[PATCH] PowerPC support for SID.

[PATCH] PowerPC support for SID.

[Johan Rydberg]

[-- Attachment #1: Type: text/plain, Size: 6891 bytes --]


Hi

Since there are a few new files, I made a tarball of them (which
is attached to this mail).  Comments are more than welcome!

[component/cgen-cpu]

2002-05-17  Johan Rydberg  <jrydberg@rtmk.org>

        * compCGEN.cxx: If SIDTARGET_PCC, add hw-cpu-ppc support.
	* configure.in: Decend into ppc if necessary.
	* configure: Regenerate.


[component/gloss]                                                
	                                             
2002-05-17  Johan Rydberg  <jrydberg@rtmk.org>

        * compGloss.cxx: If SIDTARGET_PCC, add sw-gloss-ppc/libgloss and
        sw-gloss-ppc/bug support.

        * Makefile.am (PPCGLOSS): Define if SIDTARGET_PPC is defined.
        * Makefine.in: Regenerate.                                   

[bsp]                             

2002-05-17  Johan Rydebrg  <jrydberg@rtmk.org>

        * configrun-sid.in: Add PowerPC support.
	* Makefile.am: Likewise.                
        * Makefile.in: Regenerate.

        * powerpc-elf-sid: New file.


Index: bsp/Makefile.am
===================================================================
RCS file: /cvs/src/src/sid/bsp/Makefile.am,v
retrieving revision 1.7
diff -u -u -r1.7 Makefile.am
--- bsp/Makefile.am     17 Dec 2001 09:31:17 -0000      1.7
+++ bsp/Makefile.am     17 May 2002 16:45:09 -0000
@@ -20,6 +20,10 @@
 FAM3SCRIPTS = m68k-elf-sid
 endif
 
+if SIDTARGET_PPC
+FAM4SCRIPTS = powerpc-elf-sid
+endif
+
 if SIDTARGET_XSTORMY16
 FAM10SCRIPTS = xstormy16-elf-sid
 endif
Index: bsp/configrun-sid.in
===================================================================
RCS file: /cvs/src/src/sid/bsp/configrun-sid.in,v
retrieving revision 1.34
diff -u -u -r1.34 configrun-sid.in
--- bsp/configrun-sid.in        28 Dec 2001 21:16:48 -0000      1.34
+++ bsp/configrun-sid.in        17 May 2002 16:45:10 -0000
@@ -14,6 +14,7 @@
 %cpu_comptype = (
                 "arm" => "hw-cpu-arm7t",
                 "thumb" => "hw-cpu-arm7t",
+                "ppc" => "hw-cpu-ppc",
                 "x86" => "hw-cpu-x86",
                 # INSERT NEW ENTRIES HERE
                 "m32r" => "hw-cpu-m32r/d",
@@ -21,6 +22,7 @@
                 );
 
