Commit Graph

192444 Commits

Author SHA1 Message Date
Pavel Labath 97a9ac1982 Avoid leaking log file descriptors into the inferior process.
Summary:
This commit adds a new open flag File::eOpenOptionCloseOnExec (i.e., O_CLOEXEC), and adds it to
the list of flags when opening log files (#ifndef windows). A regression test is included.

Reviewers: vharron, clayborg

Subscribers: lldb-commits

Differential Revision: http://reviews.llvm.org/D7412

llvm-svn: 228310
2015-02-05 16:44:42 +00:00
Sylvestre Ledru fe0c7ad852 revert 228308. The code has changed since the review
llvm-svn: 228309
2015-02-05 16:35:44 +00:00
Sylvestre Ledru d0ee6daffd Identical code for different branches (CID 1254883)
Reviewers: kledzik, rafael

Reviewed By: rafael

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D6303

llvm-svn: 228308
2015-02-05 16:30:25 +00:00
Oleksiy Vyalov 4536c458e1 Fix warning about the use of mktemp and make platform agnostic by adding and using PipeBase::CreateWithUniqueName - on behalf of flackr.
http://reviews.llvm.org/D7348

llvm-svn: 228307
2015-02-05 16:29:12 +00:00
Ed Maste 72b3b62fac In fd leak test report all open fds, not only first failure
llvm-svn: 228306
2015-02-05 16:11:33 +00:00
Ed Maste 7f0230fad9 Avoid leakage of file descriptors in LLDB (apply r228130 to FreeBSD)
llvm-svn: 228305
2015-02-05 16:09:03 +00:00
Hafiz Abid Qadeer ec13f825a6 Fix "-data-list-register-values" MI commands.
If register numbers are not provided then this commands is supposed
to return all the register values. But it was returning nothing in
that case.

More details about this command at
https://sourceware.org/gdb/onlinedocs/gdb/GDB_002fMI-Data-Manipulation.html#GDB_002fMI-Data-Manipulation

llvm-svn: 228304
2015-02-05 15:55:43 +00:00
Tom Stellard eea3f70432 R600/SI: Fix bug in TTI loop unrolling preferences
We should be setting UnrollingPreferences::MaxCount to MAX_UINT instead
of UnrollingPreferences::Count.

Count is a 'forced unrolling factor', while MaxCount sets an upper
limit to the unrolling factor.

Setting Count to MAX_UINT was causing the loop in the testcase to be
unrolled 15 times, when it only had a maximum of 4 iterations.

llvm-svn: 228303
2015-02-05 15:32:18 +00:00
Tom Stellard 0f29de78e6 R600/SI: Fix bug from insertion of llvm.SI.end.cf into loop headers
The llvm.SI.end.cf intrinsic is used to mark the end of if-then blocks,
if-then-else blocks, and loops.  It is responsible for updating the
exec mask to re-enable threads that had been masked during the preceding
control flow block.  For example:

s_mov_b64 exec, 0x3                 ; Initial exec mask
s_mov_b64 s[0:1], exec              ; Saved exec mask
v_cmpx_gt_u32 exec, s[2:3], v0, 0   ; llvm.SI.if
do_stuff()
s_or_b64 exec, exec, s[0:1]         ; llvm.SI.end.cf

The bug fixed by this patch was one where the llvm.SI.end.cf intrinsic
was being inserted into the header of loops.  This would happen when
an if block terminated in a loop header and we would end up with
code like this:

s_mov_b64 exec, 0x3                 ; Initial exec mask
s_mov_b64 s[0:1], exec              ; Saved exec mask
v_cmpx_gt_u32 exec, s[2:3], v0, 0   ; llvm.SI.if
do_stuff()

LOOP:                       ; Start of loop header
s_or_b64 exec, exec, s[0:1] ; llvm.SI.end.cf <-BUG: The exec mask has the
                              same value at the beginning of each loop
			      iteration.
do_stuff();
s_cbranch_execnz LOOP

The fix is to create a new basic block before the loop and insert the
llvm.SI.end.cf there.  This way the exec mask is restored before the
start of the loop instead of at the beginning of each iteration.

llvm-svn: 228302
2015-02-05 15:32:15 +00:00
Bill Schmidt 433b1c3aae [PowerPC] Implement the vclz instructions for PWR8
Patch by Kit Barton.

