bug-coreutils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#9076: coreutils-8.12 uses SA_RESETHAND and SA_RESTART unconditionall


From: Paul Eggert
Subject: bug#9076: coreutils-8.12 uses SA_RESETHAND and SA_RESTART unconditionally
Date: Fri, 15 Jul 2011 18:11:43 -0700
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2.17) Gecko/20110516 Thunderbird/3.1.10

Here are three proposed patches to coreutils to address
the SA_RESETHAND and SA_RESTART issues.  I hope that
if you combine them with the patch I already installed:

http://git.savannah.gnu.org/cgit/coreutils.git/commit/?id=837e1f55196f826b92d660808f594fde36651655

these should fix all the issues mentioned here.
Joachim, could you please try these?  Thanks.

>From bbd4c9edfa4bd5650106261b7d9b1dd434d91581 Mon Sep 17 00:00:00 2001
From: Paul Eggert <address@hidden>
Date: Fri, 15 Jul 2011 17:38:32 -0700
Subject: [PATCH 1/3] dd: port to NonStop (Bug#9076)

* src/dd.c (SA_RESETHAND): Define to 0 if not defined.
---
 src/dd.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/src/dd.c b/src/dd.c
index 3e75412..0824f6c 100644
--- a/src/dd.c
+++ b/src/dd.c
@@ -55,6 +55,11 @@
 # endif
 #endif
 
+/* NonStop circa 2011 lacks SA_RESETHAND; see Bug#9076.  */
+#ifndef SA_RESETHAND
+# define SA_RESETHAND 0
+#endif
+
 #ifndef SIGINFO
 # define SIGINFO SIGUSR1
 #endif
-- 
1.7.4.4


>From 995299ff155bef70a3b153dc8cef11ed9b1d8904 Mon Sep 17 00:00:00 2001
From: Paul Eggert <address@hidden>
Date: Fri, 15 Jul 2011 17:39:28 -0700
Subject: [PATCH 2/3] ls: port to NonStop (Bug#9076)

* src/ls.c (SA_RESTART): Define to 0 if not defined.
---
 src/ls.c |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)

diff --git a/src/ls.c b/src/ls.c
index c604e14..680a7c3 100644
--- a/src/ls.c
+++ b/src/ls.c
@@ -74,6 +74,14 @@
 # endif
 #endif
 
+/* NonStop circa 2011 lacks both SA_RESTART and siginterrupt, so don't
+   restart syscalls after a signal handler fires.  This may cause
+   colors to get messed up on the screen if 'ls' is interrupted, but
+   that's the best we can do on such a platform.  */
+#ifndef SA_RESTART
+# define SA_RESTART 0
+#endif
+
 #include "system.h"
 #include <fnmatch.h>
 
-- 
1.7.4.4


>From 638e670d76b3bf573f6a9930b376811b5663881a Mon Sep 17 00:00:00 2001
From: Paul Eggert <address@hidden>
Date: Fri, 15 Jul 2011 17:48:38 -0700
Subject: [PATCH 3/3] timeout: port to NonStop (Bug#9077)

* src/timeout.c (SA_RESTART): Define to 0 if not defined.
(main): Don't assume signal handling uses SA_RESTART.
---
 src/timeout.c |   21 ++++++++++++---------
 1 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/src/timeout.c b/src/timeout.c
index ef660a7..895d720 100644
--- a/src/timeout.c
+++ b/src/timeout.c
@@ -64,6 +64,11 @@
 # include <sys/resource.h>
 #endif
 
+/* NonStop circa 2011 lacks both SA_RESTART and siginterrupt.  */
+#ifndef SA_RESTART
+# define SA_RESTART 0
+#endif
+
 #define PROGRAM_NAME "timeout"
 
 #define AUTHORS proper_name_utf8 ("Padraig Brady", "P\303\241draig Brady")
@@ -256,7 +261,8 @@ install_signal_handlers (int sigterm)
   struct sigaction sa;
   sigemptyset (&sa.sa_mask);  /* Allow concurrent calls to handler */
   sa.sa_handler = cleanup;
-  sa.sa_flags = SA_RESTART;  /* restart syscalls (like wait() below) */
+  sa.sa_flags = SA_RESTART;   /* Restart syscalls if possible, as that's
+                                 more likely to work cleanly.  */
 
   sigaction (SIGALRM, &sa, NULL); /* our timeout.  */
   sigaction (SIGINT, &sa, NULL);  /* Ctrl-C at terminal for example.  */
@@ -354,18 +360,15 @@ main (int argc, char **argv)
     }
   else
     {
+      pid_t wait_result;
       int status;
 
       alarm (timeout);
 
-      /* We're just waiting for a single process here, so wait() suffices.
-         Note the signal() calls above on GNU/Linux and BSD at least,
-         essentially call the lower level sigaction() with the SA_RESTART flag
-         set, which ensures the following wait call will only return if the
-         child exits, not on this process receiving a signal. Also we're not
-         passing WUNTRACED | WCONTINUED to a waitpid() call and so will not get
-         indication that the child has stopped or continued.  */
-      if (wait (&status) == -1)
+      while ((wait_result = wait (&status)) < 0 && errno == EINTR)
+        continue;
+
+      if (wait_result < 0)
         {
           /* shouldn't happen.  */
           error (0, errno, _("error waiting for command"));
-- 
1.7.4.4







reply via email to

[Prev in Thread] Current Thread [Next in Thread]