 %cpu_defaultendian = (
+                     "ppc" => "-EB"
                      "x86" => "-EL",
                      "m32r" => "-EB",
                      "xstormy16" => "-EL",
@@ -29,6 +31,7 @@
 %gloss_comptype = (
                   "arm" => "sw-gloss-arm/angel",
                   "thumb" => "sw-gloss-arm/angel",
+                  "ppc" => "sw-gloss-ppc/bug",
                   "x86" => "sw-gloss-generic/libgloss",
                   # INSERT NEW ENTRIES HERE
                   "m32r" => "sw-gloss-m32r/libgloss",
@@ -38,6 +41,7 @@
 %gloss_memspecs = (
                   "arm" => "0x00000000,0x00800000",
                   "thumb" => "0x00000000,0x00800000",
+                  "ppc" => "0x00000000,0x00800000",
                   "x86" => "0x00000,0x500000",
                   # INSERT NEW ENTRIES HERE
                   "m32r" => "0x00000000,0x00800000",
Index: component/cgen-cpu/compCGEN.cxx
===================================================================
RCS file: /cvs/src/src/sid/component/cgen-cpu/compCGEN.cxx,v
retrieving revision 1.8
diff -u -u -r1.8 compCGEN.cxx
--- component/cgen-cpu/compCGEN.cxx     7 Feb 2002 17:32:28 -0000       1.8
+++ component/cgen-cpu/compCGEN.cxx     17 May 2002 16:45:10 -0000
@@ -26,6 +26,9 @@
 #if SIDTARGET_M32R
 #include "m32rbf.h"
 #endif
+#if SIDTARGET_PPC
+#include "ppcbf.h"
+#endif
 #if SIDTARGET_XSTORMY16
 #include "xstormy16.h"
 #endif
@@ -280,6 +283,9 @@
 #if SIDTARGET_ARM
   types.push_back ("hw-cpu-arm7t");
 #endif
+#if SIDTARGET_PPC
+  types.push_back ("hw-cpu-ppc");
+#endif
 #if SIDTARGET_XSTORMY16
   types.push_back ("hw-cpu-xstormy16");
 #endif
@@ -298,6 +304,10 @@
 #if SIDTARGET_ARM
   if (typeName == "hw-cpu-arm7t")
     return new arm7f::arm7f_cpu ();
+#endif
+#if SIDTARGET_PPC
+  if (typeName == "hw-cpu-ppc")
+    return new ppcbf::ppcbf_cpu ();
 #endif
 #if SIDTARGET_XSTORMY16
   if (typeName == "hw-cpu-xstormy16")
Index: component/cgen-cpu/configure.in
===================================================================
RCS file: /cvs/src/src/sid/component/cgen-cpu/configure.in,v
retrieving revision 1.3
diff -u -u -r1.3 configure.in
--- component/cgen-cpu/configure.in     17 Dec 2001 09:31:19 -0000      1.3
+++ component/cgen-cpu/configure.in     17 May 2002 16:45:10 -0000
@@ -37,6 +37,9 @@
 if test $sidtarget_m32r -eq 1; then
        cpu_subdirs="$cpu_subdirs m32r"
 fi
+if test $sidtarget_ppc -eq 1; then
+       cpu_subdirs="$cpu_subdirs ppc"
+fi
 if test $sidtarget_xstormy16 -eq 1; then
        cpu_subdirs="$cpu_subdirs xstormy16"
 fi
Index: component/gloss/Makefile.am
===================================================================
RCS file: /cvs/src/src/sid/component/gloss/Makefile.am,v
retrieving revision 1.3
diff -u -u -r1.3 Makefile.am
--- component/gloss/Makefile.am 4 Aug 2001 11:20:36 -0000       1.3
+++ component/gloss/Makefile.am 17 May 2002 16:45:10 -0000
@@ -31,9 +31,17 @@
 M32RDATA =
 endif
 
+if SIDTARGET_PPC
+PPCGLOSS = ppc.cxx
+PPCDATA =
+else
+PPCGLOSS =
+PPCDATA =
+endif
+
 libgloss_la_SOURCES = \
        compGloss.cxx gloss.cxx hostops.cxx newlib.cxx \
-       $(MIPSGLOSS) $(ARMGLOSS) $(M32RGLOSS)
+       $(MIPSGLOSS) $(ARMGLOSS) $(M32RGLOSS) $(PPCGLOSS)
 
 libgloss_la_LDFLAGS = -module -no-undefined
 
Index: component/gloss/compGloss.cxx
===================================================================
RCS file: /cvs/src/src/sid/component/gloss/compGloss.cxx,v
retrieving revision 1.1
diff -u -u -r1.1 compGloss.cxx
--- component/gloss/compGloss.cxx       7 Dec 2000 19:30:50 -0000       1.1
+++ component/gloss/compGloss.cxx       17 May 2002 16:45:10 -0000
@@ -19,6 +19,9 @@
 #if SIDTARGET_MIPS
 #include "mips.h"
 #endif
+#if SIDTARGET_PPC
+#include "ppc.h"
+#endif
 #include "gloss.h"
 
 using namespace std;
@@ -39,6 +42,10 @@
 #if SIDTARGET_MIPS
   types.push_back("sw-gloss-mips32/idt");
 #endif
+#if SIDTARGET_PPC
+  types.push_back("sw-gloss-ppc/libgloss");
+  types.push_back("sw-gloss-ppc/bug");
+#endif
   types.push_back("sw-gloss-generic/libgloss");
   return types;
 }
@@ -60,6 +67,12 @@
   if (typeName == "sw-gloss-mips32/idt")
     return new mips32_idt();
 #endif
+#if SIDTARGET_PPC
+  if (typeName == "sw-gloss-ppc/libgloss")
+    return new ppc_libgloss();
+  if (typeName == "sw-gloss-ppc/bug")
+    return new ppc_ppcbug();
+#endif
   if (typeName == "sw-gloss-generic/libgloss")
     return new gloss32();
   return 0;
@@ -83,8 +96,12 @@
   mips32_idt* g3 = dynamic_cast<mips32_idt*>(c);
   if (g3) { delete g3; return; }
 #endif
-  gloss32* g4 = dynamic_cast<gloss32*>(c);
+#if SIDTARGET_PPC
+  ppc_libgloss* g4 = dynamic_cast<ppc_libgloss*>(c);
   if (g4) { delete g4; return; }
+#endif
+  gloss32* g5 = dynamic_cast<gloss32*>(c);
+  if (g5) { delete g5; return; }
 }


[-- Attachment #2: sid-ppc.tgz --]
[-- Type: application/x-compressed-tar, Size: 110647 bytes --]

Re: [PATCH] PowerPC support for SID.

[Ben Elliston]

Hi Johan,

>>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:

  Johan> Since there are a few new files, I made a tarball of them (which
  Johan> is attached to this mail).  Comments are more than welcome!

I have a question.  Did you write testsuites for all of this stuff?

Cheers, Ben

Re: [PATCH] PowerPC support for SID.

[Johan Rydberg]

On 2002.05.17 19:03 Ben Elliston wrote:
: Hi Johan,
: 
: >>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:
: 
:   Johan> Since there are a few new files, I made a tarball of them (which
:   Johan> is attached to this mail).  Comments are more than welcome!
: 
: I have a question.  Did you write testsuites for all of this stuff?

No, I have not.  But I've done some testing against PSIM!  Frankly,
I have not had the time to write a testsuite.

Re: [PATCH] PowerPC support for SID.

[Ben Elliston]

>>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:

  Johan> : I have a question.  Did you write testsuites for all of this stuff?

  Johan> No, I have not.  But I've done some testing against PSIM!  Frankly,
  Johan> I have not had the time to write a testsuite.

Can you describe the testing methodology you used, then?  What target
software did you run on the simulator?

Ben

Re: [PATCH] PowerPC support for SID.

[Johan Rydberg]

On 2002.05.17 19:24 Ben Elliston wrote:
: >>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:
: 
:   Johan> : I have a question.  Did you write testsuites for all of this stuff?
: 
:   Johan> No, I have not.  But I've done some testing against PSIM!  Frankly,
:   Johan> I have not had the time to write a testsuite.
: 
: Can you describe the testing methodology you used, then?  What target
: software did you run on the simulator?

Just some simple tests, like hello world and such.  Also some tests to
verify the insn semantics.  

Are you not intrested in a non-verifyd port?  If not, I guess it will
do with just the CGEN cpu description.

brgds
johan

Re: [PATCH] PowerPC support for SID.

[Ben Elliston]

>>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:

  Johan> : Can you describe the testing methodology you used, then?  What target
  Johan> : software did you run on the simulator?

  Johan> Just some simple tests, like hello world and such.  Also some tests to
  Johan> verify the insn semantics.  

Where are the tests you used to verify the insn semantics?

If you are interested, there's not much work in getting a DejaGnu
board defintion in place to run make check-gcc on ppc-elf-sid.  Then
you can run the GCC execute.exp tests on the simulator.

Ben

Re: [PATCH] PowerPC support for SID.

[Johan Rydberg]

On 2002.05.17 20:13 Ben Elliston wrote:
: >>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:
: 
:   Johan> : Can you describe the testing methodology you used, then?  What target
:   Johan> : software did you run on the simulator?
: 
:   Johan> Just some simple tests, like hello world and such.  Also some tests to
:   Johan> verify the insn semantics.  
: 
: Where are the tests you used to verify the insn semantics?

Somewhere in /dev/null I think.
 
: If you are interested, there's not much work in getting a DejaGnu
: board defintion in place to run make check-gcc on ppc-elf-sid.  Then
: you can run the GCC execute.exp tests on the simulator.

I will look into that.

regards
johan

Re: [PATCH] PowerPC support for SID.

[Ben Elliston]

>>>>> "Johan" == Johan Rydberg <jrydberg@rtmk.org> writes:

  Johan> : Where are the tests you used to verify the insn semantics?

  Johan> Somewhere in /dev/null I think.

:-(

  Johan> : If you are interested, there's not much work in getting a DejaGnu
  Johan> : board defintion in place to run make check-gcc on ppc-elf-sid.  Then
  Johan> : you can run the GCC execute.exp tests on the simulator.

  Johan> I will look into that.

Consider duplicating the arm-sid.exp baseboard definition and making
the obvious changes (s/arm/ppc).  Then, for staters, build GCC for
your target and run:

	make RUNTESTFLAGS="execute.exp" check

I expect you might have a few immediate problems.  Inspect the gcc.log
DejaGnu log file for clues as to how the simulator is being invoked.
If you have any problems, you can contact me directly for help.

Cheers, Ben