Problems using -mrtd and -mshort for m68k

[Scott Howard]

Thank you for your detailed bug report.  We have identified a possible
solution which we will be testing over the next couple of days.  I will
let you know how it turns out.

Scott Howard
scott@objsw.com

Arjan Koers wrote:
> 
> Hi,
> 
> >this may be a bug in the m68k machine description for gcc.
> >
> >In order to confirm it, can you please re-run your compile and link
> >commands, but add the option '-v' to the each of the command lines, to
> >generate a verbose listing of the compiler's activity; and then please
> >post the output to the list.
> 
> I have compiled and linked with the -v option now:
> 
> D:\GCC-TEST>gcc -g -v -b 68k -mcpu32 -mshort -mrtd -c -O1 -o test.o
> test.c
> Reading specs from d:/xgcc/68k/2_8_1/specs
> gcc driver version 2.8.1 executing gcc version egcs-2.90.29
>  d:/xgcc/68k/2_8_1/cpp.exe -lang-c -v -iprefix d:/xgcc/68k/2_8_1/
> -undef -D__GNUC__=2 -D__GNUC_MINOR__=90 -Dmc68000 -D__embedded__
> -D__mc68000__ -D__embedded__ -D__mc68000 -Asystem(embedded)
> -Amachine(mc68000) -D__OPTIMIZE__ -g -D__INT_MAX__=32767 -Dmcpu32
> -D__mcpu32__ -D__mcpu32 test.c d:/xgcc/tmp/ccaaaaaa.i
> GNU CPP version egcs-2.90.29 980515 (egcs-1.0.3 release) (68k,
> Motorola syntax)
> #include "..." search starts here:
> #include <...> search starts here:
>  d:/xgcc/68k/2_8_1/include
> End of search list.
>  d:/xgcc/68k/2_8_1/cc1.exe d:/xgcc/tmp/ccaaaaaa.i -quiet -dumpbase
> test.c -mcpu32 -mshort -mrtd -g -O1 -version -o d:/xgcc/tmp/ccaaaaaa.s
> GNU C version egcs-2.90.29 980515 (egcs-1.0.3 release) (68k) compiled
> by GNU C version 2.8.1.
> test.c: In function `main':
> test.c:4: warning: return type of `main' is not `int'
>  d:/xgcc/68k/2_8_1/as.exe -mcpu32 -o test.o d:/xgcc/tmp/ccaaaaaa.s
> 
> D:\GCC-TEST>gcc -g -v -b 68k -mcpu32 -mshort -mrtd -o test.cof test.o
> -Ttest.ld
> Reading specs from d:/xgcc/68k/2_8_1/specs
> gcc driver version 2.8.1 executing gcc version egcs-2.90.29
>  d:/xgcc/68k/2_8_1/ld.exe -o test.cof
> d:/xgcc/68k/2_8_1/mshort/mrtd/mcpu32/crt0.o
> -Ld:/xgcc/68k/2_8_1/mshort/mrtd/mcpu32 -Ld:/xgcc/68k/2_8_1 -Ld:/xgcc
> test.o -lgcc -lgcc -Ttest.ld
> 
> Here's the relevant part of the generated code:
> 
> 00030026 <main>:
> double main(int i){
>    30026:       4e56 ffe0       linkw %fp,#-32
>    3002a:       48e7 3020       moveml %d2-%d3/%a2,%sp@-
>    3002e:       346e 0008       moveaw %fp@(8),%a2
>    30032:       4eb9 0003 0006  jsr 30006 <__main>
> double d;
> char ch[30];
> d = (double)i;
>    30038:       344a            moveaw %a2,%a2
>    3003a:       2f0a            movel %a2,%sp@-
>    3003c:       4eb9 0003 006c  jsr 3006c <__floatsidf>
>    30042:       588f            addql #4,%sp
>    30044:       2400            movel %d0,%d2
>    30046:       2601            movel %d1,%d3
> strcpy(&ch[0], "Just some string for testing");
>    30048:       3f3c 001d       movew #29,%sp@-
>    3004c:       486e ffe2       pea %fp@(-30)
>    30050:       487a ffb6       pea %pc@(30008 <.text>)
>    30054:       4eb9 0003 0280  jsr 30280 <bcopy>
> return d;
>    3005a:       2002            movel %d2,%d0
>    3005c:       2203            movel %d3,%d1
> }
>    3005e:       4cee 040c ffd4  moveml %fp@(-44),%d2-%d3/%a2
>    30064:       4e5e            unlk %fp
>    30066:       4e74 0002       rtd #2
> 
> 0003006c <__floatsidf>:
>    3006c:       4e56 0000       linkw %fp,#0
>    ....
>    300e6:       4e5e            unlk %fp
>    300e8:       4e74 0004       rtd #4
> 
> 00030280 <bcopy>:
>    30280:       4e56 0000       linkw %fp,#0
>    30284:       2f2e 0010       movel %fp@(16),%sp@-
>    30288:       2f2e 0008       movel %fp@(8),%sp@-
>    3028c:       2f2e 000c       movel %fp@(12),%sp@-
>    30290:       4eb9 0003 029c  jsr 3029c <memmove>
>    30296:       4e5e            unlk %fp
>    30298:       4e74 000c       rtd #12
> 
> The problems:
> 
> Argument for __floatsidf is popped twice (-mtrd problem):
>    30042:       588f            addql #4,%sp
>    300e8:       4e74 0004       rtd #4
> This problem occurrs for all fp conversion functions.
> 
> The size of the 'number of characters' argument for bcopy differs
> (-mshort problem):
>    30048:       3f3c 001d       movew #29,%sp@-
>    30284:       2f2e 0010       movel %fp@(16),%sp@-
>    30298:       4e74 000c       rtd #12
> 
> Arjan Koers.
> _______________________________________________
> New CrossGCC FAQ: http://www.objsw.com/CrossGCC
> _______________________________________________
> To remove yourself from the crossgcc list, send
> mail to crossgcc-request@cygnus.com with the
> text 'unsubscribe' (without the quotes) in the
> body of the message.
_______________________________________________
New CrossGCC FAQ: http://www.objsw.com/CrossGCC
_______________________________________________
To remove yourself from the crossgcc list, send
mail to crossgcc-request@cygnus.com with the
text 'unsubscribe' (without the quotes) in the
body of the message.