Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Pedro Alves <palves@redhat.com>
To: gdb-patches@sourceware.org
Subject: [PATCH 02/28] gcore, handle exited threads better
Date: Tue, 14 Apr 2020 18:54:08 +0100	[thread overview]
Message-ID: <20200414175434.8047-3-palves@redhat.com> (raw)
In-Reply-To: <20200414175434.8047-1-palves@redhat.com>

An early (and since discarded) version of this series tried to make
exited threads have distinct PTID between each other, and that change
exposed a problem in linux-tdep.c...  This was exposed by the
gdb.threads/gcore-stale-thread.exp testcase, which is exactly about
calling gcore with an exited thread selected:

 (gdb) [Thread 0x7ffff7fb6740 (LWP 31523) exited]
 PASS: gdb.threads/gcore-stale-thread.exp: continue to breakpoint: break-here
 gcore /home/pedro/gdb/binutils-gdb/build/gdb/testsuite/outputs/gdb.threads/gcore-stale-thread/gcore-stale-thread.core
 /home/pedro/gdb/binutils-gdb/build/../src/gdb/inferior.c:66: internal-error: void set_current_inferior(inferior*): Assertion `inf != NULL' failed.
 A problem internal to GDB has been detected,

That was find_inferior_ptid being called on the "exited" ptid, which
on that previous (and discarded attempt) had pid==-1.  The problem is
that linux-tdep.c, where it looks for the signalled thread, isn't
considering exited threads.  Also, while at it, that code isn't
considering multi-target either, since it is using
iterate_over_threads which iterates over all threads of all targets.
Fixed by switching to range-for iteration instead.

gdb/ChangeLog:
yyyy-mm-dd  Pedro Alves  <palves@redhat.com>

	* linux-tdep.c (find_signalled_thread(thread_info *,void *)):
	Delete.
	(find_signalled_thread()): New, factored out from
	linux_make_corefile_notes and adjusted to handle exited threads.
	(linux_make_corefile_notes): Adjust to use the new
	find_signalled_thread.
---
 gdb/linux-tdep.c | 60 +++++++++++++++++++++++++++++++-------------------------
 1 file changed, 33 insertions(+), 27 deletions(-)

diff --git a/gdb/linux-tdep.c b/gdb/linux-tdep.c
index f204831e39..3e39a54ba2 100644
--- a/gdb/linux-tdep.c
+++ b/gdb/linux-tdep.c
@@ -1402,18 +1402,6 @@ linux_find_memory_regions (struct gdbarch *gdbarch,
 					 &data);
 }
 
-/* Determine which signal stopped execution.  */
-
-static int
-find_signalled_thread (struct thread_info *info, void *data)
-{
-  if (info->suspend.stop_signal != GDB_SIGNAL_0
-      && info->ptid.pid () == inferior_ptid.pid ())
-    return 1;
-
-  return 0;
-}
-
 /* This is used to pass information from
    linux_make_mappings_corefile_notes through
    linux_find_memory_regions_full.  */
@@ -1861,6 +1849,30 @@ linux_fill_prpsinfo (struct elf_internal_linux_prpsinfo *p)
   return 1;
 }
 
+/* Find the signalled thread.  In case there's more than one signalled
+   thread, prefer the current thread, if it is signalled.  If no
+   thread was signalled, default to the current thread, unless it has
+   exited, in which case return NULL.  */
+
+static thread_info *
+find_signalled_thread ()
+{
+  thread_info *curr_thr = inferior_thread ();
+  if (curr_thr->state != THREAD_EXITED
+      && curr_thr->suspend.stop_signal != GDB_SIGNAL_0)
+    return curr_thr;
+
+  for (thread_info *thr : current_inferior ()->non_exited_threads ())
+    if (thr->suspend.stop_signal != GDB_SIGNAL_0)
+      return thr;
+
+  /* Default to the current thread, unless it has exited.  */
+  if (curr_thr->state != THREAD_EXITED)
+    return curr_thr;
+
+  return nullptr;
+}
+
 /* Build the note section for a corefile, and return it in a malloc
    buffer.  */
 
