record mode supposed to work on x86_64-linux?

record mode supposed to work on x86_64-linux?

[Andi Kleen]

Hi,

I was trying the new record/replay mode in CVS. Unfortunately 
it doesn't work for me on x86-64-linux. I always get SIGABRTs during
execution.  Is this supposed to work already?

Thanks,
-Andi

GNU gdb (GDB) 6.8.50.20090515-cvs
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-unknown-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
(gdb) b main
Breakpoint 1 at 0x400544: file tloop2.c, line 5.
(gdb) run
Starting program: /home2/andi/tsrc/tloop2 

Breakpoint 1, main () at tloop2.c:5
5               for (i = 0; i < 10; i++)
(gdb) record
(gdb) n
Process record: failed to record execution log.

Program received signal SIGABRT, Aborted.
main () at tloop2.c:5
5               for (i = 0; i < 10; i++)
(gdb) 
Process record: failed to record execution log.

Program received signal SIGABRT, Aborted.
main () at tloop2.c:5
5               for (i = 0; i < 10; i++)
(gdb) 

-- 
ak@linux.intel.com -- Speaking for myself only.

Re: record mode supposed to work on x86_64-linux?

[Hui Zhu]

No.  Prec just support x86 now.
I am working on x86-64 support.

Thanks,
Hui

On Sat, May 16, 2009 at 02:40, Andi Kleen <andi@firstfloor.org> wrote:
>
> Hi,
>
> I was trying the new record/replay mode in CVS. Unfortunately
> it doesn't work for me on x86-64-linux. I always get SIGABRTs during
> execution.  Is this supposed to work already?
>
> Thanks,
> -Andi
>
> GNU gdb (GDB) 6.8.50.20090515-cvs
> Copyright (C) 2009 Free Software Foundation, Inc.
> License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
> This is free software: you are free to change and redistribute it.
> There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
> and "show warranty" for details.
> This GDB was configured as "x86_64-unknown-linux-gnu".
> For bug reporting instructions, please see:
> <http://www.gnu.org/software/gdb/bugs/>...
> (gdb) b main
> Breakpoint 1 at 0x400544: file tloop2.c, line 5.
> (gdb) run
> Starting program: /home2/andi/tsrc/tloop2
>
> Breakpoint 1, main () at tloop2.c:5
> 5               for (i = 0; i < 10; i++)
> (gdb) record
> (gdb) n
> Process record: failed to record execution log.
>
> Program received signal SIGABRT, Aborted.
> main () at tloop2.c:5
> 5               for (i = 0; i < 10; i++)
> (gdb)
> Process record: failed to record execution log.
>
> Program received signal SIGABRT, Aborted.
> main () at tloop2.c:5
> 5               for (i = 0; i < 10; i++)
> (gdb)
>
> --
> ak@linux.intel.com -- Speaking for myself only.

Re: record mode supposed to work on x86_64-linux?

[Andi Kleen]

On Sat, May 16, 2009 at 09:48:09AM +0800, Hui Zhu wrote:
> No.  Prec just support x86 now.
> I am working on x86-64 support.

Ok. But should it give a error message then when the target is not
supported?

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only.

Re: record mode supposed to work on x86_64-linux?

[Hui Zhu]

[-- Attachment #1: Type: text/plain, Size: 515 bytes --]

Hi Andi,

Thanks for remind me.
I make a patch for it.  But I don't have a x86-64 close to me now.
Could you help me test it?

