point" as reference, and every
time at least as much as the limit as been transferred, we can reset
this starting point to the current position. This gets a good limiting
effect that applies to the "current speed" with instant reactivity (in
case of sudden speed burst).

Closes #971

1 files changed, 0 insertions, 54 deletions

diff --git a/lib/transfer.c b/lib/transfer.c
index e4a2835f8..5d5ee6be0 100644
--- a/lib/transfer.c
+++ b/lib/transfer.c
@@ -1264,60 +1264,6 @@ int Curl_single_getsock(const struct connectdata *conn,
   return bitmap;
 }
 
-/*
- * Determine optimum sleep time based on configured rate, current rate,
- * and packet size.
- * Returns value in milliseconds.
- *
- * The basic idea is to adjust the desired rate up/down in this method
- * based on whether we are running too slow or too fast.  Then, calculate
- * how many milliseconds to wait for the next packet to achieve this new
- * rate.
- */
-long Curl_sleep_time(curl_off_t rate_bps, curl_off_t cur_rate_bps,
-                             int pkt_size)
-{
-  curl_off_t min_sleep = 0;
-  curl_off_t rv = 0;
-
-  if(rate_bps == 0)
-    return 0;
-
-  /* If running faster than about .1% of the desired speed, slow
-   * us down a bit.  Use shift instead of division as the 0.1%
-   * cutoff is arbitrary anyway.
-   */
-  if(cur_rate_bps > (rate_bps + (rate_bps >> 10))) {
-    /* running too fast, decrease target rate by 1/64th of rate */
-    rate_bps -= rate_bps >> 6;
-    min_sleep = 1;
-  }
-  else if(cur_rate_bps < (rate_bps - (rate_bps >> 10))) {
-    /* running too slow, increase target rate by 1/64th of rate */
-    rate_bps += rate_bps >> 6;
-  }
-
-  /* Determine number of milliseconds to wait until we do
-   * the next packet at the adjusted rate.  We should wait
-   * longer when using larger packets, for instance.
-   */
-  rv = ((curl_off_t)(pkt_size * 1000) / rate_bps);
-
-  /* Catch rounding errors and always slow down at least 1ms if
-   * we are running too fast.
-   */
-  if(rv < min_sleep)
-    rv = min_sleep;
-
-  /* Bound value to fit in 'long' on 32-bit platform.  That's
-   * plenty long enough anyway!
-   */
-  if(rv > 0x7fffffff)
-    rv = 0x7fffffff;
-
-  return (long)rv;
-}
-
 /* Curl_init_CONNECT() gets called each time the handle switches to CONNECT
    which means this gets called once for each subsequent redirect etc */
 void Curl_init_CONNECT(struct Curl_easy *data)