@@ -1870,7 +1882,6 @@ linux_make_corefile_notes (struct gdbarch *gdbarch, bfd *obfd, int *note_size)
   struct linux_corefile_thread_data thread_args;
   struct elf_internal_linux_prpsinfo prpsinfo;
   char *note_data = NULL;
-  struct thread_info *curr_thr, *signalled_thr;
 
   if (! gdbarch_iterate_over_regset_sections_p (gdbarch))
     return NULL;
@@ -1898,26 +1909,21 @@ linux_make_corefile_notes (struct gdbarch *gdbarch, bfd *obfd, int *note_size)
     }
 
   /* Like the kernel, prefer dumping the signalled thread first.
-     "First thread" is what tools use to infer the signalled thread.
-     In case there's more than one signalled thread, prefer the
-     current thread, if it is signalled.  */
-  curr_thr = inferior_thread ();
-  if (curr_thr->suspend.stop_signal != GDB_SIGNAL_0)
-    signalled_thr = curr_thr;
-  else
-    {
-      signalled_thr = iterate_over_threads (find_signalled_thread, NULL);
-      if (signalled_thr == NULL)
-	signalled_thr = curr_thr;
-    }
+     "First thread" is what tools use to infer the signalled
+     thread.  */
+  thread_info *signalled_thr = find_signalled_thread ();
 
   thread_args.gdbarch = gdbarch;
   thread_args.obfd = obfd;
   thread_args.note_data = note_data;
   thread_args.note_size = note_size;
-  thread_args.stop_signal = signalled_thr->suspend.stop_signal;
+  if (signalled_thr != nullptr)
+    thread_args.stop_signal = signalled_thr->suspend.stop_signal;
+  else
+    thread_args.stop_signal = GDB_SIGNAL_0;
 
-  linux_corefile_thread (signalled_thr, &thread_args);
+  if (signalled_thr != nullptr)
+    linux_corefile_thread (signalled_thr, &thread_args);
   for (thread_info *thr : current_inferior ()->non_exited_threads ())
     {
       if (thr == signalled_thr)
-- 
2.14.5



  parent reply	other threads:[~2020-04-14 17:54 UTC|newest]

