Commit Graph

239273 Commits

Author SHA1 Message Date
Zachary Turner b6958a888e Disable TestThreadSpecificBpPlusCondition on Windows.
Test frequently times out stalling the test runner.

llvm-svn: 278529
2016-08-12 17:15:38 +00:00
Pete Cooper 980a935e27 constify InstCombine::foldAllocaCmp. NFC.
This is part of an effort to constify ValueTracking.cpp.  This change is
to methods which need const Value* instead of Value* to go with the upcoming
changes to ValueTracking.

llvm-svn: 278528
2016-08-12 17:13:28 +00:00
Todd Fiala 1dc065b659 Link LLDB only against libclang and libLLVM .a files to fix macOS build
The Xcode macOS build of LLDB is currently broken after
https://reviews.llvm.org/D23232 landed, see
http://lab.llvm.org:8080/green/job/lldb_build_test/20014/console,
because we’re trying to link against all .a files found in the
llvm-build/lib directory. Let’s be more specific in what we link
against. This patch applies a regexp to only use “libclang.*”,
“libLLVM.*” and not “libclang_rt.*” static archives.

Change by Kuba Mracek (formerly Kuba Brecka)

See review here:
https://reviews.llvm.org/D23444

Reviewers: tfiala, compnerd
llvm-svn: 278527
2016-08-12 17:01:19 +00:00
Zachary Turner bd22bf26cb Fix build on Windows.
llvm-svn: 278526
2016-08-12 16:52:31 +00:00
Duncan P. N. Exon Smith e6bc5f9b1b BugReporter: Use ilist_half_embedded_sentinel_traits, NFC
This avoids duplicated code with llvm/ADT/ilist.h.  No functionality
change.

llvm-svn: 278525
2016-08-12 16:46:25 +00:00
Greg Clayton c6c420fca1 Switch over to using socketpair for local debugserver connections as they are twice as fast as TCP sockets (on macOS at least).
This change opens a socket pair and passes the second socket pair file descriptor down to the debugserver binary using a new option: "--fd=N" where N is the file descriptor. This file descriptor gets passed via posix_spawn() so that there is no need to do any bind/listen or bind/accept calls and eliminates the hanshake unix socket that is used to pass the result of the actual port that ends up being used so it can save time on launch as well as being faster.

This is currently only enabled on __APPLE__ builds. Other OSs should try modifying the #define from ProcessGDBRemote.cpp but the first person will need to port the --fd option over to lldb-server. Any OSs that enable USE_SOCKETPAIR_FOR_LOCAL_CONNECTION in their native builds can use the socket pair stuff. The #define is Apple only right now, but looks like:

#if defined (__APPLE__)
#define USE_SOCKETPAIR_FOR_LOCAL_CONNECTION 1
#endif

<rdar://problem/27814880> 

llvm-svn: 278524
2016-08-12 16:46:18 +00:00
Duncan P. N. Exon Smith 623c4c1572 MachO: Use ilist_half_embedded_sentinel_traits, NFC
Use ilist_half_embedded_sentinel_traits for the list of
lld::mach_o::normalized::TrieEdge, rather than duplicating the code.

llvm-svn: 278523
2016-08-12 16:25:04 +00:00
Dehao Chen c0a1e432c7 Fine tuning of sample profile propagation algorithm.
Summary: The refined propagation algorithm is more accurate and robust.

Reviewers: davidxl, dnovillo

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D23224

llvm-svn: 278522
2016-08-12 16:22:12 +00:00
Duncan P. N. Exon Smith 4a1e87d9c6 Core: Use ilist_full_embedded_sentinel_traits for SimpleReference
Avoid custom code for sentinel traits in SimpleReference (the ilist node
for references to a SimpleDefinedAtom), since they'll soon/eventually
disappear from ilist entirely.

Rather than using a BumpPtrAllocator, this drops the lazy sentinel
characteristics and stores the sentinel directly in the ilist.  This
unconditionally allocates the sentinel.

At first glance, this looks like it might increase memory usage
slightly, since an unreferenced SimpleDefinedAtom pays for a
6-pointer-sized sentinel even when its list of references is empty.  In
practice, the sentinel was being lazily allocated at the first call to
DefinedAtom::begin/end anyway.  I don't expect any real memory effects
here.

