[tor-commits] [tor/release-0.3.2] Remove an erroneous 0.5 in compute_weighted_bandwidths()

nickm at torproject.org nickm at torproject.org
Mon Nov 6 14:49:10 UTC 2017


commit 01e984870a7e1db2722e85fe43af7bcb4755c2d4
Author: Nick Mathewson <nickm at torproject.org>
Date:   Fri Sep 22 15:29:15 2017 -0400

    Remove an erroneous 0.5 in compute_weighted_bandwidths()
    
    Back in 0.2.4.3-alpha (e106812a778f537), when we switched from using
    double to using uint64 for selecting by bandwidth, I got the math
    wrong: I should have used llround(x), or (uint64_t)(x+0.5), but
    instead I wrote llround(x+0.5).  That means we would always round
    up, rather than rounding to the closest integer
    
    Fixes bug 23318; bugfix on 0.2.4.3-alpha.
---
 changes/bug23318    | 7 +++++++
 src/or/routerlist.c | 2 +-
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/changes/bug23318 b/changes/bug23318
new file mode 100644
index 000000000..32c85eb19
--- /dev/null
+++ b/changes/bug23318
@@ -0,0 +1,7 @@
+  o Minor bugfixes (path selection):
+    - When selecting relays by bandwidth, avoid a rounding error that
+      could sometimes cause load to be imbalanced incorrectly. Previously,
+      we would always round upwards; now, we round towards the nearest
+      integer.  This had the biggest effect when a relay's weight adjustments
+      should have given it weight 0, but it got weight 1 instead.
+      Fixes bug 23318; bugfix on 0.2.4.3-alpha.
diff --git a/src/or/routerlist.c b/src/or/routerlist.c
index 2365f28fd..faf2eeda5 100644
--- a/src/or/routerlist.c
+++ b/src/or/routerlist.c
@@ -2713,7 +2713,7 @@ compute_weighted_bandwidths(const smartlist_t *sl,
       final_weight = weight*this_bw;
     }
 
-    bandwidths[node_sl_idx] = final_weight + 0.5;
+    bandwidths[node_sl_idx] = final_weight;
   } SMARTLIST_FOREACH_END(node);
 
   log_debug(LD_CIRC, "Generated weighted bandwidths for rule %s based "





More information about the tor-commits mailing list