[SLP] Vectorize for all-constant entries.
This should fix libc++'s iostream initialization SIGBUSing on amd64,
whenever the global cout symbol is not aligned to 16 bytes.
Some further explanation: libc++'s iostream.cpp contains the definitions
of std::cout, std::cerr and so on. These global objects are effectively
declared with an alignment of 8 bytes. When an executable is linked
against libc++.so, it can sometimes get a copy of the global object,
which is then at the same alignment.
However, with clang 3.7.0, the initialization of these global objects
will incorrectly use SSE instructions (e.g. movdqa), whenever the
optimization level is high enough, and SSE is enabled, such as on amd64.
When any of these objects is not aligned to 16 bytes, this will result
in a SIGBUS during iostream initialization. In contrast, clang 3.6.x
and earlier took the 8 byte alignment into consideration, and avoided
SSE for those particular operations.
After bisecting of upstream changes, I found that the above revision
caused the change of this behavior, so I am reverting it now as a
workaround, while a discussion and test case is being prepared for
upstream.
Highlights (not already in the FreeBSD tree):
- addr2line: Fixed multiple memory leaks related to DIE allocation
- readelf: improve sh_link validation
- various man page improvements
Sponsored by: The FreeBSD Foundation
set div/rem default values to 'expensive' in TargetTransformInfo's
cost model
...because that's what the cost model was intended to do.
As discussed in D12882, this fix has a temporary unintended
consequence for SimplifyCFG: it causes us to not speculate an fdiv.
However, two wrongs make PR24818 right, and two wrongs make PR24343
act right even though it's really still wrong.
I intend to correct SimplifyCFG and add to CodeGenPrepare to account
for this cost model change and preserve the righteousness for the bug
report cases.
https://llvm.org/bugs/show_bug.cgi?id=24818https://llvm.org/bugs/show_bug.cgi?id=24343
Differential Revision: http://reviews.llvm.org/D12882
This fixes the too-eager fdiv hoisting in pow(), which could lead to
unexpected floating point exceptions.
r288125, the required atomic library calls are available in compiler-rt.
The added stub for __libcpp_relaxed_store() can stay as a fallback; I
have also committed it upstream.
Add missing atomic libcall support.
Support for emitting libcalls for __atomic_fetch_nand and
__atomic_{add,sub,and,or,xor,nand}_fetch was missing; add it, and some
test cases.
Differential Revision: http://reviews.llvm.org/D10847
This fixes "cannot compile this atomic library call yet" errors when
compiling code which calls the above builtins, on arm < v6.
in libc++, on __ARM_ARCH < 6. Additionally, supply the missing stub
__libcpp_relaxed_store(), as proposed in http://reviews.llvm.org/D13051
NOTE: this needs to be fixed properly later on, by supplying library
functions implementing atomic operations for arm < v6. We should
probably take those from sys/arm/arm/stdatomic.c, and stuff them into
either libgcc or compiler-rt.
Some binaries (such as the FreeBSD kernel) contain a mixture of CUs
with and without debug information. Previously translate() exited upon
encountering a CU without debug information. Instead, just move on to
the next CU.
Reported by: royger
Reviewed by: royger
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D3712
both in /usr/lib and /usr/local/lib, thus simplifying the use of modules
from ports, without breaking the compat32 case again.
PR: 191151
MFC after: 3 weeks
Connect it to the build.
The code assumed that SCHED_* constants form a contiguous set of
numbers, remove the assumption by using schedulers[] array in
get_different_scheduler(). This is no-op on FreeBSD, but improves
code portability.
The selection of different priority used the min/max priority range of
the current scheduler class, instead of the priority to be changed to.
The bug caused the test failure.
Remove duplication of POSIX_SPAWN_SETSIGDEF flag and now unused
duplications of MIN/MAX definitions.
Reviewed by: jilles, pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D3533
Merge the end result of two upstream changes:
Original fix from 20141206:
+ modify MKlib_gen.sh to work around change in development version of
gcc introduced here:
https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02185.htmlhttps://gcc.gnu.org/ml/gcc-patches/2014-07/msg00236.html
(reports by Marcus Shawcroft, Maohui Lei).
Later fixed in different manner in 20150725:
+ use alternate workaround for gcc 5.x feature (adapted from patch by
Mikhail Peselnik).
"freebsd-sponsored-by" and "freebsd-mfc-after" in ~/.subversion/config,
in the "[miscellany]" section - just like you already can with
devel/subversion port/package.
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
I introduced a local copy of __multc3 in r281221, which has now been
committed upstream to compiler-rt in revision 245296. Update our version
to match the changes made there.
Sponsored by: The FreeBSD Foundation
- Use _Bool rather than bool to resolve missing type errors in malloc_np.h.
- Fix malloc manual page #include documentation.
- Add *allocm manual pages to obsolete files.
Submitted by: jbeich
Reverting r239883 and r240720:
------------------------------------------------------------------------
r239883 | echristo | 2015-06-17 00:09:32 -0700 (Wed, 17 Jun 2015) | 16 lines
Update the intel intrinsic headers to use the target attribute support.