Moreover, this is an intermediate state.  The ilist_*sentinel_traits are
being phased out.

As a preview of the final state: in lieu of a NodeTy sentinel, the ilist
will have a single, untemplated list_node_base that has next/prev
pointers.  This base node will serve both as a sentinel and as a pointer
to the head of the list (the same memory layout as
ilist_half_embedded_sentinel_traits, but without the UB).

llvm-svn: 278521
2016-08-12 16:19:34 +00:00
Artur Pilipenko 87e4038a91 [x86] X86ISelLowering zext(add_nuw(x, C)) --> add(zext(x), C_zext)
Currently X86ISelLowering has a similar transformation for sexts:
sext(add_nsw(x, C)) --> add(sext(x), C_sext)

In this change I extend this code to handle zexts as well.

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D23359

llvm-svn: 278520
2016-08-12 16:08:30 +00:00
Ehsan Amiri 17e1701075 [BasicAA] Avoid calling GetUnderlyingObject, when the result of a previous call can be reused.
Recursive calls to aliasCheck from alias[GEP|Select|PHI] may result in a second call to GetUnderlyingObject for a Value, whose underlying object is already computed. This patch ensures that in this situations, the underlying object is not computed again, and the result of the previous call is resued.

https://reviews.llvm.org/D22305

llvm-svn: 278519
2016-08-12 16:05:03 +00:00
Artur Pilipenko 2e8f82d962 [LVI] Take guards into account
Teach LVI to gather control dependant constraints from guards.

Reviewed By: sanjoy

Differential Revision: https://reviews.llvm.org/D23358

llvm-svn: 278518
2016-08-12 15:52:23 +00:00
Teresa Johnson 77a1f7566c Add move ops to satisfy MSVC.
Try to appease Windows bots after r278508:
http://lab.llvm.org:8011/builders/sanitizer-windows/builds/27250
http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/14776

llvm-svn: 278517
2016-08-12 15:39:26 +00:00
Daniel Sanders 875671185e [mips] Remove obstack.cc XFAIL added in r278504 since it broke X86 and PPC somehow.
Adding the XFAIL has caused msan to report a different line number in the call
stack (@LINE-3 rather than @LINE-30). The new line number looks more correct
at first glance since it's the line that uses uninitialized memory rather than
the first non-whitespace line of the file but this needs investigating.

llvm-svn: 278516
2016-08-12 15:34:35 +00:00
Geoff Berry 22dfbc5637 [AArch64] Re-factor code shared by AArch64LoadStoreOpt and AArch64InstrInfo.
This re-factoring could cause the following slight changes in generated
code, though none were observed during testing:

- MachineScheduler could decide not to cluster some loads/stores if
  there are other load/stores with non-pairable opcodes that have the
  same base register and offset as a pairable set of load/stores.  One
  case of different MachineScheduler pairing did show up in my testing,
  but it wasn't due to this issue, but due
  BaseMemOpClusterMutation::clusterNeighboringMemOps() being unstable
  w.r.t. the order it considers memory operations.  See PR28942.

- The ImplicitNullChecks optimization could be done for more load/store
  opcodes.  This optimization isn't done for C/C++ code, so it didn't
  show up in my testing.

Reviewers: mcrosier, t.p.northover

Subscribers: aemerson, rengolin, mcrosier, llvm-commits

Differential Revision: https://reviews.llvm.org/D23365

llvm-svn: 278515
2016-08-12 15:26:00 +00:00
Artur Pilipenko b623088abe [LVI] Fix potential memory corruption in getValueFromCondition
Rewrite Visited[Cond] = getValueFromConditionImpl(..., Visited) statement which can lead to a memory corruption since getValueFromConditionImpl changes Visited map and invalidates the iterators.

llvm-svn: 278514
2016-08-12 15:08:15 +00:00
Duncan P. N. Exon Smith 0d2ed35d3e ADT: Share code for embedded sentinel traits, NFC
Share code for the (mostly problematic) embedded sentinel traits.
- Move the LLVM_NO_SANITIZE("object-size") attribute to
  ilist_half_embedded_sentinel_traits and ilist_embedded_sentinel_traits
  (previously it spread throughout the code duplication).
- Add an ilist_full_embedded_sentinel_traits which has no UB (but has
  the downside of storing the complete node).
- Replace all the custom sentinel traits in LLVM with a declaration of
  ilist_sentinel_traits that inherits from one of the embedded sentinel
  traits classes.

