Joel Brobecker | 1 Jan 13:31 2005

Re: [RFA/pa-risc] Correct incorrect frame base

> On Fri, Dec 31, 2004 at 07:14:15PM +0400, Joel Brobecker wrote:
> > There is a small confusion in the source register for the addil
> > instruction. The actual sequence should be
> > 
> >             addil high21,%r30
> >             ldo low11(%r1),%r30
> > 
> > I think the confusion comes from the fact that the result of
> > the sum is stored in %r1.
> 
> "addil <number>,%r30" sets %r1 and not %r30?

Yes. "addil high21,%rn" adds high21 to register n, and then always
stores the results in r1.

> Whose bright idea was that, anyway?

:-)

> > 2004-12-31  Joel Brobecker  <brobecker <at> gnat.com>
> > 
> >         * hppa-tdep.c (prologue_inst_adjust_sp): Fix small confusion
> >         in register number for ldil instruction.
> > 
> > Tested on HP/UX 11.00, no regression.
> > OK to apply?
> 
> OK.  Typo in the changelog: that's addil, not ldil, isn't it?

Good catch, sorry for the confusion. Fixed and checked in.
(Continue reading)

Peter Stephens | 1 Jan 18:32 2005
Picon
Picon

Fix - remote target: Memory access error while loading section .rom_vectors

(bad programmer - spank, spank).  If you initialize the 'err' variable it
works OK.  This is probably why it has been such an intermittent problem.

Index: symfile.c
===================================================================
RCS file: /cvs/src/src/gdb/symfile.c,v
retrieving revision 1.144
diff -r1.144 symfile.c
1408c1408
<         int err;
---
>         int err=0;

Andreas Schwab | 2 Jan 23:01 2005
Picon

Don't error out when variable not available

With the introduction of DWARF location lists it became impossible to get
the full list of local variables of a function when any of them is not
available.  This is especially annoying when the first local or function
parameter is missing.  Here is an attempt at correcting this.  It is
suboptimal because it still prints the error message including the newline
which badly messes up the output.  Any idea how to fix that properly?

Andreas.

2005-01-02  Andreas Schwab  <schwab <at> suse.de>

	* findvar.c (read_var_value_no_error, read_var_value_no_error_1)
	(struct read_var_value_args): New.
	* value.h (read_var_value_no_error): Declare it.
	* printcmd.c (print_variable_value): Use read_var_value_no_error.
	* stack.c (print_frame_args): Likewise.

--- gdb/findvar.c.~1.82.~	2004-11-17 00:55:13.000000000 +0100
+++ gdb/findvar.c	2005-01-02 22:48:01.469889953 +0100
 <at>  <at>  -589,6 +589,39  <at>  <at>  addresses have not been bound by the dyn
   return v;
 }

+/* Like read_var_value, but catch errors and return NULL if any
+   occured.  */
+
+struct read_var_value_args
+{
+  struct symbol *var;
+  struct frame_info *frame;
(Continue reading)

Daniel Jacobowitz | 2 Jan 23:57 2005

Re: Don't error out when variable not available

On Sun, Jan 02, 2005 at 11:01:45PM +0100, Andreas Schwab wrote:
> With the introduction of DWARF location lists it became impossible to get
> the full list of local variables of a function when any of them is not
> available.  This is especially annoying when the first local or function
> parameter is missing.  Here is an attempt at correcting this.  It is
> suboptimal because it still prints the error message including the newline
> which badly messes up the output.  Any idea how to fix that properly?

How about reusing the LOC_OPTIMIZED_OUT handling to describe this case? 
I believe that will print "foo = <optimized out>".

--

-- 
Daniel Jacobowitz

Andreas Schwab | 3 Jan 00:49 2005
Picon

Re: Don't error out when variable not available

Daniel Jacobowitz <drow <at> false.org> writes:

> How about reusing the LOC_OPTIMIZED_OUT handling to describe this case? 
> I believe that will print "foo = <optimized out>".

LOC_OPTIMIZED_OUT is not a value, but a static property of a symbol.  I
don't think it fits into the dynamic nature of unavailable values.

Andreas.

--

-- 
Andreas Schwab, SuSE Labs, schwab <at> suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Daniel Jacobowitz | 3 Jan 01:49 2005

Re: Don't error out when variable not available

On Mon, Jan 03, 2005 at 12:49:04AM +0100, Andreas Schwab wrote:
> Daniel Jacobowitz <drow <at> false.org> writes:
> 
> > How about reusing the LOC_OPTIMIZED_OUT handling to describe this case? 
> > I believe that will print "foo = <optimized out>".
> 
> LOC_OPTIMIZED_OUT is not a value, but a static property of a symbol.  I
> don't think it fits into the dynamic nature of unavailable values.

However, we always handle it in the same case statements that handle
LOC_COMPUTED, which is how we get into the dwarf2 expression evaluator
anyway.

--

-- 
Daniel Jacobowitz

Andreas Schwab | 3 Jan 14:40 2005
Picon

Re: Don't error out when variable not available

How about this?

Andreas.

2005-01-03  Andreas Schwab  <schwab <at> suse.de>

	* stack.c (print_frame_args): Handle VALUE_OPTIMIZED_OUT.

	* dwarf2loc.c (loclist_read_variable): Return a value with
	VALUE_OPTIMIZED_OUT set instead of erring out when value not
	available.

Index: dwarf2loc.c
===================================================================
RCS file: /cvs/src/src/gdb/dwarf2loc.c,v
retrieving revision 1.18
diff -u -a -p -r1.18 dwarf2loc.c
--- dwarf2loc.c	9 Nov 2004 14:43:25 -0000	1.18
+++ dwarf2loc.c	3 Jan 2005 13:33:23 -0000
 <at>  <at>  -1,5 +1,5  <at>  <at> 
 /* DWARF 2 location expression support for GDB.
-   Copyright 2003 Free Software Foundation, Inc.
+   Copyright 2003, 2005 Free Software Foundation, Inc.
    Contributed by Daniel Jacobowitz, MontaVista Software, Inc.

    This file is part of GDB.
 <at>  <at>  -510,9 +510,14  <at>  <at>  loclist_read_variable (struct symbol *sy
   data = find_location_expression (dlbaton, &size,
 				   frame ? get_frame_pc (frame) : 0);
   if (data == NULL)
(Continue reading)

Andreas Schwab | 3 Jan 14:49 2005
Picon

Fix memory leak in break command

Found this by code inspection.

Andreas.

2005-01-03  Andreas Schwab  <schwab <at> suse.de>

	* breakpoint.c (break_command_1): Always free err_msg before
	returning when an exception was caught.

Index: gdb/breakpoint.c
===================================================================
RCS file: /cvs/src/src/gdb/breakpoint.c,v
retrieving revision 1.186
diff -u -p -a -u -p -a -r1.186 gdb/breakpoint.c
--- gdb/breakpoint.c	1 Dec 2004 06:54:56 -0000	1.186
+++ gdb/breakpoint.c	3 Jan 2005 13:45:20 -0000
 <at>  <at>  -5141,7 +5141,10  <at>  <at>  break_command_1 (char *arg, int flag, in
 	{
 	  /* If called to resolve pending breakpoint, just return error code.  */
 	  if (pending_bp)
-	    return rc;
+	    {
+	      xfree (err_msg);
+	      return rc;
+	    }

 	  error_output_message (NULL, err_msg);
 	  xfree (err_msg);
 <at>  <at>  -5169,7 +5172,10  <at>  <at>  break_command_1 (char *arg, int flag, in
 	  pending = 1;
(Continue reading)

Hans-Peter Nilsson | 3 Jan 16:30 2005
Picon

sim/cris maintainership and patch ping: pipe support and minor endian cleanup.

First of all, I'd like to renew my candidature (sp?); I
volunteer to be maintainer for the CRIS simulator (submitted,
not committed; approved by Orjan Friberg, GDB CRIS maintainer,
as redirected by Andrew Cagney).  I have pre-approval from Orjan
for all changes in sim/cris, but I think it's better to
explicitly be listed as sim maintainer.  The operative
difference would supposedly be that I can approve other people's
sim/cris changes directly (should there be any), instead of just
suggesting to Orjan that they're ok.  Andrew Cagney has
suggested I be a gdb maintainer for CRIS together with Orjan.
I'd be fine with that, but I'm not really up-to-date with GDB
and wouldn't anyway change anything non-sim-related without
getting approval from Orjan.

Over to the pending patches.

[RFA:] Minor sim callback endian cleanup.
<URL:http://sources.redhat.com/ml/gdb-patches/2004-12/msg00324.html>

sim/common: pipe syscall support
<URL:http://sources.redhat.com/ml/gdb-patches/2004-12/msg00378.html>

Ok to commit them?

brgds, H-P

Andreas Schwab | 3 Jan 16:33 2005
Picon

Parallel build fix

This fixes anonther parallel build issue.  Since device_table.h includes
the generated file hw.h, dependencies must be completely spelt out so that
hw.h is generated at the right time.  Committing as obvious.

Andreas.

2005-01-03  Andreas Schwab  <schwab <at> suse.de>

	* Makefile.in (hw_com.o, hw_eeprom.o): Depend on
	$(DEVICE_TABLE_H).

--- sim/ppc/Makefile.in.~1.18.~	2004-11-16 11:19:46.000000000 +0100
+++ sim/ppc/Makefile.in	2005-01-03 16:26:52.220425069 +0100
 <at>  <at>  -820,8 +820,10  <at>  <at>  tmp-hw: Makefile $(HW_SRC) $(srcdir)/../
 	touch tmp-hw

 hw_cpu.o: hw_cpu.c $(DEVICE_TABLE_H) $(HW_CPU_H) $(INTERRUPTS_H) $(CPU_H)
+hw_com.o: hw_com.c $(DEVICE_TABLE_H)
 hw_core.o: hw_core.c $(DEVICE_TABLE_H) $(COREFILE_H)
 hw_disk.o: hw_disk.c $(DEVICE_TABLE_H) $(PK_H)
+hw_eeprom.o: hw_eeprom.c $(DEVICE_TABLE_H)
 hw_glue.o: hw_glue.c $(DEVICE_TABLE_H)
 hw_htab.o: hw_htab.c $(DEVICE_TABLE_H) $(BFD_H)
 hw_ide.o: hw_ide.c $(DEVICE_TABLE_H)

--

-- 
Andreas Schwab, SuSE Labs, schwab <at> suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."
(Continue reading)


Gmane