1638 Commits

Author SHA1 Message Date
John Brawn
94fd963315 [ARM] Restructure cpu handling in the driver to mostly use the triple
Using the target cpu to determine some behaviour is sprinkled in
several places in the driver, but in almost all the information that
is needed can be found in the triple. Restructure things so that the
triple is used, and the cpu is only used if the exact cpu name is
needed.

Also add a check that the -mcpu argument is valid, and correct the
-march argument checking so that it handles -march=native correctly. I
would have liked to move these checks into the computation of the
triple, but the triple is calculated several times in several places
and that would lead to multiple error messages for the same thing.

Differential Revision: http://reviews.llvm.org/D9879

llvm-svn: 237894
2015-05-21 12:19:49 +00:00
Richard Smith
31d1de2229 [OpenMP] Make default OpenMP library (the one selected with just -fopenmp)
configurable in the CMake build. There shouldn't be any change in default
behavior.

Derived from a patch by Daniel Jasper!

llvm-svn: 237850
2015-05-20 22:48:44 +00:00
Alexey Bataev
db39021cee [OPENMP] -fopenmp enables OpenMP support (fix for http://llvm.org/PR23492)
-fopenmp turns on OpenMP support and links libiomp5 as OpenMP library. Also there is -fopenmp={libiomp5|libgomp} option that allows to override effect of -fopenmp and link libgomp library (if -fopenmp=libgomp is specified).
Differential Revision: http://reviews.llvm.org/D9736

llvm-svn: 237769
2015-05-20 04:24:19 +00:00
David Majnemer
8db9176d68 [clang-cl] Enable C++14 when targeting 2015 compatibility
llvm-svn: 237553
2015-05-18 04:49:30 +00:00
Ikhlas Ajbar
03404c14fe Remove unused function HasPICArg().
llvm-svn: 237374
2015-05-14 17:42:20 +00:00
Ikhlas Ajbar
522e6196f7 Factor out SmallDataThreshold
This patch factors out SmallDataThreshold code.

llvm-svn: 237364
2015-05-14 13:52:08 +00:00
Vladimir Sukharev
c6dab75bd4 [ARM] Add v8.1a architecture
Add support for ARMv8.1a architecture. 

Briefly it is described on http://community.arm.com/groups/processors/blog/2014/12/02/the-armv8-a-architecture-and-its-ongoing-development

Reviewers: 	jmolloy, rengolin

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D8799

llvm-svn: 237349
2015-05-14 08:25:18 +00:00
Renato Golin
cb3b0c6eeb Change TargetParser enum names to avoid macro conflicts (clang)
sys/time.h on Solaris (and possibly other systems) defines "SEC" as "1"
using a cpp macro.  The result is that this fails to compile.

Fixes https://llvm.org/PR23482

llvm-svn: 237113
2015-05-12 10:34:10 +00:00
Justin Bogner
c7701240ed Re-apply "Driver: Make profiling flags work with -nostdlib on Darwin"
This time without a stray "true" in an argument list.

This reverts r237077, restoring r237074.

llvm-svn: 237091
2015-05-12 05:44:36 +00:00
Justin Bogner
7b6c0e02ac Revert "Driver: Make profiling flags work with -nostdlib on Darwin"
This revert r237074. These tests are failing all over the place.

llvm-svn: 237077
2015-05-12 01:04:33 +00:00
Justin Bogner
f44ddae71e Driver: Make profiling flags work with -nostdlib on Darwin
Compiler-rt's Profiling library isn't part of the stdlib, so -nostdlib
shouldn't prevent it from being linked. This makes Darwin behave like
other toolchains, and link in the profile runtime irrespective of
-nostdlib, since the resulting program can't be run unless you link
this.

I've also added a test to show that other toolchains already behave
like this.

llvm-svn: 237074
2015-05-12 00:31:33 +00:00
Douglas Katzman
f36dddf426 [Sparc] Add support for 'sparcel' to clang.
Differential Revision: http://reviews.llvm.org/D8784

llvm-svn: 237001
2015-05-11 15:21:44 +00:00
Ikhlas Ajbar
71d19f3cd6 Factor out Hexagon code to build args.
This patch factor out the code in hexagon::Link::ConstructJob to be reused
in other functions. No functionality change intended.

llvm-svn: 236926
2015-05-09 02:13:25 +00:00
Renato Golin
0600e1ebe9 Using ARMTargetParser in Clang
This is a starting point for using the TargetParser in Clang, in a simple
enough part of the code that can be used without disrupting the crazy
platform support that we need to be compatible with other toolchains.

