memcmp, strncmp,..) out of Sema and into FunctionDecl so that the logic
could be reused in the analyzer.

llvm-svn: 148142

llvm-svn: 148141

improving the typo correction results in certain situations.

This is also the first typo correction callback conversion to affect
an existing unit test. :)

llvm-svn: 148140

Patch from Jyotsna Verma:

I have made the changes to remove assertions in the Hexagon backend
specific clang driver. Instead of asserting on invalid arch name, it has
been modified to use the default value.

I have changed the implementation of the CPU flag validation for the
Hexagon backend. Earlier, the clang driver performed the check and
asserted on invalid inputs. In the new implementation, the driver passes
the last CPU flag (or sets to "v4" if not specified) to the compiler (and
also to the assembler and linker which perform their own check) instead of
asserting on incorrect values. This patch changes the setCPU function for
the Hexagon backend in clang/lib/Basic/Targets.cpp which causes the
compiler to error out on incorrect CPU flag values.

llvm-svn: 148139

llvm-svn: 148138

llvm-svn: 148137

llvm-svn: 148136

llvm-svn: 148135

Revert r148131, it was committed before it was ready.

llvm-svn: 148134

llvm-svn: 148133

LoopUnswitch: All helper data that is collected during loop-unswitch iterations was moved to separated class (LUAnalysisCache). 

llvm-svn: 148132

llvm-svn: 148131

llvm-svn: 148128

llvm-svn: 148123

llvm-svn: 148118

llvm-svn: 148117

and switches our interceptors to using them instead of the default
vm_allocate-based approach used by mach_override_ptr.

To simplify the code, a fixed memory mapping is used for the allocation pool --
note that we can't mmap an arbitrary chunk of memory, because the shadow memory hasn't been mapped yet
(for the reasons discussed in http://code.google.com/p/address-sanitizer/issues/detail?id=24, we cannot map the shadow earlier)

The patch drops the program startup time from several second to half a second,
which speeds up the execution of ASan tests noticeably.
Because of the virtual memory size occupied by the programs it's hard
to speed up the shutdown time, which would've also helped the tests.

llvm-svn: 148116

Add __asan_mach_override_ptr_custom, which allows to inject a custom memory allocator into mach_override_ptr().

llvm-svn: 148115

in asan_linux.cc, because /proc is unavailable on Mac.
Instead the Mac version of AsanGetEnv iterates over the array of
environment variables obtained from _NSGetEnviron()

llvm-svn: 148114

Gnu hash is not supported by the Android loader.

llvm-svn: 148113

Convert SHUFPD with the same register for both sources to PSHUFD if it would prevent a register copy. Similar to SHUFPS, but requires the mask to be converted.

llvm-svn: 148112

llvm-svn: 148111

name + arguments when the data is available. It seems to work really well, 
but some more testing is needed before we make this on by default.

The new function format name is:

 ${function.name-with-args}

To see how to use these formats see the website:

    http://lldb.llvm.org/formats.html

Here is a sample backtrace of debugging LLDB with LLDB using this new format
value:

(lldb) thread backtrace all
* thread #1: tid = 0x2203, 0x00007fff88a17bca libsystem_kernel.dylib __psynch_cvwait + 10, stop reason = signal SIGINT, name = <lldb.driver.main-thread>, queue = com.apple.main-thread
    frame #0: 0x00007fff88a17bca libsystem_kernel.dylib __psynch_cvwait + 10
    frame #1: 0x00007fff884ae274 libsystem_c.dylib _pthread_cond_wait + 840
    frame #2: 0x00000001010778ea LLDB lldb_private::Condition::Wait(this=0x0000000104846770, mutex=0x0000000104846730, abstime=0x0000000000000000, timed_out=0x00007fff5fbfdea7) + 138 at Condition.cpp:92
    frame #3: 0x0000000101244c21 LLDB lldb_private::Predicate<bool>::WaitForValueEqualTo(this=0x0000000104846728, value=true, abstime=0x0000000000000000, timed_out=0x00007fff5fbfdea7) + 209 at Predicate.h:317
    frame #4: 0x0000000100f6eeb2 LLDB lldb_private::Listener::WaitForEventsInternal(this=0x0000000104846660, timeout=0x0000000000000000, broadcaster=0x0000000000000000, broadcaster_names=0x0000000000000000, num_broadcaster_names=0x00000000, event_type_mask=0x00000000, event_sp=0x00007fff5fbfe030) + 386 at Listener.cpp:388
    frame #5: 0x0000000100f6f231 LLDB lldb_private::Listener::WaitForEvent(this=0x0000000104846660, timeout=0x0000000000000000, event_sp=0x00007fff5fbfe030) + 81 at Listener.cpp:436
    frame #6: 0x0000000100098dcd LLDB lldb::SBListener::WaitForEvent(this=0x00007fff5fbff0f0, timeout_secs=0xffffffff, event=0x00007fff5fbfe430) + 685 at SBListener.cpp:181
    frame #7: 0x000000010000628c lldb Driver::MainLoop(this=0x00007fff5fbff620) + 5244 at Driver.cpp:1325
    frame #8: 0x0000000100006ca3 lldb main(argc=1, argv=0x00007fff5fbff758, envp=0x00007fff5fbff768) + 419 at Driver.cpp:1460
    frame #9: 0x0000000100000d54 lldb start + 52