Thanks,
Hui

On Sat, May 16, 2009 at 16:10, Andi Kleen <andi@firstfloor.org> wrote:
> On Sat, May 16, 2009 at 09:48:09AM +0800, Hui Zhu wrote:
>> No.  Prec just support x86 now.
>> I am working on x86-64 support.
>
> Ok. But should it give a error message then when the target is not
> supported?
>
> -Andi
>
> --
> ak@linux.intel.com -- Speaking for myself only.
>

[-- Attachment #2: prec_target_check.txt --]
[-- Type: text/plain, Size: 5223 bytes --]

---
 gdbarch.c         |   25 +++++++++++++++++++++++++
 gdbarch.h         |    6 ++++++
 gdbarch.sh        |    4 ++++
 i386-linux-tdep.c |    2 ++
 record.c          |    6 +++---
 5 files changed, 40 insertions(+), 3 deletions(-)

--- a/gdbarch.c
+++ b/gdbarch.c
@@ -240,6 +240,7 @@ struct gdbarch
   gdbarch_static_transform_name_ftype *static_transform_name;
   int sofun_address_maybe_missing;
   gdbarch_process_record_ftype *process_record;
+  int support_process_record;
   gdbarch_target_signal_from_host_ftype *target_signal_from_host;
   gdbarch_target_signal_to_host_ftype *target_signal_to_host;
   gdbarch_get_siginfo_type_ftype *get_siginfo_type;
@@ -376,6 +377,7 @@ struct gdbarch startup_gdbarch =
   0,  /* static_transform_name */
   0,  /* sofun_address_maybe_missing */
   0,  /* process_record */
+  0,  /* support_process_record */
   default_target_signal_from_host,  /* target_signal_from_host */
   default_target_signal_to_host,  /* target_signal_to_host */
   0,  /* get_siginfo_type */
@@ -633,6 +635,8 @@ verify_gdbarch (struct gdbarch *gdbarch)
   /* Skip verify of static_transform_name, has predicate */
   /* Skip verify of sofun_address_maybe_missing, invalid_p == 0 */
   /* Skip verify of process_record, has predicate */
+  if (gdbarch->support_process_record == 0)
+    fprintf_unfiltered (log, "\n\tsupport_process_record");
   /* Skip verify of target_signal_from_host, invalid_p == 0 */
   /* Skip verify of target_signal_to_host, invalid_p == 0 */
   /* Skip verify of get_siginfo_type, has predicate */
@@ -1090,6 +1094,9 @@ gdbarch_dump (struct gdbarch *gdbarch, s
                       "gdbarch_dump: static_transform_name = <%s>\n",
                       host_address_to_string (gdbarch->static_transform_name));
   fprintf_unfiltered (file,
+                      "gdbarch_dump: support_process_record = %s\n",
+                      plongest (gdbarch->support_process_record));
+  fprintf_unfiltered (file,
                       "gdbarch_dump: target_desc = %s\n",
                       host_address_to_string (gdbarch->target_desc));
   fprintf_unfiltered (file,
@@ -3283,6 +3290,24 @@ set_gdbarch_process_record (struct gdbar
   gdbarch->process_record = process_record;
 }
 
+int
+gdbarch_support_process_record (struct gdbarch *gdbarch)
+{
+  gdb_assert (gdbarch != NULL);
+  /* Check variable changed from pre-default.  */
+  gdb_assert (gdbarch->support_process_record != 0);
+  if (gdbarch_debug >= 2)
+    fprintf_unfiltered (gdb_stdlog, "gdbarch_support_process_record called\n");
+  return gdbarch->support_process_record;
+}
+
+void
+set_gdbarch_support_process_record (struct gdbarch *gdbarch,
+                                    int support_process_record)
+{
+  gdbarch->support_process_record = support_process_record;
+}
+
 enum target_signal
 gdbarch_target_signal_from_host (struct gdbarch *gdbarch, int signo)
 {
--- a/gdbarch.h
+++ b/gdbarch.h
@@ -829,6 +829,12 @@ typedef int (gdbarch_process_record_ftyp
 extern int gdbarch_process_record (struct gdbarch *gdbarch, struct regcache *regcache, CORE_ADDR addr);
 extern void set_gdbarch_process_record (struct gdbarch *gdbarch, gdbarch_process_record_ftype *process_record);
 