This involved removing the conditional inclusion and replacing them
with target attributes matching the original conditional inclusion
and checks. The testcase update removes the macro checks for each
file and replaces them with usage of the __target__ attribute, e.g.:
int __attribute__((__target__(("sse3")))) foo(int a) {
_mm_mwait(0, 0);
return 4;
}
This usage does require the enclosing function have the requisite
__target__ attribute for inlining and code generation - also for
any macro intrinsic uses in the enclosing function. There's no change
for existing uses of the intrinsic headers.
------------------------------------------------------------------------
------------------------------------------------------------------------
r240720 | silvas | 2015-06-25 16:22:11 -0700 (Thu, 25 Jun 2015) | 6 lines
Remove `requires` for x86 CPU features.
Ever since the target attributes change, we don't need to guard these
headers with `requires`. Actually it's a bit worse, because if we do
then they are included textually under the covers, causing declarations
to appear in submodules they aren't supposed to be in.
------------------------------------------------------------------------
This reverts the changes to the intrinsics headers in trunk, which could
result in some ports' configure scripts misdetecting SSE (and higher)
support.
Update to tzdata2015f:
Changes affecting future time stamps
North Korea switches to +0830 on 2015-08-15. (Thanks to Steffen Thorsen.)
The abbreviation remains "KST". (Thanks to Robert Elz.)
Uruguay no longer observes DST. (Thanks to Steffen Thorsen and Pablo Camargo.)
Changes affecting past and future time stamps
Moldova starts and ends DST at 00:00 UTC, not at 01:00 UTC. (Thanks to Roman Tudos.)
undefined behavior. The code used this macro to avoid problems on some
broken systems which define SSIZE_MAX incorrectly, but this is not
needed on FreeBSD, obviously.
MFC after: 3 days
fixes and quality of life improvements.
While there are security issues in this time frame that affect usage as a
server (eg: linked into apache), this isn't possible here.
This includes additional functions to be protected: those that
have local array definitions, or have references to local frame
addresses. This is a new option in GCC-4.9 that was relicensed
by Han Shen from Google under GPLv2 for OpenBSD.
Obtained from: OpenBSD (2014-01-14)
MFC after: 2 weeks
ELF Tool Chain built on FreeBSD's ar and elfdump, but has a number of
improvements and enhancements. Bring them into contrib in order to start
integrating into our build.
Merge upstream fix to eliminate build-breaking gcc warnings of no
importance.
commit: cab33b7a0acba7d2268a23c4383be6167106e549
Update ND_TTEST2 to fix issue 443
Add IS_NOT_NEGATIVE macro.
Avoid these warnings:
- comparison of unsigned expression >= 0 is always true [-Wtype-limits],
- comparison is always true due to limited range of data type [-Wtype-limits].
Reviewed by: adrian
Approved by: jmallett (mentor)
MFC after: 1 month
Also, the changes made in r272451 and r272653 that were lost in the
merge of 4.6.2 (r276788) have been restored.
PR: 199568
Differential Revision: https://reviews.freebsd.org/D3007
Reviewed by: brooks, hiren
Approved by: jmallett (mentor)
MFC after: 1 month
[SCCP] Turn loads of null into undef instead of zero initialized values
Surprisingly, this is a correctness issue: the mmx type exists for
calling convention purposes, LLVM doesn't have a zero representation for
them.
This partially fixes PR23999.
Pull in r241143 from upstream llvm trunk (by David Majnemer):
[LoopUnroll] Use undef for phis with no value live
We would create a phi node with a zero initialized operand instead of
undef in the case where no value was originally available. This was
problematic for x86_mmx which has no null value.
These fix a "Cannot create a null constant of that type!" error when
compiling the graphics/sdl2_gfx port with MMX enabled.
Reported by: amdmi3
Notable upstream commits (upstream revision in parens):
- Add a JSON producer to LLDB (228636)
- Don't crash on bad DWARF expression (228729)
- Add support of DWARFv3 DW_OP_form_tls_address (231342)
- Assembly profiler for MIPS64 (232619)
- Handle FreeBSD/arm64 core files (233273)
- Read/Write register for MIPS64 (233685)
- Rework LLDB system initialization (233758)
- SysV ABI for aarch64 (236098)
- MIPS software single stepping (236696)
- FreeBSD/arm live debugging support (237303)
- Assembly profiler for mips32 (237420)
- Parse function name from DWARF DW_AT_abstract_origin (238307)
- Improve LLDB prompt handling (238313)
- Add real time signals support to FreeBSDSignals (238316)
- Fix race in IOHandlerProcessSTDIO (238423)
- MIPS64 Branch instruction emulation for SW single stepping (238820)
- Improve OSType initialization in elf object file's arch_spec (239148)
- Emulation of MIPS64 floating-point branch instructions (239996)
- ABI Plugin for MIPS32 (239997)
- ABI Plugin for MIPS64 (240123)
- MIPS32 branch emulation and single stepping (240373)
- Improve instruction emulation based stack unwinding on ARM (240533)
- Add branch emulation to aarch64 instruction emulator (240769)
The import of openssl to address the FreeBSD-SA-15:10.openssl security
advisory includes a change which rejects handshakes with DH parameters
below 768 bits. sendmail releases prior to 8.15.2 (not yet released),
defaulted to a 512 bit DH parameter setting for client connections.
