[patch] add target Score to GDB

[patch] add target Score to GDB

[bjgnu]

[-- Attachment #1: Type: text/plain, Size: 2494 bytes --]

Hello Maintainers,

I'v ported GDB-6.5.90 to Score target and SunNorth has assigned the forms 
to the FSF.
The following is a patch adding Score target to GDB and i hope i could be 
the Score maintainer.
I'v posted the patch to gdb@sourceware.org on 2006.12.4, but get no reply, 
maybe it's a wrong place. 
Could anyone help me checking my patch? I really appreciate your help. 
I've checkout the current-cvs and all testpatterns run pass in DejaGNU.
Some FAILs are target dependent. I've checked these FAILs and they are not 
gdb bug. 
I hope Score target can be added to GDB-6.6 if possible.

1, Modify gdb-6.5/gdb/configure.tgt:
diff -pu -r gdb-6.5/gdb/configure.tgt gdb-6.5-score/gdb/configure.tgt
--- gdb-6.5/gdb/configure.tgt   2006-11-22 17:31:03.000000000 +0800
+++ gdb-6.5-score/gdb/configure.tgt 2006-11-22 17:48:36.000000000 +0800
@@ -21,6 +21,7 @@ m68*)         gdb_target_cpu=m68k ;;
 m88*)          gdb_target_cpu=m88k ;;
 mips*)         gdb_target_cpu=mips ;;
 powerpc*)      gdb_target_cpu=powerpc ;;
+score*)            gdb_target_cpu=score ;;
 sparc*)            gdb_target_cpu=sparc ;;
 thumb*)            gdb_target_cpu=arm ;;
 s390*)         gdb_target_cpu=s390 ;;
@@ -174,6 +175,7 @@ rs6000-*-lynxos*)   gdb_target=rs6000lynx 
 rs6000-*-aix4*)        gdb_target=aix4 ;;
 rs6000-*-*)        gdb_target=rs6000 ;;
 
+score-*-*)         gdb_target=embed ;;
 s390*-*-*)     gdb_target=s390
            build_gdbserver=yes
            ;;

2, Modify gdb-6.5/sim/configure:
diff -pu -r gdb-6.5/sim/configure gdb-6.5-score/sim/configure
--- gdb-6.5/sim/configure   2006-11-22 17:31:05.000000000 +0800
+++ gdb-6.5-score/sim/configure 2006-11-22 17:48:36.000000000 +0800
@@ -284,6 +284,7 @@ ac_subdirs_all="$ac_subdirs_all m68hc11"
 ac_subdirs_all="$ac_subdirs_all mcore"
 ac_subdirs_all="$ac_subdirs_all mips"
 ac_subdirs_all="$ac_subdirs_all mn10300"
+ac_subdirs_all="$ac_subdirs_all score"
 ac_subdirs_all="$ac_subdirs_all sh64"
 ac_subdirs_all="$ac_subdirs_all sh"
 ac_subdirs_all="$ac_subdirs_all erc32"
@@ -3510,6 +3511,14 @@ subdirs="$subdirs mn10300"
       common=yes
       igen=yes
       ;;
+       score*-*-*)
+
+
+subdirs="$subdirs score"
+
+      testsuite=yes
+      common=yes
+      ;;
        sh64-*-*)

3, Add the following directories and files:
1) gdb-6.5/gdb/config/score;
2) gdb-6.5/gdb/score-tdep.c;
3) gdb-6.5/gdb/score-tdep.h;
4) gdb-6.5/sim/score;


Best regards,
Qinwei
Mail  qinwei@sunnorth.com.cn
Phone +86-010-62981668-2708
Fax   +86-010-62985972





[-- Attachment #2: Score_GDB_patch.tar.gz --]
[-- Type: application/octet-stream, Size: 47374 bytes --]

Re: [patch] add target Score to GDB

[Jim Blandy]

bjgnu@sunnorth.com.cn writes:
> I'v ported GDB-6.5.90 to Score target and SunNorth has assigned the forms 
> to the FSF.
> The following is a patch adding Score target to GDB and i hope i could be 
> the Score maintainer.
> I'v posted the patch to gdb@sourceware.org on 2006.12.4, but get no reply, 
> maybe it's a wrong place. 
> Could anyone help me checking my patch? I really appreciate your help. 
> I've checkout the current-cvs and all testpatterns run pass in DejaGNU.
> Some FAILs are target dependent. I've checked these FAILs and they are not 
> gdb bug. 
> I hope Score target can be added to GDB-6.6 if possible.

Hi, Qinwei.  gdb-patches is the right mailing list for this.  However,
it may take us a while to review patches; there aren't very many
reviewers.  If you don't get a response within a week, then please
post a 'ping' patch.

Your work is more likely to be reviewed if you post it as a multi-file
patch which can be applied by running 'patch -p0 < foo.patch' in the
top-level source directory.  Unified diffs (diff -u) seem to be the
preferred format.

Your embed.mt file says:

    OOLIB_PATH=../../sim_lib/oolib_install
    TDEPFILES= score-tdep.o 
    SIM_OBS= remote-sim.o
    TM_CLIBS = -L $(OOLIB_PATH)/lib -lsc_gdb -lsc_tools -lsc_sim -lsc_ice -lprobe -lscore -lsocshell -lsoc -lpackages -lstdc++ -lSDL -lSDL_image
    SIM= ../sim/score/libsim.a 

It looks to me as if this means that the 'score' target depends on a
simulator located outside the GDB tree.  Is this simulator Free
software?  Please make sure your GDB satisfies the license conditions
explained in the COPYING file.  If you can release the simulator as
Free software as well, that would be best.  Otherwise, you may be able
to satisfy the license simply by building GDB without the simulator.

It seems the simulator in your post is just a stub.  We would rather
not have stub code in GDB.

I don't see a patch to 'sim/configure.in' in your post.
'sim/configure' is generated automatically from configure.in using
autoconf.  Patches should generally not include changes to generated
files, like 'configure', because they can be large.

I have some comments on score-tdep.c, but let's get the above issues
sorted out first.

Finally, I don't know how Joel Brobecker, our release manager, feels
about including new ports this late in the process.  On the one hand,
if the port is self-contained, it is unlikely to break other code.  On
the other hand, we have already produced our first release candidate.
I'll leave the decision to him.

Thanks for posting this port!

Re: Re: [patch] add target Score to GDB

[qinwei]

[-- Attachment #1: Type: text/plain, Size: 4327 bytes --]

Jim Blandy <jimb@codesourcery.com> 写于 2006-12-06 04:00:46:

> 
> bjgnu@sunnorth.com.cn writes:
> > I'v ported GDB-6.5.90 to Score target and SunNorth has assigned the 
forms 
> > to the FSF.
> > The following is a patch adding Score target to GDB and i hope i could 
be 
> > the Score maintainer.
> > I'v posted the patch to gdb@sourceware.org on 2006.12.4, but get no 
reply, 
> > maybe it's a wrong place. 
> > Could anyone help me checking my patch? I really appreciate your help. 