Add the vector count leading zeros instruction for byte, halfword,
word, and doubleword sizes.  This is a fairly straightforward addition
after the changes made for vpopcnt:

 1. Add the correct definitions for the various instructions in
    PPCInstrAltivec.td
 2. Make the CTLZ operation legal on vector types when using P8Altivec
    in PPCISelLowering.cpp 

Test Plan

Created new test case in test/CodeGen/PowerPC/vec_clz.ll to check the
instructions are being generated when the CTLZ operation is used in
LLVM.

Check the encoding and decoding in test/MC/PowerPC/ppc_encoding_vmx.s
and test/Disassembler/PowerPC/ppc_encoding_vmx.txt respectively.

llvm-svn: 228301
2015-02-05 15:24:47 +00:00
Rafael Espindola f8d662aa4b Add a FIXME.
Thanks to Eric for the suggestion.

llvm-svn: 228300
2015-02-05 14:57:47 +00:00
Hafiz Abid Qadeer 6a814ad142 Fix a typo in list-register-names command.
It was giving error when process was valid.

llvm-svn: 228299
2015-02-05 14:57:07 +00:00
Alexander Kornienko b6e1fde719 [clang-tidy] Don't ignore default set of checks when a config file is found.
Summary:
This makes clang-tidy merge the default set of checks with the one
provided in the configuration file instead of just using the checks from the
config file. This adds a way to modify the default set of checks while the
previous behavior required to always define the set of checks completely.

Reviewers: djasper

Reviewed By: djasper

Subscribers: curdeius, cfe-commits

Differential Revision: http://reviews.llvm.org/D7434

llvm-svn: 228298
2015-02-05 14:50:17 +00:00
Simon Atanasyan 15774dfc21 Delete the empty folder remaining after PowerPC target removing.
llvm-svn: 228297
2015-02-05 14:37:14 +00:00
Hafiz Abid Qadeer 98219ede4d Add missing options in -var-list-children command.
The options are checked in the following order.
    print-values
    no-values
    all-values
    simple-values

See the following link for the details of these arguments.
https://sourceware.org/gdb/onlinedocs/gdb/GDB_002fMI-Variable-Objects.html

llvm-svn: 228296
2015-02-05 14:06:30 +00:00
Aaron Ballman 94d4d33a38 Removing an unused variable warning I accidentally introduced with my last warning fix; NFC.
llvm-svn: 228295
2015-02-05 13:52:42 +00:00
Aaron Ballman 1b072b340b Silencing an MSVC warning about a switch statement with no cases; NFC.
llvm-svn: 228294
2015-02-05 13:40:04 +00:00
Bruno Cardoso Lopes ab9ae87623 [X86][MMX] Handle i32->mmx conversion using movd
Implement a BITCAST dag combine to transform i32->mmx conversion patterns
into a X86 specific node (MMX_MOVW2D) and guarantee that moves between
i32 and x86mmx are better handled, i.e., don't use store-load to do the
conversion..

llvm-svn: 228293
2015-02-05 13:23:07 +00:00
Bruno Cardoso Lopes cc6089d2e0 [X86][MMX] Add several bitcast tests
Avoid regression in previously supported MMX code by adding different
combinations of tests which exercise MMX bitcasts. Small improvements
to these patterns should come next.

llvm-svn: 228292
2015-02-05 13:22:57 +00:00
Bruno Cardoso Lopes e446aefcfe [X86][MMX] Move MMX DAG node to proper file
llvm-svn: 228291
2015-02-05 13:22:50 +00:00
Alexander Potapenko 1c4d621556 [ASan] Fix interface_symbols_darwin.c test
by manually adding __asan_mz_* to the generated interface functions list.

