This is the mail archive of the cgen@sourceware.org mailing list for the CGEN project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: "make gas-test" broken?


Dave Korn wrote:
Hi,

  I figured it was something I was doing wrong in my port, but maybe not;
seems to happen for any cpu I try, e.g.:


$ make gas-test ARCHFILE=/gnu/cgen/src/cgen/../cgen/cpu/play.cpu "`if test -f ../guile/libguile/guile ; then echo ../guile/libguile/guile; else echo guile ; fi`" -l /gnu/cgen/src/cgen/guile.scm -s /gnu/cgen/src/cgen/cgen-gas.scm \ -s /gnu/cgen/src/cgen \ -v \ -a /gnu/cgen/src/cgen/../cgen/cpu/play.cpu \ -i "all" \ -m "all" \ -B gas-build.sh \ -E gas-allinsn.exp Skipping slib/sort, already loaded. Skipping slib/random, already loaded. cgen -s /gnu/cgen/src/cgen/cgen-gas.scm -s /gnu/cgen/src/cgen -v -a /gnu/cgen/src/cgen/../cgen/cpu/play.cpu -i all -m all -B gas-build.sh -E gas-allinsn.exp Loading cpu description /gnu/cgen/src/cgen/../cgen/cpu/play.cpu ... Including file /gnu/cgen/src/cgen/../cgen/cpu/simplify.inc ... Analyzing instruction set ... Done analysis. Generating gas-build.sh ... ERROR: In procedure string-append: ERROR: Wrong type (expecting string): fp No backtrace available. make: *** [gas-test] Error 1

$ make gas-test ARCHFILE=/gnu/cgen/src/cgen/../cgen/cpu/m68k.cpu
"`if test -f ../guile/libguile/guile ; then echo ../guile/libguile/guile; else
echo guile ; fi`" -l /gnu/cgen/src/cgen/guile.scm -s
/gnu/cgen/src/cgen/cgen-gas.scm \
                -s /gnu/cgen/src/cgen \
                -v \
                -a /gnu/cgen/src/cgen/../cgen/cpu/m68k.cpu \
                -i "all" \
                -m "all" \
                -B gas-build.sh \
                -E gas-allinsn.exp
Skipping slib/sort, already loaded.
Skipping slib/random, already loaded.
cgen -s /gnu/cgen/src/cgen/cgen-gas.scm -s /gnu/cgen/src/cgen -v -a
/gnu/cgen/src/cgen/../cgen/cpu/m68k.cpu -i all -m all -B gas-build.sh -E
gas-allinsn.exp
Loading cpu description /gnu/cgen/src/cgen/../cgen/cpu/m68k.cpu ...
Including file /gnu/cgen/src/cgen/../cgen/cpu/simplify.inc ...
Analyzing instruction set ...
Done analysis.
Generating gas-build.sh ...
ERROR: In procedure string-append:
ERROR: Wrong type (expecting string): m68k
No backtrace available.
make: *** [gas-test] Error 1


Hmm, until I saw that "m68k" I thought it was always pseudo-register names, I've seen "fp" and "lr" before. Anyway, manually invoking the guile command while adding --debug doesn't help any with the backtrace. I was following the instructions in 5.7 of the manual; is there a list of interactive commands I could run (like those in 5.5 and 5.8) manually that might help shed some light on what's going wrong?

    cheers,
      DaveK


Hi. Yeah, it is broken. Bitrot. I fixed a few issues, and still have a few more to look into.

Does anyone have docs on the m32rx ISA? I can find m32r ISA docs from the Renesas link on the main webpage, but I need m32rx. The f-accs field is 2 bits yet there are only 2 h/w accumulators. This should be flagged at .cpu load time, but it's currently crashing gas-test.

btw, here's how I'm testing gas-test.

bash$ cd src/cgen
bash$ guile
guile> (load "dev.scm")
guile> (load-gtest)
guile> (cload #:arch "../cpu/m32r.cpu") ; replace with path to your cpu file of course
guile> (cgen-build.sh)
guile> (cgen-all-insn.exp)


That should run without any errors.

Also, to get a backtrace when invoking gas-test from the command line:

bash$ make gas-test CGENFLAGS='-v -b'


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]