There are still custom sentinel traits in other LLVM subprojects.  I'll
remove those in a follow-up.

Nothing at all should be changing here, this is just rearranging code.
Note that the final goal here is to remove the sentinel traits
altogether, settling on the memory layout of
ilist_half_embedded_sentinel_traits without the UB.  This intermediate
step moves the logic into ilist.h.

llvm-svn: 278513
2016-08-12 15:00:55 +00:00
Teresa Johnson ad7eb9db46 Fix type to avoid problems on 32-bit builds
lto::InputFile::Symbol::getCommonSize should return uint64_t instead of
size_t since it is returning the result of DataLayout::getTypeAllocSize
which returns uint64_t, and the result of getCommonSize is assigned to a
uint64_t variable. On 32-bit builds size_t is unsigned int and there are
type errors. This was introduced in r278338.

llvm-svn: 278512
2016-08-12 14:55:43 +00:00
James Y Knight 2cc9da9a65 Revert "[Sparc] Leon errata fix passes."
...and the two followup commits:
Revert "[Sparc][Leon] Missed resetting option flags from check-in 278489."
Revert "[Sparc][Leon] Errata fixes for various errata in different
versions of the Leon variants of the Sparc 32 bit processor."

This reverts commit r274856, r278489, and r278492.

llvm-svn: 278511
2016-08-12 14:48:09 +00:00
Tamas Berghammer 3785393def Skip 2 android test what is broken because of debuggerd
debuggerd is a crash reporting system on android what installs some
signal handler for SEGV to print a backtrace in the log. Its behavior
breaks tests where the test tries to continue after a SEGV so we skip
them as this behavior isn't required on android anyway.

llvm-svn: 278510
2016-08-12 14:17:05 +00:00
Teresa Johnson 4223dd8559 [PM] Port NameAnonFunction pass to new pass manager
Summary:
Port the NameAnonFunction pass and add a test.

Depends on D23439.

Reviewers: mehdi_amini

Subscribers: llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D23440

llvm-svn: 278509
2016-08-12 14:03:36 +00:00
Teresa Johnson f93b246f8b [PM] Port ModuleSummaryIndex analysis to new pass manager
Summary:
Port the ModuleSummaryAnalysisWrapperPass to the new pass manager.
Use it in the ported BitcodeWriterPass (similar to how we use the
legacy ModuleSummaryAnalysisWrapperPass in the legacy WriteBitcodePass).

Also, pass the -module-summary opt flag through to the new pass
manager pipeline and through to the bitcode writer pass, and add
a test that uses it.

Reviewers: mehdi_amini

Subscribers: llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D23439

llvm-svn: 278508
2016-08-12 13:53:02 +00:00
Martin Bohme 565bdd9fed [ASTMatchers] Add templateTypeParmDecl() to Registry.cpp
Summary:
This appears to have been forgotten when templateTypeParmDecl() was initially
added.

Reviewers: alexfh, aaron.ballman

Subscribers: aaron.ballman, klimek, aemerson, rengolin, samparker, cfe-commits

Differential Revision: https://reviews.llvm.org/D23448

llvm-svn: 278507
2016-08-12 13:51:00 +00:00
Davide Italiano 54eb8ffd55 [Core] Retire yet another unused member function.
The code in lib/ could use a lot of love :(

llvm-svn: 278506
2016-08-12 12:37:23 +00:00
Davide Italiano 0f672ed87a [Core] Simplify a bit. NFCI.
llvm-svn: 278505
2016-08-12 12:34:39 +00:00
Daniel Sanders 6a540c1f38 [mips] XFAIL the new mips64el compiler-rt tests that fail on clang-cmake-mipsel.
The mips64el compiler-rt build has recently been enabled. XFAIL the failing
tests to make the buildbot green again.