Declaring these functions in asan_interface_internal.h doesn't work quite well:
their prototypes must match the prototypes of zone functions in malloc/malloc.h,
but some of the types (e.g. malloc_zone_t and size_t) aren't available in
asan_interface_internal.h

llvm-svn: 228290
2015-02-05 13:07:50 +00:00
Alexander Kornienko dd2dad0d24 [clang-tidy] Detect dependent initializer_lists in google-explicit-constructor.
Summary:
Detect constructors taking a single std::initializer_list even when it
is instantiation-dependent.

Reviewers: djasper

Reviewed By: djasper

Subscribers: curdeius, cfe-commits

Differential Revision: http://reviews.llvm.org/D7431

llvm-svn: 228289
2015-02-05 12:49:07 +00:00
Daniel Jasper 30029c6b58 clang-format: Fix assert triggering on carriage returns.
llvm-svn: 228288
2015-02-05 11:05:31 +00:00
Hafiz Abid Qadeer cb8c3d789e Add and update lldb-mi tests.
This patch includes following changes:

Fix comments and code style
Add new tests for many commands
Improve existing tests
Merge MiProgramArgsTestCase and MiExecTestCase
Improve runCmd of MiTestCaseBase: add exactly option
Improve test example (make it more complicated)

Patch from ki.stfu. I xfailed some tests on Linux and also
added an empty command after -gdb-exit as it blocks without
it.

Patch was reviewed in http://reviews.llvm.org/D7410.

llvm-svn: 228286
2015-02-05 10:35:17 +00:00
Pavel Labath f81ed47a07 Clean up dependency .d.$$$$ files for tests
Summary:
Mac-only tests were leaving .d.$$$$ dependency files left in the test tree. This happened
because:
1) "make clean" was invoked although no tests in the folder were run
2) The Makefile had main.d as a dependency, which meant it tried to compile the source file (to extract
dependencies).
3) The compile failed (does not compile on linux) and left behind a main.d.$$$$ temporary file.
4) "make clean" tried to clean up these temporary files, but the expansion $(wildcard *.d.[0-9]*)
was performed before the temporary file was created, so this file was not caught.

I fix this by giving all the temporary files deterministic names, which makes it easier to clean
them up afterwards. I also make the rules for generating the dependency files more robust and
less likely to leak files in the first place.

Reviewers: vharron, zturner

Subscribers: lldb-commits

Differential Revision: http://reviews.llvm.org/D7406

llvm-svn: 228285
2015-02-05 09:52:42 +00:00
Michael Kuperstein d2b6fdbc31 Teach isDereferenceablePointer() to look through bitcast constant expressions.
This fixes a LICM regression due to the new load+store pair canonicalization.

Differential Revision: http://reviews.llvm.org/D7411

llvm-svn: 228284
2015-02-05 09:15:37 +00:00
Craig Topper 0857a1da0c [X86] Add xrstors/xsavec/xsaves/clflushopt/clwb/pcommit instructions
llvm-svn: 228283
2015-02-05 08:51:06 +00:00
Craig Topper a898c2d737 [X86] Remove two feature flags that covered sets of instructions that have no patterns or intrinsics. Since we don't check feature flags in the assembler parser for any instruction sets, these flags don't provide any value. This frees up 2 of the fully utilized feature flags.
llvm-svn: 228282
2015-02-05 08:51:02 +00:00
Dimitry Andric b1e78df0fa Fix unused private field warning in stdexcept after r207695.
Add a new _LIBCPP_UNUSED define in __config, which can be used to
indicate explicitly unused items, and apply it to the __imp__ field of
__libcpp_refstring.

Somebody who knows about Microsoft C++ and IBM C++ should fill in the
unused attribute syntax appropriate for those compilers, if there is
any.

Differential Revision: http://reviews.llvm.org/D6836

llvm-svn: 228281
2015-02-05 07:40:48 +00:00
Dimitry Andric 7e187f1be9 Test commit: remove whitespace at EOL.
llvm-svn: 228280
2015-02-05 07:26:14 +00:00
Rui Ueyama df7d133cdf PECOFF: Fix I386_DIR32 relocation to an absolute symbol
Previously, we incorrectly added the image base address to an absolute
symbol address (that calculation doesn't make any sense) if an
IMAGE_REL_I386_DIR32 relocation is applied to an absolute symbol.

