summaryrefslogtreecommitdiff
path: root/src/feature
diff options
context:
space:
mode:
authorDavid Goulet <dgoulet@torproject.org>2018-12-04 14:27:46 -0500
committerDavid Goulet <dgoulet@torproject.org>2018-12-04 14:34:04 -0500
commitcec616a0c8ff060cb722e54342fd30aeab3ad285 (patch)
treee614a2b86fc7bb0ca17f3d224532a46d8b7df7cd /src/feature
parent43bd4d7509ceab2d82a85483f08132e90b1ab10d (diff)
downloadtor-cec616a0c8ff060cb722e54342fd30aeab3ad285.tar.gz
tor-cec616a0c8ff060cb722e54342fd30aeab3ad285.zip
hs-v3: Don't BUG() if descriptor is found on SOCKS connection retry
When retrying all SOCKS connection because new directory information just arrived, do not BUG() if a connection in state AP_CONN_STATE_RENDDESC_WAIT is found to have a usable descriptor. There is a rare case when this can happen as detailed in #28669 so the right thing to do is put that connection back in circuit wait state so the descriptor can be retried. Fixes #28669 Signed-off-by: David Goulet <dgoulet@torproject.org>
Diffstat (limited to 'src/feature')
-rw-r--r--src/feature/hs/hs_client.c19
1 files changed, 13 insertions, 6 deletions
diff --git a/src/feature/hs/hs_client.c b/src/feature/hs/hs_client.c
index 22aacdfe99..5fded92fe3 100644
--- a/src/feature/hs/hs_client.c
+++ b/src/feature/hs/hs_client.c
@@ -296,12 +296,19 @@ retry_all_socks_conn_waiting_for_desc(void)
/* Order a refetch in case it works this time. */
status = hs_client_refetch_hsdesc(&edge_conn->hs_ident->identity_pk);
- if (BUG(status == HS_CLIENT_FETCH_HAVE_DESC)) {
- /* This case is unique because it can NOT happen in theory. Once we get
- * a new descriptor, the HS client subsystem is notified immediately and
- * the connections waiting for it are handled which means the state will
- * change from renddesc wait state. Log this and continue to next
- * connection. */
+ if (status == HS_CLIENT_FETCH_HAVE_DESC) {
+ /* This is a rare case where a SOCKS connection is in state waiting for
+ * a descriptor but we do have it in the cache.
+ *
+ * This can happen is tor comes back from suspend where it previously
+ * had the descriptor but the intro points were not usuable. Once it
+ * came back to life, the intro point failure cache was cleaned up and
+ * thus the descriptor became usable again leaving us in this code path.
+ *
+ * We'll mark the connection as waiting for a circuit so the descriptor
+ * can be retried. This is safe because a connection in state waiting
+ * for a descriptor can not be in the entry connection pending list. */
+ mark_conn_as_waiting_for_circuit(base_conn, approx_time());
continue;
}
/* In the case of an error, either all SOCKS connections have been