> > I've checkout the current-cvs and all testpatterns run pass in 
DejaGNU.
> > Some FAILs are target dependent. I've checked these FAILs and they are 
not 
> > gdb bug. 
> > I hope Score target can be added to GDB-6.6 if possible.
> 
> Hi, Qinwei.  gdb-patches is the right mailing list for this.  However,
> it may take us a while to review patches; there aren't very many
> reviewers.  If you don't get a response within a week, then please
> post a 'ping' patch.
> 
> Your work is more likely to be reviewed if you post it as a multi-file
> patch which can be applied by running 'patch -p0 < foo.patch' in the
> top-level source directory.  Unified diffs (diff -u) seem to be the
> preferred format.
> 
> Your embed.mt file says:
> 
>     OOLIB_PATH=../../sim_lib/oolib_install
>     TDEPFILES= score-tdep.o 
>     SIM_OBS= remote-sim.o
>     TM_CLIBS = -L $(OOLIB_PATH)/lib -lsc_gdb -lsc_tools -lsc_sim -
> lsc_ice -lprobe -lscore -lsocshell -lsoc -lpackages -lstdc++ -lSDL 
-lSDL_image
>     SIM= ../sim/score/libsim.a 
> 
> It looks to me as if this means that the 'score' target depends on a
> simulator located outside the GDB tree.  Is this simulator Free
> software?  Please make sure your GDB satisfies the license conditions
> explained in the COPYING file.  If you can release the simulator as
> Free software as well, that would be best.  Otherwise, you may be able
> to satisfy the license simply by building GDB without the simulator.
> 
> It seems the simulator in your post is just a stub.  We would rather
> not have stub code in GDB.
> 
> I don't see a patch to 'sim/configure.in' in your post.
> 'sim/configure' is generated automatically from configure.in using
> autoconf.  Patches should generally not include changes to generated
> files, like 'configure', because they can be large.
> 
> I have some comments on score-tdep.c, but let's get the above issues
> sorted out first.
> 
> Finally, I don't know how Joel Brobecker, our release manager, feels
> about including new ports this late in the process.  On the one hand,
> if the port is self-contained, it is unlikely to break other code.  On
> the other hand, we have already produced our first release candidate.
> I'll leave the decision to him.
> 
> Thanks for posting this port!

Hi Jim,

Thanks for your detailed explanation! I still have some questions.

If we won't open simulator source, can we put simulator libs(11 libs) and 
simstub.c in the dir 'gdb-6.5/sim/score'?
If we can, i will send you these libs. 
Otherwise, i post the patch in the following just containing 'score' gdb 
but without simulator.

I diff gdb-6.5 gdb-6.5-score in the top-level and the result is the 
following. 
Thanks for checking my patch and i hope to hear your advice on 
score-tdep.c.

Only in ../gdb-6.5-score/gdb/config: score
diff -r -u ./gdb/configure.tgt ../gdb-6.5-score/gdb/configure.tgt
--- ./gdb/configure.tgt 2006-11-22 17:31:03.000000000 +0800
+++ ../gdb-6.5-score/gdb/configure.tgt  2006-12-06 11:53:23.000000000 
+0800
@@ -21,6 +21,7 @@
 m88*)          gdb_target_cpu=m88k ;;
 mips*)         gdb_target_cpu=mips ;;
 powerpc*)      gdb_target_cpu=powerpc ;;
+score*)            gdb_target_cpu=score ;;
 sparc*)            gdb_target_cpu=sparc ;;
 thumb*)            gdb_target_cpu=arm ;;
 s390*)         gdb_target_cpu=s390 ;;
@@ -174,6 +175,7 @@
 rs6000-*-aix4*)        gdb_target=aix4 ;;
 rs6000-*-*)        gdb_target=rs6000 ;;
 
+score-*-*)     gdb_target=embed ;;
 s390*-*-*)     gdb_target=s390
            build_gdbserver=yes
            ;; 
Only in ../gdb-6.5-score/gdb: score-tdep.c
Only in ../gdb-6.5-score/gdb: score-tdep.h



Best regards,
Qinwei
Mail  qinwei@sunnorth.com.cn
Phone +86-010-62981668-2708
Fax   +86-010-62985972


[-- Attachment #2: Score_GDB_Patch_NoSim.tar.gz --]
[-- Type: application/octet-stream, Size: 8452 bytes --]

Re: [patch] add target Score to GDB

[Jim Blandy]

qinwei@sunnorth.com.cn writes:
> If we won't open simulator source, can we put simulator libs(11 libs) and 
> simstub.c in the dir 'gdb-6.5/sim/score'?
> If we can, i will send you these libs. 
> Otherwise, i post the patch in the following just containing 'score' gdb 
> but without simulator.

The GDB source tree cannot contain non-Free software.  We give our
users all the rights to modify and redistribute the software that we
have, on the condition that they do the same for others; that's the
founding principle of Project GNU.  Including non-free libraries with
GDB would not follow that principle.

We encourage you to free your simulator; if that's not possible, then
we'll need to leave the simulator out.

> I diff gdb-6.5 gdb-6.5-score in the top-level and the result is the 
> following. 
> Thanks for checking my patch and i hope to hear your advice on 
> score-tdep.c.

You need to run 'diff -ruN', so it shows the contents of newly added
files.  I'll look at the score-tdep.[ch] files from your post
yesterday.

It looks as if the ROUND_UP macro is not used.

In score_analyze_pdr_section, you probably want to use the following
functions declared in objfiles.h, instead of the
deprecated_obj_private pointer:

  /* Keep a registry of per-objfile data-pointers required by other GDB
     modules.  */

  extern const struct objfile_data *register_objfile_data (void);
  extern void clear_objfile_data (struct objfile *objfile);
  extern void set_objfile_data (struct objfile *objfile,
                                const struct objfile_data *data, void *value);
  extern void *objfile_data (struct objfile *objfile,
                             const struct objfile_data *data);

I see that you don't use any other deprecated features, which is good.

Do you need to define the read_pc and write_pc functions?  If you
leave them out, regcache.c's read_pc_pid and write_pc_pid should fall
back to default functions which will do the ptid-swapping for you,
making score_read_unsigned_register_pid, score_read_pc, and
score_write_pc unnecessary

Does the S+core architecture allow you to mix 32-bit and 16-bit
instructions?  If so, could score_breakpoint_from_pc be simplified to
always return a 16-bit breakpoint?  This would mean that a breakpoint
inserted at a 32-bit instruction would produce two bytes of garbage
after the breakpoint, but that's normal; on the i386, GDB uses a
one-byte breakpoint instruction regardless of the length of the
instruction it's set at.

Does score_integer_to_address do exactly the same thing that GCC does
when the user casts an integer to an address?  If they don't, then the
user could become quite confused when GDB and GCC evaluate their
expressions differently.

Future revisions of GCC change the sorts of prologues GDB will
encounter.  If you find yourself investing more time than you expected
on score_analyze_prologue, consider using the framework in
prologue-value.h, documented there.

Re: [patch] add target Score to GDB

[qinwei]

> Jim Blandy <jimb@codesourcery.com> writes:
>
> Hi, Qinwei.  The FSF confirms that we can accept your contributions,
> so you can go ahead and commit your change.  Please apply for an
> account via the form at http://sourceware.org/cgi-bin/pdw/ps_form.cgi,
> and list me as the approver.
> 
> Once that's done, follow the directions for checking out a read-write
> CVS tree at http://sourceware.org/gdb/current/.  Add yourself to the
> "Write After Approval" list in gdb/MAINTAINERS, and add yourself as
> the maintainer of the S+Core port in the "Responsible Maintainers"
> section.  Add an appropriate ChangeLog entry (see Gary Funck's on
> 2006-12-27), and commit.
> 
> Then, as a separate change, go ahead and commit the S+Core port.
> (Note that ChangeLog entries on the same day by the same person can
> share a date/name line in the ChangeLog.)
> 
> Good to have you, and thanks for the port!

Hi Jim. I've committed S+core port to the cvs.
Thanks for all your help!
And thanks for all the GDB maintainers!

Best regards,
Qinwei
Mail  qinwei@sunnorth.com.cn
Phone +86-010-62981668-2708
Fax   +86-010-62985972

Re: [patch] add target Score to GDB

[Jim Blandy]

Jim Blandy <jimb@codesourcery.com> writes:
> Hi, Qinwei.  I'm sorry we didn't make the GDB 6.6 release, but this
> version of the patch looks fine for inclusion.  I've asked the Free
> Software Foundation clerks whether the copyright paperwork is all
> ready; once that's done, we'll get you write-after-approval access to
> GDB, and you can commit the S+Core port.

Hi, Qinwei.  The FSF confirms that we can accept your contributions,
so you can go ahead and commit your change.  Please apply for an
account via the form at http://sourceware.org/cgi-bin/pdw/ps_form.cgi,
and list me as the approver.

