293 Commits

Author SHA1 Message Date
jca
6fb091b7ce loongson has moved to clang
from Brad
2020-12-22 22:29:01 +00:00
jca
469b067e18 Sync with base llvm
From Brad
2020-12-21 21:38:51 +00:00
jca
86b1e95304 Sync with base llvm
From Brad
2020-11-24 22:44:12 +00:00
jca
576530f97f Sync changes from base-clang
The previous commit was missing some files.  Ah, cvs...
From Brad
2020-11-18 21:20:00 +00:00
jca
dad7c7c329 Sync with changes in base-clang
From Brad
2020-11-18 21:13:37 +00:00
jca
7fdd7da2dd Sync -fno-ret-protector list with base Makefile.inc
From Brad
2020-11-10 12:05:50 +00:00
jca
f976bd0ffb mips64el hasn't switched to clang yet
Pointed out by Brad
2020-11-07 22:21:41 +00:00
jca
33ae21d347 Sync with base: performance optimizations CXXFLAGS
This adds:
-fno-ret-protector on powerpc and mips64el
-fomit-frame-pointer on mips64 and mips64el

Prompted by a similar diff from Brad
2020-11-07 11:45:52 +00:00
jca
945b261b77 Attempt to fix build on powerpc
Apparently -mlong-calls does more harm than good these days.
From Brad.

http://build-failures.rhaalovely.net/arm/2020-06-03/devel/llvm,,-main.log
2020-11-03 21:36:30 +00:00
sthen
350ae20cdd update lang/gcc to 8.4.0, from Brad, ok pascal@
has been through bulk builds on i386 sparc64
2020-09-04 09:55:34 +00:00
naddy
c552ea72f3 bump for plist changes 6.7 -> 6.8 2020-08-31 22:20:49 +00:00
jca
03728a3308 Zap local comment that wasn't picked up upstream
From Brad Smith.
2020-08-27 00:09:09 +00:00
jca
063b398498 Sync with base
Zap outdated powerpc alignement fix, fixed upstream in a different way.
From Brad Smith.
2020-08-26 10:33:47 +00:00
jca
f07215d215 Default to libc++ on powerpc64 2020-08-11 08:28:53 +00:00
jca
91a67452c3 Follow base and update to llvm-10.0.1 2020-08-10 11:38:43 +00:00
jca
97c836c4ba Add a comment and fix a typo 2020-08-09 14:17:26 +00:00
jca
474c1dedef Drop patch, merged upstream
The patch in cvs duplicated the content of the original testcase.
2020-08-09 14:14:42 +00:00
jca
7523bb12d6 Drop two outdated comments
from Brad
2020-08-09 14:12:30 +00:00
jca
cbd1c499a9 Bump libclang.so and friends
libclang.so is used by other ports and the update to llvm 10 warranted
a minor bump.  I haven't checked liblldb.so and libLTO.so, I bumped
their major version just in case but they aren't used in the ports tree.

Thanks to kn@ who provided an old libclang.so for comparison.
2020-08-07 22:49:06 +00:00
jca
50a2a85b0a Misc changes: x86 defaults to i586, backport a fix and tweak the PLISTs
- like base-clang, change the default i386 target cpu to i586.
Same diff from Brad
- fix a crash exposed by net/libtorrent-rasterbar.  Fix already present
in 10.0.1, analyze and diff from Nam Nguyen
- some lldb-related files snuck in the 10.0.0 PLIST-main changes, move
them to PLIST-lldb as appropriate.  Reported by daniel@
2020-08-07 11:42:02 +00:00
jca
2794487c37 Update to llvm-10.0.0
Positive feedback from landry@
2020-08-05 06:49:48 +00:00
sthen
d9cfe4113e bump REVISION; python 3 default changed to 3.8 2020-07-03 21:12:24 +00:00
jca
698137a55c Yet another python2 vs 3 fix: don't pick up libpython2.7.so at configure time
libpython2.7 could be picked up even though the python3 interpreter was
detected and used.  This lead to a funny line in configure output:
-- Found PythonLibs: /usr/local/lib/libpython2.7.so.0.0 (found version "3.7.7")

Let cmake pick a library version that matches the python interpreter
instead of declaring a preference for 2.7.

Build failure reported by naddy@
2020-06-16 00:26:23 +00:00
gkoehler
01535c2e99 Mips, PowerPC changes from base-clang
Mips: Use a distinct trapcode with retguard on mips64.

PowerPC:
 - Set max atomic size for AtomicExpandPass.
 - For OpenBSD/powerpc64: switch to ELFv2, use _mcount profiling, and
   add our OpenBSD specific defines.
 - Optimize away the frame pointer.

ok jca@ (maintainer), cwen@
2020-06-15 21:34:58 +00:00
jca
b569be790c Partial sync with base-clang 2020-05-19 10:05:47 +00:00
cwen
17f3a7653c llvm: add powerpc improvements from base clang:
- fix small pic for secure-plt mode (from kettenis@)
- modify LLVM's DAGCombiner to skip an optimization if it
  would make an illegal ISD::ADDE node, fixing a fatal backend error
  (from gkoehler@ with help from mortimer@)

OK jca@ (maintainer)
2020-05-18 22:28:22 +00:00
jca
16a1c93faf On non-clang archs, g++ is already in BUILD_DEPENDS
So don't list it in TEST_DEPENDS.
2020-04-22 09:59:45 +00:00
naddy
e8092ac1dd bump for plist changes 6.6 -> 6.7 2020-04-05 15:44:52 +00:00
jca
6ace77049c Use libc++ (not libestdc++) on powerpc
sparc64 will follow once base has done the switch
2020-04-04 21:48:39 +00:00
jca
6db569706c Switch llvm to python3
While here:
- adjust python scripts shebang lines
- also byte-compile the .py files in the lldb subpackage.  One file
  needs a brute-force conversion using 2to3

