all messages for Emacs-related lists mirrored at yhetil.org
 help / color / mirror / code / Atom feed
From: Paul Eggert <eggert@cs.ucla.edu>
To: Jorgen Schaefer <forcer@forcix.cx>
Cc: 17561@debbugs.gnu.org
Subject: bug#17561: Emacs can forget processes
Date: Fri, 23 May 2014 09:28:44 -0700	[thread overview]
Message-ID: <537F773C.8060202@cs.ucla.edu> (raw)
In-Reply-To: <8738g0bigs.fsf@loki.jorgenschaefer.de>

[-- Attachment #1: Type: text/plain, Size: 604 bytes --]

How was Emacs configured and built on your platform?  What's your platform?

Can you run the shell command 'strace -p' on an Emacs with the problem, 
and trace the system calls near the offending area?

Are you using anything involving SIGUSR1 or SIGUSR2?  I see an unlikely 
race condition there.

That's part of the problem with this code: it's so racy that it seems 
that every time I look at it I find another unlikely race condition. 
I'm attaching a patch to emacs-24 for the unlikely races I found in a 
quick look at the area, not that I think it'll fix your problem.  I 
haven't installed this.

[-- Attachment #2: racy.patch --]
[-- Type: text/plain, Size: 2791 bytes --]

=== modified file 'src/nsterm.m'
--- src/nsterm.m	2014-04-04 16:32:24 +0000
+++ src/nsterm.m	2014-05-23 16:13:50 +0000
@@ -4427,7 +4427,7 @@
 #ifdef NS_IMPL_GNUSTEP
   /* GNUstep steals SIGCHLD for use in NSTask, but we don't use NSTask.
      We must re-catch it so subprocess works.  */
-  catch_child_signal ();
+  catch_child_signal (false);
 #endif
   return dpyinfo;
 }

=== modified file 'src/process.c'
--- src/process.c	2014-05-03 20:13:10 +0000
+++ src/process.c	2014-05-23 16:17:57 +0000
@@ -6254,7 +6254,7 @@
 #ifdef NS_IMPL_GNUSTEP
   /* NSTask in GNUstep sets its child handler each time it is called.
      So we must re-set ours.  */
-  catch_child_signal();
+  catch_child_signal (true);
 #endif
 }
 
@@ -7062,16 +7062,18 @@
 /* Arrange to catch SIGCHLD if this hasn't already been arranged.
    Invoke this after init_process_emacs, and after glib and/or GNUstep
    futz with the SIGCHLD handler, but before Emacs forks any children.
-   This function's caller should block SIGCHLD.  */
+   futz with the SIGCHLD handler, but before Emacs forks any children.
+   If ALREADY_BLOCKED, this function's caller has already blocked SIGCHLD.  */
 
 #ifndef NS_IMPL_GNUSTEP
 static
 #endif
 void
-catch_child_signal (void)
+catch_child_signal (bool already_blocked)
 {
   struct sigaction action, old_action;
   emacs_sigaction_init (&action, deliver_child_signal);
+  if (!already_blocked)
   block_child_signal ();
   sigaction (SIGCHLD, &action, &old_action);
   eassert (! (old_action.sa_flags & SA_SIGINFO));
@@ -7081,6 +7083,7 @@
       = (old_action.sa_handler == SIG_DFL || old_action.sa_handler == SIG_IGN
 	 ? dummy_handler
 	 : old_action.sa_handler);
+  if (!already_blocked)
   unblock_child_signal ();
 }
 #endif	/* subprocesses */
@@ -7107,7 +7110,7 @@
 	 it into lib_child_handler.  */
       g_source_unref (g_child_watch_source_new (getpid ()));
 #endif
-      catch_child_signal ();
+      catch_child_signal (false);
     }
 
   FD_ZERO (&input_wait_mask);

=== modified file 'src/process.h'
--- src/process.h	2014-01-01 07:43:34 +0000
+++ src/process.h	2014-05-23 16:12:58 +0000
@@ -241,7 +241,7 @@
 extern void add_write_fd (int fd, fd_callback func, void *data);
 extern void delete_write_fd (int fd);
 #ifdef NS_IMPL_GNUSTEP
-extern void catch_child_signal (void);
+extern void catch_child_signal (bool);
 #endif
 
 #ifdef WINDOWSNT

=== modified file 'src/sysdep.c'
--- src/sysdep.c	2014-04-16 13:27:28 +0000
+++ src/sysdep.c	2014-05-23 15:58:51 +0000
@@ -1515,6 +1515,12 @@
 #ifdef PROFILER_CPU_SUPPORT
   sigaddset (&action->sa_mask, SIGPROF);
 #endif
+#ifdef SIGUSR1
+  sigaddset (&action->sa_mask, SIGUSR1);
+#endif
+#ifdef SIGUSR2
+  sigaddset (&action->sa_mask, SIGUSR2);
+#endif
 #ifdef SIGWINCH
   sigaddset (&action->sa_mask, SIGWINCH);
 #endif


  reply	other threads:[~2014-05-23 16:28 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-05-23 15:52 bug#17561: 24.4.50; Emacs can forget processes Jorgen Schaefer
2014-05-23 16:28 ` Paul Eggert [this message]
2014-05-23 16:44   ` bug#17561: " Jorgen Schaefer
2014-05-24 23:01     ` Paul Eggert
2014-05-25  7:57       ` Jorgen Schaefer
2014-05-26 17:08         ` Paul Eggert
2014-05-26 18:49           ` Jorgen Schaefer
2014-05-26 23:58             ` Paul Eggert
2014-05-27 18:27               ` Jorgen Schaefer
2014-05-27 21:42                 ` Paul Eggert
2014-05-27 22:16                   ` Jorgen Schaefer
2014-05-28  0:47                     ` Paul Eggert
2014-05-28 20:53                       ` Jorgen Schaefer
2014-05-28 23:00                         ` Paul Eggert
2014-05-28 23:35                           ` Jorgen Schaefer
2014-05-29  1:22                             ` Paul Eggert
2014-05-29 10:08                               ` Jorgen Schaefer
2014-05-29 23:15                                 ` Paul Eggert
2014-05-29  4:17                             ` Paul Eggert
2014-05-29 11:39                               ` Jorgen Schaefer
2014-05-29 15:09                                 ` Paul Eggert
2014-05-29 15:22                                   ` Andreas Schwab
2014-05-29 15:26                                     ` Paul Eggert
2014-05-29 17:03                                       ` Jorgen Schaefer
2014-05-29 17:55                                         ` Andreas Schwab
2014-05-29 18:23                                           ` Jorgen Schaefer
2014-05-29 19:06                                             ` Jorgen Schaefer
2014-05-29 20:27                                               ` Andreas Schwab
2014-05-29 19:15                                             ` Andreas Schwab
2014-05-30  6:07                                               ` Paul Eggert
2014-05-30 20:41                                                 ` Jorgen Schaefer
2014-05-30 21:29                                                   ` Paul Eggert
2014-05-27  4:05             ` Paul Eggert

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=537F773C.8060202@cs.ucla.edu \
    --to=eggert@cs.ucla.edu \
    --cc=17561@debbugs.gnu.org \
    --cc=forcer@forcix.cx \
    /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.
Code repositories for project(s) associated with this external index

	https://git.savannah.gnu.org/cgit/emacs.git
	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.