Once that's done, follow the directions for checking out a read-write
CVS tree at http://sourceware.org/gdb/current/.  Add yourself to the
"Write After Approval" list in gdb/MAINTAINERS, and add yourself as
the maintainer of the S+Core port in the "Responsible Maintainers"
section.  Add an appropriate ChangeLog entry (see Gary Funck's on
2006-12-27), and commit.

Then, as a separate change, go ahead and commit the S+Core port.
(Note that ChangeLog entries on the same day by the same person can
share a date/name line in the ChangeLog.)

Good to have you, and thanks for the port!

Re: [patch] add target Score to GDB

[Jim Blandy]

Hi, Qinwei.  I'm sorry we didn't make the GDB 6.6 release, but this
version of the patch looks fine for inclusion.  I've asked the Free
Software Foundation clerks whether the copyright paperwork is all
ready; once that's done, we'll get you write-after-approval access to
GDB, and you can commit the S+Core port.

Re: [patch] add target Score to GDB

[秦巍]

[-- Attachment #1: Type: text/plain, Size: 2481 bytes --]

Sorry Jim i didn't reply you yesterday for the reason that i can't 
checkout the 
current cvs caused by our network.

> Jim Blandy <jimb@codesourcery.com> wrote on 2006-12-08 04:19:03:

> >> Does score_integer_to_address do exactly the same thing that GCC does
> >> when the user casts an integer to an address?  If they don't, then 
the
> >> user could become quite confused when GDB and GCC evaluate their
> >> expressions differently.
> >
> > I didn't see clearly about this hook when i was porting score gdb.
> > I have seen your NOTES in 'value_as_address'.
> > The pointer and the address are the same on score target, 
> > and a pointer to a function always points to the entry point address 
of 
> > that function.
> > So we don't need to implement this hook. Can i say that?
> 
> If you don't need to set it to get clean test suite results, then I
> would definitely recommend leaving it out.

I will leave it out.

> Please keep all lines under 80 characters in length; for example,
> score_unwind_dummy_id needs to be reformatted.
> 
> I'm surprised by score_type_needs_double_align --- is it really true
> that struct types only need double alignment if their *first* member
> does?  Why is that?  I'd expect the code for structs and unions to be
> the same.

I have modified that.

> There should be patches to Makefile.in; could you post those?  We need
> to see the complete patch in order to approve it.
> 
> All changes need a ChangeLog entry.  You don't need to say much for
> score-tdep.c and score-tdep.h; here's the entry for the xtensa port:
> 
> 2006-11-14  Maxim Grigoriev  <maxim@tensilica.com>
> 
>    * NEWS: New port to Xtensa.
>    * Makefile.in: Add dependencies for Xtensa files.
>    * configure.tgt (xtensa*, xtensa*-*-elf*): New.
>    * configure.host (xtensa*-*-elf*): New.
>    * config/xtensa/xtensa.mt: New file.
>    * xtensa-config.c: New file.
>    * xtensa-tdep.h: New file.
>    * xtensa-tdep.c: New file.
> 
> (I would prefer that the Makefile.in entry actually listed the
> variables and rules added.)
> 
> Other than that, this looks good.  Please post a complete patch,
> generated against the current CVS sources, and with a ChangeLog entry,
> for a final review, and then we'll get it committed.

I have added S+core target to ChangeLog and dependencies to Makefile.in.
The following is a complete patch. Thanks for checking it!

Best regards,
Qinwei
Mail  qinwei@sunnorth.com.cn
Phone +86-010-62981668-2708
Fax   +86-010-62985972


[-- Attachment #2: score_gdb.patch --]
[-- Type: application/octet-stream, Size: 32824 bytes --]

diff -ruN gdb-6.6/gdb/ChangeLog gdb-6.6-score/gdb/ChangeLog
--- gdb-6.6/gdb/ChangeLog	2006-12-09 06:05:36.000000000 +0800
+++ gdb-6.6-score/gdb/ChangeLog	2006-12-09 14:03:50.000000000 +0800
@@ -1,3 +1,12 @@
+2006-12-09  Qinwei  <qinwei@sunnorth.com.cn>
+
+	* NEWS: New port to S+core.
+	* Makefile.in: Add dependencies for S+core files.
+	* configure.tgt (score*, score-*-*): New.
+	* config/score/embed.mt: New file.
+	* score-tdep.h: New file.
+	* score-tdep.c: New file.
+
 2006-12-09  Nick Roberts  <nickrob@snap.net.nz>
 
 	* varobj.c (varobj_update): Ensure frame is restored when
diff -ruN gdb-6.6/gdb/config/score/embed.mt gdb-6.6-score/gdb/config/score/embed.mt
--- gdb-6.6/gdb/config/score/embed.mt	1970-01-01 08:00:00.000000000 +0800
+++ gdb-6.6-score/gdb/config/score/embed.mt	2006-12-09 14:03:50.000000000 +0800
@@ -0,0 +1,2 @@
+# Target: S+core embedded system
+TDEPFILES= score-tdep.o 
diff -ruN gdb-6.6/gdb/configure.tgt gdb-6.6-score/gdb/configure.tgt
--- gdb-6.6/gdb/configure.tgt	2006-11-23 01:34:15.000000000 +0800
+++ gdb-6.6-score/gdb/configure.tgt	2006-12-09 14:03:50.000000000 +0800
@@ -24,6 +24,7 @@
 sparc*)			gdb_target_cpu=sparc ;;
 thumb*)			gdb_target_cpu=arm ;;
 s390*)			gdb_target_cpu=s390 ;;
+score*)			gdb_target_cpu=score ;;
 sh*)			gdb_target_cpu=sh ;;
 spu*)			gdb_target_cpu=spu ;;
 strongarm*)		gdb_target_cpu=arm ;;
@@ -179,6 +180,7 @@
 			build_gdbserver=yes
 			;;
 
