This is the mail archive of the crossgcc@sourceware.org mailing list for the crossgcc project.
See the CrossGCC FAQ for lots more information.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
On Fri, Mar 03, 2006 at 11:18:20AM -0800, Dan Kegel wrote: > > FYI, gcc-4.1.0-glibc-2.3.6 (not included in the arm.dat file) fails > > on ARM with: > > > > ../sysdeps/generic/s_fmax.c: In function `__fmax': > > ../sysdeps/generic/s_fmax.c:28: internal compiler error: in elim_reg_cond, at flow.c:3328 > > Right, that's because we build glibc-2.3.6 with gcc-3.3.6: > $ more gcc-4.1.0-20060223-glibc-2.3.6.dat > ... > GCC_CORE_DIR=gcc-3.3.6 > ... > > > This is a known gcc 3.3.x problem, IIRC. I remember there was some > > workaround but I forgot which. > > It'd be nice to have. Anyone remember? Ah, right, now I remember having to work around this for my Fedora Core 2 ARM port. This is the (ugly hackish) workaround. Not sure if we want to include this in crosstool. http://skrybele.wantstofly.org/fc2/SRPMS.patches/glibc-2.3.3-27/glibc-dontuse_builtin_float_functions.patch There's a debian bug report for this which never really got fixed: http://bugs.debian.org/298508 This seems to be the gcc bug report: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15068 (But strangely enough, the code that fixed by the included patch wasn't present in gcc 3.3.x at all, so maybe it's just a different bug than the one that #15068 talks about but in the same function?) crosstool 0.38 gcc-3.3.6-glibc-2.3.5 also had this issue: http://kegel.com/crosstool/crosstool-0.38/buildlogs/arm-softfloat-gcc-3.3.6-glibc-2.3.5.log.txt cheers, Lennert -- For unsubscribe information see http://sourceware.org/lists.html#faq
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |