[HEADS-UP] gdb segfault

[HEADS-UP] gdb segfault

[Michael Snyder]

Hey yo --

gdb is seg-faulting immediately whenever its run with --noasync.
I have an old build from the beginning of March that doesn't
do it (nor do even older ones), so it's probably a change
from between then and now.

It crashes in display_gdb_prompt, as called from
tui_command_loop.  Here's a backtrace.


#0  0x40106d33 in strlen () from /lib/tls/libc.so.6
#1  0x08114c6a in display_gdb_prompt (new_prompt=0x0)
     at /home/msnyder/src/gdb/event-top.c:280
#2  0x080bf1f5 in tui_command_loop (data=0x0)
     at /home/msnyder/src/gdb/tui/tui-interp.c:140
#3  0x0811168a in current_interp_command_loop ()
     at /home/msnyder/src/gdb/interps.c:277
#4  0x0807aafb in captured_command_loop (data=0x0)
     at /home/msnyder/src/gdb/main.c:97
#5  0x080801d9 in do_catch_errors (uiout=0x830b0c0, data=0xbfffbe18)
     at /home/msnyder/src/gdb/top.c:524
#6  0x0807ffdc in catcher (func=0x80801bd <do_catch_errors>,
     func_uiout=0x830b0c0, func_args=0xbfffbe18, func_val=0xbfffbe24,
     func_caught=0xbfffbe20, errstring=0x822cc48 "", gdberrmsg=0x0, mask=6)
     at /home/msnyder/src/gdb/top.c:431
#7  0x08080214 in catch_errors (func=0x807aaf0 <captured_command_loop>,
     func_args=0x0, errstring=0x822cc48 "", mask=6)
     at /home/msnyder/src/gdb/top.c:536
#8  0x0807b918 in captured_main (data=0xbfffc140)
     at /home/msnyder/src/gdb/main.c:805
#9  0x080801d9 in do_catch_errors (uiout=0x82b8220, data=0xbfffc0f8)
     at /home/msnyder/src/gdb/top.c:524
#10 0x0807ffdc in catcher (func=0x80801bd <do_catch_errors>,
     func_uiout=0x82b8220, func_args=0xbfffc0f8, func_val=0xbfffc104,
     func_caught=0xbfffc100, errstring=0x822cc48 "", gdberrmsg=0x0, mask=6)
     at /home/msnyder/src/gdb/top.c:431
#11 0x08080214 in catch_errors (func=0x807ab2e <captured_main>,
     func_args=0xbfffc140, errstring=0x822cc48 "", mask=6)
     at /home/msnyder/src/gdb/top.c:536
#12 0x0807b942 in gdb_main (args=0xbfffc140)
     at /home/msnyder/src/gdb/main.c:814
#13 0x0807aae9 in main (argc=2, argv=0xbfffc1e4)
     at /home/msnyder/src/gdb/gdb.c:35

Re: [HEADS-UP] gdb segfault

[Elena Zannoni]

Andrew Cagney writes:
 > > Hey yo --
 > > 
 > > gdb is seg-faulting immediately whenever its run with --noasync.
 > > I have an old build from the beginning of March that doesn't
 > > do it (nor do even older ones), so it's probably a change
 > > from between then and now.
 > > 
 > > It crashes in display_gdb_prompt, as called from
 > > tui_command_loop.  Here's a backtrace. 
 > 
 > Sounds like it's time to delete --noasync ....
 > 
 > 

I think so, yes.

Re: [HEADS-UP] gdb segfault

[Michael Snyder]

Elena Zannoni wrote:
> Andrew Cagney writes:
>  > > Hey yo --
>  > > 
>  > > gdb is seg-faulting immediately whenever its run with --noasync.
>  > > I have an old build from the beginning of March that doesn't
>  > > do it (nor do even older ones), so it's probably a change
>  > > from between then and now.
>  > > 
>  > > It crashes in display_gdb_prompt, as called from
>  > > tui_command_loop.  Here's a backtrace. 
>  > 
>  > Sounds like it's time to delete --noasync ....
> 
> I think so, yes.