+score-*-*)		gdb_target=embed ;;
 sh*-*-pe)		gdb_target=wince ;;
 sh-*-coff*)		gdb_target=embed ;;
 sh-*-elf*)		gdb_target=embed ;;
diff -ruN gdb-6.6/gdb/Makefile.in gdb-6.6-score/gdb/Makefile.in
--- gdb-6.6/gdb/Makefile.in	2006-11-29 06:14:31.000000000 +0800
+++ gdb-6.6-score/gdb/Makefile.in	2006-12-09 14:03:50.000000000 +0800
@@ -782,6 +782,7 @@
 s390_tdep_h = s390-tdep.h
 scm_lang_h = scm-lang.h $(scm_tags_h)
 scm_tags_h = scm-tags.h
+score_tdep_h = score-tdep.h
 sentinel_frame_h = sentinel-frame.h
 serial_h = serial.h
 ser_base_h = ser-base.h
@@ -1490,6 +1491,7 @@
 	remote-st.c remote-utils.c dcache.c \
 	rs6000-nat.c rs6000-tdep.c \
 	s390-tdep.c s390-nat.c \
+	score-tdep.c \
 	ser-go32.c ser-pipe.c ser-tcp.c ser-mingw.c \
 	sh-tdep.c sh64-tdep.c shnbsd-tdep.c shnbsd-nat.c \
 	sol2-tdep.c \
@@ -2580,6 +2582,10 @@
 scm-valprint.o: scm-valprint.c $(defs_h) $(symtab_h) $(gdbtypes_h) \
 	$(expression_h) $(parser_defs_h) $(language_h) $(value_h) \
 	$(scm_lang_h) $(valprint_h) $(gdbcore_h) $(c_lang_h)
+score-tdep.o: score-tdep.c $(defs_h) $(gdb_assert_h) $(inferior_h) \
+	$(symtab_h) $(objfiles_h) $(gdbcore_h) $(target_h) \
+	$(arch_utils_h) $(regcache_h) $(dis_asm_h) $(frame_unwind_h) \
+	$(frame_base_h) $(trad_frame_h) $(dwarf2_frame_h) $(score_tdep_h)
 sentinel-frame.o: sentinel-frame.c $(defs_h) $(regcache_h) \
 	$(sentinel_frame_h) $(inferior_h) $(frame_unwind_h)
 ser-e7kpc.o: ser-e7kpc.c $(defs_h) $(serial_h) $(gdb_string_h)
diff -ruN gdb-6.6/gdb/score-tdep.c gdb-6.6-score/gdb/score-tdep.c
--- gdb-6.6/gdb/score-tdep.c	1970-01-01 08:00:00.000000000 +0800
+++ gdb-6.6-score/gdb/score-tdep.c	2006-12-09 14:03:50.000000000 +0800
@@ -0,0 +1,905 @@
+/* Target-dependent code for the S+core architecture, for GDB, the GNU Debugger.
+
+   Copyright (C) 2006
+   Free Software Foundation, Inc.
+
+   Contributed by Qinwei (qinwei@sunnorth.com.cn)
+   Contributed by Ching-Peng Lin (cplin@sunplus.com)
+
+   This file is part of GDB.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 51 Franklin Street, Fifth Floor,
+   Boston, MA 02110-1301, USA.  */
+
+#include "defs.h"
+#include "gdb_assert.h"
+#include "inferior.h"
+#include "symtab.h"
+#include "objfiles.h"
+#include "gdbcore.h"
+#include "target.h"
+#include "arch-utils.h"
+#include "regcache.h"
+#include "dis-asm.h"
+#include "frame-unwind.h"
+#include "frame-base.h"
+#include "trad-frame.h"
+#include "dwarf2-frame.h"
+#include "score-tdep.h"
+
+#define G_FLD(_i,_ms,_ls)   (((_i) << (31 - (_ms))) >> (31 - (_ms) + (_ls)))
+#define RM_PBITS(_raw)      ((G_FLD(_raw, 31, 16) << 15) | G_FLD(_raw, 14, 0))
+
+typedef struct{
+    unsigned int v;
+    unsigned int raw;
+    char         is15;
+}inst_t;
+
+struct score_frame_cache
+{
+  CORE_ADDR base;
+  struct trad_frame_saved_reg *saved_regs;
+};
+
+#if 0
+/* If S+core GCC will generate these instructions in the prologue:
+
+   lw   rx, imm1
+   addi rx, -imm2
+   mv!  r2, rx
+
+   then .pdr section is used.  */
+
+#define P_SIZE          8
+#define PI_SYM          0
+#define PI_R_MSK        1
+#define PI_R_OFF        2
+#define PI_R_LEF        4
+#define PI_F_OFF        5
+#define PI_F_REG        6
+#define PI_RAREG        7
+
+typedef struct frame_extra_info
+{
+  CORE_ADDR p_frame;
+  unsigned int pdr[P_SIZE];
+} extra_info_t;
+
+struct obj_priv
+{
+  bfd_size_type size;
+  char *contents;
+};
+
+static bfd *the_bfd;
+
+static int
+score_compare_pdr_entries (const void *a, const void *b)
+{
+  CORE_ADDR lhs = bfd_get_32 (the_bfd, (bfd_byte *) a);
+  CORE_ADDR rhs = bfd_get_32 (the_bfd, (bfd_byte *) b);
+  if (lhs < rhs)
+    return -1;
+  else if (lhs == rhs)
+    return 0;
+  else
+    return 1;
+}
+
+static void
+score_analyze_pdr_section (CORE_ADDR startaddr, CORE_ADDR pc,
+                           struct frame_info *next_frame,
+                           struct score_frame_cache *this_cache)
+{
+  struct symbol *sym;
+  struct obj_section *sec;
+  extra_info_t *fci_ext;
+  CORE_ADDR leaf_ra_stack_addr = -1;
+
+  gdb_assert (startaddr <= pc);
+  gdb_assert (this_cache != NULL);
+
+  fci_ext = frame_obstack_zalloc (sizeof (extra_info_t));
+  if ((sec = find_pc_section (pc)) == NULL)
+    {
+      error ("Can't find section in file:%s, line:%d!", __FILE__, __LINE__);
+      return;
+    }
+
+  /* Anylyze .pdr section and get coresponding fields.  */
+  {
+    static struct obj_priv *priv = NULL;
+
+    if (priv == NULL)
+      {
+        asection *bfdsec;
+        priv = obstack_alloc (&sec->objfile->objfile_obstack,
+                              sizeof (struct obj_priv));
+        if ((bfdsec = bfd_get_section_by_name (sec->objfile->obfd, ".pdr")))
+          {
+            priv->size = bfd_section_size (sec->objfile->obfd, bfdsec);
+            priv->contents = obstack_alloc (&sec->objfile->objfile_obstack,
+                                            priv->size);
+            bfd_get_section_contents (sec->objfile->obfd, bfdsec,
+                                      priv->contents, 0, priv->size);
+            the_bfd = sec->objfile->obfd;
+            qsort (priv->contents, priv->size / 32, 32,
+                   score_compare_pdr_entries);
+            the_bfd = NULL;
+          }
+        else
+          priv->size = 0;
+      }
+    if (priv->size != 0)
+      {
+        int low = 0, mid, high = priv->size / 32;
+        char *ptr;
+        do
+
+          {
+            CORE_ADDR pdr_pc;
+            mid = (low + high) / 2;
+            ptr = priv->contents + mid * 32;
+            pdr_pc = bfd_get_signed_32 (sec->objfile->obfd, ptr);
+            pdr_pc += ANOFFSET (sec->objfile->section_offsets,
+                                SECT_OFF_TEXT (sec->objfile));
+            if (pdr_pc == startaddr)
+              break;
+            if (pdr_pc > startaddr)
+              high = mid;
+            else
+              low = mid + 1;
+          }
+        while (low != high);
+
+        if (low != high)
+          {
+            gdb_assert (bfd_get_32 (sec->objfile->obfd, ptr) == startaddr);
+#define EXT_PDR(_pi)    bfd_get_32(sec->objfile->obfd, ptr+((_pi)<<2))
+            fci_ext->pdr[PI_SYM] = EXT_PDR (PI_SYM);
+            fci_ext->pdr[PI_R_MSK] = EXT_PDR (PI_R_MSK);
+            fci_ext->pdr[PI_R_OFF] = EXT_PDR (PI_R_OFF);
+            fci_ext->pdr[PI_R_LEF] = EXT_PDR (PI_R_LEF);
+            fci_ext->pdr[PI_F_OFF] = EXT_PDR (PI_F_OFF);
+            fci_ext->pdr[PI_F_REG] = EXT_PDR (PI_F_REG);
+            fci_ext->pdr[PI_RAREG] = EXT_PDR (PI_RAREG);
+#undef EXT_PDR
+          }
+      }
+  }
+}
+#endif
+
+static struct type *
+score_register_type (struct gdbarch *gdbarch, int regnum)
+{
+  gdb_assert (regnum >= 0 && regnum < SCORE_NUM_REGS);
+  return builtin_type_uint32;
+}
+
+static LONGEST
+score_read_unsigned_register (int regnum)
+{
+  LONGEST val;
+  regcache_cooked_read_unsigned (current_regcache, regnum, &val);
+  return val;
+}
+
+static CORE_ADDR
+score_read_sp (void)
+{
+  return score_read_unsigned_register (SCORE_SP_REGNUM);
+}
+
+static CORE_ADDR
+score_unwind_pc (struct gdbarch *gdbarch, struct frame_info *next_frame)
+{
+  return frame_unwind_register_unsigned (next_frame, SCORE_PC_REGNUM);
+}
+
+static const char *
+score_register_name (int regnum)
+{
+  const char *score_register_names[] = {
+    "r0",  "r1",  "r2",  "r3",  "r4",  "r5",  "r6",  "r7",
+    "r8",  "r9",  "r10", "r11", "r12", "r13", "r14", "r15",
+    "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
+    "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31",
+
+    "PSR",   "COND",   "ECR",    "EXCPVEC",