Also adding a few FIXME on obvious places that need replacing, but those
cases will indeed break a few of the platform assumptions, as arch/cpu names
change multiple times in the driver.

Finally, I'm changing the "neon-vfpv3" behaviour to match standard NEON, since
-mfpu=neon implies vfpv3 by default in both Clang and LLVM. That option
string is still supported as an alias to "neon".

llvm-svn: 236901
2015-05-08 21:04:50 +00:00
Renato Golin
4045f66175 Revert "Allow case-insensitive values for -mcpu for ARM and AArch64"
This reverts commit r236859, as it broke multiple builds. I'll investigate
and reapply when safe.

llvm-svn: 236869
2015-05-08 15:44:36 +00:00
Renato Golin
9e36fb4c18 Allow case-insensitive values for -mcpu for ARM and AArch64
GCC allows case-insensitive values for -mcpu, -march and -mtune options.
This patch implements the same behaviour for the -mcpu option.

Patch by Gabor Ballabas.

llvm-svn: 236859
2015-05-08 14:50:32 +00:00
John Brawn
a95c1a8315 [ARM] Give an error on invalid -march values
llvm::Triple::getARMCPUForArch now returns nullptr for invalid -march
values, instead of silently translating it to arm7tdmi. Use this to
give an error message, which is consistent with how gcc behaves.

Differential Revision: http://reviews.llvm.org/D9602

llvm-svn: 236846
2015-05-08 12:52:18 +00:00
Artem Belevich
ba558951d8 [driver] Cosmetic change to use Input instead of Inputs[0].
Differential Revision: http://reviews.llvm.org/D9506

llvm-svn: 236621
2015-05-06 18:20:23 +00:00
Ulrich Weigand
66ff51b4ea [SystemZ] Add support for z13 and its vector facility
This patch adds support for the z13 architecture type.  For compatibility
with GCC, a pair of options -mvx / -mno-vx can be used to selectively
enable/disable use of the vector facility.

When the vector facility is present, we default to the new vector ABI.
This is characterized by two major differences:
- Vector types are passed/returned in vector registers
  (except for unnamed arguments of a variable-argument list function).
- Vector types are at most 8-byte aligned.

The reason for the choice of 8-byte vector alignment is that the hardware
is able to efficiently load vectors at 8-byte alignment, and the ABI only
guarantees 8-byte alignment of the stack pointer, so requiring any higher
alignment for vectors would require dynamic stack re-alignment code.

However, for compatibility with old code that may use vector types, when
*not* using the vector facility, the old alignment rules (vector types
are naturally aligned) remain in use.

These alignment rules are not only implemented at the C language level,
but also at the LLVM IR level.  This is done by selecting a different
DataLayout string depending on whether the vector ABI is in effect or not.

Based on a patch by Richard Sandiford.

llvm-svn: 236531
2015-05-05 19:35:52 +00:00
Justin Bogner
a71e681792 InstrProf: Support for setting profile output from command line
This change is the third of 3 patches to add support for specifying
the profile output from the command line via -fprofile-instr-generate=<path>,
where the specified output path/file will be overridden by the
LLVM_PROFILE_FILE environment variable.

This patch adds the necessary support to the clang frontend, and adds a
new test.

The compiler-rt and llvm parts are r236055 and r236288, respectively.

Patch by Teresa Johnson. Thanks!

llvm-svn: 236289
2015-04-30 23:49:42 +00:00
Nico Weber
6e0ebaea1f Revert r236060, it caused PR23375.
llvm-svn: 236159
2015-04-29 21:16:40 +00:00
Eric Christopher
52fa24172f Stop emitting the soft-float and soft-float-abi target features
for ARM while the backend will only ignore them. No functional
change intended.

llvm-svn: 236060
2015-04-28 23:18:33 +00:00
Alexey Samsonov
381845d54f Introduce tsan_cxx and msan_cxx libraries (Clang part).
For now tsan_cxx and msan_cxx contain only operator new/delete
replacements. In the future, when we add support for running UBSan+TSan
and UBSan+MSan, they will also contain bits ubsan_cxx runtime.

llvm-svn: 235924
2015-04-27 22:07:50 +00:00
Artem Belevich
c3e2fd7bea Revert "PR21000: pass -I options to assembler" as the test was failing on hexagon.
llvm-svn: 235919
2015-04-27 21:11:08 +00:00
Artem Belevich
8fe8ed56ee PR21000: pass -I options to assembler
Pass -I options to assembly so it can find files included with
.include.