The two asan tests require the integrated assembler. This will be fixed soon
for Debian mips64el but not for any other mips64el targets since doing so
requires triple-related issues to be fixed..
The msan tests are largely failing because caused by a kernel update (a patch
has already been posted for this).
I'm not sure why the dfsan test fails yet.

llvm-svn: 278504
2016-08-12 11:56:36 +00:00
Simon Pilgrim 728134c277 Fix Wdocumentation unknown parameter warning
llvm-svn: 278503
2016-08-12 11:43:57 +00:00
Simon Pilgrim 687d71e877 [X86][SSE] Add support for combining target shuffles to PSLLDQ/PSRLDQ byte shifts
llvm-svn: 278502
2016-08-12 11:24:34 +00:00
Andrey Bokhanko 1d31e45a8b Fix For pr28288 - Error message in shift of vector values
This fixes an error in type checking of shift of vector values.

Patch by Vladimir Yakovlev.

Differential Revision: https://reviews.llvm.org/D21678

llvm-svn: 278501
2016-08-12 11:22:12 +00:00
Pavel Labath fdb2d99eaf Fix-up r278299 for windows
FD_SETSIZE on windows limits the number of file descriptors, rather than their individual
magnitude (the underlying implementation uses an array rather than a bitset). This meant that the
assert in the SelectHelper was incorrect, and failing all the time. Fix that.

I am not sure whether this should be #ifdef MSVC, or #ifdef WINDOWS, but my feeling is that a
more posix-conforming implementation on windows would choose the bitset implementation, so I'm
sticking with the former.

llvm-svn: 278500
2016-08-12 11:20:21 +00:00
Krzysztof Parzyszek be976d4ea9 [Hexagon] Standardize pseudo-instructions for calls and returns
- CALLv3nr        PS_call_nr
- CALLRv3nr       PS_callr_nr
- CALLstk         PS_call_stk

- TCRETURNi       PS_tailcall_i
- TCRETURNr       PS_tailcall_r

- JMPret          PS_jmpret
- JMPrett         PS_jmprett
- JMPretf         PS_jmpretf
- JMPrettnew      PS_jmprettnew
- JMPretfnew      PS_jmpretfnew
- JMPrettnewpt    PS_jmprettnewpt
- JMPretfnewpt    PS_jmpretfnewpt

llvm-svn: 278499
2016-08-12 11:12:02 +00:00
Krzysztof Parzyszek ab9127ca3c [Hexagon] Treat non-returning indirect calls as scheduling boundaries
llvm-svn: 278498
2016-08-12 11:01:10 +00:00
Jonas Hahnfeld d5fa61ee4a Fix cuda-detect.cu when CLANG_DEFAULT_CXX_STDLIB is set
Reported by Ismail Donmez!

llvm-svn: 278497
2016-08-12 10:36:04 +00:00
Artur Pilipenko 6669f253d5 [LVI] Take range metadata into account while calculating icmp condition constraints
Take range metadata into account for conditions like this:

%length = load i32, i32* %length_ptr, !range !{i32 0, i32 2147483647}
%cmp = icmp ult i32 %a, %length

This is a common pattern for range checks where the length of the array is dynamically loaded.

Reviewed By: sanjoy

Differential Revision: https://reviews.llvm.org/D23267

llvm-svn: 278496
2016-08-12 10:14:11 +00:00
Benjamin Kramer 05e760ec4b [Webassembly] disable unstable test.
It reads uninitialized memory and crashes randomly.

llvm-svn: 278495
2016-08-12 10:13:45 +00:00
Simon Pilgrim ed96b9adfb [X86][SSE] Fixed PALIGNR target shuffle decode
The PALIGNR target shuffle decode was not taking into account that DecodePALIGNRMask (rather oddly) expects the operands to be in reverse order, nor was it detecting unary patterns, causing combines to combine with the incorrect input.

The cgbuiltin, auto upgrade and instruction comments code correctly swap the operands so are not affected.

llvm-svn: 278494
2016-08-12 10:10:51 +00:00
Artur Pilipenko 635625855f [LVI] Handle any predicate in comparisons like icmp <pred> (add Val, Offset), ...
Currently LVI can only gather value constraints from comparisons like:

* icmp <pred> Val, ...
* icmp ult (add Val, Offset), ...

In fact we can handle any predicate in latter comparisons.

