[gdb/testsuite] Fix gdb.threads/access-mem-running-thread-exit.exp

In OBS (Open Build Service), with a 15.2 based gdb package, occasionally I run
into:
...
(gdb) inferior 2
[Switching to inferior 2 [process 31372] (access-mem-running-thread-exit)]
[Switching to thread 2.1 (Thread 0xf7db9700 (LWP 31372))](running)
(gdb) print global_var = 555
$1 = 555
(gdb) print global_var
$2 = 556
(gdb) FAIL: $exp: all-stop: access mem \
  (print global_var after writing, inf=2, iter=1)
...

I managed to reproduce this on current trunk using a reproducer patch (posted
in the PR).

The problem is due to commit 31c21e2c13d ("[gdb/testsuite] Fix
gdb.threads/access-mem-running-thread-exit.exp with clang"), which introduced
an increment of global_var at the start of main.

This created a race between:
- gdb modifying global_var, and
- the inferior modifying global_var.

Fix this by:
- adding a new empty function setup_done,
- adding a call to setup_done after the increment of global_var, and
- rather than running to main, running to setup_done.

Tested on x86_64-linux.

PR testsuite/32822
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=32822

diff --git a/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.c b/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.c
index af05b13c76398e3e826876b17f00192f45c98a82..e22bf12df752961e5a8758f226a38ed61919f5fb 100644 (file)
--- a/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.c
+++ b/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.c
@@ -97,6 +97,11 @@ thread_fn (void *arg)
   return NULL;
 }
 
+static void
+setup_done (void)
+{
+}
+
 int
 main (void)
 {
@@ -104,6 +109,8 @@ main (void)
 
   global_var++;
 
+  setup_done ();
+
   for (i = 0; i < 4; i++)
     {
       struct thread_arg *p;

diff --git a/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.exp b/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.exp
index 784f17ff3b20b80b90d888f7008d53b7f696b97a..42222c0fb354f74b5073d0dddd29fa146c4d3658 100644 (file)
--- a/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.exp
+++ b/gdb/testsuite/gdb.threads/access-mem-running-thread-exit.exp
@@ -54,7 +54,7 @@ proc test { non_stop } {
       clean_restart ${binfile}
     }
 
-    if ![runto_main] {
+    if ![runto setup_done] {
        return -1
     }
 
@@ -76,7 +76,7 @@ proc test { non_stop } {
     # Start the second inferior.
     with_test_prefix "second inferior" {
        # With stub targets that do reload on run, if we let the new
-       # inferior share inferior 1's connection, runto_main would
+       # inferior share inferior 1's connection, runto would
        # fail because GDB is already connected to something, like
        # e.g. with --target_board=native-gdbserver:
        #
@@ -86,10 +86,10 @@ proc test { non_stop } {
        #  Already connected to a remote target.  Disconnect? (y or n)
        #
        # Instead, start the inferior with no connection, and let
-       # gdb_load/runto_main spawn a new remote connection/gdbserver.
+       # gdb_load/runto spawn a new remote connection/gdbserver.
        #
        # OTOH, with extended-remote, we must let the new inferior
-       # reuse the current connection, so that runto_main below can
+       # reuse the current connection, so that runto below can
        # issue the "run" command, and have the inferior run on the
        # remote target.  If we forced no connection, then "run" would
        # either fail if "set auto-connect-native-target" is on, like
@@ -108,7 +108,7 @@ proc test { non_stop } {
 
        gdb_load $binfile
 
-       if ![runto_main] {
+       if ![runto setup_done] {
            return -1
        }
     }