Handle gdb SIGTERM by throwing / catching gdb_exception_force_quit

When a GDB process receives the SIGTERM signal, handle_sigterm() in
event-top.c is called.  The global variable 'sync_quit_force_run' is
set by this signal handler.  It does some other things too, but the
setting of this global is the important bit for the SIGTERM part of
this discussion.

GDB will periodically check to see whether a Ctrl-C or SIGTERM has
been received.  This is performed via use of the QUIT macro in
GDB's code.  QUIT is defined to invoke maybe_quit(), which will be
periodically called during any lengthy operation.  This is supposed to
ensure that the user won't have to wait too long for a Ctrl-C or
SIGTERM to be acted upon.

When a Ctrl-C / SIGINT is received, quit_handler() will decide whether
to pass the SIGINT onto the inferior or to call quit() which causes
gdb_exception_quit to be thrown.  This exception (usually) propagates
to the top level.  Control is then returned to the top level event
loop.

At the moment, SIGTERM is handled very differently.  Instead of
throwing an exception, quit_force() is called.  This does eventually
cause GDB to exit(), but prior to that happening, the inferiors
are killed or detached and other target related cleanup occurs.
As shown in this discussion between Pedro Alves and myself...

https://sourceware.org/pipermail/gdb-patches/2021-July/180802.html
https://sourceware.org/pipermail/gdb-patches/2021-July/180902.html
https://sourceware.org/pipermail/gdb-patches/2021-July/180903.html

...we found that it is possible for inferior_ptid and current_thread_
to get out of sync.  When that happens, the "current_thread_ != nullptr"
assertion in inferior_thread() can fail resulting in a GDB internal
error.

Pedro recommended that we "let the normal quit exception propagate all
the way to the top level, and then have the top level call quit_force
if sync_quit_force_run is set."  However, after the v2 series for this
patch set, we tweaked that idea by introducing a new exception for
handling SIGTERM.

This commit implements the obvious part of Pedro's suggestion:
Instead of calling quit_force from quit(), throw_forced_quit() is now
called instead.  This causes the new exception 'gdb_exception_forced_quit'
to be thrown.

At the top level, I changed catch_command_errors() and captured_main()
to catch gdb_exception_forced_quit and then call quit_force() from the
catch block.  I also changed start_event_loop() to also catch
gdb_exception_forced_quit; while we could also call quit_force() from
that catch block, it's sufficient to simply rethrow the exception
since it'll be caught by the newly added code in captured_main().

Making these changes fixed the failure / regression that I was seeing
for gdb.base/gdb-sigterm.exp when run on a machine with glibc-2.34.
However, there are many other paths back to the top level which this
test case does not test.  I did an audit of all of the try / catch
code in GDB in which calls in the try-block might (eventually) call
QUIT.  I found many cases where gdb_exception_quit and the new
gdb_exception_forced_quit will be swallowed.  (When using GDB, have
you ever hit Ctrl-C and not have it do anything; if so, it could be
due to a swallowed gdb_exception_quit in one of the cases I've
identified.)  The rest of the patches in this series deal with this
concern.

Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=26761
Tested-by: Tom de Vries <tdevries@suse.de>
Approved-by: Pedro Alves <pedro@palves.net>

diff --git a/gdb/main.c b/gdb/main.c
index c04d37a45f9165234c5ce41cc9ed546523776da9..0bf6f98a797bfaabdaf124bfd068b10e2b67ea96 100644 (file)
--- a/gdb/main.c
+++ b/gdb/main.c
@@ -410,6 +410,10 @@ start_event_loop ()
        {
          result = gdb_do_one_event ();
        }
+      catch (const gdb_exception_forced_quit &ex)
+       {
+         throw;
+       }
       catch (const gdb_exception &ex)
        {
          exception_print (gdb_stderr, ex);
@@ -518,6 +522,10 @@ catch_command_errors (catch_command_errors_const_ftype command,
       if (do_bp_actions)
        bpstat_do_actions ();
     }
+  catch (const gdb_exception_forced_quit &e)
+    {
+      quit_force (NULL, 0);
+    }
   catch (const gdb_exception &e)
     {
       return handle_command_errors (e);
@@ -1309,6 +1317,10 @@ captured_main (void *data)
        {
          captured_command_loop ();
        }
+      catch (const gdb_exception_forced_quit &ex)
+       {
+         quit_force (NULL, 0);
+       }
       catch (const gdb_exception &ex)
        {
          exception_print (gdb_stderr, ex);

diff --git a/gdb/utils.c b/gdb/utils.c
index 91e6974b97640b802f1ff1b0449231d1d37daec8..08cc41b4ef875779792afec89dd2a53a204be09c 100644 (file)
--- a/gdb/utils.c
+++ b/gdb/utils.c
@@ -642,7 +642,7 @@ quit (void)
   if (sync_quit_force_run)
     {
       sync_quit_force_run = 0;
-      quit_force (NULL, 0);
+      throw_forced_quit ("SIGTERM");
     }
 
 #ifdef __MSDOS__