#13 Error: opcode `unpkbl' not supported for target ev56

open
Mans Rullgard
None
8
2005-03-06
2005-03-02
Will L G
No

Receiving the error below compiling on the following
hardware/software:
RedHat 7.2
GCC-3.2.3
Binutils-2.15
Make-3.80

gcc-3.2.3 -c -O3 -mtune=ev56 -mcpu=ev56 -mieee -
funroll-loops -mmax -ffast-math -fstrict-aliasing -pipe -
fexpensive-optimizations -ffloat-store -finline -Winline -
foptimize-register-move -foptimize-sibling-calls -fforce-
addr -fforce-mem -fomit-frame-pointer -I/usr/include -
D_REENTRANT -D_LARGEFILE_SOURCE -
D_FILE_OFFSET_BITS=64 -I. -I.. -Wall -o
postprocess.o postprocess.c
In file included from postprocess.c:10:
postprocess.h:99:7: warning: no newline at end of file
gcc-3.2.3 -c -O3 -mtune=ev56 -mcpu=ev56 -mieee -
funroll-loops -mmax -ffast-math -fstrict-aliasing -pipe -
fexpensive-optimizations -ffloat-store -finline -Winline -
foptimize-register-move -foptimize-sibling-calls -fforce-
addr -fforce-mem -fomit-frame-pointer -I/usr/include -
D_REENTRANT -D_LARGEFILE_SOURCE -
D_FILE_OFFSET_BITS=64 -I. -I.. -Wall -o swscale.o
swscale.c
swscale.c: In function `yuv2packedXinC':
swscale.c:651: warning: unused variable `r'
swscale.c:651: warning: unused variable `b'
swscale.c:651: warning: unused variable `g'
swscale.c:659: warning: unused variable `r'
swscale.c:659: warning: unused variable `b'
swscale.c:659: warning: unused variable `g'
In file included from swscale.c:707:
swscale_template.c: In function `yuv2packedX_C':
swscale_template.c:853: warning: unused variable
`dummy'
In file included from swscale.c:707:
swscale_template.c: In function `swScale_C':
swscale_template.c:2549: warning: unused variable
`lumMmxFilter'
swscale_template.c:2550: warning: unused variable
`chrMmxFilter'
gcc-3.2.3 -c -O3 -mtune=ev56 -mcpu=ev56 -mieee -
funroll-loops -mmax -ffast-math -fstrict-aliasing -pipe -
fexpensive-optimizations -ffloat-store -finline -Winline -
foptimize-register-move -foptimize-sibling-calls -fforce-
addr -fforce-mem -fomit-frame-pointer -I/usr/include -
D_REENTRANT -D_LARGEFILE_SOURCE -
D_FILE_OFFSET_BITS=64 -I. -I.. -Wall -o rgb2rgb.o
rgb2rgb.c
In file included from rgb2rgb.c:16:
../libvo/fastmemcpy.h: In function `mem2agpcpy_pic':
../libvo/fastmemcpy.h:29: warning: implicit declaration of
function `memcpy'
{standard input}: Assembler messages:
{standard input}:3430: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3431: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3442: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3446: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3450: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3451: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3461: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3465: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3470: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3475: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3486: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3490: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3497: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3498: Error: opcode `unpkbl' not
supported for target ev56
{standard input}:3514: Error: opcode `unpkbw' not
supported for target ev56
{standard input}:3519: Error: opcode `unpkbw' not
supported for target ev56
make[2]: *** [rgb2rgb.o] Error 1
make[2]: Leaving directory `/usr2/www/pub/alpha-
RH7/multimedia/mplayerxp/mplayerxp-
0.5.0/mplayerxp/postproc'
make[1]: *** [postproc/libpostproc.a] Error 2
make[1]: Leaving directory `/usr2/www/pub/alpha-
RH7/multimedia/mplayerxp/mplayerxp-0.5.0/mplayerxp'
make: *** [all] Error 2

Discussion

  • Will L G
    Will L G
    2005-03-02

    • priority: 5 --> 8
     
  • Will L G
    Will L G
    2005-03-02

    Logged In: YES
    user_id=312992

    Well managed to get past that error by setting my CFLAGS
    to include 'mcpu=pca56'. The PCA56 and faster cpu,
    supports multimedia extensions. THen I promptly received a
    GP_RELATIVE error in regards to RTjpegN.c. I then added '-
    fpic' to my CFLAGS and then recompiled again. This time I
    received the error below:

