Reid Kleckner [Thu, 13 Nov 2014 23:32:52 +0000 (23:32 +0000)]
Fix symbol resolution of floating point libc builtins in MCJIT
Fix for LLI failure on Windows\X86: http://llvm.org/PR5053
LLI.exe crashes on Windows\X86 when single precession floating point
intrinsics like the following are used: acos, asin, atan, atan2, ceil,
copysign, cos, cosh, exp, floor, fmin, fmax, fmod, log, pow, sin, sinh,
sqrt, tan, tanh
The above intrinsics are defined as inline-expansions in math.h, and are
not exported by msvcr120.dll (Win32 API GetProcAddress returns null).
For an FREM instruction, the JIT compiler generates a call to a stub for
the fmodf() intrinsic, and adds a relocation to fixup at load time. The
loader searches the libraries for the function, but fails because the
symbol is not exported. So, the call target remains NULL and the
execution crashes.
Since the math functions are loaded at JIT/runtime, the JIT can patch
CALL instruction directly instead of the searching the libraries'
exported symbols. However, this fix caused build failures due to
unresolved symbols like _fmodf at link time.
Therefore, the current fix defines helper functions in the Runtime
link/load library to perform the above operations. The address of these
helper functions are used to patch up the CALL instruction at load time.
Reviewers: lhames, rnk
Reviewed By: rnk
Differential Revision: http://reviews.llvm.org/D5387
Patch by Swaroop Sridhar!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221947
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 13 Nov 2014 23:23:02 +0000 (23:23 +0000)]
ADT: Use perfect forwarding in StringMapEntry::Create()
Now you can pass references into constructors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221946
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 13 Nov 2014 23:07:55 +0000 (23:07 +0000)]
Relax the gcov version.ll test to check '.' instead of '\*'
The escaping of the '\*' doesn't work with my combination of testing
tools.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221944
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 13 Nov 2014 23:07:22 +0000 (23:07 +0000)]
Silence MSVC warning on missing return after fully covered switch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221943
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 23:03:09 +0000 (23:03 +0000)]
R600/SI: Fix fmin_legacy / fmax_legacy matching for SI
select_cc is expanded on SI, so this was never matched.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221941
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 13 Nov 2014 22:55:19 +0000 (22:55 +0000)]
Use nullptr instead of NULL for variadic sentinels
Windows defines NULL to 0, which when used as an argument to a variadic
function, is not a null pointer constant. As a result, Clang's
-Wsentinel fires on this code. Using '0' would be wrong on most 64-bit
platforms, but both MSVC and Clang make it work on Windows. Sidestep the
issue with nullptr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221940
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 13 Nov 2014 22:54:59 +0000 (22:54 +0000)]
Revert "[GVN] Perform Scalar PRE on gep indices that feed loads before doing Load PRE."
This reverts commit r221924. It appears the commit was a bit premature and is causing
bot failures that need further investigation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221939
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 13 Nov 2014 22:49:44 +0000 (22:49 +0000)]
[x86] Add some tests for specific patterns of lane-flips combined with
in-lane shuffles that aren't always handled well by the current vector
shuffle lowering.
No functionality change yet, that will follow in a subsequent commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221938
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 22:29:02 +0000 (22:29 +0000)]
Move calls to push_back out of readAbbreviated(Literal|Field).
These functions always return a single value and not all callers want to
push them into a SmallVector.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221934
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 13 Nov 2014 22:09:56 +0000 (22:09 +0000)]
Avoid usage of char16_t as MSVC "14" doesn't appear to support it
Fixes the MSVC "14" build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221932
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Thu, 13 Nov 2014 21:56:57 +0000 (21:56 +0000)]
Fix nested namespace with decltype to hopefully work with MSVC
Build failed here:
http://lab.llvm.org:8011/builders/lld-x86_64-win7/builds/14629/steps/build_Lld/logs/stdio
So I'm taking a shot in the dark that MSVC (whatever version that is)
can't cope with nested name specifiers with a decltype prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221931
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 21:54:59 +0000 (21:54 +0000)]
Make a few helper functions static. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221930
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Thu, 13 Nov 2014 21:40:02 +0000 (21:40 +0000)]
Use unique_ptr to handle ownership of TreePatterns in CodeGenDAGPatterns::PatternFragments
We might be able to use unique_ptr to handle ownership of the
TreePatternNodes too - looking into that next.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221928
91177308-0d34-0410-b5e6-
96231b3b80d8
Aditya Nandakumar [Thu, 13 Nov 2014 21:29:21 +0000 (21:29 +0000)]
We can get the TLOF from the TargetMachine - so constructor no longer requires TargetLoweringObjectFile to be passed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221926
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 13 Nov 2014 21:17:58 +0000 (21:17 +0000)]
[GVN] Perform Scalar PRE on gep indices that feed loads before doing Load PRE.