It bothers me that ^C doesn't work with async, and I can't fall back
to noasync to make it work.  I think async should work (wrt. ^C, at
least), before we delete noasync.

Re: [HEADS-UP] gdb segfault

[Andrew Cagney]

[-- Attachment #1: Type: text/plain, Size: 1102 bytes --]

> Elena Zannoni wrote:
> 
>> Andrew Cagney writes:
>>  > > Hey yo --
>>  > >  > > gdb is seg-faulting immediately whenever its run with --noasync.
>>  > > I have an old build from the beginning of March that doesn't
>>  > > do it (nor do even older ones), so it's probably a change
>>  > > from between then and now.
>>  > >  > > It crashes in display_gdb_prompt, as called from
>>  > > tui_command_loop.  Here's a backtrace.  >  > Sounds like it's time to delete --noasync ....
>>
>> I think so, yes.
> 
> 
> It bothers me that ^C doesn't work with async, and I can't fall back
> to noasync to make it work.  I think async should work (wrt. ^C, at
> least), before we delete noasync.

Hmm, ^C with noasync doesn't work either, it dumps core ;-)

Lets all take responsibility for this problem and see through the proper 
fix:
http://sources.redhat.com/ml/gdb-patches/2004-05/msg00320.html
http://sources.redhat.com/ml/gdb-patches/2004-02/msg00530.html

I've attached is the first of two initial steps - disable the broken 
async code ready for its removal.  Then we can fix QUIT.

Eli, Elena?

Andrew


[-- Attachment #2: diffs --]
[-- Type: text/plain, Size: 4409 bytes --]

Index: doc/ChangeLog
2004-06-10  Andrew Cagney  <cagney@gnu.org>

	* gdb.texinfo (Mode Options): Delete documentation on "-async" and
	"-noasync".

2004-06-10  Andrew Cagney  <cagney@gnu.org>

	* defs.h (event_loop_p): Replace variable declaration with macro,
	always 1.
	* main.c (captured_main): Delete options "-async" and "-noasync".
	(event_loop_p): Delete variable.
	* NEWS: Mention that "-async" and "-noasync" were removed.

Index: NEWS
===================================================================
RCS file: /cvs/src/src/gdb/NEWS,v
retrieving revision 1.151
diff -p -u -r1.151 NEWS
--- NEWS	26 May 2004 06:32:01 -0000	1.151
+++ NEWS	10 Jun 2004 16:18:01 -0000
@@ -3,6 +3,11 @@
 
 *** Changes since GDB 6.1:
 
+* "-async" and "-noasync" options removed.
+
+Support for the broken "-noasync" option has been removed (it caused
+GDB to dump core).
+
 * New ``start'' command.
 
 This command runs the program until the begining of the main procedure.
Index: defs.h
===================================================================
RCS file: /cvs/src/src/gdb/defs.h,v
retrieving revision 1.146
diff -p -u -r1.146 defs.h
--- defs.h	8 May 2004 23:02:10 -0000	1.146
+++ defs.h	10 Jun 2004 16:18:01 -0000
@@ -1216,8 +1216,11 @@ struct target_waitstatus;
 struct cmd_list_element;
 
 /* Should the asynchronous variant of the interpreter (using the
-   event-loop) be enabled? */
-extern int event_loop_p;
+   event-loop) be enabled?  */
+/* NOTE: cagney/2004-06-10: GDB changed to always uses the event-loop.
+   A follow-on cleanup is to eliminate references to
+   "event_loop_p".  */
+#define event_loop_p 1
 
 extern void (*deprecated_init_ui_hook) (char *argv0);
 extern void (*deprecated_command_loop_hook) (void);
Index: main.c
===================================================================
RCS file: /cvs/src/src/gdb/main.c,v
retrieving revision 1.41
diff -p -u -r1.41 main.c
--- main.c	21 Apr 2004 23:52:20 -0000	1.41
+++ main.c	10 Jun 2004 16:18:01 -0000
@@ -49,12 +49,6 @@ int display_time;
 
 int display_space;
 
