Commit Graph

853 Commits

Author SHA1 Message Date
mrg
955b6c1f27 commonize the netbsd local version string. (the native one wasn't
updated, and now that won't be a problem again.)
2014-05-29 07:40:37 +00:00
mrg
1b1872cba4 make LINK_EH_SPEC match other usage. fixes powerpc build. 2014-05-29 07:05:10 +00:00
mrg
04408297e6 move the -std=gnu++0x code from Makefile to Makefile.common.
fixes the remaining 2/3 link issues.
2014-05-28 17:09:18 +00:00
mrg
271c711fb0 add -pthread to compiler/linker flags. fixes 1/3 of the link issues. 2014-05-28 16:41:06 +00:00
mrg
a878343648 re-run mknative-gcc after threading fixes by skrll and myself. 2014-05-28 09:09:35 +00:00
mrg
295c54e235 apply patch from skrll@: allow netbsd host=target to configure
GLIBCXX_IS_NATIVE.
2014-05-28 07:03:27 +00:00
mrg
5b741f074a rerun mknative-gcc for GCC 4.8.3 release.
XXX:  still not 100% right, but we're getting close.
2014-05-28 06:54:48 +00:00
skrll
6fac5056c2 Resolve conflicts 2014-05-27 09:18:53 +00:00
skrll
9c8f586f64 Import gcc 4.8.3 which has 98 bugs fixed on gcc-4-8-3-pre-r208254 2014-05-27 08:50:03 +00:00
skrll
0f2e98845a Import gcc 4.8.3 which has 98 bugs fixed on gcc-4-8-3-pre-r208254 2014-05-27 08:39:52 +00:00
skrll
5f4cdc7d41 Import gcc 4.8.3 which has 98 bugs fixed on gcc-4-8-3-pre-r208254 2014-05-27 08:29:46 +00:00
matt
bc43ee9904 Add LEGITIMATE_PIC_OPERAND_P(X) to prevent global symbol + offset from
being a legitimate pic operand (because it's not).
2014-05-23 13:35:26 +00:00
mrg
c16f810189 apply -fimplicit-templates instead of forcing -O2. fix the
build fix on platforms that don't use -O2 by default, and
also enables building with -O0.
2014-05-16 09:36:51 +00:00
joerg
9af9cc70aa Skip assembler helpers when using HAVE_LIBGCC_EH=no. Always use the
generic unwind.h version in that case, even on ARM.
2014-05-15 14:37:08 +00:00
martin
0b6a7ee607 Fix comment pointing at the source of the list of functions and add
two missing ones: _clrsbsi2 and _clrsbdi2
2014-05-03 16:43:10 +00:00
dogcow
cd6db90cfd fix munged patch: xcompilation on linux/non-arm systems -> netbsd/arm works again. 2014-04-30 02:04:11 +00:00
joerg
0af4f3b8a9 Don't use inline in C sources, it will certainly not work correctly with
a C99 compiler.
2014-04-26 22:38:09 +00:00
martin
17480df656 Pull up upstream revision 209789: remove the *ashlsi3_extend pattern.
This fixes gcc [Bug target/60941] and resolves a firefox crash for me.
2014-04-25 11:06:52 +00:00
christos
2d59a755a9 Returning in the middle of a TRY_CATCH macro gives heartburns since the
jump buffer is not being reset.
2014-04-21 17:32:14 +00:00
apb
3ff1cdf566 Set ac_install_sh="$SHELL $ac_aux_dir/$ac_t -c"
instead of ac_install_sh="$ac_aux_dir/$ac_t -c"
in all configure scripts in the gcc.old/dist tree.
This ensures that the install-sh script is always run
under the correct shell.
2014-04-19 15:57:46 +00:00
apb
7f226bea47 Set ac_install_sh="$SHELL $ac_aux_dir/$ac_t -c"
instead of ac_install_sh="$ac_aux_dir/$ac_t -c"
in all configure scripts in the gcc/dist tree.
This ensures that the install-sh script is always run
under the correct shell.