  thread #3: tid = 0x2703, 0x00007fff88a17df2 libsystem_kernel.dylib select$DARWIN_EXTSN + 10, name = <lldb.comm.debugger.input>
    frame #0: 0x00007fff88a17df2 libsystem_kernel.dylib select$DARWIN_EXTSN + 10
    frame #1: 0x0000000100f3f072 LLDB lldb_private::ConnectionFileDescriptor::BytesAvailable(this=0x000000010524d040, timeout_usec=0x004c4b40, error_ptr=0x0000000105640a18) + 722 at ConnectionFileDescriptor.cpp:542
    frame #2: 0x0000000100f3e6dd LLDB lldb_private::ConnectionFileDescriptor::Read(this=0x000000010524d040, dst=0x0000000105640a60, dst_len=1024, timeout_usec=0x004c4b40, status=0x0000000105640a14, error_ptr=0x0000000105640a18) + 301 at ConnectionFileDescriptor.cpp:273
    frame #3: 0x0000000100f3b8f7 LLDB lldb_private::Communication::ReadFromConnection(this=0x0000000104846270, dst=0x0000000105640a60, dst_len=1024, timeout_usec=0x004c4b40, status=0x0000000105640a14, error_ptr=0x0000000105640a18) + 167 at Communication.cpp:317
    frame #4: 0x0000000100f3b197 LLDB lldb_private::Communication::ReadThread(p=0x0000000104846270) + 327 at Communication.cpp:344
    frame #5: 0x0000000101078923 LLDB ThreadCreateTrampoline(arg=0x00000001045f6650) + 227 at Host.cpp:549
    frame #6: 0x00007fff884aa8bf libsystem_c.dylib _pthread_start + 335
    frame #7: 0x00007fff884adb75 libsystem_c.dylib thread_start + 13

  thread #4: tid = 0x2803, 0x00007fff88a17df2 libsystem_kernel.dylib select$DARWIN_EXTSN + 10, name = <lldb.comm.driver.editline>
    frame #0: 0x00007fff88a17df2 libsystem_kernel.dylib select$DARWIN_EXTSN + 10
    frame #1: 0x0000000100f3f072 LLDB lldb_private::ConnectionFileDescriptor::BytesAvailable(this=0x0000000105700370, timeout_usec=0x004c4b40, error_ptr=0x00000001056c3a18) + 722 at ConnectionFileDescriptor.cpp:542
    frame #2: 0x0000000100f3e6dd LLDB lldb_private::ConnectionFileDescriptor::Read(this=0x0000000105700370, dst=0x00000001056c3a60, dst_len=1024, timeout_usec=0x004c4b40, status=0x00000001056c3a14, error_ptr=0x00000001056c3a18) + 301 at ConnectionFileDescriptor.cpp:273
    frame #3: 0x0000000100f3b8f7 LLDB lldb_private::Communication::ReadFromConnection(this=0x0000000105700000, dst=0x00000001056c3a60, dst_len=1024, timeout_usec=0x004c4b40, status=0x00000001056c3a14, error_ptr=0x00000001056c3a18) + 167 at Communication.cpp:317
    frame #4: 0x0000000100f3b197 LLDB lldb_private::Communication::ReadThread(p=0x0000000105700000) + 327 at Communication.cpp:344
    frame #5: 0x0000000101078923 LLDB ThreadCreateTrampoline(arg=0x0000000105700430) + 227 at Host.cpp:549
    frame #6: 0x00007fff884aa8bf libsystem_c.dylib _pthread_start + 335
    frame #7: 0x00007fff884adb75 libsystem_c.dylib thread_start + 13

