Eli Zaretskii | 1 Jul 14:35 2006

Re: [commit] Fix MI prompt in manual (was: MI Doco)

> Date: Sat, 10 Jun 2006 10:32:32 +0300
> From: Eli Zaretskii <eliz <at> gnu.org>
> > Date: Fri, 9 Jun 2006 15:57:43 -0400
> > From: Daniel Jacobowitz <drow <at> false.org>
> > 
> > > Are you saying that all the MI examples that use  <at> value{GDBP} are
> > > incorrect?
> > 
> > Yes, I'm afraid so.
> I guess we will need to fix that ASAP.

Done with the following patch:

2006-07-01  Eli Zaretskii  <eliz <at> gnu.org>

	* gdb.texinfo (GDB/MI Output Syntax, GDB/MI Simple Examples)
	(GDB/MI Breakpoint Commands, GDB/MI Program Context)
	(GDB/MI Thread Commands, GDB/MI Program Execution)
	(GDB/MI Stack Manipulation, GDB/MI Variable Objects)
	(GDB/MI Data Manipulation, GDB/MI Symbol Query)
	(GDB/MI File Commands, GDB/MI Target Manipulation)
	(GDB/MI Miscellaneous Commands): Change ( <at> value{GDBP})->(gdb),
	since the MI prompt is hard-wired into the code.

Index: gdb.texinfo
RCS file: /cvs/src/src/gdb/doc/gdb.texinfo,v
retrieving revision 1.341
(Continue reading)

Nathan J. Williams | 1 Jul 18:37 2006

[RFA] bsd-kvm.c: open read-write when write_files is set.

This should be a pretty minor approval; it struck me as a bit shy of
"obvious", though.

        - Nathan

2006-07-01  Nathan J. Williams  <nathanw <at> wasabisystems.com>

	* bsd-kvm.c (bsd_kvm_open): Open the KVM interface read-write if
	write_files is set.

Index: bsd-kvm.c
RCS file: /cvs/src/src/gdb/bsd-kvm.c,v
retrieving revision 1.16
diff -u -r1.16 bsd-kvm.c
--- bsd-kvm.c	17 Dec 2005 22:33:59 -0000	1.16
+++ bsd-kvm.c	1 Jul 2006 16:37:24 -0000
 <at>  <at>  -81,7 +81,8  <at>  <at> 

   execfile = get_exec_file (0);
-  temp_kd = kvm_openfiles (execfile, filename, NULL, O_RDONLY, errbuf);
+  temp_kd = kvm_openfiles (execfile, filename, NULL,
+			   write_files ? O_RDWR : O_RDONLY, errbuf);
   if (temp_kd == NULL)
     error (("%s"), errbuf);

Mark Kettenis | 1 Jul 19:06 2006

Re: [RFA] bsd-kvm.c: open read-write when write_files is set.

> From: "Nathan J. Williams" <nathanw <at> wasabisystems.com>
> Date: 01 Jul 2006 12:37:45 -0400
> This should be a pretty minor approval; it struck me as a bit shy of
> "obvious", though.

Looks sane to me.  Suppose this allows you to do very scary things ;-).

> 2006-07-01  Nathan J. Williams  <nathanw <at> wasabisystems.com>
> 	* bsd-kvm.c (bsd_kvm_open): Open the KVM interface read-write if
> 	write_files is set.

Daniel Jacobowitz | 3 Jul 17:48 2006

Re: cleanup some tests

On Fri, Jun 09, 2006 at 09:48:33PM +0100, Nathan Sidwell wrote:
> This patch cleans up a number of spurious test failures I encountered
> When using a remote stub to talk to a board, I found it necessary to run to
> main, in order to syncronize the debugger correctly.

This change is wrong.  After the patch:

gdb_test "break main" ".*Breakpoint.* at .*" "set breakpoint at main"

# initial sal should be first statement in main
gdb_test "list" \
    "v0 = 0;" \
    "list before run"

You're no longer testing list-before-run.  Probably the trouble comes
from the fact that gdb_load starts the program running on remote
targets.  I bet that what you want is to replace the gdb_load with
gdb_file_cmd, and then do a gdb_load before the runto_main.