-/* Whether this is the async version or not.  The async version is
-   invoked on the command line with the -nw --async options.  In this
-   version, the usual command_loop is substituted by and event loop which
-   processes UI events asynchronously. */
-int event_loop_p = 1;
-
 /* The selected interpreter.  This will be used as a set command
    variable, so it should always be malloc'ed - since
    do_setshow_command will free it. */
@@ -253,8 +247,6 @@ captured_main (void *data)
     };
     static struct option long_options[] =
     {
-      {"async", no_argument, &event_loop_p, 1},
-      {"noasync", no_argument, &event_loop_p, 0},
 #if defined(TUI)
       {"tui", no_argument, 0, OPT_TUI},
 #endif
Index: doc/gdb.texinfo
===================================================================
RCS file: /cvs/src/src/gdb/doc/gdb.texinfo,v
retrieving revision 1.204
diff -p -u -r1.204 gdb.texinfo
--- doc/gdb.texinfo	24 May 2004 23:53:58 -0000	1.204
+++ doc/gdb.texinfo	10 Jun 2004 16:18:02 -0000
@@ -1065,30 +1065,6 @@ that control @value{GDBN}, and level 2 h
 The annotation mechanism has largely been superseeded by @sc{gdb/mi}
 (@pxref{GDB/MI}).
 
-@item -async
-@cindex @code{--async}
-Use the asynchronous event loop for the command-line interface.
-@value{GDBN} processes all events, such as user keyboard input, via a
-special event loop.  This allows @value{GDBN} to accept and process user
-commands in parallel with the debugged process being
-run@footnote{@value{GDBN} built with @sc{djgpp} tools for
-MS-DOS/MS-Windows supports this mode of operation, but the event loop is
-suspended when the debuggee runs.}, so you don't need to wait for
-control to return to @value{GDBN} before you type the next command.
-(@emph{Note:} as of version 5.1, the target side of the asynchronous
-operation is not yet in place, so @samp{-async} does not work fully
-yet.)
-@c FIXME: when the target side of the event loop is done, the above NOTE
-@c should be removed.
-
-When the standard input is connected to a terminal device, @value{GDBN}
-uses the asynchronous event loop by default, unless disabled by the
-@samp{-noasync} option.
-
-@item -noasync
-@cindex @code{--noasync}
-Disable the asynchronous event loop for the command-line interface.
-
 @item --args
 @cindex @code{--args}
 Change interpretation of command line so that arguments following the

Re: [HEADS-UP] gdb segfault

[Eli Zaretskii]

> Date: Thu, 10 Jun 2004 12:21:48 -0400
> From: Andrew Cagney <cagney@gnu.org>
> 
> I've attached is the first of two initial steps - disable the broken 
> async code ready for its removal.  Then we can fix QUIT.
> 
> Eli, Elena?

Okay with me.

Re: [HEADS-UP] gdb segfault

[Andrew Cagney]

> Index: doc/ChangeLog
> 2004-06-10  Andrew Cagney  <cagney@gnu.org>
> 
> 	* gdb.texinfo (Mode Options): Delete documentation on "-async" and
> 	"-noasync".
> 
> 2004-06-10  Andrew Cagney  <cagney@gnu.org>
> 
> 	* defs.h (event_loop_p): Replace variable declaration with macro,
> 	always 1.
> 	* main.c (captured_main): Delete options "-async" and "-noasync".
> 	(event_loop_p): Delete variable.
> 	* NEWS: Mention that "-async" and "-noasync" were removed.

Oops, dropped.  I've now checked this in (mainline and 6.2).  Oked by 
Eli on, and Elena off list.

Andrew