Thread overview: 72+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-14 17:54 [PATCH 00/28] Decouple inferior_ptid/inferior_thread(); dup ptids in thread list (PR/25412) Pedro Alves
2020-04-14 17:54 ` [PATCH 01/28] Don't write to inferior_ptid in linux_get_siginfo_data Pedro Alves
2020-04-14 17:54 ` Pedro Alves [this message]
2020-04-14 17:54 ` [PATCH 03/28] Refactor delete_program_space as a destructor Pedro Alves
2020-04-15 15:54   ` Simon Marchi
2020-04-16 14:47     ` Pedro Alves
2020-04-14 17:54 ` [PATCH 04/28] Don't write to inferior_ptid in gdbarch-selftests.c, mock address_space too Pedro Alves
2020-04-14 17:54 ` [PATCH 05/28] Don't write to inferior_ptid in inf-ptrace.c Pedro Alves
2020-04-14 17:54 ` [PATCH 06/28] Don't write to inferior_ptid in target.c Pedro Alves
2020-04-14 17:54 ` [PATCH 07/28] Don't write to inferior_ptid in infrun.c Pedro Alves
2020-04-14 17:54 ` [PATCH 08/28] Don't write to inferior_ptid in procfs.c Pedro Alves
2020-04-14 17:54 ` [PATCH 09/28] Don't write to inferior_ptid in tracefile-tfile.c Pedro Alves
2020-04-14 17:54 ` [PATCH 10/28] Don't write to inferior_ptid in tracectf.c Pedro Alves
2020-04-14 17:54 ` [PATCH 11/28] Don't write to inferior_ptid in remote.c Pedro Alves
2020-04-14 17:54 ` [PATCH 12/28] Don't write to inferior_ptid in remote-sim.c Pedro Alves
2020-04-14 17:54 ` [PATCH 13/28] Don't write to inferior_ptid in nto-procfs.c Pedro Alves
2020-04-14 17:54 ` [PATCH 14/28] Don't write to inferior_ptid in go32-nat.c Pedro Alves
2020-04-14 17:54 ` [PATCH 15/28] Don't write to inferior_ptid in gnu-nat.c Pedro Alves
2020-04-14 17:54 ` [PATCH 16/28] Don't write to inferior_ptid in darwin-nat.c Pedro Alves
2020-04-16  1:33   ` Simon Marchi
2020-04-16 19:23     ` Pedro Alves
2020-04-14 17:54 ` [PATCH 17/28] Don't write to inferior_ptid in corelow.c Pedro Alves
2020-04-14 17:54 ` [PATCH 18/28] Don't write to inferior_ptid in bsd-kvm.c Pedro Alves
2020-04-14 17:54 ` [PATCH 19/28] Don't write to inferior_ptid in btrace_fetch Pedro Alves
2020-04-15  4:52   ` Metzger, Markus T
2020-04-15 14:13     ` Pedro Alves
2020-04-15 15:17       ` Metzger, Markus T
2020-04-14 17:54 ` [PATCH 20/28] Don't write to inferior_ptid in bsd-kvm.c Pedro Alves
2020-04-14 17:54 ` [PATCH 21/28] Don't write to inferior_ptid in fork-child.c Pedro Alves
2020-04-14 17:54 ` [PATCH 22/28] Don't write to inferior_ptid in go32-nat.c Pedro Alves
2020-04-14 17:54 ` [PATCH 23/28] Don't write to inferior_ptid in remote-sim.c Pedro Alves
2020-04-16  0:53   ` Simon Marchi
2020-04-16 14:58     ` Pedro Alves
2020-04-14 17:54 ` [PATCH 24/28] Don't write to inferior_ptid in windows-nat.c, part I Pedro Alves
2020-04-14 17:54 ` [PATCH 25/28] Don't write to inferior_ptid in windows-nat.c, part II Pedro Alves
2020-04-14 22:41   ` Hannes Domani
2020-04-15 15:08     ` Pedro Alves
2020-04-15 15:32       ` Hannes Domani
2020-04-14 17:54 ` [PATCH 26/28] Don't write to inferior_ptid in ravenscar-thread.c Pedro Alves
2020-04-17 18:45   ` Tom Tromey
2020-06-18 20:00     ` Pedro Alves
2020-06-18 21:38       ` Tom Tromey
2020-04-14 17:54 ` [PATCH 27/28] Don't write to inferior_ptid in aix-thread.c Pedro Alves
2020-04-14 17:54 ` [PATCH 28/28] Decouple inferior_ptid/inferior_thread(); dup ptids in thread list (PR/25412) Pedro Alves
2020-04-16 19:39   ` Simon Marchi
2020-04-16 20:12     ` Pedro Alves
2020-04-16 20:38       ` Simon Marchi
2020-04-17 10:29         ` Pedro Alves
2020-04-17 14:06           ` Simon Marchi
2020-04-17 16:46             ` Pedro Alves
2020-04-17 18:53   ` Tom Tromey
2020-06-18 19:59     ` Pedro Alves
2020-06-23 13:37   ` Andrew Burgess
2020-06-23 14:26     ` Pedro Alves
2020-06-23 15:38       ` [PATCH] Fix "maint selftest" regression, add struct, scoped_mock_context Pedro Alves
2020-06-23 16:34         ` Andrew Burgess
2020-06-23 17:58           ` Pedro Alves
2020-04-14 18:46 ` [PATCH 00/28] Decouple inferior_ptid/inferior_thread(); dup ptids in thread list (PR/25412) Hannes Domani
2020-04-14 19:24   ` Pedro Alves
2020-04-15 15:04     ` Simon Marchi
2020-04-16 13:41       ` Pedro Alves
2020-04-15 14:46 ` Simon Marchi
2020-04-15 15:33   ` Pedro Alves
2020-04-15 15:42     ` Simon Marchi
2020-04-17 20:20 ` Tom Tromey
2020-06-18 20:00   ` Pedro Alves
2020-06-18 22:30 ` Pedro Alves
2020-07-07 23:16 ` John Baldwin
2020-07-07 23:53   ` Pedro Alves
2020-07-08  0:19     ` John Baldwin
2020-07-08  0:10   ` Multiprocess on FreeBSD John Baldwin
2020-07-08  0:34     ` John Baldwin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200414175434.8047-3-palves@redhat.com \
    --to=palves@redhat.com \
    --cc=gdb-patches@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox