[Bug gdb/31832] [gdb] FAIL: gdb.threads/attach-many-short-lived-threads.exp: iter 3: attach (timeout)

["vries at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=31832

--- Comment #6 from Tom de Vries <vries at gcc dot gnu.org> ---
Created attachment 15582
  --> https://sourceware.org/bugzilla/attachment.cgi?id=15582&action=edit
gdb.log (with statistics-gathering patch)

I applied the following statistics-gathering patch:
...
diff --git a/gdb/nat/linux-procfs.c b/gdb/nat/linux-procfs.c
index 5efc86cca1f..f6847257b85 100644
--- a/gdb/nat/linux-procfs.c
+++ b/gdb/nat/linux-procfs.c
@@ -22,6 +22,7 @@
 #include <sys/stat.h>
 #include <unordered_set>
 #include <utility>
+#include "gdbsupport/scope-exit.h"

 /* Return the TGID of LWPID from /proc/pid/status.  Returns -1 if not
    found.  */
@@ -377,17 +378,42 @@ linux_proc_attach_tgid_threads (pid_t pid,
      threads, new threads may be spawned.  Cycle through the list of
      threads until we have done two iterations without finding new
      threads.  */
+
+  unsigned long long total_iterations = 0;
+  unsigned long long dir_entries = 0;
+  unsigned long long no_lwp = 0;
+  unsigned long long lookup = 0;
+  unsigned long long skipped = 0;
+  unsigned long long insert = 0;
+  unsigned long long attach = 0;
+  unsigned long long start_over = 0;
+  SCOPE_EXIT
+    {
+      printf ("total_iterations: %llu\n", total_iterations);
+      printf ("dir_entries: %llu\n", dir_entries);
+      printf ("no_lwp: %llu\n", no_lwp);
+      printf ("lookup: %llu\n", lookup);
+      printf ("skipped: %llu\n", skipped);
+      printf ("insert: %llu\n", insert);
+      printf ("attach: %llu\n", attach);
+      printf ("start_over: %llu\n", start_over);
+    };
+
   for (iterations = 0; iterations < 2; iterations++)
     {
+      total_iterations++;
       struct dirent *dp;

       new_threads_found = 0;
       while ((dp = readdir (dir.get ())) != NULL)
        {
+         dir_entries++;
          unsigned long lwp;

          /* Fetch one lwp.  */
          lwp = strtoul (dp->d_name, NULL, 10);
+         if (lwp == 0)
+           no_lwp++;
          if (lwp != 0)
            {
              ptid_t ptid = ptid_t (pid, lwp);
@@ -399,12 +425,18 @@ linux_proc_attach_tgid_threads (pid_t pid,
                  std::pair<unsigned long, ULONGEST> key (lwp, *starttime);

                  /* If we already visited this LWP, skip it this time.  */
+                 lookup++;
                  if (visited_lwps.find (key) != visited_lwps.cend ())
-                   continue;
+                   {
+                     skipped++;
+                     continue;
+                   }

+                 insert++;
                  visited_lwps.insert (key);
                }

+             attach++;
              if (attach_lwp (ptid))
                new_threads_found = 1;
            }
@@ -414,6 +446,7 @@ linux_proc_attach_tgid_threads (pid_t pid,
        {
          /* Start over.  */
          iterations = -1;
+         start_over++;
        }

       rewinddir (dir.get ());
...
and ran the test-case.

The first iteration gives us:
...
(gdb) builtin_spawn
/home/vries/gdb/build/gdb/testsuite/outputs/gdb.threads/attach-many-short-lived-threads/attach-many-short-lived-threads^M
attach 1301317^M
Attaching to program:
/home/vries/gdb/build/gdb/testsuite/outputs/gdb.threads/attach-many-short-lived-threads/attach-many-short-lived-threads,
process 1301317^M
FAIL: gdb.threads/attach-many-short-lived-threads.exp: iter 1: attach (timeout)
total_iterations: 2092^M
dir_entries: 594518^M
no_lwp: 4412^M
lookup: 119037^M
skipped: 118355^M
insert: 682^M
attach: 471751^M
start_over: 2091^M
Cannot attach to lwp 2340832: Operation not permitted (1)^M
...

I'm not sure what this means, but I do notice the big difference between
dir_entries and lookup.  So only 20% of the time we find the starttime and can
use the cache.

-- 
You are receiving this mail because:
You are on the CC list for the bug.