Benjamin Kramer [Tue, 4 Aug 2015 13:34:50 +0000 (13:34 +0000)]
[AST] Really allocate a SmallVector to the right size.
set_size only resets the end pointer and asserts if it is used to grow
the buffer. This would crash when mangling a float with more than 80 bits,
add a test with a ppc double double (128 bits).
Found by inspection.
llvm-svn: 243979
Bruce Mitchener [Tue, 4 Aug 2015 13:25:23 +0000 (13:25 +0000)]
[lldb-mi] Fix non-OS X builds.
llvm-svn: 243978
Duncan P. N. Exon Smith [Tue, 4 Aug 2015 13:24:26 +0000 (13:24 +0000)]
Fix 80-column
llvm-svn: 243977
Duncan P. N. Exon Smith [Tue, 4 Aug 2015 13:23:30 +0000 (13:23 +0000)]
Linker: Fix ASan failure from r243961
r243883 and r243961 made a use-after-free far more likely:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/6041/steps/check-llvm%20asan/logs/stdio
Unresolved nodes get inserted into the `Cycles` array. If they later
get resolved through RAUW, we need to update the reference. It's
interesting that this never hit before (maybe an asan-ified clang
bootstrap with `-flto -g` would have hit it, but I admit I haven't tried
anything quite that crazy).
llvm-svn: 243976
Bruce Mitchener [Tue, 4 Aug 2015 13:12:31 +0000 (13:12 +0000)]
[lldb-mi] Simplify MICmnMIOutOfBandRecord implementation.
Summary:
* Remove extraneous members that were just storing temporary
values.
* OutOfBand_e parameters don't need to be const as they are
scalars.
* Switch from a map with CMIUtilString values to using a mapping
function. This uses a switch statement which will generate
a warning if a new result class is added.
* Make BuildAsyncRecord a static function rather than a private
member function so that we can construct the result text
correctly and avoid having extra stuff in the header.
Reviewers: abidh, ki.stfu
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D11751
llvm-svn: 243975
Benjamin Kramer [Tue, 4 Aug 2015 12:59:41 +0000 (12:59 +0000)]
[AST] Simplify code by letting ArrayRef do the copying.
No functionality change intended.
llvm-svn: 243974
Benjamin Kramer [Tue, 4 Aug 2015 12:34:30 +0000 (12:34 +0000)]
[CodeGen] Don't write into a SmallVector's uninitialized memory.
That's just ugly and not needed in this case. No functionality change.
llvm-svn: 243973
Benjamin Kramer [Tue, 4 Aug 2015 12:34:23 +0000 (12:34 +0000)]
[ASTContext] Add a templated convenience wrapper for Allocate.
This brings ASTContext closer to LLVM's Allocator concept. Ideally we
would just derive ASTContext from llvm::AllocatorBase, but that does
not work because ASTContext's allocator is mutable and we allocate using
const ASTContext& everywhere.
llvm-svn: 243972
Tobias Grosser [Tue, 4 Aug 2015 11:49:39 +0000 (11:49 +0000)]
Run ScopInfo after the ScopViewers
This change is required to see the detected scops even in cases where there is
no other ScopInfo user after the ScopViewers. Before this change, when
running with -polly-optimizer=none -polly-code-generator=none detected scops
have not been shown.
llvm-svn: 243971
Benjamin Kramer [Tue, 4 Aug 2015 11:27:08 +0000 (11:27 +0000)]
[FileManager] Use StringRef::copy to store canonical paths.
No functional change intended.
llvm-svn: 243970
Alexey Bataev [Tue, 4 Aug 2015 11:18:19 +0000 (11:18 +0000)]
[OPENMP 4.1] Support for 'linear' clause in loop directives.
OpenMP 4.1 allows 'linear' clause in loop directives. Patch adds support for it.
llvm-svn: 243969
Benjamin Kramer [Tue, 4 Aug 2015 11:18:09 +0000 (11:18 +0000)]
[Edit] Use StringRef's copy method. No functional change intended.
llvm-svn: 243968
Bruce Mitchener [Tue, 4 Aug 2015 10:24:20 +0000 (10:24 +0000)]
[lldb-mi] Use empty arg lists instead of (void).
Summary: This brings the code more in line with the usual LLDB style. NFC.
Reviewers: abidh, ki.stfu
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D11746
llvm-svn: 243967
Benjamin Kramer [Tue, 4 Aug 2015 10:22:38 +0000 (10:22 +0000)]
[AST] Use StringRef's convenient copy method. No functionality change.
llvm-svn: 243966
David Majnemer [Tue, 4 Aug 2015 08:21:40 +0000 (08:21 +0000)]
Drive-by fixes for LandingPad -> EHPad
This change was done as an audit and is by inspection. The new EH
system is still very much a work in progress. NFC for the landingpad
case.
llvm-svn: 243965
Alexey Bataev [Tue, 4 Aug 2015 08:10:48 +0000 (08:10 +0000)]
[OPENMP] Fix compiler crash during data-sharing attributes analysis.
If a global variable is marked as private in OpenMP construct and then is used in of the private clauses of the same construct, it might cause compiler crash because of incorrect capturing.
llvm-svn: 243964
Simon Pilgrim [Tue, 4 Aug 2015 08:05:27 +0000 (08:05 +0000)]
[InstCombine] Split off SSE2/AVX2 vector shift tests.
These aren't vector demanded bits tests. More tests to follow.
llvm-svn: 243963
Simon Pilgrim [Tue, 4 Aug 2015 07:49:58 +0000 (07:49 +0000)]
[InstCombine] Moved SSE vector shift constant folding into its own helper function. NFCI.
This will make some upcoming bugfixes + improvements easier to manage.
llvm-svn: 243962
Duncan P. N. Exon Smith [Tue, 4 Aug 2015 06:42:31 +0000 (06:42 +0000)]
Linker: Fix references to uniqued nodes after r243883
r243883 started moving 'distinct' nodes instead of duplicated them in
lib/Linker. This had the side-effect of sometimes not cloning uniqued
nodes that reference them. I missed a corner case:
!named = !{!0}
!0 = !{!1}
!1 = distinct !{!0}
!0 is the entry point for "remapping", and a temporary clone (say,
!0-temp) is created and mapped in case we need to model a uniquing
cycle.
Recursive descent into !1. !1 is distinct, so we leave it alone,
but update its operand to !0-temp.
Pop back out to !0. Its only operand, !1, hasn't changed, so we don't
need to use !0-temp. !0-temp goes out of scope, and we're finished
remapping, but we're left with:
!named = !{!0}
!0 = !{!1}
!1 = distinct !{null} ; uh oh...
Previously, if !0 and !0-temp ended up with identical operands, then
!0-temp couldn't have been referenced at all. Now that distinct nodes
don't get duplicated, that assumption is invalid. We need to
!0-temp->replaceAllUsesWith(!0) before freeing !0-temp.
I found this while running an internal `-flto -g` bootstrap. Strangely,
there was no case of this in the open source bootstrap I'd done before
commit...
llvm-svn: 243961
Mehdi Amini [Tue, 4 Aug 2015 06:32:54 +0000 (06:32 +0000)]
Update test suite to make "ninja check" succeed without native backend builtin
Requires "native" feature in most places that were failing.
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 243960
Mehdi Amini [Tue, 4 Aug 2015 06:32:45 +0000 (06:32 +0000)]
Move generic MIR tests in their own subdir, requires "native" as well
These tests rely on the native backend to be built-in.
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 243959
Mehdi Amini [Tue, 4 Aug 2015 06:32:31 +0000 (06:32 +0000)]
Improve lit "native" feature to check if the native backend is builtin
The goal is to have 'ninja check' passing even if the X86 backend is
not built.
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 243958
Justin Bogner [Tue, 4 Aug 2015 06:29:58 +0000 (06:29 +0000)]
Remove the configure and cmake checks for sys/wait.h
If we don't have sys/wait.h and we're on a unix system there's no way
that several of the llvm tools work at all. This includes clang.
Just remove the configure and cmake checks entirely - we'll get a
build error instead of building something broken now.
llvm-svn: 243957
Hal Finkel [Tue, 4 Aug 2015 06:29:12 +0000 (06:29 +0000)]
[SDAG] Fix a result chain in ExpandUnalignedLoad
On the code path in ExpandUnalignedLoad which expands an unaligned vector/fp
value in terms of a legal integer load of the same size, the ChainResult needs
to be the chain result of the integer load.
No in-tree test case is currently available.
Patch by Jan Hranac!
llvm-svn: 243956
Adam Nemet [Tue, 4 Aug 2015 05:16:20 +0000 (05:16 +0000)]
[LAA] Remove unused pointer partition argument from addRuntimeCheck, NFC
This variant of addRuntimeCheck is only used now from the LoopVectorizer
which does not use this parameter.
llvm-svn: 243955
Chen Li [Tue, 4 Aug 2015 04:41:34 +0000 (04:41 +0000)]
Introduce enum value for previously defined metadata -- make.implicit
Summary: This patch adds enum value for an existing metadata type -- make.implicit. Using preassigned enum will be helpful to get compile time type checking and avoid string construction and comparison. The patch also changes uses of make.implicit from string metadata to enum metadata. There is no functionality change.