    gcc -c -O3 -mtune=pca56 -mcpu=pca56 -mieee -funroll-
    loops -mmax -ffast-math -fstrict-aliasing -pipe -fexpensive-
    optimizations -ffloat-store -finline -Winline -foptimize-register-
    move -foptimize-sibling-calls -fforce-addr -fforce-mem -fomit-
    frame-pointer -I/usr/include -fpic -o sr1.o sr1.c
    In file included from sr1.c:24:
    mpg123.h:19: warning: useless keyword or type name in
    empty declaration
    mpg123.h:19: warning: empty declaration
    sr1.c:437: warning: `__stdcall__' attribute directive ignored
    sr1.c:438: warning: `__stdcall__' attribute directive ignored
    sr1.c:439: warning: `__stdcall__' attribute directive ignored
    sr1.c:440: warning: `__stdcall__' attribute directive ignored
    sr1.c:441: warning: `__stdcall__' attribute directive ignored
    gcc -c -O3 -mtune=pca56 -mcpu=pca56 -mieee -funroll-
    loops -mmax -ffast-math -fstrict-aliasing -pipe -fexpensive-
    optimizations -ffloat-store -finline -Winline -foptimize-register-
    move -foptimize-sibling-calls -fforce-addr -fforce-mem -fomit-
    frame-pointer -I/usr/include -fpic -o decode_MMX.o
    decode_MMX.c
    decode_MMX.c: In function `synth_1to1_MMX_s_16':
    decode_MMX.c:54: error: unknown register name `%ebx' in
    `asm'
    decode_MMX.c:54: error: unknown register name `%esi' in
    `asm'
    decode_MMX.c:54: error: unknown register name `%edi' in
    `asm'
    make[2]: *** [decode_MMX.o] Error 1
    make[2]: Leaving directory `/usr2/www/pub/alpha-
    RH7/multimedia/mplayerxp/mplayerxp-0.5.0/codecs/mp3lib'
    make[1]: *** [all] Error 2
    make[1]: Leaving directory `/usr2/www/pub/alpha-
    RH7/multimedia/mplayerxp/mplayerxp-0.5.0/codecs'
    make: *** [all] Error 2

    Unfortunately my computer doesn't support Intel MMX
    multimedia extensions :-(

     
  • nickols_k
    nickols_k
    2005-03-04

    • assigned_to: nobody --> mru
    • status: open --> open-accepted
     
  • nickols_k
    nickols_k
    2005-03-04

    Logged In: YES
    user_id=5556

    >{standard input}:3430: Error: opcode `unpkbl' not supported
    for target ev56
    These optimizations were added by Mans Rullgard! You should
    investigate why configure produces '#define HAVE_MVI' for ev56.
    IMHO this block should be avoided by right configured mplayerxp.
    For detail see: rgb2rgb_template.c:1502

    >sr1.c:437: warning: `__stdcall__' attribute directive ignored
    Fixed!

    >decode_MMX.c
    >decode_MMX.c: In function `synth_1to1_MMX_s_16':
    >decode_MMX.c:54: error: unknown register name `%ebx' in

    My codecs/mp3lib/Makefile constains:
    ifeq ($(TARGET_ARCH_X86),yes)
    SRCS += decode_MMX.c dct64_MMX.c tabinit_MMX.c

    So it's strange why your compiler tries to compile these
    sources :(

     
  • Mans Rullgard
    Mans Rullgard
    2005-03-04

    Logged In: YES
    user_id=148780

    If configure sets both HAVE_MVI and
    TARGET_ARCH_X86, something is badly broken.
    Are you using a clean copy of the source?

     
  • Will L G
    Will L G
    2005-03-04

    Logged In: YES
    user_id=312992

    The section immediately following the lines you pointed out,
    appear to be compiling all files with the suffixe *.c and
    outputting them to *.o...

    OBJS += dct36_3dnow.o dct64_3dnow.o decode_3dnow.o
    SRCS += dct64_k7.c
    OBJS += dct64_k7.o
    endif