  thread #5: tid = 0x2903, 0x00007fff88a17df2 libsystem_kernel.dylib select$DARWIN_EXTSN + 10, name = <lldb.comm.driver.editline_output>
    frame #0: 0x00007fff88a17df2 libsystem_kernel.dylib select$DARWIN_EXTSN + 10
    frame #1: 0x0000000100f3f072 LLDB lldb_private::ConnectionFileDescriptor::BytesAvailable(this=0x00000001057178f0, timeout_usec=0x004c4b40, error_ptr=0x0000000105980a18) + 722 at ConnectionFileDescriptor.cpp:542
    frame #2: 0x0000000100f3e6dd LLDB lldb_private::ConnectionFileDescriptor::Read(this=0x00000001057178f0, dst=0x0000000105980a60, dst_len=1024, timeout_usec=0x004c4b40, status=0x0000000105980a14, error_ptr=0x0000000105980a18) + 301 at ConnectionFileDescriptor.cpp:273
    frame #3: 0x0000000100f3b8f7 LLDB lldb_private::Communication::ReadFromConnection(this=0x0000000105717580, dst=0x0000000105980a60, dst_len=1024, timeout_usec=0x004c4b40, status=0x0000000105980a14, error_ptr=0x0000000105980a18) + 167 at Communication.cpp:317
    frame #4: 0x0000000100f3b197 LLDB lldb_private::Communication::ReadThread(p=0x0000000105717580) + 327 at Communication.cpp:344
    frame #5: 0x0000000101078923 LLDB ThreadCreateTrampoline(arg=0x00000001057179b0) + 227 at Host.cpp:549
    frame #6: 0x00007fff884aa8bf libsystem_c.dylib _pthread_start + 335
    frame #7: 0x00007fff884adb75 libsystem_c.dylib thread_start + 13

  thread #6: tid = 0x2a03, 0x00007fff88a18af2 libsystem_kernel.dylib read + 10, name = <lldb.driver.commandline_io>
    frame #0: 0x00007fff88a18af2 libsystem_kernel.dylib read + 10
    frame #1: 0x0000000100050c3b libedit.3.dylib read_init + 247
    frame #2: 0x0000000100050e96 libedit.3.dylib el_wgetc + 155
    frame #3: 0x000000010005115d libedit.3.dylib el_wgets + 578
    frame #4: 0x000000010005debc libedit.3.dylib el_gets + 37
    frame #5: 0x000000010000d409 lldb IOChannel::LibeditGetInput(this=0x0000000105700490, new_line=0x0000000105a03db0) + 89 at IOChannel.cpp:311
    frame #6: 0x000000010000d8b6 lldb IOChannel::Run(this=0x0000000105700490) + 806 at IOChannel.cpp:391
    frame #7: 0x000000010000d57d lldb IOChannel::IOReadThread(ptr=0x0000000105700490) + 29 at IOChannel.cpp:345
    frame #8: 0x0000000101078923 LLDB ThreadCreateTrampoline(arg=0x00000001057179f0) + 227 at Host.cpp:549
    frame #9: 0x00007fff884aa8bf libsystem_c.dylib _pthread_start + 335
    frame #10: 0x00007fff884adb75 libsystem_c.dylib thread_start + 13
(lldb) 

llvm-svn: 148110

llvm-svn: 148109

Make X86 instruction selection use 256-bit VPXOR for build_vector of all ones if AVX2 is enabled. This gives the ExeDepsFix pass a chance to choose FP vs int as appropriate. Also use v8i32 as the type for getZeroVector if AVX2 is enabled. This is consistent with SSE2 using prefering v4i32.

llvm-svn: 148108

llvm-svn: 148107

llvm-svn: 148106

llvm-svn: 148105

llvm-svn: 148104

llvm-svn: 148103

llvm-svn: 148102

llvm-svn: 148101

llvm-svn: 148100

llvm-svn: 148098

Fixed the new __apple_types to be able to accept a DW_TAG_structure_type
forward declaration and then find a DW_TAG_class_type definition, or vice
versa.

llvm-svn: 148097

Fix typo in PerformAddCombine that caused any vector type to be checked for horizontal add/sub if AVX2 is enabled. This caused an assert to fail for non 128/256-bit vectors when done before type legalizing. Fixes PR11749.

llvm-svn: 148096

llvm-svn: 148095

base classes.

llvm-svn: 148094

A base subobject type doesn't make sense for unions; don't try to compute it.  Based on patch by Yin Ma.  Fixes PR11751.

llvm-svn: 148093

The code type was always identical to a string anyway. Now it is simply
a synonym. The code literal syntax [{...}] is still valid.

llvm-svn: 148092