+/* If the target of architecture support process record and replay, This
+   variable will be set to true. */
+
+extern int gdbarch_support_process_record (struct gdbarch *gdbarch);
+extern void set_gdbarch_support_process_record (struct gdbarch *gdbarch, int support_process_record);
+
 /* Signal translation: translate inferior's signal (host's) number into
    GDB's representation. */
 
--- a/gdbarch.sh
+++ b/gdbarch.sh
@@ -715,6 +715,10 @@ v:int:sofun_address_maybe_missing:::0:0:
 # Return -1 if something goes wrong, 0 otherwise.
 M:int:process_record:struct regcache *regcache, CORE_ADDR addr:regcache, addr
 
+# If the target of architecture support process record and replay, This
+# variable will be set to true.
+v:int:support_process_record:::0:0
+
 # Signal translation: translate inferior's signal (host's) number into
 # GDB's representation.
 m:enum target_signal:target_signal_from_host:int signo:signo::default_target_signal_from_host::0
--- a/i386-linux-tdep.c
+++ b/i386-linux-tdep.c
@@ -613,6 +613,8 @@ i386_linux_init_abi (struct gdbarch_info
   tdep->sc_reg_offset = i386_linux_sc_reg_offset;
   tdep->sc_num_regs = ARRAY_SIZE (i386_linux_sc_reg_offset);
 
+  set_gdbarch_support_process_record (gdbarch, 1);
+
   /* Initialize the i386_linux_record_tdep.  */
   i386_linux_record_tdep.size__old_kernel_stat =
     I386_LINUX_RECORD_SIZE__old_kernel_stat;
--- a/record.c
+++ b/record.c
@@ -427,9 +427,9 @@ record_open (char *name, int from_tty)
     error (_("Process record target can't debug inferior in asynchronous "
 	     "mode (target-async)."));
 
-  if (!gdbarch_process_record_p (current_gdbarch))
-    error (_("Process record: the current architecture doesn't support "
-	     "record function."));
+  if (!gdbarch_support_process_record (current_gdbarch))
+    error (_("Process record: the current target of architecture doesn't "
+             "support record function."));
 
   /* Check if record target is already running.  */
   if (current_target.to_stratum == record_stratum)

Re: record mode supposed to work on x86_64-linux?

[Andi Kleen]

On Sat, May 16, 2009 at 05:09:12PM +0800, Hui Zhu wrote:
> Hi Andi,
> 
> Thanks for remind me.
> I make a patch for it.  But I don't have a x86-64 close to me now.
> Could you help me test it?

Gets

../../src/gdb/gdbarch.c:648: internal-error: verify_gdbarch: the following are invalid ...
        support_process_record
A problem internal to GDB has been detected,
further debugging may prove unreliable.

at startup. I haven't tried further.

-Andi

Re: record mode supposed to work on x86_64-linux?

[Hui Zhu]

[-- Attachment #1: Type: text/plain, Size: 647 bytes --]

Sorry.  I made a mistake in gdbarch.sh.
How about this one?

Thanks,
Hui

On Sat, May 16, 2009 at 17:27, Andi Kleen <andi@firstfloor.org> wrote:
> On Sat, May 16, 2009 at 05:09:12PM +0800, Hui Zhu wrote:
>> Hi Andi,
>>
>> Thanks for remind me.
>> I make a patch for it.  But I don't have a x86-64 close to me now.
>> Could you help me test it?
>
> Gets
>
> ../../src/gdb/gdbarch.c:648: internal-error: verify_gdbarch: the following are invalid ...
>        support_process_record
> A problem internal to GDB has been detected,
> further debugging may prove unreliable.
>
> at startup. I haven't tried further.
>
> -Andi
>

[-- Attachment #2: prec_target_check.txt --]
[-- Type: text/plain, Size: 5139 bytes --]

---
 gdbarch.c         |   23 +++++++++++++++++++++++
 gdbarch.h         |    6 ++++++
 gdbarch.sh        |    4 ++++
 i386-linux-tdep.c |    2 ++
 record.c          |    6 +++---
 5 files changed, 38 insertions(+), 3 deletions(-)