Differential Revision: http://reviews.llvm.org/D7472

llvm-svn: 235915
2015-04-27 20:51:44 +00:00
Paul Robinson
d7214a7651 Support generating NMake/Jom-style depfiles.
NMake is a Make-like builder that comes with Microsoft Visual Studio.
Jom (https://wiki.qt.io/Jom) is an NMake-compatible build tool.
Dependency files for NMake/Jom need to use double-quotes to wrap
filespecs containing special characters, instead of the backslash
escapes that GNU Make wants.

Adds the -MV option, which specifies to use double-quotes as needed
instead of backslash escapes when writing the dependency file.

Differential Revision: http://reviews.llvm.org/D9260

llvm-svn: 235903
2015-04-27 18:14:32 +00:00
Vladimir Sukharev
e851e04262 [AArch64] Add v8.1a architecture
Add support for AArch64 v8.1 architecture. Briefly it is described on http://community.arm.com/groups/processors/blog/2014/12/02/the-armv8-a-architecture-and-its-ongoing-development

Reviewers: jmolloy

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D8493

llvm-svn: 235110
2015-04-16 15:53:09 +00:00
Duncan P. N. Exon Smith
bb61825cd5 uselistorder: -mllvm -preserve-bc-use-list-order => -emit-llvm-uselists
Stop relying on `cl::opt` to pass along the driver's decision to
preserve use-lists.  Create a new `-cc1` option called
`-emit-llvm-uselists` that does the right thing (when -emit-llvm-bc).
Note that despite its generic name, it *doesn't* do the right thing when
-emit-llvm (LLVM assembly) yet.  I'll hook that up soon.

This doesn't really change the behaviour of the driver.  The default is
still to preserve use-lists for `clang -emit-llvm` and `clang
-save-temps`, and nothing else.  But it stops relying on global state
(and also is a nicer interface for hackers using `clang -cc1`).

llvm-svn: 234962
2015-04-15 01:16:18 +00:00
Duncan P. N. Exon Smith
d863f6f250 IR: Change clang to set -preserve-bc-uselistorder
Change `clang` to set `-preserve-bc-uselistorder` for the driver options
`-emit-llvm` and `-save-temps`.  The former is useful for reproducing
results from `clang` in `opt` or `llc`, while the latter prevents
`-save-temps` from affecting the output.  This is part of PR5680.

`-preserve-bc-uselistorder=true` is currently on by default, but a
follow-up commit in LLVM will reverse it.

llvm-svn: 234920
2015-04-14 18:30:13 +00:00
Petar Jovanovic
1dbc317736 [Mips] Generate warning for invalid '-mnan' and '-march' combinations
This patch generates a warning for invalid combination of '-mnan' and
'-march' options, it properly sets NaN encoding for a given '-march',
and it passes a proper NaN encoding to the assembler.

Patch by Vladimir Radosavljevic.

Differential Revision: http://reviews.llvm.org/D8170

llvm-svn: 234882
2015-04-14 12:49:08 +00:00
Ahmed Bougacha
256a869d31 [Driver] Properly support -mglobal-merge using explicit options.
Follow-up to r234666.  With this, the -m[no-]global-merge options
have the expected behavior. Previously, -mglobal-merge was ignored,
and there was no way of enabling the optimization.

llvm-svn: 234668
2015-04-11 00:10:44 +00:00
Derek Schuff
a419e1ce6c NaCl ARM: fix assembler float abi flags
Summary:
tools::arm::getARMFloatABI() was falling back to guessing soft-float because
it wasn't seeing the GNUEABIHF environment from ComputeEffectivClangTriple
when it was called from gnutools::Assemble::ConstructJob.

Fix by using the effective clang triple in gnutools::Assemble, which now
matches the -triple flag used by cc1 and ClangAs jobs.

Reviewers: jvoung

Subscribers: rengolin, jfb, aemerson, cfe-commits

Differential Revision: http://reviews.llvm.org/D8902

llvm-svn: 234661
2015-04-10 23:07:19 +00:00
Sanjay Patel
359b105745 Process the -freciprocal-math optimization flag (PR20912)
The driver currently accepts but ignores the -freciprocal-math flag.
This patch passes the flag through and enables 'arcp' fast-math-flag
generation in IR.

Note that this change does not actually enable the optimization for
any target. The reassociation optimization that this flag specifies
was implemented by http://reviews.llvm.org/D6334 :
http://llvm.org/viewvc/llvm-project?view=revision&revision=222510

Because the optimization is done in the backend rather than IR,
the backend must be modified to understand instruction-level
fast-math-flags or a new function-level attribute must be created.

Also note that -freciprocal-math is independent of any target-specific
usage of reciprocal estimate hardware instructions. That requires
its own flag ('-mrecip').

https://llvm.org/bugs/show_bug.cgi?id=20912

llvm-svn: 234493
2015-04-09 15:03:23 +00:00
Javed Absar
879d18b843 [ARM] add support for Cortex-R4/R4F
Adds ARM Cortex-R4 and R4F support and tests in Clang. Though Cortex-R4
support was present, the support for hwdiv in thumb-mode was not defined
or tested properly. This has also been added.

llvm-svn: 234488
2015-04-09 14:12:10 +00:00
Rafael Espindola
298059a4ac This reverts commit r234104, bringing back 233393 now that ARM is fixed.
Original message:

Don't use unique section names by default if using the integrated as.

This saves some IO and ccache space by not creating long section names. It
should work with every ELF linker.

llvm-svn: 234143
2015-04-06 04:36:45 +00:00
Rafael Espindola
d7329ef99f Revert "Revert "Revert "Don't use unique section names by default if using the integrated as."""
This reverts commit r234101. I will debug what went wrong with ARM.

llvm-svn: 234104
2015-04-04 19:32:41 +00:00
Rafael Espindola
321fabfd8e Revert "Revert "Don't use unique section names by default if using the integrated as.""
This reverts commit r233398, bringing back 233393 now that LLVM is fixed.

Original message:

Don't use unique section names by default if using the integrated as.

This saves some IO and ccache space by not creating long section names. It
should work with every ELF linker.

llvm-svn: 234101
2015-04-04 18:21:14 +00:00
Eric Christopher
269c2a29ef Complete comment. Reflow conditional.
llvm-svn: 234083
2015-04-04 03:34:43 +00:00
Alexey Samsonov
31095469aa [UBSan] Embed UBSan into ASan runtime (Clang part).
Summary:
Change the way we use ASan and UBSan together. Instead of keeping two
separate runtimes (libclang_rt.asan and libclang_rt.ubsan), embed UBSan
into ASan and get rid of libclang_rt.ubsan. If UBSan is not supported on
a platform, all UBSan sources are just compiled into dummy empty object
files. UBSan initialization code (e.g. flag parsing) is directly called
from ASan initialization, so we are able to enforce correct
initialization order.

This mirrors the approach we already use for ASan+LSan. This change doesn't
modify the way we use standalone UBSan.

Test Plan: regression test suite

Reviewers: kubabrecka, zaks.anna, kcc, rsmith

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D8645

llvm-svn: 233860
2015-04-01 22:42:25 +00:00
Ulrich Weigand
3a610ebf1e [SystemZ] Support transactional execution on zEC12
The zEC12 provides the transactional-execution facility.  This is exposed
to users via a set of builtin routines on other compilers.  This patch
adds clang support to enable those builtins.  In partciular, the patch:

- enables the transactional-execution feature by default on zEC12
- allows to override presence of that feature via the -mhtm/-mno-htm options
- adds a predefined macro __HTM__ if the feature is enabled
- adds support for the transactional-execution GCC builtins
- adds Sema checking to verify the __builtin_tabort abort code
- adds the s390intrin.h header file (for GCC compatibility)
- adds s390 sections to the htmintrin.h and htmxlintrin.h header files

Since this is first use of target-specific intrinsics on the platform,
the patch creates the include/clang/Basic/BuiltinsSystemZ.def file and
hooks it up in TargetBuiltins.h and lib/Basic/Targets.cpp.

An associated LLVM patch adds the required LLVM IR intrinsics.

For reference, the transactional-execution instructions are documented
in the z/Architecture Principles of Operation for the zEC12:
http://publibfp.boulder.ibm.com/cgi-bin/bookmgr/download/DZ9ZR009.pdf
The associated builtins are documented in the GCC manual:
http://gcc.gnu.org/onlinedocs/gcc/S_002f390-System-z-Built-in-Functions.html
The htmxlintrin.h intrinsics provided for compatibility with the IBM XL
compiler are documented in the "z/OS XL C/C++ Programming Guide".

llvm-svn: 233804
2015-04-01 12:54:25 +00:00
Craig Topper
a8bd6008ee [X86] Use getHostCPUFeatures when 'native' is specified for cpu.
This is necessary because not aall Sandybridge, Ivybrige, Haswell, and Broadwell CPUs support AVX. Currently we modify the CPU name back to Nehalem for this case, but that turns off additional features for these CPUs.

llvm-svn: 233672
2015-03-31 05:45:00 +00:00
Derek Schuff
6ab52fabcf Add driver support for Native Client SDK
Add Tool and ToolChain support for clang to target the NaCl OS using the NaCl
SDK for x86-32, x86-64 and ARM.

Includes nacltools::Assemble and Link which are derived from gnutools. They
are similar to Linux but different enought that they warrant their own class.
Also includes a NaCl_TC in ToolChains derived from Generic_ELF with library
and include paths suitable for an SDK and independent of the system tools.

Differential Revision: http://reviews.llvm.org/D8590

llvm-svn: 233594
2015-03-30 20:31:33 +00:00
Rafael Espindola
d053d20d0d Revert "Don't use unique section names by default if using the integrated as."
This reverts commit r233393 while a debug a bot failure.

llvm-svn: 233398
2015-03-27 19:01:14 +00:00
Rafael Espindola
463a0afe15 Don't use unique section names by default if using the integrated as.
This saves some IO and ccache space by not creating long section names. It
should work with every ELF linker.

llvm-svn: 233393
2015-03-27 17:55:23 +00:00
Ed Schouten
6e57615ac8 Enable -ffunction-sections and -fdata-sections for CloudABI by default.
Unlike most of the other platforms supported by Clang, CloudABI only
supports static linkage, for the reason that global filesystem access is
prohibited. Functions provided by dlfcn.h are not present. As we know
that applications will not try to do any symbol lookups at run-time, we
can garbage collect unused code quite aggressively. Because of this, it
makes sense to enable -ffunction-sections and -fdata-sections by
default.

Object files will be a bit larger than usual, but the resulting binary
will not be affected, as the sections are merged again. However, when
--gc-sections is used, the linker is able to remove unused code far more
more aggressively. It also has the advantage that transitive library
dependencies only need to be provided to the linker in case that
functionality is actually used.

Differential Revision:	http://reviews.llvm.org/D8635
Reviewed by:	echristo

llvm-svn: 233299
2015-03-26 17:50:28 +00:00
Ed Schouten
3c3e58c42d Let Clang invoke CloudABI's linker.
Now that CloudABI's target information and header search logic for Clang
has been submitted, the only thing that remains to be done is adding
support for CloudABI's linker.

CloudABI uses Binutils ld, although there is some work to use lld
instead. This means that this code is largely based on what we use on
FreeBSD. There are some exceptions, however:

- Only static linking is performed. CloudABI does not support any
  dynamically linked executables.
- CloudABI uses compiler-rt, libc++ and libc++abi unconditionally. Link
  in these libraries instead of using libgcc_s, libstdc++, etc.
- We must ensure that the .eh_frame_hdr is present to make C++
  exceptions work properly.


Differential Revision:	http://reviews.llvm.org/D8250

llvm-svn: 233269
2015-03-26 11:13:44 +00:00
Alexey Samsonov
3ae177cdb3 [UBSan] Introduce "ubsan_standalone" library (Clang part).
Get rid of "libclang_rt.san" library that used to contain
sanitizer_common pieces required by UBSan if it's used in a standalone
mode. Instead, build two variants of UBSan runtime: "ubsan" and
"ubsan_standalone" (same for "ubsan_cxx" and "ubsan_standalone_cxx").

Later "ubsan" and "ubsan_cxx" libraries will go away, as they will
embedded it into corresponding ASan runtimes.

llvm-svn: 233010
2015-03-23 20:05:42 +00:00
Eric Christopher
c54920a123 Fix and update comments and a small reformatting.
llvm-svn: 232997
2015-03-23 19:26:05 +00:00
Scott Douglass
3205f52198 handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob
Differential Revision: http://reviews.llvm.org/D8196

llvm-svn: 232940
2015-03-23 10:54:24 +00:00
Scott Douglass
d4cfffccc7 Convert cascading if-else-if to switch. NFC
Differential Revision: http://reviews.llvm.org/D8485

llvm-svn: 232939
2015-03-23 10:46:28 +00:00