Secondly, gdb_run_cmd starts the program running.  If you want to issue
another command, you've got to wait for it to stop first.  Which boils
down to runto_main, already used just below.

> The remote.exp changes are obvious, when one looks at gdb's source.

No it's not :-)  I bet you made this patch against the csl branch.  I
hacked download_write_size on that branch; Vlad's been looking at this
a little closer.  512 is right for mainline.
(Continue reading)

Nick Roberts | 4 Jul 01:22 2006

[PATCH] Re-instating an annotation.

Having removed many of the annotations in Level 3, I'd like to put one back.
Just in case anyone is interested, one reason I want it back because for some
reason GDB prints out the frame details _before_ the stopped annotation.

I would really like to call the "--annotate=3" option "--emacs" to discourage
others from using it.  In any case adding one annotation shouldn't break
a front end because it should just ignore it.

OK to commit?


Nick                                           http://www.inet.net.nz/~nickrob

2006-07-04  Nick Roberts  <nickrob <at> snap.net.nz>

	* annotate.c (annotate_frame_begin): Re-instate frame-begin
	annotation for level 3 annotations.

--- annotate.c	20 Jun 2006 08:48:51 +1200	
*************** annotate_source (char *filename, int lin
*** 430,436 ****
  annotate_frame_begin (int level, CORE_ADDR pc)
!   if (annotation_level == 2)
        printf_filtered (("\n\032\032frame-begin %d 0x"), level);
        deprecated_print_address_numeric (pc, 0, gdb_stdout);
--- 430,436 ----
(Continue reading)

Nathan Sidwell | 4 Jul 11:41 2006

Re: cleanup some tests

Daniel Jacobowitz wrote:

> Want to post and commit the approved bits, and then we can look at the
> rest?

I've committed this patch for the approved bits.  I'll investigate the other 
points and repost once I have a solution.


Nathan Sidwell    ::   http://www.codesourcery.com   ::         CodeSourcery
nathan <at> codesourcery.com    ::     http://www.planetfall.pwp.blueyonder.co.uk

2006-07-04  Nathan Sidwell  <nathan <at> codesourcery.com>

	* gdb.base/huge.c: Allow CRASH_GDB to be set from command
	* gdb.base/huge.exp: Loop over compilation to find a size that
	is acceptable.
	* gdb.base/remote.exp: Correct expected strings.
	* gdb.dwarf2/mac-fileno.S: Avoid # comments.

Index: testsuite/gdb.base/huge.c
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/huge.c,v
retrieving revision 1.3
diff -c -3 -p -r1.3 huge.c
(Continue reading)

Daniel Jacobowitz | 4 Jul 15:32 2006

[ob] Fix a memory leak from my qSupported patch

Noticed while merging the patch to another branch.  Tested, committed as

Daniel Jacobowitz

2006-07-04  Daniel Jacobowitz  <dan <at> codesourcery.com>

	* remote.c (init_remote_state): Use xrealloc instead of xmalloc.

Index: gdb/remote.c
RCS file: /cvs/src/src/gdb/remote.c,v
retrieving revision 1.217
diff -u -p -r1.217 remote.c
--- gdb/remote.c	22 Jun 2006 14:06:37 -0000	1.217
+++ gdb/remote.c	4 Jul 2006 13:31:10 -0000
 <at>  <at>  -334,7 +334,7  <at>  <at>  init_remote_state (struct gdbarch *gdbar
   if (rs->buf_size < rsa->remote_packet_size)
       rs->buf_size = 2 * rsa->remote_packet_size;
-      rs->buf = xmalloc (rs->buf_size);
+      rs->buf = xrealloc (rs->buf, rs->buf_size);

   return rsa;

Daniel Jacobowitz | 5 Jul 20:40 2006

[rfa] ARM/Linux signal handling update

The latest kernel.org kernels use new layouts for signal frames.
This updates GDB to cope with both the old and new layouts.