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11698
llvm-svn: 243954
Bruce Mitchener [Tue, 4 Aug 2015 04:01:16 +0000 (04:01 +0000)]
[debugserver] Fix "control may reach end of non-void function" warnings.
Reviewers: clayborg, jingham
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D11745
llvm-svn: 243953
Saleem Abdulrasool [Tue, 4 Aug 2015 03:57:56 +0000 (03:57 +0000)]
ARM: support windows division routines
This adds the software division routines for the Windows RTABI. These are not
expected to be used often though as most modern Windows ARM capable targets
support hardware division. In the case that the target CPU doesnt support
hardware division, this will be the fallback.
llvm-svn: 243952
Saleem Abdulrasool [Tue, 4 Aug 2015 03:57:52 +0000 (03:57 +0000)]
ARM: make Darwin libcall registration table driven (NFC)
Make the libcall updating table driven similar to the approach that the Linux
and Windows codepath does below. NFC.
llvm-svn: 243951
Chandler Carruth [Tue, 4 Aug 2015 03:53:04 +0000 (03:53 +0000)]
[UB] Avoid a really broken call to realloc that would later result in
a bad call to memcpy.
When we only have a buffer from one of the two reparse calls, we can
just return that buffer rather than going through the realloc/memcpy
dance.
Found with UBsan.
llvm-svn: 243950
Chandler Carruth [Tue, 4 Aug 2015 03:53:01 +0000 (03:53 +0000)]
[UB] Guard two calls to memcpy in generated attribute code to handle
null StringRef objects as inputs.
Found by UBSan.
llvm-svn: 243949
Chandler Carruth [Tue, 4 Aug 2015 03:53:00 +0000 (03:53 +0000)]
[UB] Another place where we were trying to put string data into
a BumpPtrAllocator. This at least now handles the case where there is no
concatentation without calling memcpy on a null pointer. It might be
interesting to handle the case where everything is empty without
round-tripping through the allocator, but it wasn't clear to me if the
pointer returned is significant in any way, so I've left it in
a conservatively more-correct state.
Again, found with UBSan.
llvm-svn: 243948
Chandler Carruth [Tue, 4 Aug 2015 03:52:58 +0000 (03:52 +0000)]
[UB] When attaching empty strings to the AST, use an empty StringRef
rather than forcing the bump pointer allocator to produce a viable
pointer. This also fixes UB when we would try to memcpy from the null
incoming StringRef.
llvm-svn: 243947
Chandler Carruth [Tue, 4 Aug 2015 03:52:56 +0000 (03:52 +0000)]
[UB] Fix the two ways that we would try to memcpy from a null buffer in
the nested name specifier code.
First, skip the entire thing when the input is empty.
Next, handle the case where we started off with a null buffer and a zero
capacity to skip copying and freeing.
This was found with UBSan.
llvm-svn: 243946
Chandler Carruth [Tue, 4 Aug 2015 03:52:52 +0000 (03:52 +0000)]
[UB] Fix two cases of UB in copy/pasted code from SmallVector.
We should really stop copying and pasting code around. =/
Found by UBSan.
llvm-svn: 243945
Chandler Carruth [Tue, 4 Aug 2015 03:48:26 +0000 (03:48 +0000)]
[UB] Don't allocate space for contained types and then try to copy the
contained types into the space when we have no contained types. This
fixes the UB stemming from a call to memcpy with a null pointer. This
also reduces the calls to allocate because this actually happens in
a notable client - Clang.
Found by UBSan.
llvm-svn: 243944
Jason Molenda [Tue, 4 Aug 2015 03:14:49 +0000 (03:14 +0000)]
Bump the version # in the xcode proj file from major number 340 to 350.
The SB API for major version 340 won't be changing any more.
llvm-svn: 243943
Sean Silva [Tue, 4 Aug 2015 03:12:33 +0000 (03:12 +0000)]
[GettingStarted.rst] Commit the right patch.
Looks like the rebased version that Mehdi committed didn't incorporate
the latest changes.