Phabricator Revision: http://reviews.llvm.org/D6103
Patch by "Balaram Makam" <bmakam@codeaurora.org>!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221924
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Thu, 13 Nov 2014 20:50:44 +0000 (20:50 +0000)]
[FastISel][AArch64] Don't bail during simple GEP instruction selection.
The generic FastISel code would bail, because it can't emit a sign-extend for
AArch64. This copies the code over and uses AArch64 specific emit functions.
This is not ideal and 'computeAddress' should handles this, so it can fold the
address computation into the memory operation.
I plan to clean up 'computeAddress' anyways, so I will add that in a future
commit.
Related to rdar://problem/
18962471.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221923
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 20:44:23 +0000 (20:44 +0000)]
R600/SI: Use s_movk_i32
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221922
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 20:23:36 +0000 (20:23 +0000)]
R600/SI: Fix definition for s_cselect_b32
These were directly using the old base instruction
class, and specifying the wrong register classes
for operands. The operands can be the other special
inputs besides SGPRs. The op name was also being
directly used for the asm string, so this was printed
without any operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221921
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 20:07:40 +0000 (20:07 +0000)]
R600: Fix assert on empty function
If a function is just an unreachable, this would hit a
"this is not a MachO target" assertion because of setting
HasSubsectionViaSymbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221920
91177308-0d34-0410-b5e6-
96231b3b80d8
Rui Ueyama [Thu, 13 Nov 2014 20:07:06 +0000 (20:07 +0000)]
Un-break the big-endian buildbots
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221919
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 19:56:13 +0000 (19:56 +0000)]
R600: Error on initializer for LDS.
Also give a proper error for other address spaces.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221917
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 19:49:04 +0000 (19:49 +0000)]
R600/SI: Get rid of FCLAMP_SI pseudo
It's not necessary. Also use complex patterns to allow
src modifier usage.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221916
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 19:48:56 +0000 (19:48 +0000)]
Object, Mach-O: Refactor and clean code up
Don't assert if we can return an error code, reuse existing
functionality like is64Bit().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221915
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Thu, 13 Nov 2014 19:45:27 +0000 (19:45 +0000)]
Use -Wcast-qual in cmake builds, not only autoconfo ones.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221913
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 19:26:50 +0000 (19:26 +0000)]
R600/SI: Allow commuting with src2_modifiers
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221911
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 13 Nov 2014 19:26:47 +0000 (19:26 +0000)]
R600/SI: Allow commuting some 3 op instructions
e.g. v_mad_f32 a, b, c -> v_mad_f32 b, a, c
This simplifies matching v_madmk_f32.
This looks somewhat surprising, but it appears to be
OK to do this. We can commute src0 and src1 in all
of these instructions, and that's all that appears
to matter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221910
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 18:44:53 +0000 (18:44 +0000)]
Return word_t from read.
This removes the need for a special Read64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221909
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 13 Nov 2014 17:58:53 +0000 (17:58 +0000)]
ARM: allow constpool entry to be moved to the user's block in all cases.
Normally entries can only move to a lower address, but when that wasn't viable,
the user's block was considered anyway. Unfortunately, it went via
createNewWater which wasn't designed to handle the case where there's already
an island after the block.
Unfortunately, the test we have is slow and fragile, and I couldn't reduce it
to anything sane even with the @llvm.arm.space intrinsic. The test change here
is recreating the previous one after the change.
rdar://problem/
18545506
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221905
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 13 Nov 2014 17:58:51 +0000 (17:58 +0000)]
ARM: avoid duplicating branches during constant islands.