Tested in a bulk by ajacoutot@, prompted by and ok kn@
2020-03-22 00:58:36 +00:00
jca
5342f82d30 Drop gtest from RUN_DEPENDS
The installed headers don't need it, except for a testsuite header not
used by the ecosystem.
Also drop two outdated PLIST lines from back when llvm tried to install
libgtest.a

Spotted while discussing the deps of the various subpackages with kn@
Tested in a bulk by ajacoutot@, ok kn@
2020-03-22 00:57:51 +00:00
jca
aa5673c526 Drop bogus test target
Long overdue...
2020-03-15 19:33:32 +00:00
naddy
47517a7414 bump the remaining subpackages; all PLISTs changed 2020-03-06 22:22:03 +00:00
jca
373996e012 Some PLIST tweaks
- don't install six.py, turns out there's a cmake config knob for that
- drop misplaced, @commented out header
- add @so and @static-lib markers

Don't use DEBUG_PACKAGES for now, as the size increase of ${WRKBUILD} is
huge.  I tried to use -gline-tables-only (I think?) as advised by kmos@
but the size of the resulting non-debug packages increased.  To be
revisited...
2020-03-06 14:39:57 +00:00
gkoehler
9322b5bc94 Add -msvr4-struct-return to ports-clang for powerpc.
This is the same change as in base-clang.
ok jca@ (maintainer)
2020-03-05 00:50:00 +00:00
semarie
07b38a5196 devel/llvm: Fix a crash bug when lowering unordered loads to cmpxchg
backport of 2153c4b828

needs for upcoming lang/rust 1.41.0 which trigger the bug on i386

ok jca@
2020-02-09 06:14:21 +00:00
jca
af04cfdc9c Bring in some retguard improvements from base
- Add retguard for octeon/mips64.
- Move the hashed __retguard_* symbols into individual sections and mark
  them as COMDATs so that the linker can individually discard them, instead
  of just ignoring duplicate symbols but keep the (duplicate) space.

from Brad
2020-01-07 22:59:43 +00:00
jca
3e91ee7a2f Merge two improvements from base-clang/lld
- Make clang emit the proper path to our libcompiler_rt.a when asked to.
- Sort relative relocations (and relocations against the same symbol) by offset
  to optimize the cache and UVM faulting behavior

from Brad
2019-12-27 00:15:46 +00:00
jca
38048dc03d Don't add a dep on all headers in ${Backtrace_INCLUDE_DIRS}
since ${Backtrace_INCLUDE_DIRS} is likely to be a system dir.
TODO: push upstream
2019-12-25 22:31:01 +00:00
jca
b106a6c4ab Error out if we end up adding a dep on all headers in /usr/local/include
Just in case an upcoming update introduces the same problem as what
happened with libexecinfo/Backtrace.
2019-12-25 21:58:23 +00:00
jca
178b2ee6a0 Properly disable libexecinfo to fix up random build failures seen by many
Passing -DBacktrace_LIBRARY='' doesn't work since the flag has been
renamed to Backtrace_LIBRARIES and isn't enough anyway to prevent the
llvm cmake files to *add a dep on all headers* in
${Backtrace_INCLUDE_DIRS}, ie /usr/local/include.

... which of course leads to builds blowing up whenever dpb(1) junking
kicks in.

Bulk builders: sorry it took me so long to get my head out of the sand.

Last report by naddy@
2019-12-25 21:55:03 +00:00
jca
3eaafeecdd Build all targets on powerpc
Supposedly ld -relax fixes the relocation overflow problems.
Diff from Brad tested by cwen@
2019-12-05 15:24:58 +00:00
jca
ab2e72f705 Use libcxx on mips64, like base clang
from Brad some time ago.
2019-11-08 17:11:50 +00:00
jca
c64dc40473 Sort PLIST
from Brad
2019-09-09 21:15:16 +00:00
jca
cbb02a6e7f Update to llvm-8.0.1
Similar diff from Brad; bulk build by ajacoutot@
2019-09-08 10:40:39 +00:00
naddy
15934c06f0 bump for plist changes 6.5 -> 6.6 2019-08-10 19:41:07 +00:00
jca
1af85b1214 Fix invalid regex
Reported by bcallah@, upstream fix pointed out by sthen@
2019-07-17 20:15:53 +00:00
jca
3335a863b8 Merge various improvements from base
o clang:
- add back kernel printf %b length specifier support (%llb, etc, lost in
the update to 8.0.0)
o lld:
- Restore previous section after setting the MIPS ABI marker
- Fix output section alignement when entry size isn't a power of two
o arm64, amd64:
- Do not store the retguard cookie in frame in leaf functions if possible
- Emit variable length trap padding in retguard epilogue
o amd64:
- move code that selects retpoline by default to a different source file
o mips64:
- Fix a bug in memory operand handling
- Implement SGE pseudo-instructions
- Implement .cplocal directive
- Fix instruction guard
- Implement the 'h' register constraint on mips64
o sparc64:
- Remove cast that truncates immediate operands to 32 bits
2019-07-09 13:21:37 +00:00
jca
f14c55be2c Major fix for -stdlib=libstdc++, properly register libOptRemarks.so
I botched the library search paths when merging OpenBSD.cpp, repair
that.  Also always add -L${LOCALBASE}/lib like was done before.
Looks like I committed an old diff which lacked SHARED_LIBS+=OptRemarks.

While here, force-disable Z3 support from math/z3.
2019-07-08 07:56:52 +00:00
jca
58f7f768f9 Whitespace and comments sync with base 2019-07-07 14:04:07 +00:00