Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Karel Gardas]

Hello,

todays gcc-3_4-branch sources fails to compile for powerpc-eabisim target.
The failure looks:

gcc   -g -O2 -DIN_GCC -DCROSS_COMPILE  -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long
-DHAVE_CONFIG_H  -o cc1 \
        c-parse.o c-lang.o c-pretty-print.o stub-objc.o attribs.o
c-errors.o c-lex.o c-pragma.o c-decl.o c-typeck.o c-convert.o c-aux-info.o
c-common.o c-opts.o c-format.o c-semantics.o c-incpath.o cppdefault.o
c-ppoutput.o c-cppbuiltin.o prefix.o c-objc-common.o c-dump.o c-pch.o
libcpp.a rs6000-c.o main.o libbackend.a ../libiberty/libiberty.a
libbackend.a(function.o)(.text+0x6861): In function `assign_parms':
../../combined/gcc/function.c:5245: undefined reference to
`set_decl_incoming_rtl'
collect2: ld returned 1 exit status
make[1]: *** [cc1] Error 1
make[1]: Leaving directory `/mnt/karel/gcc340-sim/obj/gcc'
make: *** [all-gcc] Error 2


The compiler and environment was configured with:

thinkpad:/mnt/karel/gcc340-sim/obj$ ../combined/configure
--target=powerpc-eabisim --prefix=/mnt/karel/gcc340-sim/powerpc-eabisim
--with-newlib --disable-gdbtk --enable-languages=c++


Cheers,

Karel
--
Karel Gardas                  kgardas@objectsecurity.com
ObjectSecurity Ltd.           http://www.objectsecurity.com

Re: Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Andrew Pinski]

On Feb 19, 2004, at 08:00, Karel Gardas wrote:

> Hello,
>
> todays gcc-3_4-branch sources fails to compile for powerpc-eabisim 
> target.
> The failure looks:
>
> gcc   -g -O2 -DIN_GCC -DCROSS_COMPILE  -W -Wall -Wwrite-strings
> -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long
> -DHAVE_CONFIG_H  -o cc1 \
>         c-parse.o c-lang.o c-pretty-print.o stub-objc.o attribs.o
> c-errors.o c-lex.o c-pragma.o c-decl.o c-typeck.o c-convert.o 
> c-aux-info.o
> c-common.o c-opts.o c-format.o c-semantics.o c-incpath.o cppdefault.o
> c-ppoutput.o c-cppbuiltin.o prefix.o c-objc-common.o c-dump.o c-pch.o
> libcpp.a rs6000-c.o main.o libbackend.a ../libiberty/libiberty.a
> libbackend.a(function.o)(.text+0x6861): In function `assign_parms':
> ../../combined/gcc/function.c:5245: undefined reference to
> `set_decl_incoming_rtl'
> collect2: ld returned 1 exit status
> make[1]: *** [cc1] Error 1
> make[1]: Leaving directory `/mnt/karel/gcc340-sim/obj/gcc'
> make: *** [all-gcc] Error 2
>
>
> The compiler and environment was configured with:
>
> thinkpad:/mnt/karel/gcc340-sim/obj$ ../combined/configure
> --target=powerpc-eabisim --prefix=/mnt/karel/gcc340-sim/powerpc-eabisim
> --with-newlib --disable-gdbtk --enable-languages=c++


This was caused by:
2004-02-19  Alan Modra  <amodra@bigpond.net.au>

         * function.c (assign_parms): When building decl_rtl for
         SPLIT_COMPLEX_ARGS, ensure inner modes of concat match outer.

Basically set_decl_incoming_rtl is not on the branch but on the 
mainline.


Thanks,
Andrew Pinski

Re: Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Karel Gardas]

On Thu, 19 Feb 2004, Andrew Pinski wrote:

> > The compiler and environment was configured with:
> >
> > thinkpad:/mnt/karel/gcc340-sim/obj$ ../combined/configure
> > --target=powerpc-eabisim --prefix=/mnt/karel/gcc340-sim/powerpc-eabisim
> > --with-newlib --disable-gdbtk --enable-languages=c++
>
>
> This was caused by:
> 2004-02-19  Alan Modra  <amodra@bigpond.net.au>
>
>          * function.c (assign_parms): When building decl_rtl for
>          SPLIT_COMPLEX_ARGS, ensure inner modes of concat match outer.
>
> Basically set_decl_incoming_rtl is not on the branch but on the
> mainline.

FYI: When I tried to revert last revision of function.c file, the build
still fails with:

gcc -c   -g -O2 -DIN_GCC -DCROSS_COMPILE  -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long
-DHAVE_CONFIG_H    -I. -I. -I../../combined/gcc -I../../combined/gcc/.
-I../../combined/gcc/../include  ../../combined/gcc/function.c -o
function.o
../../combined/gcc/function.c: In function `aggregate_value_p':
../../combined/gcc/function.c:4269: error: `hard_regno_nregs' undeclared
(first use in this function)
../../combined/gcc/function.c:4269: error: (Each undeclared identifier is
reported only once
../../combined/gcc/function.c:4269: error: for each function it appears
in.)
../../combined/gcc/function.c: In function `assign_parms':
../../combined/gcc/function.c:4654: warning: implicit declaration of
function `set_decl_incoming_rtl'
../../combined/gcc/function.c: In function `allocate_struct_function':
../../combined/gcc/function.c:6417: warning: implicit declaration of
function `DECL_STRUCT_FUNCTION'
../../combined/gcc/function.c:6417: error: invalid lvalue in assignment
../../combined/gcc/function.c: In function `prepare_function_start':
../../combined/gcc/function.c:6443: warning: assignment makes pointer from
integer without a cast
../../combined/gcc/function.c: In function `keep_stack_depressed':
../../combined/gcc/function.c:7501: error: `hard_regno_nregs' undeclared
(first use in this function)
make[1]: *** [function.o] Error 1
make[1]: Leaving directory `/mnt/karel/gcc340-sim/obj/gcc'
make: *** [all-gcc] Error 2


Is there anything wrong with my assumption that this patch touches only
this one file?

Thanks,

Karel
--
Karel Gardas                  kgardas@objectsecurity.com
ObjectSecurity Ltd.           http://www.objectsecurity.com

Re: Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Jim Wilson]

Karel Gardas wrote:
> FYI: When I tried to revert last revision of function.c file, the build
> still fails with:

I thnk you accidentally reverted to a mainline copy of function.c, 
instead of the previous revision on the gcc-3.4 branch.  The gcc-3.4 
branch is now fixed so you shouldn't need any work around here.
-- 
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com

Re: Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Karel Gardas]

On Fri, 20 Feb 2004, Jim Wilson wrote:

> Karel Gardas wrote:
> > FYI: When I tried to revert last revision of function.c file, the build
> > still fails with:
>
> I thnk you accidentally reverted to a mainline copy of function.c,
> instead of the previous revision on the gcc-3.4 branch.  The gcc-3.4
> branch is now fixed so you shouldn't need any work around here.

Yes, you are right, I've been able to build powerpc-eabisim build after
next update.

BTW: It seems I've forgotten CVS basics after nearly two years of using
Arch version control for our own project.

Cheers,

Karel
--
Karel Gardas                  kgardas@objectsecurity.com
ObjectSecurity Ltd.           http://www.objectsecurity.com

Re: Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Alan Modra]

On Thu, Feb 19, 2004 at 08:12:08AM -0800, Andrew Pinski wrote:
> On Feb 19, 2004, at 08:00, Karel Gardas wrote:
> >../../combined/gcc/function.c:5245: undefined reference to
> >`set_decl_incoming_rtl'
> 
> This was caused by:
> 2004-02-19  Alan Modra  <amodra@bigpond.net.au>

Sorry.  Obvious fix applied.

	* function.c (assign_parms): Correct leakage of mainline code
	in last commit.

Index: gcc/function.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/function.c,v
retrieving revision 1.483.4.3
diff -u -p -r1.483.4.3 function.c
--- gcc/function.c	19 Feb 2004 08:05:46 -0000	1.483.4.3
+++ gcc/function.c	19 Feb 2004 22:34:39 -0000
@@ -5242,7 +5242,7 @@ assign_parms (tree fndecl)
 		  imag = gen_lowpart_SUBREG (inner, imag);
 		}
 	      tmp = gen_rtx_CONCAT (DECL_MODE (parm), real, imag);
-	      set_decl_incoming_rtl (parm, tmp);
+	      DECL_INCOMING_RTL (parm) = tmp;
 	      fnargs = TREE_CHAIN (fnargs);
 	    }
 	  else

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre

Re: Bootstrap failure on gcc-3_4-branch for powerpc-eabisim target.

[Alan Modra]

On Fri, Feb 20, 2004 at 09:16:03AM +1030, Alan Modra wrote:
> On Thu, Feb 19, 2004 at 08:12:08AM -0800, Andrew Pinski wrote:
> > On Feb 19, 2004, at 08:00, Karel Gardas wrote:
> > >../../combined/gcc/function.c:5245: undefined reference to
> > >`set_decl_incoming_rtl'
> > 
> > This was caused by:
> > 2004-02-19  Alan Modra  <amodra@bigpond.net.au>
> 
> Sorry.  Obvious fix applied.
> 
> 	* function.c (assign_parms): Correct leakage of mainline code
> 	in last commit.

Oh dear, I see I also leaked code from my INIT_CUMULATIVE_ARGS patch.

	* function.c (assign_parms): ....
	....  Also leakage from INIT_CUMULATIVE_ARGS patch.

Committed.

Index: gcc/function.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/function.c,v
retrieving revision 1.483.4.4
diff -u -p -r1.483.4.4 function.c
--- gcc/function.c	19 Feb 2004 22:45:14 -0000	1.483.4.4
+++ gcc/function.c	19 Feb 2004 22:54:30 -0000
@@ -4351,7 +4351,7 @@ assign_parms (tree fndecl)
 #ifdef INIT_CUMULATIVE_INCOMING_ARGS
   INIT_CUMULATIVE_INCOMING_ARGS (args_so_far, fntype, NULL_RTX);
 #else
-  INIT_CUMULATIVE_ARGS (args_so_far, fntype, NULL_RTX, fndecl, -1);
+  INIT_CUMULATIVE_ARGS (args_so_far, fntype, NULL_RTX, fndecl);
 #endif
 
   /* We haven't yet found an argument that we must push and pretend the

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre