Commit Graph

193 Commits

Author SHA1 Message Date
Alexey Samsonov 84e0223a9d Fix configure+make build of profile runtime library
llvm-svn: 212665
2014-07-10 00:53:20 +00:00
Alexey Samsonov 00997e29e5 Unify the name of compiler-rt builtins library on Linux.
Call it "libclang_rt.builtins-<arch>.a" to be consistent
with sanitizers/profile libraries naming. Modify Makefile
and CMake build systems and Clang driver accordingly.

Fixes PR19822.

llvm-svn: 209473
2014-05-22 21:12:43 +00:00
Alp Toker 1ee7fc7a1a Fix typos
llvm-svn: 208841
2014-05-15 02:22:34 +00:00
Alexey Samsonov 1866b58885 [ASan] Split static ASan runtime in two parts: asan and asan_cxx.
asan_cxx containts replacements for new/delete operators, and should
only be linked in C++ mode. We plan to start building this part
with exception support to make new more standard-compliant.

See https://code.google.com/p/address-sanitizer/issues/detail?id=295
for more details.

llvm-svn: 208609
2014-05-12 18:39:22 +00:00
Alexey Samsonov 9c1a57b64f [DFSan] Update build rules for Makefile build
llvm-svn: 208268
2014-05-07 23:48:59 +00:00
Duncan P. N. Exon Smith be0a5e176b InstrProf: Reorganize files; no functionality change
Move functions around to prepare for some other changes.

  - Merge InstrProfilingExtras.h with InstrProfiling.h.  There's no
    benefit to having these split.

  - Rename InstrProfilingExtras.c to InstrProfilingFile.c.

  - Split actual buffer writing code out of InstrProfiling.c into
    InstrProfilingBuffer.c.

  - Drive-by corrections of a couple of header comments.

<rdar://problem/15943240>

llvm-svn: 204497
2014-03-21 18:29:15 +00:00
Duncan P. N. Exon Smith 775c178a90 PGO: Add explicit static initialization
Instead of relying on explicit static initialization from translation
units, create a new file, InstrProfilingRuntime.cc, with an
__llvm_pgo_runtime variable.  After this commit (and its pair in clang),
the driver will create a use of this variable.  Unless the user defines
their own version, the new object file will get pulled in, including
that C++ static initialization that calls
__llvm_pgo_register_write_atexit.

The result is that, at least on Darwin, static initialization typically
consists of a single function call, which registers a writeout functino
atexit.  Furthermore, users can skip even this behaviour by defining
their own __llvm_pgo_runtime.

<rdar://problem/15943240>

llvm-svn: 204380
2014-03-20 19:23:53 +00:00
Duncan P. N. Exon Smith 9cee5e0e92 PGO: Moving files for clarity
<rdar://problem/15943240>

llvm-svn: 204373
2014-03-20 18:43:09 +00:00
Duncan P. N. Exon Smith ecf2256dfe PGO: Implement Darwin linker magic for instrumentation
Use Darwin linker magic to find bounds of instrumentation data sections
at link time instead of runtime.

<rdar://problem/15943240>

llvm-svn: 204302
2014-03-20 03:57:33 +00:00
Duncan P. N. Exon Smith da0de8a237 PGO: Split out initialization of section bounds
Currently we register instrumentation data at runtime to determine the
bounds of the sections where the data lives.  Soon we'll implement
platform-specific linker magic to determine this at link time.

Move this logic to a separate file, so that our build system can choose
the correct platform-specific code.

No functionality change intended.

<rdar://problem/15943240>

llvm-svn: 204299
2014-03-20 03:23:10 +00:00
Duncan P. N. Exon Smith fb4447688f PGO: Appease buildbots after r204268
llvm-svn: 204276
2014-03-19 22:28:32 +00:00
Alexey Samsonov a7b8472397 One more attempt to fix Makefile build
llvm-svn: 204241
2014-03-19 15:26:54 +00:00
Alexey Samsonov 725122eca3 [Make] Build sanitizer runtimes with -std=c++11
llvm-svn: 204239
2014-03-19 14:50:10 +00:00
Evgeniy Stepanov 0b5b3295e5 Move Android ucontext.h out of third_party.
Google is re-licensing this code under the standard dual license of
compiler-rt.

llvm-svn: 204128
2014-03-18 08:32:14 +00:00
Tim Northover 1683caab3a ARM: implement __sync_fetch_and_* operations
Since these are primarily useful for deeply embedded targets where code size is
very important, they are each in a separate file making use of infrastructure
in sync-ops.h. This allows a linker to include just the functions that are
actually used.

rdar://problem/14736665

llvm-svn: 202812
2014-03-04 10:10:17 +00:00
Evgeniy Stepanov 2821fb7d18 [asan] Fix Android configure/make build.
Mirror recent cmake changes in the configure/make build system.

llvm-svn: 202793
2014-03-04 07:17:38 +00:00
Tim Northover 23a2fb59b6 Darwin: use the compiler-provided linker
When actually compiling we use clang for the final .dylib creation stage, so we
should ask it what linker it'll use when detecting support for our
architectures.

llvm-svn: 201845
2014-02-21 10:19:22 +00:00
Alexey Samsonov 599290e1ca [compiler-rt] Fix CompilerTargetTriple for non-English locales.
When the locale of a shell is set other than English locales or the C locale,
The word 'Target' may be translated. Thus, with e.g. ja_JP locale, compiler-rt
couldn't be built properly. Forcing LANG=C fixes the problem.

Patch by Ogino Masanori.

llvm-svn: 201506
2014-02-17 09:14:01 +00:00
Alexey Samsonov a6b264b51d Move original compiler-rt functions (libgcc replacement) to lib/builtins directory
llvm-svn: 201393
2014-02-14 09:20:33 +00:00
Evgeniy Stepanov aeefe8c1bb Remove -arm-enable-ehabi from Android build rules. It's enabled by default.
llvm-svn: 200392
2014-01-29 12:36:45 +00:00
Tim Northover 5a8af3e136 Darwin: add x86_64h slice for Haswell.
llvm-svn: 199996
2014-01-24 13:10:55 +00:00
Evgeniy Stepanov 350fe4b1af [asan] Android logging.
This change duplicates all ASan output to system log on Android.

llvm-svn: 199887
2014-01-23 11:34:41 +00:00
Tim Northover bea1cafde8 Macho-embedded: update build script for newer Clang behaviour.
Clang now uses the *-apple-macho triples for embedded targets using MachO. The
environment makes less pretences to be running an actual Mach kernel with all
the support that would imply, so a few changes are needed to the build scripts:
    
+ The Apple way of detecting endianness applies in all cases so drop the
  __MACH__ test.
+ A quick rename since "darwin_embedded" no longer makes any sense. If there's
  no Mach going on, there's certainly no Darwin.
+ Change how we test for support for a platform. The functions we build don't
  need any SDK support (stdio.h, ...) so we don't need to check for it.

llvm-svn: 199731
2014-01-21 12:38:02 +00:00
Evgeniy Stepanov 755fc5c313 [asan] Link runtime library with -lm -Wl,-z,defs to ensure that it declares all its dependencies.
llvm-svn: 199483
2014-01-17 09:38:46 +00:00
Evgeniy Stepanov e259b658e4 [asan] Nuke ASAN_FLEXIBLE_MAPPING_AND_OFFSET.
llvm-svn: 199370
2014-01-16 10:16:19 +00:00
Bob Wilson e38cc76c32 Remove support for armv7f slice. <rdar://problem/12478440>
This was never used for anything so we should just get rid of it.

llvm-svn: 199333
2014-01-15 21:05:20 +00:00
Justin Bogner 65bb804f05 profile: Rudimentary suppport for PGO instrumentation
This is fairly minimal support for instrumentation based PGO. The data
format is inefficient, and the output file name is hardcoded to
pgo-data.

llvm-svn: 198638
2014-01-06 22:27:03 +00:00
Tim Northover 638a424b56 Only build embedded darwin variants if the compiler supports them
We need to filter out architectures that the compiler hasn't been built to
target (most likely the ARM ones) before attemptint to build a version of
libcompiler_rt.

This can result in a completely empty library (e.g. soft-float doesn't have any
x86 variants), in which case we shouldn't even try the build

llvm-svn: 197028
2013-12-11 11:25:53 +00:00
Alexey Samsonov 58e44a3447 Revert r196490 and fix include paths in makefile-based build
llvm-svn: 196492
2013-12-05 10:40:11 +00:00
Duncan P. N. Exon Smith 2f234f4908 Fixing ordering of definition of IOSSIM_SDK_PATH
llvm-svn: 195216
2013-11-20 05:36:52 +00:00
Alexander Potapenko 9ff321d18d [ASan] Don't put asan_iossim_dynamic in the list of configurations to build
if the iOS Simulator SDK is missing on the machine.

llvm-svn: 195125
2013-11-19 13:24:29 +00:00
Tim Northover 3ae5a369d8 Revert premature instructions to build __sync_fetch_*
My commits got a bit twisted when I was working on both embedded support and
__sync stuff on the same branch. Sorry about that.

llvm-svn: 194886
2013-11-15 23:26:29 +00:00
Tim Northover 084647d4c0 Add clang_darwin_embedded platform for embedded projects
This should build a separate set of embedded runtime libraries, supporting the
option product:

    {PIC, static} x { Hard-float, soft-float }

The emphasis is on ARM platforms (Cortex-M4F, Cortex-M3, Cortex-M0) but X86
variants are also built where possible.

rdar://problem/10817205

llvm-svn: 194873
2013-11-15 23:00:42 +00:00
Alexander Potapenko bf50fbfc21 [ASan] Add the configure+make rules for building the ASan runtime for iOS simulator.
llvm-svn: 194813
2013-11-15 13:13:01 +00:00
Alexander Potapenko 0706e5dd36 [ASan] Remove the stale dependency on the Foundation framework from the dynamic ASan runtime.
llvm-svn: 194583
2013-11-13 15:52:39 +00:00
Bob Wilson 64b4abd07b Push back various changes for building Apple's compiler-rt dylib.
These changes had accumulated internally at Apple for no good reason.

llvm-svn: 193944
2013-11-03 07:06:29 +00:00
Evgeniy Stepanov 364f70c728 [asan] Don't require X86 to build Android runtime.
Patch by Greg Fitzgerald.

llvm-svn: 191883
2013-10-03 08:31:14 +00:00
Sergey Matveev e38dd3ff4e [lsan] Follow-up to r189010 - do the same in Makefile build.
llvm-svn: 189011
2013-08-22 13:30:01 +00:00
Bill Wendling 4f20f51090 Add missing ']'.
llvm-svn: 188927
2013-08-21 20:12:27 +00:00
Bob Wilson e7ca6f9359 For Darwin builds, locate tools using xcrun when available.
The make/platform/darwin_bni.mk file already has similar code but we
apparently neglected to add it to the clang_darwin.mk file.

llvm-svn: 188864
2013-08-21 04:42:56 +00:00
Alexey Samsonov 287fca44af [LSan] Add support for building standalone LSan runtime to Make build (compiler-rt part)
llvm-svn: 188804
2013-08-20 14:49:01 +00:00
Peter Collingbourne 5cbab07d02 DataFlowSanitizer; compiler-rt changes.
DataFlowSanitizer is a generalised dynamic data flow analysis.

Unlike other Sanitizer tools, this tool is not designed to detect a
specific class of bugs on its own.  Instead, it provides a generic
dynamic data flow analysis framework to be used by clients to help
detect application-specific issues within their own code.

Differential Revision: http://llvm-reviews.chandlerc.com/D967

llvm-svn: 187924
2013-08-07 22:47:26 +00:00
Alexey Samsonov af4806f719 [ASan] Bump min supported Mac OS X version to 10.6 - makefile build
llvm-svn: 186410
2013-07-16 13:16:15 +00:00
Bob Wilson c4b26ad006 Revert "Temporarily disable building the armv7 variants of profile_ios on"
This reverts commit r184816.  With Chandler's change in r184929, this should
no longer be necessary.

llvm-svn: 184961
2013-06-26 16:35:16 +00:00
Chandler Carruth fc9d259c98 Revert the Darwin half of r184805: "Remove the sysroot (or isysroot)
restriction from the GCDAProfile.c"

Bill added a much more effective way of coping with the use of errno and
other system interactions for Darwin for now by disabling it on that
platform. Building against the SDK stuff is more fundamental to the
darwin build process.

llvm-svn: 184929
2013-06-26 10:05:37 +00:00
Chandler Carruth 825abd8649 Temporarily disable building the armv7 variants of profile_ios on
darwin.

After talking with Jim Grosbach pretty extensively, he was OK with me
punting on this to the Apple folks.

We agreed that the correct fix is either to extend the fake SDK headers
in compiler-rt to support the rest of libc needed by GCDAProfiling.c or
to teach the make build to test for the existence of a suitable SDK on
the system prior to building it. Both of these require someone with
access to the SDK which I don't have, and the latter (my preferred
solution) requires understanding the strange compiler-rt make build
system. Punting to Dan Dunbar or one of the others who support this
stuff on ios to write the appropriate fix, and we can let the build bots
proceed in the mean time.

llvm-svn: 184816
2013-06-25 02:15:14 +00:00
Chandler Carruth ffdf342728 Remove the sysroot (or isysroot) restriction from the GCDAProfile.c
runtime in the Makefile build system as well. Sorry for the temporary
fallout, it took me a while to find these bits.

Bill, I'm not at all confident of the change for Darwin and iOS, but as
discussed we're completely blocked on fixing this. Anyways, please
review and let me know if this will work for you guys. If necessary,
I can work with you to rig up an errno.h stub for Darwin, but I expect
that to be... moderately challenging.

llvm-svn: 184805
2013-06-25 00:57:06 +00:00
Sylvestre Ledru 6e0a254ec9 When the compiler is not providing a clean triple, we should fail
the build right from the start. 
For now, it is only done with DEBUGMAKE is provided. That makes
things harder to debug during ports of compiler-rt on other systems.

llvm-svn: 184295
2013-06-19 12:59:56 +00:00
Alexey Samsonov 49eb5700e2 Revert r182465 and add lsan-common library to makefile-based build
llvm-svn: 182470
2013-05-22 13:20:37 +00:00
Nick Kledzik 69e25c45e4 <rdar://problem/13806954> enable __atomic_* functions for Darwin
llvm-svn: 182435
2013-05-21 23:02:04 +00:00