+    "CCR",   "EPC",    "EMA",    "TLBLOCK",
+    "TLBPT", "PEADDR", "TLBRPT", "PEVN",
+    "PECTX", "LIMPFN", "LDMPFN", "PREV",
+    "DREG",  "PC",     "DSAVE",  "COUNTER",
+    "LDCR",  "STCR",   "CEH",    "CEL",
+  };
+
+  gdb_assert (regnum >= 0 && regnum < SCORE_NUM_REGS);
+  return score_register_names[regnum];
+}
+
+static int
+score_register_sim_regno (int regnum)
+{
+  gdb_assert (regnum >= 0 && regnum < SCORE_NUM_REGS);
+  return regnum;
+}
+
+static int
+score_print_insn (bfd_vma memaddr, struct disassemble_info *info)
+{
+  if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG)
+    return print_insn_big_score (memaddr, info);
+  else
+    return print_insn_little_score (memaddr, info);
+}
+
+static const gdb_byte *
+score_breakpoint_from_pc (CORE_ADDR *pcptr, int *lenptr)
+{
+  gdb_byte buf[SCORE_INSTLEN] = { 0 };
+  int ret;
+  unsigned int raw;
+
+  if ((ret = target_read_memory (*pcptr & ~0x3, buf, SCORE_INSTLEN)) != 0)
+    {
+      memory_error (ret, *pcptr);
+    }
+  raw = extract_unsigned_integer (buf, SCORE_INSTLEN);
+
+  if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG)
+    {
+      if (!(raw & 0x80008000))
+        {
+          /* 16bits instruction.  */
+          static gdb_byte big_breakpoint16[] = { 0x60, 0x02 };
+          *pcptr &= ~0x1;
+          *lenptr = sizeof (big_breakpoint16);
+          return big_breakpoint16;
+        }
+      else
+        {
+          /* 32bits instruction.  */
+          static gdb_byte big_breakpoint32[] = { 0x80, 0x00, 0x80, 0x06 };
+          *pcptr &= ~0x3;
+          *lenptr = sizeof (big_breakpoint32);
+          return big_breakpoint32;
+        }
+    }
+  else
+    {
+      if (!(raw & 0x80008000))
+        {
+          /* 16bits instruction.  */
+          static gdb_byte little_breakpoint16[] = { 0x02, 0x60 };
+          *pcptr &= ~0x1;
+          *lenptr = sizeof (little_breakpoint16);
+          return little_breakpoint16;
+        }
+      else
+        {
+          /* 32bits instruction.  */
+          static gdb_byte little_breakpoint32[] = { 0x06, 0x80, 0x00, 0x80 };
+          *pcptr &= ~0x3;
+          *lenptr = sizeof (little_breakpoint32);
+          return little_breakpoint32;
+        }
+    }
+}
+
+static CORE_ADDR
+score_frame_align (struct gdbarch *gdbarch, CORE_ADDR addr)
+{
+  return align_down (addr, 16);
+}
+
+static void
+score_xfer_register (struct regcache *regcache, int regnum, int length,
+                     enum bfd_endian endian, gdb_byte *readbuf,
+                     const gdb_byte *writebuf, int buf_offset)
+{
+  int reg_offset = 0;
+  gdb_assert (regnum >= 0 && regnum < SCORE_NUM_REGS);
+
+  switch (endian)
+    {
+    case BFD_ENDIAN_BIG:
+      reg_offset = SCORE_REGSIZE - length;
+      break;
+    case BFD_ENDIAN_LITTLE:
+      reg_offset = 0;
+      break;
+    case BFD_ENDIAN_UNKNOWN:
+      reg_offset = 0;
+      break;
+    default:
+      internal_error (__FILE__, __LINE__, _("score_xfer_register error!"));
+    }
+
+  if (readbuf != NULL)
+    regcache_cooked_read_part (regcache, regnum, reg_offset, length,
+                               readbuf + buf_offset);
+  if (writebuf != NULL)
+    regcache_cooked_write_part (regcache, regnum, reg_offset, length,
+                                writebuf + buf_offset);
+}
+
+static enum return_value_convention
+score_return_value (struct gdbarch *gdbarch, struct type *type,
+                    struct regcache *regcache,
+                    gdb_byte * readbuf, const gdb_byte * writebuf)
+{
+  if (TYPE_CODE (type) == TYPE_CODE_STRUCT
+      || TYPE_CODE (type) == TYPE_CODE_UNION
+      || TYPE_CODE (type) == TYPE_CODE_ARRAY)
+    return RETURN_VALUE_STRUCT_CONVENTION;
+  else
+    {
+      int offset;
+      int regnum;
+      for (offset = 0, regnum = SCORE_A0_REGNUM;
+           offset < TYPE_LENGTH (type);
+           offset += SCORE_REGSIZE, regnum++)
+        {
+          int xfer = SCORE_REGSIZE;
+          if (offset + xfer > TYPE_LENGTH (type))
+            xfer = TYPE_LENGTH (type) - offset;
+          score_xfer_register (regcache, regnum, xfer, TARGET_BYTE_ORDER,
+                               readbuf, writebuf, offset);
+        }
+      return RETURN_VALUE_REGISTER_CONVENTION;
+    }
+}
+
+static struct frame_id
+score_unwind_dummy_id (struct gdbarch *gdbarch, struct frame_info *next_frame)
+{
+  return frame_id_build (
+           frame_unwind_register_unsigned (next_frame, SCORE_SP_REGNUM),
+           frame_pc_unwind (next_frame));
+}
+
+static int
+score_type_needs_double_align (struct type *type)
+{
+  enum type_code typecode = TYPE_CODE (type);
+
+  if (typecode == TYPE_CODE_INT && TYPE_LENGTH (type) == 8)
+    return 1;
+  if (typecode == TYPE_CODE_FLT && TYPE_LENGTH (type) == 8)
+    return 1;
+  else if (typecode == TYPE_CODE_STRUCT || typecode == TYPE_CODE_UNION)
+    {
+      int i, n;
+
+      n = TYPE_NFIELDS (type);
+      for (i = 0; i < n; i++)
+        if (score_type_needs_double_align (TYPE_FIELD_TYPE (type, i)))
+          return 1;
+      return 0;
+    }
+  return 0;
+}
+
+static CORE_ADDR
+score_push_dummy_call (struct gdbarch *gdbarch, struct value *function,
+                       struct regcache *regcache, CORE_ADDR bp_addr,
+                       int nargs, struct value **args, CORE_ADDR sp,
+                       int struct_return, CORE_ADDR struct_addr)
+{
+  int argnum;
+  int argreg;
+  int arglen = 0;
+  CORE_ADDR stack_offset = 0;
+  CORE_ADDR addr = 0;
+
+  /* Step 1, Save RA.  */
+  regcache_cooked_write_unsigned (regcache, SCORE_RA_REGNUM, bp_addr);
+
+  /* Step 2, Make space on the stack for the args.  */
+  struct_addr = align_down (struct_addr, 16);
+  sp = align_down (sp, 16);
+  for (argnum = 0; argnum < nargs; argnum++)
+    arglen += align_up (TYPE_LENGTH (value_type (args[argnum])),
+                        SCORE_REGSIZE);
+  sp -= align_up (arglen, 16);
+
+  argreg = SCORE_BEGIN_ARG_REGNUM;
+
+  /* Step 3, Check if struct return then save the struct address to r4 and
+     increase the stack_offset by 4.  */
+  if (struct_return)
+    {
+      regcache_cooked_write_unsigned (regcache, argreg++, struct_addr);
+      stack_offset += SCORE_REGSIZE;
+    }
+
+  /* Step 4, Load arguments:
+     If arg length is too long (> 4 bytes),
+     then split the arg and save every parts.  */
+  for (argnum = 0; argnum < nargs; argnum++)
+    {
+      struct value *arg = args[argnum];
+      struct type *arg_type = check_typedef (value_type (arg));
+      arglen = TYPE_LENGTH (arg_type);
+      enum type_code typecode = TYPE_CODE (arg_type);
+      const gdb_byte *val = value_contents (arg);
+      int downward_offset = 0;
+
+      int odd_sized_struct_p = (arglen > SCORE_REGSIZE
+                                && arglen % SCORE_REGSIZE != 0);
+      int arg_last_part_p = 0;
+
+      /* If a arg should be aligned to 8 bytes (long long or double),
+         the value should be put to even register numbers.  */
+      if (score_type_needs_double_align (arg_type))
+        {
+          if (argreg & 1)
+            argreg++;
+        }
+
+      /* If sizeof a block < SCORE_REGSIZE, then Score GCC will chose
+         the default "downward"/"upward" method:
+
+         Example:
+
+         struct struc
+         {
+           char a; char b; char c;
+         } s = {'a', 'b', 'c'};
+
+         Big endian:    s = {X, 'a', 'b', 'c'}
+         Little endian: s = {'a', 'b', 'c', X}
+
+         Where X is a hole.  */
+
+      if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG
+          && (typecode == TYPE_CODE_STRUCT
