[tor-commits] [tor/master] Lower the delay before saving guard status to disk

[nickm at torproject.org]

commit fb68f507618ab5479471384cc4a1fa8b4bfa99fe
Author: Nick Mathewson <nickm at torproject.org>
Date:   Wed Jan 7 10:39:44 2015 -0500

    Lower the delay before saving guard status to disk
    
    "Maybe this time should be reduced, since we are considering
    guard-related changes as quite important? It would be a pity to
    settle on a guard node, then close the Tor client fast and lose that
    information."
    
    Closes 12485.
---
 changes/bug12485    |    4 ++++
 src/or/entrynodes.c |    4 ++--
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/changes/bug12485 b/changes/bug12485
new file mode 100644
index 0000000..53ce33e
--- /dev/null
+++ b/changes/bug12485
@@ -0,0 +1,4 @@
+  o Minor features (Guard nodes):
+    - Reduce the time delay before saving guard status to disk from 10
+      minute to 30 seconds (or from one hour to 10 minutes if
+      AvoidDiskWrites is set).  Closes ticket 12485.
diff --git a/src/or/entrynodes.c b/src/or/entrynodes.c
index 05ed560..116f047 100644
--- a/src/or/entrynodes.c
+++ b/src/or/entrynodes.c
@@ -1525,10 +1525,10 @@ entry_guards_parse_state(or_state_t *state, int set, char **msg)
 
 /** How long will we let a change in our guard nodes stay un-saved
  * when we are trying to avoid disk writes? */
-#define SLOW_GUARD_STATE_FLUSH_TIME 3600
+#define SLOW_GUARD_STATE_FLUSH_TIME 600
 /** How long will we let a change in our guard nodes stay un-saved
  * when we are not trying to avoid disk writes? */
-#define FAST_GUARD_STATE_FLUSH_TIME 600
+#define FAST_GUARD_STATE_FLUSH_TIME 30
 
 /** Our list of entry guards has changed, or some element of one
  * of our entry guards has changed. Write the changes to disk within