When we moved channel_matches_target_addr_for_extend() into a separate
function, its sense was inverted from what one might expect, and we
didn't have a ! in one place where we should have.
Found by skruffy.
--- /dev/null
+ o Minor bugfixes:
+ - When evaluating whether to use a connection that we haven't
+ decided is canonical using a recent link protocol version,
+ decide that it's canonical only if it used address _does_
+ match the desired address. Fixes bug 9309; bugfix on
+ 0.2.4.4-alpha. Reported by skruffy.
if (chan->state != CHANNEL_STATE_OPEN) {
/* If the address matches, don't launch a new connection for this
* circuit. */
- if (!channel_matches_target_addr_for_extend(chan, target_addr))
+ if (channel_matches_target_addr_for_extend(chan, target_addr))
++n_inprogress_goodaddr;
continue;
}
}
/**
- * Check if a channel matches a given target address
+ * Check if a channel matches a given target address; return true iff we do.
*
* This function calls into the lower layer and asks if this channel thinks
* it matches a given target address for circuit extension purposes.
}
/**
- * Check if we match a target address
+ * Check if we match a target address; return true iff we do.
*
* This implements the matches_target method for channel_tls t_; the upper
* layer wants to know if this channel matches a target address when extending
tor_assert(target);
tor_assert(tlschan->conn);
- return tor_addr_compare(&(tlschan->conn->real_addr),
- target, CMP_EXACT);
+ return tor_addr_eq(&(tlschan->conn->real_addr), target);
}
/**