Reviewed By: sanjoy

Differential Revision: https://reviews.llvm.org/D23357

llvm-svn: 278493
2016-08-12 10:05:11 +00:00
Chris Dewhurst 5247af24c3 [Sparc][Leon] Missed resetting option flags from check-in 278489.
llvm-svn: 278492
2016-08-12 09:54:39 +00:00
Pavel Labath 709269153f XFAIL TestNamespaceDefinitions on gcc-4.8 and below
llvm-svn: 278491
2016-08-12 09:52:14 +00:00
Pavel Labath b491ef4317 Make TestCallStopAndContinue clang-format-resilient
llvm-svn: 278490
2016-08-12 09:39:22 +00:00
Chris Dewhurst 829f8efe55 [Sparc][Leon] Errata fixes for various errata in different versions of the Leon variants of the Sparc 32 bit processor.
The nature of the errata are listed in the comments preceding the errata fix passes. Relevant unit tests are implemented for each of these.

These changes update older versions of these errata fixes with improvements to code and unit tests.

Differential Revision: https://reviews.llvm.org/D21960

llvm-svn: 278489
2016-08-12 09:34:26 +00:00
Benjamin Kramer 6604e9638b Prune unused diagnostics. NFC.
llvm-svn: 278488
2016-08-12 09:23:14 +00:00
Benjamin Kramer df8bf140dd [C++1z] Fix crash when decomposing structs with anonymous members.
The diagnostic format was invalid.

llvm-svn: 278487
2016-08-12 09:19:34 +00:00
George Rimar 06ae683620 [ELF] - Linkerscript: reimplemented output sections constrains matching functionality.
Previously filtering that was used worked incorrectly. 
For example for next script it would just remove both sections completely:

SECTIONS { 
. = 0x1000;
.aaa : ONLY_IF_RW { *(.aaa.*) } 
. = 0x2000;
.aaa : ONLY_IF_RO { *(.aaa.*) } 
}

Patch fixes above issues and adds testcase showing the issue. Testcase is a subset of
FreeBSD script which has:

.eh_frame       : ONLY_IF_RO { KEEP (*(.eh_frame)) }
...
.eh_frame       : ONLY_IF_RW { KEEP (*(.eh_frame)) }

Differential revision: https://reviews.llvm.org/D23326

llvm-svn: 278486
2016-08-12 09:07:57 +00:00
Benjamin Kramer bbff9c6130 [Coroutines] Move class into anonymous namespace.
Hopefully fixes visibility warnings from GCC. No functionality change.

llvm-svn: 278485
2016-08-12 08:47:13 +00:00
Haicheng Wu d9cbb1608f Revert "[BranchFolding] Restrict tail merging loop blocks after MBP"
This reverts commit r278463 because it hits the bot.

llvm-svn: 278484
2016-08-12 08:40:24 +00:00
Roger Ferrer Ibanez 722a4db198 This patch implements PR#22821.
Taking the address of a packed member is dangerous since the reduced
alignment of the pointee is lost. This can lead to memory alignment
faults in some architectures if the pointer value is dereferenced.

This change adds a new warning to clang emitted when taking the address
of a packed member. A packed member is either a field/data member
declared as attribute((packed)) or belonging to a struct/class
declared as such. The associated flag is -Waddress-of-packed-member.
Conversions (either implicit or via a valid casting) to pointer types
with lower or equal alignment requirements (e.g. void* or char*)
will silence the warning.

Differential Revision: https://reviews.llvm.org/D20561

llvm-svn: 278483
2016-08-12 08:04:13 +00:00
Simon Atanasyan 85c6b44817 [ELF][MIPS] Support .MIPS.abiflags section
This section supersedes .reginfo and .MIPS.options sections. But for now
we have to support all three sections for ABI transition period.

llvm-svn: 278482
2016-08-12 06:28:49 +00:00
Gor Nishanov 0f303accde [Coroutines]: Part6b: Add coro.id intrinsic.
Summary:
1. Make coroutine representation more robust against optimization that may duplicate instruction by introducing coro.id intrinsics that returns a token that will get fed into coro.alloc and coro.begin. Due to coro.id returning a token, it won't get duplicated and can be used as reliable indicator of coroutine identify when a particular coroutine call gets inlined.
2. Move last three arguments of coro.begin into coro.id as they will be shared if coro.begin will get duplicated.
3. doc + test + code updated to support the new intrinsic.

Reviewers: mehdi_amini, majnemer

Subscribers: mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D23412

llvm-svn: 278481
2016-08-12 05:45:49 +00:00
Simon Atanasyan 6be57b3fb1 [ELF][MIPS] Fix the comment
llvm-svn: 278480
2016-08-12 05:43:42 +00:00