Patch by Erik de Castro Lopo <erikd@mega-nerd.com>!
llvm-svn: 243942
Bruce Mitchener [Tue, 4 Aug 2015 02:41:49 +0000 (02:41 +0000)]
Fix cmake build on non-Windows platforms.
r243914 introduced a change which mistakenly tried to build the
Windows minidump code on all platforms rather than only on Windows.
llvm-svn: 243941
Richard Smith [Tue, 4 Aug 2015 02:05:09 +0000 (02:05 +0000)]
[modules] Make IndirectFieldDecl mergeable to avoid lookup ambiguity when the same anonymous union is defined across multiple modules.
llvm-svn: 243940
Sanjoy Das [Tue, 4 Aug 2015 01:52:05 +0000 (01:52 +0000)]
Revert "[LSR] Generate and use zero extends"
This reverts commit r243348 and r243357. They caused PR24347.
llvm-svn: 243939
Argyrios Kyrtzidis [Tue, 4 Aug 2015 01:44:13 +0000 (01:44 +0000)]
Remove unused header include.
llvm-svn: 243938
Marshall Clow [Tue, 4 Aug 2015 01:38:34 +0000 (01:38 +0000)]
Change char_traits<char16_t>::eof() to return 0xFFFF instead of 0xDFFF. Fixes PR#24342
llvm-svn: 243937
Ahmed Bougacha [Tue, 4 Aug 2015 01:38:08 +0000 (01:38 +0000)]
[AArch64] Rename FP formats to be more consistent. NFC.
Some are named "FP", others "SD", others still "FP*SD".
Rename all this to just use "FP", which, except for conversions
(which don't use this format naming scheme), implies "SD" anyway.
llvm-svn: 243936
Ahmed Bougacha [Tue, 4 Aug 2015 01:29:38 +0000 (01:29 +0000)]
[AArch64] Add isel support for f16 indexed LD/ST.
llvm-svn: 243935
Chandler Carruth [Tue, 4 Aug 2015 01:00:56 +0000 (01:00 +0000)]
[UB] Fix yet another use of memcpy with a null pointer argument. I think
this is the last of them in my build of LLVM. Haven't tried Clang yet.
Found via UBSan.
llvm-svn: 243934
Ahmed Bougacha [Tue, 4 Aug 2015 00:55:11 +0000 (00:55 +0000)]
[AArch64][v8.1a] The "pan" sysreg isn't MSR-specific. NFCI.
It's already in SysRegMappings, no need to also have it in MSRMappings:
the latter is only used if we didn't find a match in the former.
llvm-svn: 243933
Chandler Carruth [Tue, 4 Aug 2015 00:53:01 +0000 (00:53 +0000)]
[UB] Fix another place where we would pass a null pointer to memcpy.
This too was found by UBSan. Down to 35 failures for me.
llvm-svn: 243932
Ahmed Bougacha [Tue, 4 Aug 2015 00:49:08 +0000 (00:49 +0000)]
[AArch64] Remove unnecessary "break". NFC.
llvm-svn: 243931
Ahmed Bougacha [Tue, 4 Aug 2015 00:48:02 +0000 (00:48 +0000)]
[AArch64] Use SDValue bool operator. NFC.
llvm-svn: 243930
Hans Wennborg [Tue, 4 Aug 2015 00:47:58 +0000 (00:47 +0000)]
Add a -revert option to utils/release/merge.sh
llvm-svn: 243929
Chandler Carruth [Tue, 4 Aug 2015 00:44:07 +0000 (00:44 +0000)]
[UB] Fix a nasty place where we would pass null pointers to memcpy.
This happens to work, but is not guaranteed to work. Indeed, most memcpy
interfaces in Linux-land annotate these arguments as nonnull, and GCC
and LLVM both can and do optimized based upon that. When they do so,
they might legitimately have miscompiled code calling this routine with
two valid iterators, 'nullptr' and 'nullptr'. There was even code doing
precisely this because StringRef().begin() and StringRef().end() both
produce null pointers.
This was found by UBSan.
llvm-svn: 243927
Ahmed Bougacha [Tue, 4 Aug 2015 00:42:34 +0000 (00:42 +0000)]
[AArch64] Vector FCOPYSIGN supports Custom-lowering: mark it as such.
There's a bunch of code in LowerFCOPYSIGN that does smart lowering, and
is actually already vector-aware; let's use it instead of scalarizing!
The only interesting change is that for v2f32, we previously always used
use v4i32 as the integer vector type.
Use v2i32 instead, and mark FCOPYSIGN as Custom.
llvm-svn: 243926
Ahmed Bougacha [Tue, 4 Aug 2015 00:32:55 +0000 (00:32 +0000)]
[CodeGen] Fix FCOPYSIGN legalization to account for mismatched types.
We used to legalize it like it's any other binary operations. It's not,
because it accepts mismatched operand types. Because of that, we used
to hit various asserts and miscompiles.
Specialize vector legalizations to, in the worst case, unroll, or, when
possible, to just legalize the operand that needs legalization.