The first fix committed last week changed the default to 1024 bits.
This commit fixes the case where the DHParameters option is set to a
file which doesn't exist, which is the case on newer versions of
FreeBSD which enable STARTTLS by default by auto-creating TLS
certificates.
MFC after: 2 days
Some point after gcc-4.2 the MIPS inline assembly restrictions changed -
=h (hi register) disappeared from the list of restrictions and can no
longer be used.
So, until someone requires an assembly version of this function,
just use a non-assembly version and let the compiler sort it out.
Suggested by: kan
advisory includes a change which rejects handshakes with DH parameters
below 768 bits. sendmail releases prior to 8.15.2 (not yet released),
defaulted to a 512 bit DH parameter setting for client connections.
This commit chages that default to 1024 bits. sendmail 8.15.2, when
released well use a default of 2048 bits.
MFC after: 1 day
Update to tzdata2015e:
Changes affecting future time stamps
Morocco will suspend DST from 2015-06-14 03:00 through 2015-07-19 02:00,
not 06-13 and 07-18 as we had guessed. (Thanks to Milamber.)
Assume Cayman Islands will observe DST starting next year, using US rules.
Although it isn't guaranteed, it is the most likely.
The xo_format_string_direct function loops forever never advancing the
processed string pointer when it encounters a character that makes
mbrtowc fail. Make it emit '?' character instead, as it seems this is
what the code intent was, sans bugs.
Differential Revision: https://reviews.freebsd.org/D2802
Reviewed by: marcel
As of LLVM revision 238073, LLVM stores symbols and section names in
the same string table. From the upstream commit mesage:
With the scheme of naming sections like ".text.foo" where foo is a
symbol, there is a big potential saving in using a single one.
This is a cherry-pick of ELF Tool Chain revision 3225.
Sponsored by: The FreeBSD Foundation
contrib/top/top.local.H to contrib/top/top.local.hs.
This fixes a build breakage when src is on a case-
insensitive file system -- we never properly create
top.x nor top.local.h. Change the makefile accordingly.
MFC after: 3 days
Highlights (upstream revisions):
- Fix SHT_GROUP handling in elfcopy/strip (3206 3220 3221)
- Misc elfcopy / strip bug fixes (3215 3216 3217)
- Many C++ demangler improvements (3199 3200 3201 3202 3203 3204 3205
3208 3210 3211 3212)
- Improve GNU binutils compatibility in elfcopy / strip (3213 3214)
- Add -g option to readelf(1): dump contents of section groups (3219)
- Add EM_IAMCU 32-bit Intel MCU (3198)
Also add a compat #define for building with older FreeBSD ELF headers.
The GRP_COMDAT flag was added to elf_common.h in r283110, but it's not
available during the bootstrap build. It is also convenient to be able
to build on older hosts.
Thanks to antoine@ for tracking down issues through multiple exp-runs
and to kaiw@ for fixing.
PR: 198611 (exp-run), 200350
Sponsored by: The FreeBSD Foundation
This function originated in glibc, and this matches their behaviour
(and NetBSD, OpenBSD, and musl).
An empty big string (arg "l") is handled by the existing
l_len < s_len test.
Reviewed by: bapt, ngie
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D2657
atf-sh/atf_check_test:flush_stdout_on_timeout
Many thanks for jmmv for the fix!
MFC after: 1 week
PR: 197060
Original commit message:
From 0e546407567ea858e261e72f75c5ed61e07d0ddf Mon Sep 17 00:00:00 2001
From: Julio Merino <jmmv@google.com>
Date: Tue, 17 Feb 2015 18:10:11 -0500
Subject: [PATCH] Fix atf-sh/atf_check_test:flush_stdout_on_death
The test atf-sh/atf_check_test:flush_stdout_on_timeout was flaky as it
was playing solely with time. Fix this by making the test more robust
and rename it while we are at it: there is nothing left about "timeouts"
in this test, considering that ATF itself does not enforce deadlines
any longer.
Fixes FreeBSD PR 197060.
Update to tzdata2015d:
Release 2015d - 2015-04-24 08:09:46 -0700
Changes affecting future time stamps
Egypt will not observe DST in 2015 and will consider canceling it
permanently. For now, assume no DST indefinitely.
(Thanks to Ahmed Nazmy and Tim Parenti.)
Change affecting past and future time zone abbreviations
The abbreviations for Hawaii-Aleutian standard and daylight times
have been changed from HAST/HADT to HST/HDT, as per US Government
Printing Office style. This affects only America/Adak since 1983,
as America/Honolulu was already using the new style.
Update to tzdata2015c:
Release 2015c - 2015-04-11 08:55:55 -0700
Changes affecting future time stamps
Egypt's spring-forward transition is at 24:00 on April's last Thursday,
not 00:00 on April's last Friday. 2015's transition will therefore be on
Thursday, April 30 at 24:00, not Friday, April 24 at 00:00. Similar fixes
apply to 2026, 2037, 2043, etc. (Thanks to Steffen Thorsen.)
