https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
Richard Henderson rth at gcc dot gnu.org changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #22 from howarth at bromo dot med.uc.edu ---
I did a quick and dirty test in the i386/libffi build directory of rebuilding
and rerunning make check as I reverted each source file and header being
compiled on darwin to the pre-r184997
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #23 from howarth at bromo dot med.uc.edu ---
The failures in Comment 8 at -m32 on x86_64 darwin appear to be eliminated with
this change...
Index: libffi/include/ffi.h.in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #26 from howarth at bromo dot med.uc.edu ---
(In reply to Richard Henderson from comment #25)
Thanks for applying the fix to trunk. FYI, the current state of libffi head on
x86_64-apple-darwin14 is described in...
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #25 from Richard Henderson rth at gcc dot gnu.org ---
Author: rth
Date: Thu Jan 15 21:01:21 2015
New Revision: 219688
URL: https://gcc.gnu.org/viewcvs?rev=219688root=gccview=rev
Log:
PR libffi/64572
* include/ffi.h.in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
Richard Henderson rth at gcc dot gnu.org changed:
What|Removed |Added
Status|WAITING |ASSIGNED
---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
Richard Henderson rth at gcc dot gnu.org changed:
What|Removed |Added
Status|ASSIGNED|WAITING
---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #13 from howarth at bromo dot med.uc.edu ---
(In reply to Dominique d'Humieres from comment #8)
=== libffi Summary for unix/-m32 ===
# of expected passes 1870
# of unexpected failures 20
# of
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #18 from Dominique d'Humieres dominiq at lps dot ens.fr ---
What about the other files used by darwin?
darwin.S is also unchanged.
Those are files specific to darwin, but I don't think they are the only one
used to build libffi.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #17 from Richard Henderson rth at gcc dot gnu.org ---
(In reply to Dominique d'Humieres from comment #15)
What about the other files used by darwin?
darwin.S is also unchanged.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #15 from Dominique d'Humieres dominiq at lps dot ens.fr ---
Running target unix-m64
ERROR: couldn't load description file for unix-m64
I think it is pr64607.
Well, then you'll have to figure out what else could have changed
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #16 from H.J. Lu hjl.tools at gmail dot com ---
This may be a dup of PR 64607 or PR 64581.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #19 from Richard Henderson rth at gcc dot gnu.org ---
(In reply to Dominique d'Humieres from comment #18)
Those are files specific to darwin, but I don't think they are the only one
used to build libffi.
99%.
Anyway, it doesn't
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #21 from howarth at bromo dot med.uc.edu ---
One potential issue could be that src/x86/ffitarget.h has changed
FFI_TRAMPOLINE_SIZE from 10 to 12 for X86_DARWIN which is used in darwin.S...
#define RAW_CLOSURE_CIF_OFFSET
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #12 from Dominique d'Humieres dominiq at lps dot ens.fr ---
I can confirm these now. Executing 'make -k check' in
/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/x86_64-apple-
darwin14.1.0/i386/libffi doesn't cause -m32 to be
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #11 from howarth at bromo dot med.uc.edu ---
(In reply to Dominique d'Humieres from comment #8)
I can confirm these now. Executing 'make -k check' in
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #8 from Dominique d'Humieres dominiq at lps dot ens.fr ---
Confirmed that the proposed patch eliminates both the bootstrap failure
on x86_64-apple-darwin14 and the libffi testsuite failures at both -m32 and
-m64.
When running
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #9 from Richard Henderson rth at gcc dot gnu.org ---
Author: rth
Date: Tue Jan 13 15:29:47 2015
New Revision: 219538
URL: https://gcc.gnu.org/viewcvs?rev=219538root=gccview=rev
Log:
PR libffi/64572
* src/x86/ffitarget.h
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #10 from howarth at bromo dot med.uc.edu ---
(In reply to Dominique d'Humieres from comment #8)
I tested with a brute force 'make -k check' in both
/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/x86_64-apple-darwin14.1.0/libffi
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #1 from howarth at bromo dot med.uc.edu ---
Observed with...
../gcc-5-20150112/configure --prefix=/sw --prefix=/sw/lib/gcc5.0
--mandir=/sw/share/man --infodir=/sw/lib/gcc5.0/info
--enable-languages=c,c++,fortran,lto,objc,obj-c++,java
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
Richard Henderson rth at gcc dot gnu.org changed:
What|Removed |Added
Status|UNCONFIRMED |WAITING
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #3 from howarth at bromo dot med.uc.edu ---
(In reply to Richard Henderson from comment #2)
At the bottom of libffi/configure.host, try
X86_64_DARWIN)
SOURCES=darwin64_c.c darwin64.S
+ TARGET=X86_64
;;
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #4 from howarth at bromo dot med.uc.edu ---
The same failures are seen in the libffi test suite at -m32 and -m64. The first
failure appears as...
./cls_align_complex_float.exe
48 0.00,-0.00i 255 0 0.00,0.00i 0: 48
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
Richard Henderson rth at gcc dot gnu.org changed:
What|Removed |Added
Status|WAITING |ASSIGNED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #7 from howarth at bromo dot med.uc.edu ---
(In reply to Richard Henderson from comment #6)
Created attachment 34432 [details]
proposed patch
Should fix both the original report and the testsuite failures.
Confirmed that the
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64572
--- Comment #6 from Richard Henderson rth at gcc dot gnu.org ---
Created attachment 34432
-- https://gcc.gnu.org/bugzilla/attachment.cgi?id=34432action=edit
proposed patch
Should fix both the original report and the testsuite failures.
26 matches
Mail list logo
