This is the mail archive of the
mailing list for the GCC project.
Re: PATCH RFC: Remove fork from collect2
- From: Andrew Pinski <pinskia at physics dot uc dot edu>
- To: Zack Weinberg <zack at codesourcery dot com>
- Cc: Caroline Tice <ctice at apple dot com>, "gcc at gcc dot gnu dot org list" <gcc at gcc dot gnu dot org>, Ian Lance Taylor <ian at wasabisystems dot com>, Andrew Pinski <apinski at apple dot com>, Andrew Pinski <pinskia at physics dot uc dot edu>
- Date: Sat, 20 Mar 2004 16:36:56 -0800
- Subject: Re: PATCH RFC: Remove fork from collect2
- References: <email@example.com> <20040317092131.GA95030@lorien.int.gecko.de> <firstname.lastname@example.org> <20040317133116.GA15850@lorien.int.gecko.de> <email@example.com> <20040317170336.GA16667@lorien.int.gecko.de> <firstname.lastname@example.org> <email@example.com> <A7570576-79EC-11D8-9F81-000393BB90B6@apple.com> <009F425B-79EE-11D8-92E2-00039372607E@apple.com> <firstname.lastname@example.org> <email@example.com> <347BCA1C-7ACA-11D8-973A-000393A6D2F2@physics.uc.edu> <firstname.lastname@example.org>
On Mar 20, 2004, at 16:18, Zack Weinberg wrote:
Looks good in principle, but two questions:
1) Is private_extern really appropriate for routines that will go into
libgcc_s.so? (There is a LIB2FUNCS_STATIC_EXTRA variable to use
if the routines should go only into libgcc.a.)
You are right they should go into LIB2FUNCS_STATIC_EXTRA, I did not know
about that variable, I should have looked more, I had looked a little
only found LIB2ADDEH which was definitely wrong.
2) Are the .private_extern directive and -force_cpusubtype_ALL command
line option accepted by all Darwin assemblers? (Note I have no
idea if there is more than one assembler for Darwin.)
Yes they are accepted by all Darwin assemblers from 1.0 (Mac OS X 10.0)
7.2.0 (Mac OS X 10.3.2). There is only one assembler for Darwin right
a heavy modified gas 1.38. Also the FSF GCC uses private_extern for
hidden and -force_cpusubtype_ALL when you supply -maltivec respectively
Also, a meta-note: please use Content-Type text/plain and text/x-patch
for attached source files and patches (respectively), not
application/octet-stream. And it is easier for someone who needs to
dig a patch out of the mailing list archives, years after the fact, if
you include new files in the patch (use 'cvs add' and 'cvs diff -N' --
you can safely do 'cvs add' for files [not directories] before a patch
I had forgot about this as I was used to making patches without cvs
access and that was hard to do a "cvs add".
Ok, here is the patch which I am testing right now.
* config/rs6000/t-darwin (LIB2FUNCS_STATIC_EXTRA): Add
darwin-vecsave.asm, and darwin-world.asm.
(TARGET_LIBGCC2_CFLAGS): Add -Wa,-force_cpusubtype_ALL as the asm files
contain altivec instructions.
* config/rs6000/darwin-fpsave.asm: New file.
* config/rs6000/darwin-vecsave.asm: New file.
* config/rs6000/darwin-world.asm: New file.
Description: Text document