We were using a naive heuristic to determine whether a basic block already had
an unconditional branch at the end. This mostly corresponded to reality
(assuming branches got optimised) because there's not much point in a branch to
the next block, but could go wrong.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221904
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 13 Nov 2014 17:58:48 +0000 (17:58 +0000)]
ARM: add @llvm.arm.space intrinsic for testing ConstantIslands.
Creating tests for the ConstantIslands pass is very difficult, since it depends
on precise layout details. Having the ability to precisely inject a number of
bytes into the stream helps greatly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221903
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 17:12:19 +0000 (17:12 +0000)]
Fix the other build system.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221901
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 16:52:07 +0000 (16:52 +0000)]
Fix a regression on the disassembling C API.
The fix is easy. Unfortunately, we had 0 tests, so adding one was somewhat
complicated.
Thanks to Kevin Enderby for the report.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221899
91177308-0d34-0410-b5e6-
96231b3b80d8
Colin LeMahieu [Thu, 13 Nov 2014 16:36:30 +0000 (16:36 +0000)]
[Hexagon]
NFC Renaming reserved identifier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221898
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 13 Nov 2014 15:40:20 +0000 (15:40 +0000)]
[Reassociate] Update comment. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221894
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 14:45:22 +0000 (14:45 +0000)]
Simplify code a bit. NFC.
Thanks to Sean Silva for the suggestion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221892
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 14:37:51 +0000 (14:37 +0000)]
Small optimization: once the size is know, we don't have to call fillCurWord.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221891
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Thu, 13 Nov 2014 13:55:13 +0000 (13:55 +0000)]
Fixing -Wtype-limits warnings with the asserts (the expression would always evaluate to true). Also fixing a -Wcast-qual warning, where the cast expression isn't required.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221888
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Thu, 13 Nov 2014 13:39:49 +0000 (13:39 +0000)]
Fixing some sign comparison warnings from MSVC; NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221887
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 13 Nov 2014 13:17:47 +0000 (13:17 +0000)]
IR: Create the Metadata class
This will become the root of a new class hierarchy separate from
`Value`. As a first step, stick it between `Value` and `MDNode`.
This is part of PR21532.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221886
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Thu, 13 Nov 2014 11:46:16 +0000 (11:46 +0000)]
AVX-512: SINT_TO_FP cost model and some bugfixes
Checked some corner cases, for example translation
of <8 x i1> to <8 x double>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221883
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 09:50:18 +0000 (09:50 +0000)]
Object, COFF: Refactor code to get relocation iterators
No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221880
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Thu, 13 Nov 2014 09:29:30 +0000 (09:29 +0000)]
OCAMLFLAGS can contain =, don't use = with sed
Like HOST_LDFLAGS, etc. OCAMLFLAGS can contain =, so use ! as the substitution
separator instead of = (otherwise, sed might error).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221879
91177308-0d34-0410-b5e6-
96231b3b80d8
Aditya Nandakumar [Thu, 13 Nov 2014 09:26:31 +0000 (09:26 +0000)]
This patch changes the ownership of TLOF from TargetLoweringBase to TargetMachine so that different subtargets could share the TLOF effectively
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221878
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Thu, 13 Nov 2014 09:16:54 +0000 (09:16 +0000)]
Revert r219432 - "Revert "[BasicAA] Revert "Revert r218714 - Make better use of zext and sign information."""
Let's try this again...
This reverts r219432, plus a bug fix.
Description of the bug in r219432 (by Nick):
The bug was using AllPositive to break out of the loop; if the loop break
condition i != e is changed to i != e && AllPositive then the
test_modulo_analysis_with_global test I've added will fail as the Modulo will
be calculated incorrectly (as the last loop iteration is skipped, so Modulo
isn't updated with its Scale).
Nick also adds this comment:
ComputeSignBit is safe to use in loops as it takes into account phi nodes, and
the == EK_ZeroEx check is safe in loops as, no matter how the variable changes
between iterations, zero-extensions will always guarantee a zero sign bit. The
isValueEqualInPotentialCycles check is therefore definitely not needed as all
the variable analysis holds no matter how the variables change between loop
iterations.
And this patch also adds another enhancement to GetLinearExpression - basically
to convert ConstantInts to Offsets (see test_const_eval and
test_const_eval_scaled for the situations this improves).
Original commit message:
This reverts r218944, which reverted r218714, plus a bug fix.