    .SUFFIXES: .c .o

    # .PHONY: all clean

    .c.o:
    $(CC) -c $(CFLAGS) -o $@ $<

    .s.o:
    $(CC) -c $(CFLAGS) -o $@ $<

    $(LIBNAME): $(OBJS)
    $(LD) --shared -soname $@ -o $@ $(OBJS) -lc -lm

     
  • Will L G
    Will L G
    2005-03-04

    Logged In: YES
    user_id=312992

    Oh, one other thing... I ran 'make -d' to toss the output to the
    terminal and this is what I got:
    No need to remake target `sr1.c'.
    Finished prerequisites of target file `sr1.o'.
    Prerequisite `sr1.c' is older than target `sr1.o'.
    No need to remake target `sr1.o'.
    Considering target file `decode_MMX.o'.
    File `decode_MMX.o' does not exist.
    Looking for an implicit rule for `decode_MMX.o'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.c'.
    Found an implicit rule for `decode_MMX.o'.
    Considering target file `decode_MMX.c'.
    Looking for an implicit rule for `decode_MMX.c'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.y'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.l'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.w'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.w'.
    Trying pattern rule with stem `decode_MMX.c'.
    Trying implicit prerequisite `decode_MMX.c,v'.
    Trying pattern rule with stem `decode_MMX.c'.
    Trying implicit prerequisite `RCS/decode_MMX.c,v'.
    Trying pattern rule with stem `decode_MMX.c'.
    Trying implicit prerequisite `RCS/decode_MMX.c'.
    Trying pattern rule with stem `decode_MMX.c'.
    Trying implicit prerequisite `s.decode_MMX.c'.
    Trying pattern rule with stem `decode_MMX.c'.
    Trying implicit prerequisite `SCCS/s.decode_MMX.c'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.y'.
    Looking for a rule with intermediate file `decode_MMX.y'.
    Avoiding implicit rule recursion.
    Trying pattern rule with stem `decode_MMX.y'.
    Trying implicit prerequisite `decode_MMX.y,v'.
    Trying pattern rule with stem `decode_MMX.y'.
    Trying implicit prerequisite `RCS/decode_MMX.y,v'.
    Trying pattern rule with stem `decode_MMX.y'.
    Trying implicit prerequisite `RCS/decode_MMX.y'.
    Trying pattern rule with stem `decode_MMX.y'.
    Trying implicit prerequisite `s.decode_MMX.y'.
    Trying pattern rule with stem `decode_MMX.y'.
    Trying implicit prerequisite `SCCS/s.decode_MMX.y'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.l'.
    Looking for a rule with intermediate file `decode_MMX.l'.
    Avoiding implicit rule recursion.
    Trying pattern rule with stem `decode_MMX.l'.
    Trying implicit prerequisite `decode_MMX.l,v'.
    Trying pattern rule with stem `decode_MMX.l'.
    Trying implicit prerequisite `RCS/decode_MMX.l,v'.
    Trying pattern rule with stem `decode_MMX.l'.
    Trying implicit prerequisite `RCS/decode_MMX.l'.
    Trying pattern rule with stem `decode_MMX.l'.
    Trying implicit prerequisite `s.decode_MMX.l'.
    Trying pattern rule with stem `decode_MMX.l'.
    Trying implicit prerequisite `SCCS/s.decode_MMX.l'.
    Trying pattern rule with stem `decode_MMX'.
    Trying implicit prerequisite `decode_MMX.w'.
    Looking for a rule with intermediate file `decode_MMX.w'.
    Avoiding implicit rule recursion.
    Trying pattern rule with stem `decode_MMX.w'.
    Trying implicit prerequisite `decode_MMX.w,v'.
    Trying pattern rule with stem `decode_MMX.w'.
    Trying implicit prerequisite `RCS/decode_MMX.w,v'.
    Trying pattern rule with stem `decode_MMX.w'.
    Trying implicit prerequisite `RCS/decode_MMX.w'.
    Trying pattern rule with stem `decode_MMX.w'.
    Trying implicit prerequisite `s.decode_MMX.w'.
    Trying pattern rule with stem `decode_MMX.w'.
    Trying implicit prerequisite `SCCS/s.decode_MMX.w'.
    Trying pattern rule with stem `decode_MMX'.
    Rejecting impossible implicit prerequisite
    `decode_MMX.w'.
    No implicit rule found for `decode_MMX.c'.
    Finished prerequisites of target file `decode_MMX.c'.
    No need to remake target `decode_MMX.c'.
    Finished prerequisites of target file `decode_MMX.o'.
    Must remake target `decode_MMX.o'.

     
  • Will L G
    Will L G
    2005-03-04

    Logged In: YES
    user_id=312992

    You're right it is over looking the "ifeq" statement. When I
    remarked the offending lines out, it continued with the build...
    check 'other' error msg...

     
  • Will L G
    Will L G
    2005-03-04

    • status: open-accepted --> closed-accepted
     
  • nickols_k
    nickols_k
    2005-03-05

    Logged In: YES
    user_id=5556

    >You're right it is over looking the "ifeq" statement. When I
    >remarked the offending lines out, it continued with the
    build...
    >check 'other' error msg...
    Which version of bash, coreutils, glibc and so on do you use?
    mplayerxp requires GNU stuff to be compiled!
    It seems that we will fix your system instead of mplayerxp
    if you will send such reports ;)

     
  • Will L G
    Will L G
    2005-03-05

    Logged In: YES
    user_id=312992

    I have the following versions:
    Bash GNU bash, version 2.05.8
    ls (coreutils) 5.2.1
    glibc-2.3.2-5

    Already reinstalled GLIBC, I'm going to try reinstalling
    Coreutils and see what happens...

     
  • Will L G
    Will L G
    2005-03-05

    Logged In: YES
    user_id=312992

    Well I'm going to try updating GLIBC-2.3.4 and see what
    happens. I just have to get this mutha to compile properly ;-)
    Talk to you in a bit Mns

     
  • Will L G
    Will L G
    2005-03-06

    Logged In: YES
    user_id=312992

    Well I reinstalled :
    Coreutils-5.2.1 + GLIB-2.3.2 + Make-3.80 and GTK-2.6.4 for
    that matter...

    The problem still persists... I wouldn't mind taking Mans up
    on his offer [news:yw1xy8d2j39u.fsf@ford.inprovide.com] but I
    don't like installing a fist full of dependencies and the like. I
    tend to be 'window-isk' and like my proggies self contained.

     
  • Will L G
    Will L G
    2005-03-06

    • status: closed-accepted --> open
     
  • nickols_k
    nickols_k
    2005-03-08

    Logged In: YES
    user_id=5556

    Could you please investigate by studying configure.log or
    similar things -
    why configure script defines HAVE_MVI for ev56?
    >You're right it is over looking the "ifeq" statement. When I
    >remarked the offending lines out, it continued with the
    build...
    >check 'other' error msg...
    It's not a way :(