+              || typecode == TYPE_CODE_UNION)
+          && argreg > SCORE_LAST_ARG_REGNUM
+          && arglen < SCORE_REGSIZE)
+        downward_offset += (SCORE_REGSIZE - arglen);
+
+      while (arglen > 0)
+        {
+          int partial_len = arglen < SCORE_REGSIZE ? arglen : SCORE_REGSIZE;
+          ULONGEST regval = extract_unsigned_integer (val, partial_len);
+
+          /* The last part of a arg should shift left when
+             TARGET_BYTE_ORDER is BFD_ENDIAN_BIG.  */
+          if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG
+              && arg_last_part_p == 1
+              && (typecode == TYPE_CODE_STRUCT
+                  || typecode == TYPE_CODE_UNION))
+            regval <<= ((SCORE_REGSIZE - partial_len) * TARGET_CHAR_BIT);
+
+          /* Always increase the stack_offset and save args to stack.  */
+          addr = sp + stack_offset + downward_offset;
+          write_memory (addr, val, partial_len);
+
+          if (argreg <= SCORE_LAST_ARG_REGNUM)
+            {
+              regcache_cooked_write_unsigned (regcache, argreg++, regval);
+              if (arglen > SCORE_REGSIZE && arglen < SCORE_REGSIZE * 2)
+                arg_last_part_p = 1;
+            }
+
+          val += partial_len;
+          arglen -= partial_len;
+          stack_offset += align_up (partial_len, SCORE_REGSIZE);
+        }
+    }
+
+  /* Step 5, Save SP.  */
+  regcache_cooked_write_unsigned (regcache, SCORE_SP_REGNUM, sp);
+
+  return sp;
+}
+
+static inst_t *
+score_fetch_instruction (CORE_ADDR addr)
+{
+  static inst_t inst = { 0, 0 };
+  char buf[SCORE_INSTLEN];
+  int big;
+  int ret = target_read_memory (addr & ~0x3, buf, SCORE_INSTLEN);
+  unsigned int raw;
+
+  if (ret)
+    {
+      memory_error (ret, addr);
+      return 0;
+    }
+  inst.raw = extract_unsigned_integer (buf, SCORE_INSTLEN);
+  inst.is15 = !(inst.raw & 0x80008000);
+  inst.v = RM_PBITS (inst.raw);
+  big = (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG);
+
+  if (inst.is15)
+    {
+      if (big ^ ((addr & 0x2) == 2))
+        inst.v = G_FLD (inst.v, 29, 15);
+      else
+        inst.v = G_FLD (inst.v, 14, 0);
+    }
+  return &inst;
+}
+
+static CORE_ADDR
+score_skip_prologue (CORE_ADDR pc)
+{
+  CORE_ADDR cpc = pc;
+  int iscan = 32, stack_sub = 0;
+  while (iscan-- > 0)
+    {
+      inst_t *inst = score_fetch_instruction (cpc);
+      if (!inst)
+        break;
+      if (!inst->is15 && !stack_sub
+          && (G_FLD (inst->v, 29, 25) == 0x1
+              && G_FLD (inst->v, 24, 20) == 0x0))
+        {
+          /* addi r0, offset */
+          pc = stack_sub = cpc + SCORE_INSTLEN;
+        }
+      else if (!inst->is15
+               && inst->v == RM_PBITS (0x8040bc56))
+        {
+          /* mv r2, r0  */
+          pc = cpc + SCORE_INSTLEN;
+          break;
+        }
+      else if (inst->is15
+               && inst->v == RM_PBITS (0x0203))
+        {
+          /* mv! r2, r0 */
+          pc = cpc + SCORE16_INSTLEN;
+          break;
+        }
+      else if (inst->is15
+               && ((G_FLD (inst->v, 14, 12) == 3)    /* j15 form */
+                   || (G_FLD (inst->v, 14, 12) == 4) /* b15 form */
+                   || (G_FLD (inst->v, 14, 12) == 0x0
+                       && G_FLD (inst->v, 3, 0) == 0x4))) /* br! */
+        break;
+      else if (!inst->is15
+               && ((G_FLD (inst->v, 29, 25) == 2)    /* j32 form */
+                   || (G_FLD (inst->v, 29, 25) == 4) /* b32 form */
+                   || (G_FLD (inst->v, 29, 25) == 0x0
+                       && G_FLD (inst->v, 6, 1) == 0x4)))  /* br */
+        break;
+
+      cpc += inst->is15 ? SCORE16_INSTLEN : SCORE_INSTLEN;
+    }
+  return pc;
+}
+
+static int
+score_in_function_epilogue_p (struct gdbarch *gdbarch, CORE_ADDR cur_pc)
+{
+  inst_t *inst = score_fetch_instruction (cur_pc);
+
+  if (inst->v == 0x23)
+    return 1;   /* mv! r0, r2 */
+  else if (G_FLD (inst->v, 14, 12) == 0x2
+           && G_FLD (inst->v, 3, 0) == 0xa)
+    return 1;   /* pop! */
+  else if (G_FLD (inst->v, 14, 12) == 0x0
+           && G_FLD (inst->v, 7, 0) == 0x34)
+    return 1;   /* br! r3 */
+  else if (G_FLD (inst->v, 29, 15) == 0x2
+           && G_FLD (inst->v, 6, 1) == 0x2b)
+    return 1;   /* mv r0, r2 */
+  else if (G_FLD (inst->v, 29, 25) == 0x0
+           && G_FLD (inst->v, 6, 1) == 0x4
+           && G_FLD (inst->v, 19, 15) == 0x3)
+    return 1;   /* br r3 */
+  else
+    return 0;
+}
+
+static void
+score_analyze_prologue (CORE_ADDR startaddr, CORE_ADDR pc,
+                        struct frame_info *next_frame,
+                        struct score_frame_cache *this_cache)
+{
+  CORE_ADDR sp;
+  CORE_ADDR cur_pc = startaddr;
+
+  int sp_offset = 0;
+  int ra_offset = 0;
+  int fp_offset = 0;
+  int ra_offset_p = 0;
+  int fp_offset_p = 0;
+  int inst_len = 0;
+
+  sp = frame_unwind_register_unsigned (next_frame, SCORE_SP_REGNUM);
+
+  for (; cur_pc < pc; cur_pc += inst_len)
+    {
+      inst_t *inst = score_fetch_instruction (cur_pc);
+      if (inst->is15 == 1)
+        {
+          inst_len = SCORE16_INSTLEN;
+
+          if (G_FLD (inst->v, 14, 12) == 0x2
+              && G_FLD (inst->v, 3, 0) == 0xe)
+            {
+              /* push! */
+              sp_offset += 4;
+
+              if (G_FLD (inst->v, 11, 7) == 0x6
+                  && ra_offset_p == 0)
+                {
+                  /* push! r3, [r0] */
+                  ra_offset = sp_offset;
+                  ra_offset_p = 1;
+                }
+              else if (G_FLD (inst->v, 11, 7) == 0x4
+                       && fp_offset_p == 0)
+                {
+                  /* push! r2, [r0] */
+                  fp_offset = sp_offset;
+                  fp_offset_p = 1;
+                }
+            }
+          else if (G_FLD (inst->v, 14, 12) == 0x2
+                   && G_FLD (inst->v, 3, 0) == 0xa)
+            {
+              /* pop! */
+              sp_offset -= 4;
+            }
+          else if (G_FLD (inst->v, 14, 7) == 0xc1
+                   && G_FLD (inst->v, 2, 0) == 0x0)
+            {
+              /* subei! r0, n */
+              sp_offset += (int) pow (2, G_FLD (inst->v, 6, 3));
+            }
+          else if (G_FLD (inst->v, 14, 7) == 0xc0
+                   && G_FLD (inst->v, 2, 0) == 0x0)
+            {
+              /* addei! r0, n */
+              sp_offset -= (int) pow (2, G_FLD (inst->v, 6, 3));
+            }
+        }
+      else
+        {
+          inst_len = SCORE_INSTLEN;
+
+          if (G_FLD (inst->v, 29, 15) == 0xc60
+              && G_FLD (inst->v, 2, 0) == 0x4)
+            {
+              /* sw r3, [r0, offset]+ */
+              sp_offset += SCORE_INSTLEN;
+              if (ra_offset_p == 0)
+                {
+                  ra_offset = sp_offset;
+                  ra_offset_p = 1;
+                }
+            }
+          if (G_FLD (inst->v, 29, 15) == 0xc40
+              && G_FLD (inst->v, 2, 0) == 0x4)
+            {
+              /* sw r2, [r0, offset]+ */
+              sp_offset += SCORE_INSTLEN;
+              if (fp_offset_p == 0)
+                {
+                  fp_offset = sp_offset;
+                  fp_offset_p = 1;
+                }
+            }
+          else if (G_FLD (inst->v, 29, 15) == 0x1c60
+                   && G_FLD (inst->v, 2, 0) == 0x0)
+            {
+              /* lw r3, [r0]+, 4 */
+              sp_offset -= SCORE_INSTLEN;
+              ra_offset_p = 1;
+            }
+          else if (G_FLD (inst->v, 29, 15) == 0x1c40
