BUG/MEDIUM: connection: properly leave stopping list on error
The stopping-list management introduced by commit d3a88c1c3 ("MEDIUM:
connection: close front idling connection on soft-stop") missed two
error paths in the H1 and H2 muxes. The effect is that if a stream
or HPACK table couldn't be allocated for these incoming connections,
we would leave with the connection freed still attached to the
stopping_list and it would never leave it, resulting in use-after-free
hence either a crash or a data corruption.
This is marked as medium as it only happens under extreme memory pressure
or when playing with tune.fail-alloc. Other stability issues remain in
such a case so that abnormal behaviors cannot be explained by this bug
alone.
This must be backported to 2.4.
diff --git a/src/mux_h1.c b/src/mux_h1.c
index a00ac60..1f8cad2 100644
--- a/src/mux_h1.c
+++ b/src/mux_h1.c
@@ -990,6 +990,8 @@
tasklet_free(h1c->wait_event.tasklet);
pool_free(pool_head_h1c, h1c);
fail_h1c:
+ if (!conn_is_back(conn))
+ LIST_DEL_INIT(&conn->stopping_list);
conn->ctx = conn_ctx; // restore saved context
TRACE_DEVEL("leaving in error", H1_EV_H1C_NEW|H1_EV_H1C_END|H1_EV_H1C_ERR);
return -1;
diff --git a/src/mux_h2.c b/src/mux_h2.c
index 6338cf0..ddc19c1 100644
--- a/src/mux_h2.c
+++ b/src/mux_h2.c
@@ -1042,6 +1042,8 @@
tasklet_free(h2c->wait_event.tasklet);
pool_free(pool_head_h2c, h2c);
fail_no_h2c:
+ if (!conn_is_back(conn))
+ LIST_DEL_INIT(&conn->stopping_list);
conn->ctx = conn_ctx; /* restore saved ctx */
TRACE_DEVEL("leaving in error", H2_EV_H2C_NEW|H2_EV_H2C_END|H2_EV_H2C_ERR);
return -1;