Re[2]: [Gc]: FreeBSD Itanium: gc-7.1 build failure...

Ivan Maidanski ivmai at
Thu Aug 29 05:24:51 PDT 2013

Hi Vitaly,

Среда, 28 августа 2013, 19:11 +03:00 от Vitaly Magerya <vmagerya at>:
> On 08/25/2013 11:15, Ivan Maidanski wrote:
> >  Hi Vitaly,
> >
> > I've applied yet more changes:
> > * fix old_segv/bus_act variables initialization:
> > * towards better portability of GC_RETURN_ADDR_PARENT/GC_ADD_CALLER:
> >
> > GC_ADD_CALLER is used in case of absence of GC_CAN_SAVE_CALL_STACKS. The latter macro is defined on Linux but not on FreeBSD, and GC_RETURN_ADDR_PARENT and GC_caller_func_offset are used (as the only alternative to missing support of GC_CAN_SAVE_CALL_STACKS) to record the caller of GC_debug_malloc/realloc_replacement. Unfortunately, __builtin_return_address(1) is not supported on all H/W (e.g., ARM or MIPS), so I currently enabled it for x86/x64.
> >
> > The branch ( ) is ready to be merged to gc7_2d and to master (excluding changes to and doc/ that are not present in  devel/ boehm - gc -threaded ). Please test it before I merge it.
> Yup, seems to work fine for me. Note that I only have FreeBSD on amd64,
> so I can't test changes related to ARM, SPARC or MIPS platforms.
> Also, could you add the test for 'enable_parallel_mark' into the
> '*-*-freebsd*)' part of the 'case $host' in It's there in
> 7_3 branch, but not in the 7_2 or 'freebsd-ports-devel-boehm-gc'. Note
> that '*-*-kfreebsd*-gnu)' has that test for both 7_3 and 7_2 branches.

Done (in release-7_2 branch).


> _______________________________________________
> Gc mailing list
> Gc at


Отправлено из мобильной Почты Mail.Ru

More information about the Gc mailing list