+                   && G_FLD (inst->v, 2, 0) == 0x0)
+            {
+              /* lw r2, [r0]+, 4 */
+              sp_offset -= SCORE_INSTLEN;
+              fp_offset_p = 1;
+            }
+
+          else if (G_FLD (inst->v, 29, 17) == 0x100
+                   && G_FLD (inst->v, 0, 0) == 0x0)
+            {
+              /* addi r0, -offset */
+              sp_offset += 65536 - G_FLD (inst->v, 16, 1);
+            }
+          else if (G_FLD (inst->v, 29, 17) == 0x110
+                   && G_FLD (inst->v, 0, 0) == 0x0)
+            {
+              /* addi r2, offset */
+              if (pc - cur_pc > 4)
+                {
+                  unsigned int save_v = inst->v;
+                  inst_t *inst2 =
+                    score_fetch_instruction (cur_pc + SCORE_INSTLEN);
+                  if (inst2->v == 0x23)
+                    /* mv! r0, r2 */
+                    sp_offset -= G_FLD (save_v, 16, 1);
+                }
+            }
+        }
+    }
+
+  /* Save RA.  */
+  if (ra_offset_p == 1)
+    {
+      if (this_cache->saved_regs[SCORE_PC_REGNUM].addr == -1)
+        this_cache->saved_regs[SCORE_PC_REGNUM].addr =
+          sp + sp_offset - ra_offset;
+    }
+  else
+    {
+      this_cache->saved_regs[SCORE_PC_REGNUM] =
+        this_cache->saved_regs[SCORE_RA_REGNUM];
+    }
+
+  /* Save FP.  */
+  if (fp_offset_p == 1)
+    {
+      if (this_cache->saved_regs[SCORE_FP_REGNUM].addr == -1)
+        this_cache->saved_regs[SCORE_FP_REGNUM].addr =
+          sp + sp_offset - fp_offset;
+    }
+
+  /* Save SP.  */
+  this_cache->base =
+    frame_unwind_register_unsigned (next_frame, SCORE_SP_REGNUM) + sp_offset;
+}
+
+static struct score_frame_cache *
+score_make_prologue_cache (struct frame_info *next_frame, void **this_cache)
+{
+  struct score_frame_cache *cache;
+
+  if ((*this_cache) != NULL)
+    return (*this_cache);
+
+  cache = FRAME_OBSTACK_ZALLOC (struct score_frame_cache);
+  (*this_cache) = cache;
+  cache->saved_regs = trad_frame_alloc_saved_regs (next_frame);
+
+  /* Analyze the prologue.  */
+  {
+    const CORE_ADDR pc = frame_pc_unwind (next_frame);
+    CORE_ADDR start_addr;
+
+    find_pc_partial_function (pc, NULL, &start_addr, NULL);
+    if (start_addr == 0)
+      return cache;
+    score_analyze_prologue (start_addr, pc, next_frame, *this_cache);
+  }
+
+  /* Save SP.  */
+  trad_frame_set_value (cache->saved_regs, SCORE_SP_REGNUM, cache->base);
+
+  return (*this_cache);
+}
+
+static void
+score_prologue_this_id (struct frame_info *next_frame, void **this_cache,
+                        struct frame_id *this_id)
+{
+  struct score_frame_cache *info = score_make_prologue_cache (next_frame,
+                                                              this_cache);
+  (*this_id) = frame_id_build (info->base, frame_func_unwind (next_frame));
+}
+
+static void
+score_prologue_prev_register (struct frame_info *next_frame,
+                              void **this_cache,
+                              int regnum, int *optimizedp,
+                              enum lval_type *lvalp, CORE_ADDR * addrp,
+                              int *realnump, gdb_byte * valuep)
+{
+  struct score_frame_cache *info = score_make_prologue_cache (next_frame,
+                                                              this_cache);
+  trad_frame_get_prev_register (next_frame, info->saved_regs, regnum,
+                                optimizedp, lvalp, addrp, realnump, valuep);
+}
+
+static const struct frame_unwind score_prologue_unwind =
+{
+  NORMAL_FRAME,
+  score_prologue_this_id,
+  score_prologue_prev_register
+};
+
+static const struct frame_unwind *
+score_prologue_sniffer (struct frame_info *next_frame)
+{
+  return &score_prologue_unwind;
+}
+
+static CORE_ADDR
+score_prologue_frame_base_address (struct frame_info *next_frame,
+                                   void **this_cache)
+{
+  struct score_frame_cache *info =
+    score_make_prologue_cache (next_frame, this_cache);
+  return info->base;
+}
+
+static const struct frame_base score_prologue_frame_base =
+{
+  &score_prologue_unwind,
+  score_prologue_frame_base_address,
+  score_prologue_frame_base_address,
+  score_prologue_frame_base_address,
+};
+
+static const struct frame_base *
+score_prologue_frame_base_sniffer (struct frame_info *next_frame)
+{
+  return &score_prologue_frame_base;
+}
+
+static struct gdbarch *
+score_gdbarch_init (struct gdbarch_info info, struct gdbarch_list *arches)
+{
+  struct gdbarch *gdbarch;
+
+  arches = gdbarch_list_lookup_by_info (arches, &info);
+  if (arches != NULL)
+    {
+      return (arches->gdbarch);
+    }
+  gdbarch = gdbarch_alloc (&info, 0);
+
+  set_gdbarch_short_bit (gdbarch, 16);
+  set_gdbarch_int_bit (gdbarch, 32);
+  set_gdbarch_float_bit (gdbarch, 32);
+  set_gdbarch_double_bit (gdbarch, 64);
+  set_gdbarch_long_double_bit (gdbarch, 64);
+  set_gdbarch_register_sim_regno (gdbarch, score_register_sim_regno);
+  set_gdbarch_pc_regnum (gdbarch, SCORE_PC_REGNUM);
+  set_gdbarch_sp_regnum (gdbarch, SCORE_SP_REGNUM);
+  set_gdbarch_num_regs (gdbarch, SCORE_NUM_REGS);
+  set_gdbarch_register_name (gdbarch, score_register_name);
+  set_gdbarch_breakpoint_from_pc (gdbarch, score_breakpoint_from_pc);
+  set_gdbarch_register_type (gdbarch, score_register_type);
+  set_gdbarch_frame_align (gdbarch, score_frame_align);
+  set_gdbarch_inner_than (gdbarch, core_addr_lessthan);
+  set_gdbarch_read_sp (gdbarch, score_read_sp);
+  set_gdbarch_unwind_pc (gdbarch, score_unwind_pc);
+  set_gdbarch_print_insn (gdbarch, score_print_insn);
+  set_gdbarch_skip_prologue (gdbarch, score_skip_prologue);
+  set_gdbarch_in_function_epilogue_p (gdbarch, score_in_function_epilogue_p);
+  set_gdbarch_call_dummy_location (gdbarch, AT_ENTRY_POINT);
+  set_gdbarch_return_value (gdbarch, score_return_value);
+  set_gdbarch_unwind_dummy_id (gdbarch, score_unwind_dummy_id);
+  set_gdbarch_push_dummy_call (gdbarch, score_push_dummy_call);
+
+  frame_unwind_append_sniffer (gdbarch, dwarf2_frame_sniffer);
+  frame_base_append_sniffer (gdbarch, dwarf2_frame_base_sniffer);
+  frame_unwind_append_sniffer (gdbarch, score_prologue_sniffer);
+  frame_base_append_sniffer (gdbarch, score_prologue_frame_base_sniffer);
+
+  return gdbarch;
+}
+
+extern initialize_file_ftype _initialize_score_tdep;
+
+void
+_initialize_score_tdep (void)
+{
+  gdbarch_register (bfd_arch_score, score_gdbarch_init, NULL);
+}
diff -ruN gdb-6.6/gdb/score-tdep.h gdb-6.6-score/gdb/score-tdep.h
--- gdb-6.6/gdb/score-tdep.h	1970-01-01 08:00:00.000000000 +0800
+++ gdb-6.6-score/gdb/score-tdep.h	2006-12-09 14:03:50.000000000 +0800
@@ -0,0 +1,50 @@
+/* Target-dependent code for the S+core architecture, for GDB, the GNU Debugger.
+
+   Copyright (C) 2006
+   Free Software Foundation, Inc.
+
+   Contributed by Qinwei (qinwei@sunnorth.com.cn)
+   Contributed by Ching-Peng Lin (cplin@sunplus.com)
+
+   This file is part of GDB.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 51 Franklin Street, Fifth Floor,
+   Boston, MA 02110-1301, USA.  */
+
+#ifndef SCORE_TDEP_H
+#define SCORE_TDEP_H
+#include "math.h"
+
+enum gdb_regnum
+{
+  SCORE_SP_REGNUM = 0,
+  SCORE_FP_REGNUM = 2,
+  SCORE_RA_REGNUM = 3,
+  SCORE_A0_REGNUM = 4,
+  SCORE_AL_REGNUM = 7,
+  SCORE_PC_REGNUM = 49,
+};
+
+#define SCORE_A0_REGNUM        4
+#define SCORE_A1_REGNUM        5
+#define SCORE_REGSIZE          4
+#define SCORE_NUM_REGS         56
+#define SCORE_BEGIN_ARG_REGNUM 4
+#define SCORE_LAST_ARG_REGNUM  7
+
+#define SCORE_INSTLEN          4
+#define SCORE16_INSTLEN        2
+
+#endif /* SCORE_TDEP_H */