There's actually more useful information in the new layouts: iWMMXt
registers and eventually Crunch/VFP registers can now be saved on the stack
in a predictable location.  I didn't implement GDB support for that yet.
I hope I'll be back to it later.

Look OK?

Daniel Jacobowitz

2006-07-05  Daniel Jacobowitz  <dan <at> codesourcery.com>

	* arm-linux-tdep.c: Doc fixes.
	(arm_linux_sigreturn_init, arm_linux_rt_sigreturn_init): Support
	the new signal frame layouts.

Index: src/gdb/arm-linux-tdep.c
--- src.orig/gdb/arm-linux-tdep.c	2006-06-20 15:35:06.000000000 -0400
+++ src/gdb/arm-linux-tdep.c	2006-06-24 17:13:54.000000000 -0400
 <at>  <at>  -227,9 +227,10  <at>  <at>  arm_linux_extract_return_value (struct t
 #define ARM_LINUX_SIGRETURN_INSTR	0xef900077
(Continue reading)

Daniel Jacobowitz | 5 Jul 21:06 2006

Re: [rfc] Correct semantics of target_read_partial, add target_read_whole

Here is an improved version of my previous patch.  According to
suggestions I received, I've renamed target_read_whole to the hopefully
clearer target_read_object_alloc (since it reads an entire object, and
allocates memory for it).  I've eliminated the extra use of
target_read; the code to use target_read_partial directly is slightly
more complicated, but more efficient and overall more sensible.  And:
the _partial methods are no longer exported!  Roughly no one ever wants
to use them; use target_read or target_read_object_alloc instead.

Any further comments on this patch?


Daniel Jacobowitz

2006-07-05  Daniel Jacobowitz  <dan <at> codesourcery.com>

	* target.c (target_read): Stop if target_read_partial returns 0
	when some bytes have already been read.
	(target_write): Likewise for target_write_partial.
	(target_read_partial, target_write_partial): Make static.
	(target_read_object_alloc): New.
	* target.h: Doc fixes.
	(target_read_partial, target_write_partial): Delete prototypes.
	(target_read_object_alloc): New prototype.

	* auxv.c (target_auxv_read): Delete.
	(target_auxv_search, fprint_target_auxv): Use target_read_object_alloc.
	* auxv.h (target_auxv_read): Delete prototype.
	* avr-tdep.c (avr_io_reg_read_command): Use target_read_object_alloc.
(Continue reading)

Daniel Jacobowitz | 5 Jul 21:06 2006

[commit] Remove KOD

Back in January, I got a consensus that Kernel Object Display should be
removed.  This patch does so.  While undoubtedly a useful feature at the
time, there are no longer any targets available which implement it, and
the protocol was undocumented.

I've built this and checked it in.

Daniel Jacobowitz

2006-07-05  Daniel Jacobowitz  <dan <at> codesourcery.com>

	* remote.c (remote_xfer_partial): Remove KOD support.
	* target.h (enum target_object): Remove TARGET_OBJECT_KOD.
	* MAINTAINERS: Move Kernel Object Display entry to past maintainers.
	* Makefile.in (SFILES, kod_h, COMMON_OBS, kod.o, kod-cisco.o): Remove KOD
	* mi/mi-cmds.c (mi_cmds): Remove dummy KOD commands.
	* NEWS: Mention KOD.
	* kod.h, kod.c, kod-cisco.c: Delete files.

2006-07-05  Daniel Jacobowitz  <dan <at> codesourcery.com>

	* doc/gdb.texinfo (KOD): Remove node.
	(GDB/MI Kod Commands): Remove commented out node.

Index: src/gdb/remote.c
--- src.orig/gdb/remote.c	2006-07-05 11:34:35.000000000 -0400
(Continue reading)