--- a/gdbarch.c
+++ b/gdbarch.c
@@ -240,6 +240,7 @@ struct gdbarch
   gdbarch_static_transform_name_ftype *static_transform_name;
   int sofun_address_maybe_missing;
   gdbarch_process_record_ftype *process_record;
+  int support_process_record;
   gdbarch_target_signal_from_host_ftype *target_signal_from_host;
   gdbarch_target_signal_to_host_ftype *target_signal_to_host;
   gdbarch_get_siginfo_type_ftype *get_siginfo_type;
@@ -376,6 +377,7 @@ struct gdbarch startup_gdbarch =
   0,  /* static_transform_name */
   0,  /* sofun_address_maybe_missing */
   0,  /* process_record */
+  0,  /* support_process_record */
   default_target_signal_from_host,  /* target_signal_from_host */
   default_target_signal_to_host,  /* target_signal_to_host */
   0,  /* get_siginfo_type */
@@ -633,6 +635,7 @@ verify_gdbarch (struct gdbarch *gdbarch)
   /* Skip verify of static_transform_name, has predicate */
   /* Skip verify of sofun_address_maybe_missing, invalid_p == 0 */
   /* Skip verify of process_record, has predicate */
+  /* Skip verify of support_process_record, invalid_p == 0 */
   /* Skip verify of target_signal_from_host, invalid_p == 0 */
   /* Skip verify of target_signal_to_host, invalid_p == 0 */
   /* Skip verify of get_siginfo_type, has predicate */
@@ -1090,6 +1093,9 @@ gdbarch_dump (struct gdbarch *gdbarch, s
                       "gdbarch_dump: static_transform_name = <%s>\n",
                       host_address_to_string (gdbarch->static_transform_name));
   fprintf_unfiltered (file,
+                      "gdbarch_dump: support_process_record = %s\n",
+                      plongest (gdbarch->support_process_record));
+  fprintf_unfiltered (file,
                       "gdbarch_dump: target_desc = %s\n",
                       host_address_to_string (gdbarch->target_desc));
   fprintf_unfiltered (file,
@@ -3283,6 +3289,23 @@ set_gdbarch_process_record (struct gdbar
   gdbarch->process_record = process_record;
 }
 
+int
+gdbarch_support_process_record (struct gdbarch *gdbarch)
+{
+  gdb_assert (gdbarch != NULL);
+  /* Skip verify of support_process_record, invalid_p == 0 */
+  if (gdbarch_debug >= 2)
+    fprintf_unfiltered (gdb_stdlog, "gdbarch_support_process_record called\n");
+  return gdbarch->support_process_record;
+}
+
+void
+set_gdbarch_support_process_record (struct gdbarch *gdbarch,
+                                    int support_process_record)
+{
+  gdbarch->support_process_record = support_process_record;
+}
+
 enum target_signal
 gdbarch_target_signal_from_host (struct gdbarch *gdbarch, int signo)
 {
--- a/gdbarch.h
+++ b/gdbarch.h
@@ -829,6 +829,12 @@ typedef int (gdbarch_process_record_ftyp
 extern int gdbarch_process_record (struct gdbarch *gdbarch, struct regcache *regcache, CORE_ADDR addr);
 extern void set_gdbarch_process_record (struct gdbarch *gdbarch, gdbarch_process_record_ftype *process_record);
 
+/* If the target of architecture support process record and replay, This
+   variable will be set to true. */
+
+extern int gdbarch_support_process_record (struct gdbarch *gdbarch);
+extern void set_gdbarch_support_process_record (struct gdbarch *gdbarch, int support_process_record);
+
 /* Signal translation: translate inferior's signal (host's) number into
    GDB's representation. */
 
--- a/gdbarch.sh
+++ b/gdbarch.sh
@@ -715,6 +715,10 @@ v:int:sofun_address_maybe_missing:::0:0:
 # Return -1 if something goes wrong, 0 otherwise.
 M:int:process_record:struct regcache *regcache, CORE_ADDR addr:regcache, addr
 
