sssd/SOURCES/0002-sbus-arm-watchdog-for-sbus_connect_init_send.patch
2023-09-21 20:30:23 +00:00

54 lines
1.9 KiB
Diff

From 55564defec8fdbb4d9df6b0124a8b18b31743230 Mon Sep 17 00:00:00 2001
From: Sumit Bose <sbose@redhat.com>
Date: Fri, 9 Jun 2023 13:01:47 +0200
Subject: [PATCH 2/2] sbus: arm watchdog for sbus_connect_init_send()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
There seem to be conditions where the reply in the
sbus_call_DBus_Hello_send() request gets lost and the backend cannot
properly initialize its sbus/DBus server. Since the backend cannot be
connected by the frontends in this state the best way to recover would
be a restart. Since the event-loop is active in this state, e.g. waiting
for the reply, the watchdog will not consider the process as hung and
will not restart the process.
To make the watchdog handle this case arm_watchdog() and
disarm_watchdog() are called before and after the request, respectively.
Resolves: https://github.com/SSSD/sssd/issues/6803
Reviewed-by: Alexey Tikhonov <atikhono@redhat.com>
Reviewed-by: Pavel Březina <pbrezina@redhat.com>
(cherry picked from commit cca9361d92501e0be34d264d370fe897a0c970af)
---
src/sbus/connection/sbus_connection_connect.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/src/sbus/connection/sbus_connection_connect.c b/src/sbus/connection/sbus_connection_connect.c
index 45a0fa491..edc090e15 100644
--- a/src/sbus/connection/sbus_connection_connect.c
+++ b/src/sbus/connection/sbus_connection_connect.c
@@ -67,6 +67,8 @@ sbus_connect_init_send(TALLOC_CTX *mem_ctx,
tevent_req_set_callback(subreq, sbus_connect_init_hello_done, req);
+ arm_watchdog();
+
return req;
}
@@ -111,6 +113,8 @@ static void sbus_connect_init_done(struct tevent_req *subreq)
uint32_t res;
errno_t ret;
+ disarm_watchdog();
+
req = tevent_req_callback_data(subreq, struct tevent_req);
ret = sbus_call_DBus_RequestName_recv(subreq, &res);
--
2.38.1