Scalarization isn't covered, because I can't think of a target where
some but not all of the 1-element vector types are to be scalarized.
llvm-svn: 243924
Alex Lorenz [Tue, 4 Aug 2015 00:24:45 +0000 (00:24 +0000)]
MIR Serialization: Serialize the 'volatile' machine memory operand flag.
llvm-svn: 243923
Chaoren Lin [Mon, 3 Aug 2015 23:59:41 +0000 (23:59 +0000)]
XFAIL TestInferiorAssert for Android API <= 16.
llvm-svn: 243922
Adam Nemet [Mon, 3 Aug 2015 23:33:03 +0000 (23:33 +0000)]
[LAA] Remove unused needsAnyChecking(), NFC
llvm-svn: 243921
Adam Nemet [Mon, 3 Aug 2015 23:32:57 +0000 (23:32 +0000)]
[LoopVer] Remove unused needsRuntimeChecks(), NFC
The previous commits moved this functionality into the client.
Also remove the now unused member variable.
llvm-svn: 243920
Mehdi Amini [Mon, 3 Aug 2015 23:25:46 +0000 (23:25 +0000)]
Docs: s/Sanitiser/Sanitizer/ for consistency
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 243918
Mehdi Amini [Mon, 3 Aug 2015 23:17:47 +0000 (23:17 +0000)]
GettingStarted.rst: Add info about building sanitizers
From: Erik de Castro Lopo <erikd@qti.qualcomm.com>
llvm-svn: 243917
Mehdi Amini [Mon, 3 Aug 2015 23:17:44 +0000 (23:17 +0000)]
docs/GettingStarted.rst: Whitespace only
From: Erik de Castro Lopo <erikd@qti.qualcomm.com>
llvm-svn: 243916
Alex Lorenz [Mon, 3 Aug 2015 23:08:19 +0000 (23:08 +0000)]
MIR Serialization: Initial serialization of the machine memory operands.
Reviewers: Duncan P. N. Exon Smith
llvm-svn: 243915
Adrian McCarthy [Mon, 3 Aug 2015 23:01:51 +0000 (23:01 +0000)]
Create a Windows mini-dump target Differential Revision: reviews.llvm.org/D11611
llvm-svn: 243914
Naomi Musgrave [Mon, 3 Aug 2015 23:01:19 +0000 (23:01 +0000)]
Runtime check of poisoning derived class members.
Summary: Simple test case to verify that an instance of a derived class with virtual base is properly poisoned
Reviewers: eugenis, kcc
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D11733
modified test to be more concise, and check the local pointer to the destroyed object
revised test to not examine padding- only explicit object members
llvm-svn: 243913
Naomi Musgrave [Mon, 3 Aug 2015 22:53:11 +0000 (22:53 +0000)]
Dtor callback emitted when msan attribute not repressed for this function.
Summary: In addition to checking compiler flags, the front-end also examines the attributes of the destructor definition to ensure that the SanitizeMemory attribute is attached.
Reviewers: eugenis, kcc
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D11727
refactored test into new file, revised how function attribute examined
modified test to examine default dtor with and without attribute
removed attribute check
llvm-svn: 243912
Justin Bogner [Mon, 3 Aug 2015 22:43:14 +0000 (22:43 +0000)]
lto: Avoid relying on the environment for this test
It's better to pass libLTO to ld64 via the command line flag than rely
on setting DYLD_LIBRARY_PATH.
llvm-svn: 243911
David Blaikie [Mon, 3 Aug 2015 22:33:50 +0000 (22:33 +0000)]
Update/correct comment.
llvm-svn: 243910
David Blaikie [Mon, 3 Aug 2015 22:30:24 +0000 (22:30 +0000)]
-Wdeprecated-clean: Fix cases of violating the rule of 5 in ways that are deprecated in C++11
Various value handles needed to be copy constructible and copy
assignable (mostly for their use in DenseMap). But to avoid an API that
might allow accidental slicing, make these members protected in the base
class and make derived classes final (the special members become
implicitly public there - but disallowing further derived classes that
might be sliced to the intermediate type).
Might be worth having a warning a bit like -Wnon-virtual-dtor that
catches public move/copy assign/ctors in classes with virtual functions.
(suppressable in the same way - by making them protected in the base,
and making the derived classes final) Could be fancier and only diagnose
them when they're actually called, potentially.