+# If the target of architecture support process record and replay, This
+# variable will be set to true.
+v:int:support_process_record:::0:0::0
+
 # Signal translation: translate inferior's signal (host's) number into
 # GDB's representation.
 m:enum target_signal:target_signal_from_host:int signo:signo::default_target_signal_from_host::0
--- a/i386-linux-tdep.c
+++ b/i386-linux-tdep.c
@@ -613,6 +613,8 @@ i386_linux_init_abi (struct gdbarch_info
   tdep->sc_reg_offset = i386_linux_sc_reg_offset;
   tdep->sc_num_regs = ARRAY_SIZE (i386_linux_sc_reg_offset);
 
+  set_gdbarch_support_process_record (gdbarch, 1);
+
   /* Initialize the i386_linux_record_tdep.  */
   i386_linux_record_tdep.size__old_kernel_stat =
     I386_LINUX_RECORD_SIZE__old_kernel_stat;
--- a/record.c
+++ b/record.c
@@ -427,9 +427,9 @@ record_open (char *name, int from_tty)
     error (_("Process record target can't debug inferior in asynchronous "
 	     "mode (target-async)."));
 
-  if (!gdbarch_process_record_p (current_gdbarch))
-    error (_("Process record: the current architecture doesn't support "
-	     "record function."));
+  if (!gdbarch_support_process_record (current_gdbarch))
+    error (_("Process record: the current target of architecture doesn't "
+             "support record function."));
 
   /* Check if record target is already running.  */
   if (current_target.to_stratum == record_stratum)

Re: record mode supposed to work on x86_64-linux?

[Andi Kleen]

On Sat, May 16, 2009 at 05:51:42PM +0800, Hui Zhu wrote:
> Sorry.  I made a mistake in gdbarch.sh.
> How about this one?

Works -- i get a error message. Thanks.

The english of the message could be probably improved:

"Process recording: The current architecture target does not support recording functionality"

-Andi

Re: record mode supposed to work on x86_64-linux?

[Pedro Alves]

On Saturday 16 May 2009 10:51:42, Hui Zhu wrote:
> Sorry.  I made a mistake in gdbarch.sh.
> How about this one?

There's no need for a new gdbarch method if you
implement what I suggested before:

 http://sourceware.org/ml/gdb-patches/2009-05/msg00267.html

> 
> Thanks,
> Hui
> 
> On Sat, May 16, 2009 at 17:27, Andi Kleen <andi@firstfloor.org> wrote:
> > On Sat, May 16, 2009 at 05:09:12PM +0800, Hui Zhu wrote:
> >> Hi Andi,
> >>
> >> Thanks for remind me.
> >> I make a patch for it.  But I don't have a x86-64 close to me now.
> >> Could you help me test it?
> >
> > Gets
> >
> > ../../src/gdb/gdbarch.c:648: internal-error: verify_gdbarch: the following are invalid ...
> >        support_process_record
> > A problem internal to GDB has been detected,
> > further debugging may prove unreliable.
> >
> > at startup. I haven't tried further.
> >
> > -Andi
> >
> 



-- 
Pedro Alves

Re: record mode supposed to work on x86_64-linux?

[Hui Zhu]

[-- Attachment #1: Type: text/plain, Size: 1474 bytes --]

Hi Pedro,

What do you think about this one?

2009-05-17  Hui Zhu  <teawater@gmail.com>
	
	* i386-tdep.c (i386_process_record): Remove the static of it.
	(i386_gdbarch_init): Remove the set_gdbarch_process_record.
	* i386-tdep.c (i386_process_record): New extern.
	* i386-linux-tdep.c (i386_linux_init_abi): Call
	set_gdbarch_process_record.

