Re: libffi on m68k failures

[Alan Hourihane <alanh@fairlite.co.uk>]

On 14/01/12 20:04, Thorsten Glaser wrote:
> Andreas Schwab dixit:
>
> > Thorsten Glaser<tg@mirbsd.de>  writes:
> >
> > > The floating point errors are to be expected (on the emulator),
> > No, they aren't.
> OK. With that patch, we’re here:

Attached a patch to fix the err_bad_abi.c failure.

The return_sc problem looks like it could be a GCC bug with 
optimizations as it doesn't happen with -O0.

Not sure what the nested_struct2.c problem is.

Alan.

> Test Run By root on Sat Jan 14 18:04:27 2012
> Native configuration is m68k-unknown-linux-gnu
>
>                  === libffi tests ===
>
> Schedule of variations:
>      unix
>
> Running target unix
> Using /usr/share/dejagnu/baseboards/unix.exp as board description file for target.
> Using /usr/share/dejagnu/config/unix.exp as generic interface file for target.
> Using /tmp/buildd/libffi-3.0.10/testsuite/config/default.exp as tool-and-target-specific interface file.
> Running /tmp/buildd/libffi-3.0.10/testsuite/libffi.call/call.exp ...
> FAIL: libffi.call/err_bad_abi.c -O0 -W -Wall execution test
> FAIL: libffi.call/nested_struct2.c -O0 -W -Wall execution test
> FAIL: libffi.call/err_bad_abi.c -O2 execution test
> FAIL: libffi.call/nested_struct2.c -O2 execution test
> FAIL: libffi.call/return_sc.c -O2 execution test
> FAIL: libffi.call/err_bad_abi.c -O3 execution test
> FAIL: libffi.call/nested_struct2.c -O3 execution test
> FAIL: libffi.call/return_sc.c -O3 execution test
> FAIL: libffi.call/err_bad_abi.c -Os execution test
> FAIL: libffi.call/nested_struct2.c -Os execution test
> FAIL: libffi.call/return_sc.c -Os execution test
> FAIL: libffi.call/err_bad_abi.c -O2 -fomit-frame-pointer execution test
> FAIL: libffi.call/nested_struct2.c -O2 -fomit-frame-pointer execution test
> FAIL: libffi.call/return_sc.c -O2 -fomit-frame-pointer execution test
> Running /tmp/buildd/libffi-3.0.10/testsuite/libffi.special/special.exp ...
>
>                  === libffi Summary ===
>
> # of expected passes            1630
> # of unexpected failures        14
> # of unsupported tests          15
>
>
> HTH&  HAND,
> //mirabilos

--- src/m68k/ffi.c.old	2012-01-16 12:35:40.383137560 +0000
+++ src/m68k/ffi.c	2012-01-16 12:36:10.036224809 +0000
@@ -261,7 +261,8 @@
 		      void *user_data,
 		      void *codeloc)
 {
-  FFI_ASSERT (cif->abi == FFI_SYSV);
+  if (cif->abi != FFI_SYSV)
+    return FFI_BAD_ABI;
 
   *(unsigned short *)closure->tramp = 0x207c;
   *(void **)(closure->tramp + 2) = codeloc;