This patch fixes the issue. With this fix, we can link Bochs using LLD.
(Choosing Bochs has no special meaining -- I just picked it up as a
test program and found it didn't work.) This also fixes one of the
issues we currently have to link Chromium using LLD.

llvm-svn: 228279
2015-02-05 07:22:53 +00:00
Justin Hibbits f9ec0d1ea5 Implement initial Altivec support
Summary:
This adds the register plumbing, as well as register reading in FreeBSD core
dumps.  Further work on the POSIX/FreeBSD ProcessMonitor is required in order to
support ptrace access to these registers.

Reviewers: tfiala, emaste

Reviewed By: emaste

Subscribers: emaste, lldb-commits

Differential Revision: http://reviews.llvm.org/D7039

llvm-svn: 228278
2015-02-05 07:12:01 +00:00
Justin Hibbits b07ee8ded9 Add PowerPC FPR access to the process monitor
Summary: This adds reading and writing to the POSIX PowerPC ProcessMonitor.

Reviewers: emaste

Reviewed By: emaste

Subscribers: emaste, lldb-commits

Differential Revision: http://reviews.llvm.org/D7040

llvm-svn: 228277
2015-02-05 07:10:11 +00:00
Richard Smith ed83ebd77e PR22465: when performing list-initialization for a class type C, if we see an
initializer of the form {x}, where x is of type C or a type derived from C,
perform *non-list* initialization of the entity from x, but create a
CXXConstructExpr that knows that we used list-initialization syntax.

Plus some fixes to ensure we mangle correctly in this and related cases.

llvm-svn: 228276
2015-02-05 07:02:11 +00:00
Alexey Bataev 6956e2e683 [OPENMP] Initial codegen for 'single' directive.
This patch emits the following code for the single directive:

#pragma omp single
<body>
<---->

if(__kmpc_single(...)) {
    <body>
      __kmpc_end_single(...);
}
Differential Revision: http://reviews.llvm.org/D7045

llvm-svn: 228275
2015-02-05 06:35:41 +00:00
Richard Smith 520449d55e Various fixes to mangling of list-initialization.
llvm-svn: 228274
2015-02-05 06:15:50 +00:00
Matt Arsenault abd271b4e8 R600/SI: Fix i64 truncate to i1
llvm-svn: 228273
2015-02-05 06:05:13 +00:00
Alexey Bataev 9f797f32e2 [OPENMP] Codegen for 'taskyield' directive
For 'taskyield' directive emit call to kmp_int32 __kmpc_omp_taskyield(ident_t *,
kmp_int32 global_tid, int end_part); runtime function call with end_part arg set
to 0 (it is ignored).
Differential Revision: http://reviews.llvm.org/D7047

llvm-svn: 228272
2015-02-05 05:57:51 +00:00
Larisse Voufo bc9f12e7bc Disable enumeral mismatch warning when compiling llvm with gcc.
Tested with gcc 4.9.2.
Compiling with -Werror was producing:
.../llvm/lib/Target/X86/X86ISelLowering.cpp: In function 'llvm::SDValue lowerVectorShuffleAsBitMask(llvm::SDLoc, llvm::MVT, llvm::SDValue, llvm::SDValue, llvm::ArrayRef<int>, llvm::SelectionDAG&)':
.../llvm/lib/Target/X86/X86ISelLowering.cpp:7771:40: error: enumeral mismatch in conditional expression: 'llvm::X86ISD::NodeType' vs 'llvm::ISD::NodeType' [-Werror=enum-compare]
   V = DAG.getNode(VT.isFloatingPoint() ? X86ISD::FAND : ISD::AND, DL, VT, V,
                                        ^

llvm-svn: 228271
2015-02-05 04:54:51 +00:00
Matt Arsenault f28cf0cbaf Add addrspacecast node to tablegen
The node is still defined oddly so that the
address spaces are not operands and not accessible
from tablegen, but as-is this can now be used to write
a ComplexPattern with an addrspacecast root node.