Sorry for my always understand your mean so slow.  :)

Thanks,
Hui

On Sun, May 17, 2009 at 00:35, Pedro Alves <pedro@codesourcery.com> wrote:
> On Saturday 16 May 2009 10:51:42, Hui Zhu wrote:
>> Sorry.  I made a mistake in gdbarch.sh.
>> How about this one?
>
> There's no need for a new gdbarch method if you
> implement what I suggested before:
>
>  http://sourceware.org/ml/gdb-patches/2009-05/msg00267.html
>
>>
>> Thanks,
>> Hui
>>
>> On Sat, May 16, 2009 at 17:27, Andi Kleen <andi@firstfloor.org> wrote:
>> > On Sat, May 16, 2009 at 05:09:12PM +0800, Hui Zhu wrote:
>> >> Hi Andi,
>> >>
>> >> Thanks for remind me.
>> >> I make a patch for it.  But I don't have a x86-64 close to me now.
>> >> Could you help me test it?
>> >
>> > Gets
>> >
>> > ../../src/gdb/gdbarch.c:648: internal-error: verify_gdbarch: the following are invalid ...
>> >        support_process_record
>> > A problem internal to GDB has been detected,
>> > further debugging may prove unreliable.
>> >
>> > at startup. I haven't tried further.
>> >
>> > -Andi
>> >
>>
>
>
>
> --
> Pedro Alves
>

[-- Attachment #2: prec_move_to_target.txt --]
[-- Type: text/plain, Size: 1580 bytes --]

---
 i386-linux-tdep.c |    2 ++
 i386-tdep.c       |    4 +---
 i386-tdep.h       |    3 +++
 3 files changed, 6 insertions(+), 3 deletions(-)

--- a/i386-linux-tdep.c
+++ b/i386-linux-tdep.c
@@ -613,6 +613,8 @@ i386_linux_init_abi (struct gdbarch_info
   tdep->sc_reg_offset = i386_linux_sc_reg_offset;
   tdep->sc_num_regs = ARRAY_SIZE (i386_linux_sc_reg_offset);
 
+  set_gdbarch_process_record (gdbarch, i386_process_record);
+
   /* Initialize the i386_linux_record_tdep.  */
   i386_linux_record_tdep.size__old_kernel_stat =
     I386_LINUX_RECORD_SIZE__old_kernel_stat;
--- a/i386-tdep.c
+++ b/i386-tdep.c
@@ -2982,7 +2982,7 @@ i386_record_lea_modrm (struct i386_recor
    memory that will be changed in current instruction to "record_arch_list".
    Return -1 if something wrong. */
 
-static int
+int
 i386_process_record (struct gdbarch *gdbarch, struct regcache *regcache,
 		     CORE_ADDR addr)
 {
@@ -5277,8 +5277,6 @@ i386_gdbarch_init (struct gdbarch_info i
   set_gdbarch_skip_permanent_breakpoint (gdbarch,
 					 i386_skip_permanent_breakpoint);
 
-  set_gdbarch_process_record (gdbarch, i386_process_record);
-
   return gdbarch;
 }
 
--- a/i386-tdep.h
+++ b/i386-tdep.h
@@ -220,6 +220,9 @@ extern void i386_elf_init_abi (struct gd
 
 /* Initialize a SVR4 architecture variant.  */
 extern void i386_svr4_init_abi (struct gdbarch_info, struct gdbarch *);
+
+extern int i386_process_record (struct gdbarch *gdbarch,
+                                struct regcache *regcache, CORE_ADDR addr);
 \f
 
 /* Functions and variables exported from i386bsd-tdep.c.  */

Re: record mode supposed to work on x86_64-linux?

[Pedro Alves]

On Sunday 17 May 2009 01:15:00, Hui Zhu wrote:
> Hi Pedro,
> 
> What do you think about this one?

I think it's good, thanks.

> 2009-05-17  Hui Zhu  <teawater@gmail.com>
>         
>         * i386-tdep.c (i386_process_record): Remove the static of it.
>         (i386_gdbarch_init): Remove the set_gdbarch_process_record.
>         * i386-tdep.c (i386_process_record): New extern.
>         * i386-linux-tdep.c (i386_linux_init_abi): Call
>         set_gdbarch_process_record.