Also allow a few default implementations where custom implementations
(especially with non-standard return types) were implemented.
llvm-svn: 243909
Justin Bogner [Mon, 3 Aug 2015 22:26:06 +0000 (22:26 +0000)]
X86: Teach X86 builtins which target features they require
This adds the required target feature names to x86 builtins that need
particular features. Most have exactly one ("avx", "aes", etc), but some
of the avx512 features have multiple requirements, eg "avx512vl,avx512bw".
llvm-svn: 243908
Tim Northover [Mon, 3 Aug 2015 22:19:08 +0000 (22:19 +0000)]
ARM: remove horrible printf left over from debugging
llvm-svn: 243907
Alexander Kornienko [Mon, 3 Aug 2015 22:02:08 +0000 (22:02 +0000)]
[clang-tidy] Improve the misc-unused-alias-decl message
"this namespace alias decl is unused" -> "namespace alias decl '...' is unused"
llvm-svn: 243906
Davide Italiano [Mon, 3 Aug 2015 21:46:32 +0000 (21:46 +0000)]
[llvm-objdump] Range-loopify. NFC intended.
llvm-svn: 243905
Derek Schuff [Mon, 3 Aug 2015 21:23:51 +0000 (21:23 +0000)]
Fix memory leak in unit test of Bitcode/BitReaderTest.cpp
Fixes obvious memory leak in test
TestForEofAfterReadFailureOnDataStreamer. Also removes constexpr use
from same test.
Patch by Karl Schimpf.
Differential Revision: http://reviews.llvm.org/D11735
llvm-svn: 243904
David Blaikie [Mon, 3 Aug 2015 20:55:00 +0000 (20:55 +0000)]
Fix with a bit more care. (but only a bit)
llvm-svn: 243903
Enrico Granata [Mon, 3 Aug 2015 20:47:19 +0000 (20:47 +0000)]
Jim suggested to use eArgTypeAddressOrExpression for the addresses that 'memory find' takes
llvm-svn: 243902
Douglas Katzman [Mon, 3 Aug 2015 20:41:31 +0000 (20:41 +0000)]
Grammar: Don't imply that a program wouldn't want a person as its output.
(Because, hey, who wouldn't?)
llvm-svn: 243901
Chandler Carruth [Mon, 3 Aug 2015 20:32:27 +0000 (20:32 +0000)]
[Unroll] Improve the brute force loop unroll estimate by propagating
through PHI nodes across iterations.
This patch teaches the new advanced loop unrolling heuristics to propagate
constants into the loop from the preheader and around the backedge after
simulating each iteration. This lets us brute force solve simple recurrances
that aren't modeled effectively by SCEV. It also makes it more clear why we
need to process the loop in-order rather than bottom-up which might otherwise
make much more sense (for example, for DCE).
This came out of an attempt I'm making to develop a principled way to account
for dead code in the unroll estimation. When I implemented
a forward-propagating version of that it produced incorrect results due to
failing to propagate *cost* between loop iterations through the PHI nodes, and
it occured to me we really should at least propagate simplifications across
those edges, and it is quite easy thanks to the loop being in canonical and
LCSSA form.
Differential Revision: http://reviews.llvm.org/D11706
llvm-svn: 243900
David Blaikie [Mon, 3 Aug 2015 20:30:53 +0000 (20:30 +0000)]
Try to fix the build for C++ standard libraries missing std::map::emplace
llvm-svn: 243899
Renato Golin [Mon, 3 Aug 2015 20:19:35 +0000 (20:19 +0000)]
[Release Script] Check for correct symlink name
While checking for the existence of the clang-tools-extra directory,
the script was not checking for its destination name, "extra", and
the script was failing when re-running without checking out new
sources.
llvm-svn: 243898
David Blaikie [Mon, 3 Aug 2015 20:12:58 +0000 (20:12 +0000)]
-Wdeprecated-clean: Fix cases of violating the rule of 5 in ways that are deprecated in C++11
Some functions return concrete ByteStreamers by value - explicitly
support that in the base class. (dtor can be virtual, no one seems to be
polymorphically owning/destroying them)
llvm-svn: 243897
David Blaikie [Mon, 3 Aug 2015 20:08:41 +0000 (20:08 +0000)]
Recommit r243824: -Wdeprecated-clean: Fix cases of violating the rule of 5 in ways that are deprecated in C++11
This reverts commit r243888, recommitting r243824.
This broke the Windows build due to a difference in the C++ standard
library implementation. Using emplace/forward_as_tuple should ensure
there's no need to copy ValIDs.
llvm-svn: 243896
Reid Kleckner [Mon, 3 Aug 2015 19:51:18 +0000 (19:51 +0000)]
[asan] Print VAs instead of RVAs for module offsets on Windows
Summary:
This is consistent with binutils and ASan behavior on other platforms,
and makes it easier to use llvm-symbolizer with WinASan. The
--relative-address flag to llvm-symbolizer is also no longer needed.
An RVA is a "relative virtual address", meaning it is the address of
something inside the image minus the base of the mapping at runtime.
A VA in this context is an RVA plus the "preferred base" of the module,
and not a real runtime address. The real runtime address of a symbol
will equal the VA iff the module is loaded at its preferred base at
runtime.
