laforge submitted this change.
lchan/gscon: always clear both cross ref pointers
During specific release scenarios, it became clear that an lchan still
pointed at a conn even after it had been deallocated. That was due to
setting conn->lchan = NULL but not lchan->conn = NULL. Fix that.
Do lchan_forget_conn() first, because during gscon_forget_lchan() we may
enter the gscon clearing dance, which in case of no SCCP conn being
present will soon / should immediately deallocate the conn.
Related: OS#5337
Related: I8c8537acf6b47b121903197608636c43ae601a57 (osmo-bsc)
Change-Id: Idbfe4672233ba8105eff5ba77ee07fd871358255
---
M src/osmo-bsc/lchan_fsm.c
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/src/osmo-bsc/lchan_fsm.c b/src/osmo-bsc/lchan_fsm.c
index 5ae6df6..84f8dc5 100644
--- a/src/osmo-bsc/lchan_fsm.c
+++ b/src/osmo-bsc/lchan_fsm.c
@@ -478,8 +478,11 @@
{
LOG_LCHAN(lchan, LOGL_DEBUG, "Clearing lchan state\n");
- if (lchan->conn)
- gscon_forget_lchan(lchan->conn, lchan);
+ if (lchan->conn) {
+ struct gsm_subscriber_connection *conn = lchan->conn;
+ lchan_forget_conn(lchan);
+ gscon_forget_lchan(conn, lchan);
+ }
if (lchan->rqd_ref) {
talloc_free(lchan->rqd_ref);
@@ -1392,8 +1395,9 @@
* lchan_reset(), we make sure it does. But in case of releases from error handling, the
* conn might as well notice now already that its lchan is becoming unusable. */
if (lchan->conn) {
- gscon_forget_lchan(lchan->conn, lchan);
+ struct gsm_subscriber_connection *conn = lchan->conn;
lchan_forget_conn(lchan);
+ gscon_forget_lchan(conn, lchan);
}
rc = rsl_tx_rf_chan_release(lchan);
To view, visit change 26610. To unsubscribe, or for help writing mail filters, visit settings.