Re: [patch] add target Score to GDB

[Jim Blandy]

秦巍<qinwei@sunnorth.com.cn> writes:
> Jim Blandy <jimb@codesourcery.com> wrote on 2006-12-07 03:41:54:
>  
>> qinwei@sunnorth.com.cn writes:
>> > If we won't open simulator source, can we put simulator libs(11 libs) 
> and 
>> > simstub.c in the dir 'gdb-6.5/sim/score'?
>> > If we can, i will send you these libs. 
>> > Otherwise, i post the patch in the following just containing 'score' 
> gdb 
>> > but without simulator.
>> 
>> The GDB source tree cannot contain non-Free software.  We give our
>> users all the rights to modify and redistribute the software that we
>> have, on the condition that they do the same for others; that's the
>> founding principle of Project GNU.  Including non-free libraries with
>> GDB would not follow that principle.
>> 
>> We encourage you to free your simulator; if that's not possible, then
>> we'll need to leave the simulator out.
>
> Hi Jim, 
> Sorry our company SunNorth doesn't want to open simulator source, please 
> leave it out.

Okay.

>> Does the S+core architecture allow you to mix 32-bit and 16-bit
>> instructions?  If so, could score_breakpoint_from_pc be simplified to
>> always return a 16-bit breakpoint?  This would mean that a breakpoint
>> inserted at a 32-bit instruction would produce two bytes of garbage
>> after the breakpoint, but that's normal; on the i386, GDB uses a
>> one-byte breakpoint instruction regardless of the length of the
>> instruction it's set at.
>
> Yes, score supports the mixture of 32-bit and 16-bit instructions.
> But score uses the word address to fetch instruction, and then identify if 
> the 32-bits data 
> is a 32-bit instruction or 2 16-bit instructions.
> Only setting 16-bit breakpoint instruction will affect score indentifing 
> instruction types
> and cause error both on hardware and simulator.

I see --- there's much more going on here than there is in most other
processors.

>> Does score_integer_to_address do exactly the same thing that GCC does
>> when the user casts an integer to an address?  If they don't, then the
>> user could become quite confused when GDB and GCC evaluate their
>> expressions differently.
>
> I didn't see clearly about this hook when i was porting score gdb.
> I have seen your NOTES in 'value_as_address'.
> The pointer and the address are the same on score target, 
> and a pointer to a function always points to the entry point address of 
> that function.
> So we don't need to implement this hook. Can i say that?

If you don't need to set it to get clean test suite results, then I
would definitely recommend leaving it out.

>> Future revisions of GCC change the sorts of prologues GDB will
>> encounter.  If you find yourself investing more time than you expected
>> on score_analyze_prologue, consider using the framework in
>> prologue-value.h, documented there.
>
> The next step i will survey other methods such as using 
> prologue-value.[ch] or 
> using .eh_frame section and will post new patches.
>
> Can this patch be checked in first?

Using prologue-value certainly isn't required; I just wanted to
suggest it.

There should be patches to Makefile.in; could you post those?  We need
to see the complete patch in order to approve it.

All changes need a ChangeLog entry.  You don't need to say much for
score-tdep.c and score-tdep.h; here's the entry for the xtensa port:

2006-11-14  Maxim Grigoriev  <maxim@tensilica.com>

	* NEWS: New port to Xtensa.
	* Makefile.in: Add dependencies for Xtensa files.
	* configure.tgt (xtensa*, xtensa*-*-elf*): New.
	* configure.host (xtensa*-*-elf*): New.
	* config/xtensa/xtensa.mt: New file.
	* xtensa-config.c: New file.
	* xtensa-tdep.h: New file.
	* xtensa-tdep.c: New file.

(I would prefer that the Makefile.in entry actually listed the
variables and rules added.)

Please keep all lines under 80 characters in length; for example,
score_unwind_dummy_id needs to be reformatted.

I'm surprised by score_type_needs_double_align --- is it really true
that struct types only need double alignment if their *first* member
does?  Why is that?  I'd expect the code for structs and unions to be
the same.

Other than that, this looks good.  Please post a complete patch,
generated against the current CVS sources, and with a ChangeLog entry,
for a final review, and then we'll get it committed.