Re: [PATCH v4 5/8] migration: converts exec backend to accept MigrateAddress struct.

[Het Gala]

On 15/05/23 3:59 pm, Daniel P. Berrangé wrote:
On Fri, May 12, 2023 at 02:32:37PM +0000, Het Gala wrote:
Exec transport backend for 'migrate'/'migrate-incoming' QAPIs accept
new wire protocol of MigrateAddress struct.

It is achived by parsing 'uri' string and storing migration parameters
required for exec connection into strList struct.

Suggested-by: Aravind Retnakaran <aravind.retnaka...@nutanix.com>
Signed-off-by: Het Gala <het.g...@nutanix.com>
---
  migration/exec.c      | 58 ++++++++++++++++++++++++++++++++-----------
  migration/exec.h      |  4 +--
  migration/migration.c | 10 +++-----
  3 files changed, 50 insertions(+), 22 deletions(-)

diff --git a/migration/exec.c b/migration/exec.c
index c4a3293246..210f4e9400 100644
--- a/migration/exec.c
+++ b/migration/exec.c
@@ -39,22 +39,51 @@ const char *exec_get_cmd_path(void)
  }
  #endif
  
-void exec_start_outgoing_migration(MigrationState *s, const char *command,
+/* provides the length of strList */
+static int
+str_list_length(strList *list)
+{
+    int len = 0;
+    strList *elem;
+
+    for (elem = list; elem != NULL; elem = elem->next) {
+        len++;
+    }
+
+    return len;
+}
+

Juan for your comment : "I can't believe tat we have a list type and we 
don't have a length() function for that type."

I had seen come patches regarding adding utility for finding length of a 
strList in util.h file as a MACRO, but I think the patches have still 
not gone in.

+static void
+init_exec_array(strList *command, const char **argv, Error **errp)
+{
+    int i = 0;
+    strList *lst;
+
+    for (lst = command; lst; lst = lst->next) {
+        argv[i++] = lst->value;
+    }
+
+    argv[i] = NULL;
+    return;
+}
+
+void exec_start_outgoing_migration(MigrationState *s, strList *command,
                                     Error **errp)
  {
      QIOChannel *ioc;
  
-#ifdef WIN32
-    const char *argv[] = { exec_get_cmd_path(), "/c", command, NULL };
-#else
-    const char *argv[] = { "/bin/sh", "-c", command, NULL };
-#endif
+    int length = str_list_length(command);
+    const char **argv = g_malloc0(length * sizeof(const char *));
g_malloc0 is almost never desirable to use, instead:

         g_new0(const char *, length);
Ack. Will change that.
  
-    trace_migration_exec_outgoing(command);
+    init_exec_array(command, argv, errp);
+    char *new_command = g_strjoinv(" ", (char **)argv);
Never freed - use

    g_autofree char *new_command...
Ack.
+    trace_migration_exec_outgoing(new_command);

      ioc = QIO_CHANNEL(qio_channel_command_new_spawn(argv,
                                                      O_RDWR,
                                                      errp));
      if (!ioc) {
+        g_free(argv);
          return;
      }
argv needs freeing in success too. Simpler to declare it
with

     g_auto(GStrv) argv = .....
Yes, Ack.
  
@@ -72,21 +101,22 @@ static gboolean exec_accept_incoming_migration(QIOChannel *ioc,
      return G_SOURCE_REMOVE;
  }
  
-void exec_start_incoming_migration(const char *command, Error **errp)
+void exec_start_incoming_migration(strList *command, Error **errp)
  {
      QIOChannel *ioc;
  
-#ifdef WIN32
-    const char *argv[] = { exec_get_cmd_path(), "/c", command, NULL };
-#else
-    const char *argv[] = { "/bin/sh", "-c", command, NULL };
-#endif
+    int length = str_list_length(command);
+    const char **argv = g_malloc0(length * sizeof(const char *));
+
+    init_exec_array(command, argv, errp);
+    char *new_command = g_strjoinv(" ", (char **)argv);
  
-    trace_migration_exec_incoming(command);
+    trace_migration_exec_incoming(new_command);
      ioc = QIO_CHANNEL(qio_channel_command_new_spawn(argv,
                                                      O_RDWR,
                                                      errp));
      if (!ioc) {
+        g_free(argv);
          return;
      }
All the same comments as the outgoing case.
Ack. Thanks Daniel for the inputs.
With regards,
Daniel
Regards,
Het Gala