qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 3/4] qga: Don't daemonize before channel is initialized


From: Konstantin Kostiuk
Subject: Re: [PATCH 3/4] qga: Don't daemonize before channel is initialized
Date: Wed, 4 Dec 2024 11:44:16 +0200



On Mon, Nov 4, 2024 at 11:54 AM Michal Privoznik <mprivozn@redhat.com> wrote:
If the agent is set to daemonize but for whatever reason fails to
init the channel, the error message is lost. Worse, the agent
daemonizes needlessly and returns success. For instance:

  # qemu-ga -m virtio-serial \
            -p /dev/nonexistent_device \
            -f /run/qemu-ga.pid \
            -t /run \
            -d
  # echo $?
  0

This makes it needlessly hard for init scripts to detect a
failure in qemu-ga startup. Though, they shouldn't pass '-d' in
the first place.

Let's open the channel first and only after that become a daemon.

Related bug: https://bugs.gentoo.org/810628

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
---
 qga/main.c | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/qga/main.c b/qga/main.c
index c003aacbe0..6240845f39 100644
--- a/qga/main.c
+++ b/qga/main.c
@@ -1430,7 +1430,6 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
         if (config->daemonize) {
             /* delay opening/locking of pidfile till filesystems are unfrozen */
             s->deferred_options.pid_filepath = config->pid_filepath;
-            become_daemon(NULL);
         }
         if (config->log_filepath) {
             /* delay opening the log file till filesystems are unfrozen */
@@ -1438,9 +1437,6 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)
         }
         ga_disable_logging(s);
     } else {
-        if (config->daemonize) {
-            become_daemon(config->pid_filepath);
-        }
         if (config->log_filepath) {
             FILE *log_file = ga_open_logfile(config->log_filepath);
             if (!log_file) {
@@ -1487,6 +1483,20 @@ static GAState *initialize_agent(GAConfig *config, int socket_activation)

     ga_apply_command_filters(s);

+    if (!channel_init(s, s->config->method, s->config->channel_path,
+                      s->socket_activation ? FIRST_SOCKET_ACTIVATION_FD : -1)) {
+        g_critical("failed to initialize guest agent channel");
+        return NULL;
+    }
+
+    if (config->daemonize) {
+        if (ga_is_frozen(s)) {
+            become_daemon(NULL);
+        } else {
+            become_daemon(config->pid_filepath);
+        }
+    }
+
     ga_state = s;
     return s;
 }
@@ -1513,12 +1523,6 @@ static void cleanup_agent(GAState *s)

 static int run_agent_once(GAState *s)
 {
-    if (!channel_init(s, s->config->method, s->config->channel_path,
-                      s->socket_activation ? FIRST_SOCKET_ACTIVATION_FD : -1)) {
-        g_critical("failed to initialize guest agent channel");
-        return EXIT_FAILURE;
-    }
-

The old flow:
run_agent call run_agent_once in loop
run_agent_once initialize channel for every run

after your changes
you expect to initialize the channel only once
this can cause issues on Windows during driver update
the default configuration on Windows is QGA + VirtioSerial and in CLI --retry-path
during driver update (that can happen via Windows Update) the channel will be closed
so QGA must reinitialize the channel

Theoretically, the same can happen on Linux with a UNIX socket

 
     g_main_loop_run(s->main_loop);

     if (s->channel) {
--
2.45.2


reply via email to

[Prev in Thread] Current Thread [Next in Thread]