On Windows, the preferred base is stored in the ImageBase field of one
of the PE file header, and this change adds the necessary code to
extract it. On Linux, this offset is typically included in program and
section headers of executables.
ELF shared objects typically use a preferred base of zero, meaning the
smallest p_vaddr field in the program headers is zero. This makes it so
that PIC and PIE module offsets come out looking like RVAs, but they're
actually VAs. The difference between them simply happens to be zero.
Reviewers: samsonov, majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11681
llvm-svn: 243895
Pete Cooper [Mon, 3 Aug 2015 19:04:32 +0000 (19:04 +0000)]
Convert some AArch64 code to foreach loops. NFC.
Also converted a cast<> to dyn_cast while i was working on the same
line of code.
llvm-svn: 243894
Enrico Granata [Mon, 3 Aug 2015 18:51:39 +0000 (18:51 +0000)]
Fix the memory find command such that it can actually take an expression
llvm-svn: 243893
Lang Hames [Mon, 3 Aug 2015 18:03:40 +0000 (18:03 +0000)]
[MCJIT] Fix a cast warning in the unit-test introduced in r243589.
Thanks to Aaron Ballman for spotting this.
llvm-svn: 243891
Derek Schuff [Mon, 3 Aug 2015 18:01:50 +0000 (18:01 +0000)]
Fix testing for end of stream in bitstream reader.
This fixes a bug found while working on the bitcode reader. In
particular, the method BitstreamReader::AtEndOfStream doesn't always
behave correctly when processing a data streamer. The method
fillCurWord doesn't properly set CurWord/BitsInCurWord if the data
streamer was already at eof, but GetBytes had not yet set the
ObjectSize field of the streaming memory object.
This patch fixes this problem, and provides a test to show that
this problem has been fixed.
Patch by Karl Schimpf.
Differential Revision: http://reviews.llvm.org/D11391
llvm-svn: 243890
Tobias Grosser [Mon, 3 Aug 2015 17:53:21 +0000 (17:53 +0000)]
Do not add spaces into memory-access identifiers
llvm-svn: 243889
Reid Kleckner [Mon, 3 Aug 2015 17:36:22 +0000 (17:36 +0000)]
Revert "-Wdeprecated-clean: Fix cases of violating the rule of 5 in ways that are deprecated in C++11"
This reverts commit r243824.
It broke the build on Windows.
llvm-svn: 243888
Duncan P. N. Exon Smith [Mon, 3 Aug 2015 17:28:43 +0000 (17:28 +0000)]
Update testcases after LLVM r243885
llvm-svn: 243887
Duncan P. N. Exon Smith [Mon, 3 Aug 2015 17:26:41 +0000 (17:26 +0000)]
DI: Disallow uniquable DICompileUnits
Since r241097, `DIBuilder` has only created distinct `DICompileUnit`s.
The backend is liable to start relying on that (if it hasn't already),
so make uniquable `DICompileUnit`s illegal and automatically upgrade old
bitcode. This is a nice cleanup, since we can remove an unnecessary
`DenseSet` (and the associated uniquing info) from `LLVMContextImpl`.
Almost all the testcases were updated with this script:
git grep -e '= !DICompileUnit' -l -- test |
grep -v test/Bitcode |
xargs sed -i '' -e 's,= !DICompileUnit,= distinct !DICompileUnit,'
I imagine something similar should work for out-of-tree testcases.
llvm-svn: 243885
Tim Northover [Mon, 3 Aug 2015 17:20:10 +0000 (17:20 +0000)]
ARM: prefer allocating VFP regs at stride 4 on Darwin.
This is necessary for WatchOS support, where the compact unwind format assumes
this kind of layout. For now we only want this on Swift-like CPUs though, where
it's been the Xcode behaviour for ages. Also, since it can expand the prologue
we don't want it at -Oz.
llvm-svn: 243884
Duncan P. N. Exon Smith [Mon, 3 Aug 2015 17:09:38 +0000 (17:09 +0000)]
Linker: Move distinct MDNodes instead of cloning
Instead of cloning distinct `MDNode`s when linking in a module, just
move them over. The module linker destroys the source module, so the
old node would otherwise just be leaked on the context. Create the new
node in place. This also reduces the number of cloned uniqued nodes
(since it's less likely their operands have changed).
This mapping strategy is only correct when we're discarding the source,
so the linker turns it on via a ValueMapper flag, `RF_MoveDistinctMDs`.
There's nothing observable in terms of `llvm-link` output here: the
linked module should be semantically identical.
I'll be adding more 'distinct' nodes to the debug info metadata graph in
order to break uniquing cycles, so the benefits of this will partly come
in future commits. However, we should get some gains immediately, since
we have a fair number of 'distinct' `DILocation`s being linked in.
llvm-svn: 243883
Tobias Grosser [Mon, 3 Aug 2015 16:39:56 +0000 (16:39 +0000)]
Add option -polly-view-only
If set, this option instructs -view-scops and -polly-show to only print
functions that contain the specified string in their name. This allows to
look at the scops of a specific function in a large .ll file, without flooding
the screen with .dot graphs.
llvm-svn: 243882
Tobias Grosser [Mon, 3 Aug 2015 16:37:12 +0000 (16:37 +0000)]
Allow derived DOTViewers to choose the functions to illustrate
Instead of always showing/printing all functions, a class derived from
the DOTViewer class can overwrite the set of functions that will be
processed.
This will be used (and tested) by Polly's scop viewers, but other users
can be imagined as well.
llvm-svn: 243881
JF Bastien [Mon, 3 Aug 2015 15:29:47 +0000 (15:29 +0000)]
Refactor AtomicExpand::expandAtomicRMWToCmpXchg into a standalone function.
Summary:
This is useful for PNaCl's `RewriteAtomics` pass. NaCl intrinsics don't exist for some of the more exotic RMW instructions, so by refactoring this function into its own, `RewriteAtomics` can share code rewriting those atomics with `AtomicExpand` while additionally saving a few cycles by generating the `cmpxchg` NaCl-specific intrinsic with the callback. Without this patch, `RewriteAtomics` would require two extra passes over functions, by first requiring use of the full `AtomicExpand` pass to just expand the leftover exotic RMWs and then running itself again to expand resulting `cmpxchg`s.
NFC
Reviewers: jfb
Subscribers: jfb, llvm-commits
Differential Revision: http://reviews.llvm.org/D11422
llvm-svn: 243880
Kuba Brecka [Mon, 3 Aug 2015 14:48:59 +0000 (14:48 +0000)]
[asan] Fix dyld version detection on OS X
We currently have a dyld check in DyldNeedsEnvVariable that detects whether we are on a new OS X (10.11+) where we don't need to re-exec. For iOS simulator, we have a dlsym() hack that checks for a specific symbol, but this turns out to be fragile and problematic, because dlsym can sometimes call malloc(), which is not a good idea this early in the process runtime.
Let's instead of this do a direct comparison of dyld's version, which is exported in a public symbol `dyldVersionNumber`.
Differential Revision: http://reviews.llvm.org/D11719
llvm-svn: 243879
Douglas Katzman [Mon, 3 Aug 2015 14:34:22 +0000 (14:34 +0000)]
[SHAVE] Pass all -f, -g, -O, -W options through directly to moviCompile.
Differential Revision: http://reviews.llvm.org/D11581
llvm-svn: 243878
Artur Pilipenko [Mon, 3 Aug 2015 14:31:49 +0000 (14:31 +0000)]
Currently string attributes on function arguments/return values can be generated using LLVM API. However they are not supported in parser. So, the following scenario will fail:
* generate function with string attribute using API,
* dump it in LL format,
* try to parse.
Add parser support for string attributes to fix the issue.
Reviewed By: reames, hfinkel
Differential Revision: http://reviews.llvm.org/D11058
llvm-svn: 243877
Nathan Wilson [Mon, 3 Aug 2015 14:25:45 +0000 (14:25 +0000)]
[CONCEPTS] Add concept to VarDecl and diagnostic for uninitialized variable concept
Summary: Add IsConcept bit to VarDecl::NonParmVarDeclBitfields and associated isConcept/setConcept member functions. Set IsConcept to true when 'concept' specifier is in variable declaration. Create diagnostic when variable concept is not initialized.
Reviewers: fraggamuffin, hubert.reinterpretcast, faisalv, aaron.ballman, rsmith
Subscribers: aemerson, cfe-commits
Differential Revision: http://reviews.llvm.org/D11600
llvm-svn: 243876
Silviu Baranga [Mon, 3 Aug 2015 14:00:58 +0000 (14:00 +0000)]
[TTI] Fix default costs for interleaved accesses
Summary:
Modify the cost calculation function for interleaved accesses
to use the target-specific costs for insert/extract element and
memory operations.
This better models the case where the backend can't match
the interleaved group, and we are forced to use a wide load
and shuffle vectors.
Interleaved accesses are not enabled by default, so this shouldn't
cause a performance change.
Reviewers: jmolloy
Subscribers: jmolloy, llvm-commits
Differential Revision: http://reviews.llvm.org/D11718
llvm-svn: 243875