Please consider the alternative ChangeLog suggestions below instead.

> 	* i386-tdep.c (i386_process_record): Remove the static of it.

 	* i386-tdep.c (i386_process_record): Make it extern.

> 	(i386_gdbarch_init): Remove the set_gdbarch_process_record.

From this entry, it isn't clear what you're removing.  Actually, it
sounds like you're removing the function implementation.  Consider
instead:

	(i386_gdbarch_init): Remove the set_gdbarch_process_record call.
                                                                   ^^^^

or even:

	(i386_gdbarch_init): Don't call set_gdbarch_process_record here.


> 	* i386-tdep.c (i386_process_record): New extern.

                    ^ typo

 	* i386-tdep.h (i386_process_record): Declare.

(`new extern' sounds like adding a new variable or function,
but you're just changing the linkage of an existing function.)

-- 
Pedro Alves

Re: record mode supposed to work on x86_64-linux?

[Hui Zhu]

Thank you very much Pedro and Andi.

I changed the changelog to:
2009-05-17  Hui Zhu  <teawater@gmail.com>

	* i386-tdep.c (i386_process_record): Make it extern.
	(i386_gdbarch_init): Don't call set_gdbarch_process_record
	here.
	* i386-tdep.c (i386_process_record): Declare.
	* i386-linux-tdep.c (i386_linux_init_abi): Call
	set_gdbarch_process_record.

And checked-in.

Hui

On Sun, May 17, 2009 at 19:52, Pedro Alves <pedro@codesourcery.com> wrote:
> On Sunday 17 May 2009 01:15:00, Hui Zhu wrote:
>> Hi Pedro,
>>
>> What do you think about this one?
>
> I think it's good, thanks.
>
>> 2009-05-17  Hui Zhu  <teawater@gmail.com>
>>
>>         * i386-tdep.c (i386_process_record): Remove the static of it.
>>         (i386_gdbarch_init): Remove the set_gdbarch_process_record.
>>         * i386-tdep.c (i386_process_record): New extern.
>>         * i386-linux-tdep.c (i386_linux_init_abi): Call
>>         set_gdbarch_process_record.
>
> Please consider the alternative ChangeLog suggestions below instead.
>
>>       * i386-tdep.c (i386_process_record): Remove the static of it.
>
>        * i386-tdep.c (i386_process_record): Make it extern.
>
>>       (i386_gdbarch_init): Remove the set_gdbarch_process_record.
>
> From this entry, it isn't clear what you're removing.  Actually, it
> sounds like you're removing the function implementation.  Consider
> instead:
>
>        (i386_gdbarch_init): Remove the set_gdbarch_process_record call.
>                                                                   ^^^^
>
> or even:
>
>        (i386_gdbarch_init): Don't call set_gdbarch_process_record here.
>
>
>>       * i386-tdep.c (i386_process_record): New extern.
>
>                    ^ typo
>
>        * i386-tdep.h (i386_process_record): Declare.
>
> (`new extern' sounds like adding a new variable or function,
> but you're just changing the linkage of an existing function.)
>
> --
> Pedro Alves
>

Re: record mode supposed to work on x86_64-linux?

[Pedro Alves]

On Sunday 17 May 2009 13:28:01, Hui Zhu wrote:
>         * i386-tdep.c (i386_process_record): Declare.

                      ^ typo still present.

-- 
Pedro Alves

Re: record mode supposed to work on x86_64-linux?

[Hui Zhu]

Oops.  fixed.  :)

Hui

On Sun, May 17, 2009 at 20:31, Pedro Alves <pedro@codesourcery.com> wrote:
> On Sunday 17 May 2009 13:28:01, Hui Zhu wrote:
>>         * i386-tdep.c (i386_process_record): Declare.
>
>                      ^ typo still present.
>
> --
> Pedro Alves
>