llvm-svn: 228270
2015-02-05 03:35:34 +00:00
Matt Arsenault d931642cc7 Add support for double / float to EndianStream
Also add new unit tests for endian::Writer

llvm-svn: 228269
2015-02-05 03:30:08 +00:00
Shankar Easwaran d67bcb5f5c [ELF] Dont discard sections in the input file.
The reader was discarding certain types of sections from the input file.

llvm-svn: 228268
2015-02-05 02:56:06 +00:00
Dan Albert 3be4efa341 Some more -Wundef issues.
This should be all of them for Linux. Might be some for the others.

llvm-svn: 228267
2015-02-05 02:44:50 +00:00
Dan Albert 8645ea65e5 Fix some -Wundef issues.
llvm-svn: 228266
2015-02-05 02:34:59 +00:00
Michael Zolotukhin a9aadd2903 Implement new heuristic for complete loop unrolling.
Complete loop unrolling can make some loads constant, thus enabling a
lot of other optimizations. To catch such cases, we look for loads that
might become constants and estimate number of instructions that would be
simplified or become dead after substitution.

Example:
Suppose we have:
int a[] = {0, 1, 0};
v = 0;
for (i = 0; i < 3; i ++)
  v += b[i]*a[i];

If we completely unroll the loop, we would get:
v = b[0]*a[0] + b[1]*a[1] + b[2]*a[2]

Which then will be simplified to:
v = b[0]* 0 + b[1]* 1 + b[2]* 0

And finally:
v = b[1]

llvm-svn: 228265
2015-02-05 02:34:00 +00:00
Greg Clayton 68c00bd205 Keep the user data for compile units up to date since we often create lldb_private::CompileUnit objects without creating the DWARFCompileUnit objects when we do DWARF in .o files.
Now we make sure to update our DWARFCompileUnit -> lldb_private::CompileUnit user data when it isn't set to ensure quick transitions between the two.

<rdar://problem/18371367>

llvm-svn: 228264
2015-02-05 02:10:29 +00:00
Cameron Esfahani 17177d1e84 Value soft float calls as more expensive in the inliner.
Summary: When evaluating floating point instructions in the inliner, ask the TTI whether it is an expensive operation.  By default, it's not an expensive operation.  This keeps the default behavior the same as before.  The ARM TTI has been updated to return back TCC_Expensive for targets which don't have hardware floating point.

Reviewers: chandlerc, echristo

Reviewed By: echristo

Subscribers: t.p.northover, aemerson, llvm-commits

Differential Revision: http://reviews.llvm.org/D6936

llvm-svn: 228263
2015-02-05 02:09:33 +00:00
Nico Weber 94b2368c24 Revert r228258.
It caused a chromium base unittest that tests throwing and catching SEH
exceptions to fail (http://crbug.com/455488) and I suspect it might also
be the cause of the chromium clang win 64-bit shared release builder timing
out during compiles.  So revert to see if that's true.

llvm-svn: 228262
2015-02-05 02:08:50 +00:00
Greg Clayton 08928f308b Don't wait for the dynamic loader to set a module as a dynamic link editor, figure it out through the ObjectFile.
Background: dyld binaries often have extra symbols in their symbol table like "malloc" and "free" for the early bringup of dyld and we often don't want to set breakpoints in dynamic linker binaries. We also don't want to call the "malloc" or "free" function in dyld when a user writes an expression like "(void *)malloc(123)" so we need to avoid doing name lookups in dyld. We mark Modules as being dynamic link editors and this helps do correct lookups for breakpoints by name and function lookups.

<rdar://problem/19716267>

llvm-svn: 228261
2015-02-05 02:01:34 +00:00
Shankar Easwaran 50ad727051 [ELF] Dont release ownership of MemoryBuffer.
MemoryBuffer is being released too early. ELFFile owns MemoryBuffer.

llvm-svn: 228260
2015-02-05 02:00:03 +00:00