2014-04-19 13:00:57 +00:00
apb
441c2cf2f8 Use $(SHELL) to run move-if-change. 2014-04-19 12:41:31 +00:00
christos
035ceafe9d It is not nice to fake things we don't have; instead check explicitly
that we can get CS.
2014-04-16 02:14:44 +00:00
christos
07009f7010 take constant expressions out of the loop. 2014-04-16 00:34:10 +00:00
christos
b8c0fcebe9 supply a fake cs that has has not have SEL_UPL set, so that our trapframe
sniffer works.
2014-04-16 00:32:33 +00:00
joerg
68aeaa009b Use correct library path for -m32 on sparc64. 2014-04-14 21:09:02 +00:00
christos
ebffd37d96 remove extra 2014-04-04 01:19:57 +00:00
christos
19f79caf6f Add NOCTF 2014-04-04 01:19:14 +00:00
joerg
0e80d9e124 Before messing with macros from stdio.h, include cstdio. Some parts of
the backend would include it afterwards and the macro changes conflict
with libc++'s version of cstdio.
2014-03-29 21:33:39 +00:00
matt
1a3430eb55 Bias the cfi pc expr by nbytes since vax pcrel are relative to end of the
relocation, not the beginning.  By adding the size of relocation, it will
become relative to the start of the relocation.
2014-03-25 16:17:33 +00:00
joerg
9fe3724ad4 Cast to _Unwind_Ptr here as well. 2014-03-25 10:03:51 +00:00
ozaki-r
9a1cd017d1 Cast ptr to the correct return type (_Unwind_Ptr) 2014-03-25 00:50:45 +00:00
joerg
3b142775a7 Cast context->ldsa to the correct return type. 2014-03-24 22:39:06 +00:00
joerg
47a96e3a4d _Unwind_GetLanguageSpecificData is supposed to return a _Unwind_Ptr.
This makes a difference on M68K as it uses a different register for
integer vs pointer values.
2014-03-24 21:26:01 +00:00
martin
e5db261cf2 Avoid undefined behaviour (which causes traps on VAX). 2014-03-22 20:36:45 +00:00
tron
adbf0b7734 Remove (harmless) debug target committed by accident.
Pointed out by Matthew Green in private e-mail.
2014-03-22 19:52:21 +00:00
tron
0bb5c0ea6c Fix build with "HAVE_GCC=48" and "USE_SSP=yes". 2014-03-22 17:57:47 +00:00
joerg
b204cbfda6 For HAVE_LIBGCC_EH=no, skip libgcc's unwind code. 2014-03-21 20:54:13 +00:00
joerg
c629e629e3 Annotation for the Processor Status Word. Sensible DWARF frame layout. 2014-03-20 16:53:55 +00:00
skrll
584cd3c172 Remove a file that should be generated and fix the build 2014-03-20 15:21:46 +00:00
mrg
44125d6efb regenerate after recent vax specific fixes. 2014-03-20 10:15:13 +00:00
joerg
89ecf27180 Always use --eh-frame-hdr, even when linking statically. 2014-03-20 01:46:19 +00:00
matt
b47c61b618 Add support for the .cfi pseudo-ops for VAX. 2014-03-18 23:41:36 +00:00
matt
95d660c12b Make readelf know what the vax pcrel32 reloc number is. 2014-03-18 23:41:02 +00:00
skrll
9bf4ee7d90 In fill_fpregset correct the address to grab the register from 2014-03-18 08:40:05 +00:00
uwe
fb3a642893 G/c extraneous TAB. 2014-03-16 00:07:58 +00:00
mrg
9c0370d59c add missing -DGENERATOR_FILE to HOST_CPPFLAGS. 2014-03-15 23:12:58 +00:00
mrg
4d51ddd922 make sure we use the C compiler (not C++) for CC= when running the
configure for auto-build.h.  fixes problems on linux.
2014-03-15 20:46:38 +00:00
christos
14e72322f6 Don't SSP a file that uses alloca. 2014-03-14 12:17:53 +00:00
martin
74a4cee79e Help the hack verifying malloc alignment expectations and force the
word size definition on sh* to be a compile time constant. We do not
provide a single library supporting both variants, so we pick the
smaller alignment for now.
2014-03-13 19:33:23 +00:00