Description of the bug in r218714 (by Nick):
The original patch forgot to check if the Scale in VariableGEPIndex flipped the
sign of the variable. The BasicAA pass iterates over the instructions in the
order they appear in the function, and so BasicAliasAnalysis::aliasGEP is
called with the variable it first comes across as parameter GEP1. Adding a
%reorder label puts the definition of %a after %b so aliasGEP is called with %b
as the first parameter and %a as the second. aliasGEP later calculates that %a
== %b + 1 - %idxprom where %idxprom >= 0 (if %a was passed as the first
parameter it would calculate %b == %a - 1 + %idxprom where %idxprom >= 0) -
ignoring that %idxprom is scaled by -1 here lead the patch to incorrectly
conclude that %a > %b.
Revised patch by Nick White, thanks! Thanks to Lang to isolating the bug.
Slightly modified by me to add an early exit from the loop and avoid
unnecessary, but expensive, function calls.
Original commit message:
Two related things:
1. Fixes a bug when calculating the offset in GetLinearExpression. The code
previously used zext to extend the offset, so negative offsets were converted
to large positive ones.
2. Enhance aliasGEP to deduce that, if the difference between two GEP
allocations is positive and all the variables that govern the offset are also
positive (i.e. the offset is strictly after the higher base pointer), then
locations that fit in the gap between the two base pointers are NoAlias.
Patch by Nick White!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221876
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 08:46:37 +0000 (08:46 +0000)]
Object, COFF: Increase code reuse
Split getObject's smarts into checkOffset, use this to replace the
handwritten check in getSectionContents. Similarly, replace checks in
section_rel_begin/section_rel_end with getNumberOfRelocations.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221873
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 07:54:05 +0000 (07:54 +0000)]
llvm-readobj: relocAddressLess could potentially lie
On error conditions, relocAddressLess might claim that a value is less
than itself. Instead, abort llvm-readobj. No functionality change
intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221872
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 07:42:13 +0000 (07:42 +0000)]
llvm-readobj, COFF: Remove an unused variable
printRelocation doesn't use the section contents. No functionality
change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221871
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 07:42:11 +0000 (07:42 +0000)]
Object, COFF: getRelocationSymbol shouldn't assert
lib/Object is supposed to be robust to malformed object files. Don't
assert if we don't have a symbol table. I'll try to come up with a test
case later.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221870
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 07:42:09 +0000 (07:42 +0000)]
Object, COFF: Cleanup some code in getSectionName
Use StringRef::startswith to tidy up some code, no functionality change
intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221869
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 07:42:07 +0000 (07:42 +0000)]
Object, COFF: Fix some theoretical bugs
getObject didn't consider the case where a pointer came before the start
of the object file. No test is included, trying to come up with
something reasonable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221868
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 13 Nov 2014 07:42:05 +0000 (07:42 +0000)]
Object, COFF: Clean up formatting in hasExtendedRelocations
No functionality changed intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221867
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 13 Nov 2014 07:23:22 +0000 (07:23 +0000)]
Read 64 bits at a time in the bitcode reader.
The reading of 64 bit values could still be optimized, but at least this cuts
down on the number of virtual calls to fetch more data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221865
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 13 Nov 2014 04:56:41 +0000 (04:56 +0000)]
Update \param(s) in MemoryObject::readBytes(). [-Wdocumentation]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221863
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 13 Nov 2014 04:06:10 +0000 (04:06 +0000)]
[x86] Teach the vector shuffle lowering to make a more nuanced decision
between splitting a vector into 128-bit lanes and recombining them vs.
decomposing things into single-input shuffles and a final blend.
This handles a large number of cases in AVX1 where the cross-lane
shuffles would be much more expensive to represent even though we end up
with a fast blend at the root. Instead, we can do a better job of
shuffling in a single lane and then inserting it into the other lanes.
This fixes the remaining bits of Halide's regression captured in PR21281
for AVX1. However, the bug persists in AVX2 because I've made this
change reasonably conservative. The cases where it makes sense in AVX2
to split into 128-bit lanes are much more rare because we can often do
full permutations across all elements of the 256-bit vector. However,
the particular test case in PR21281 is an example of one of the rare
cases where it is *always* better to work in a single 128-bit lane. I'm
going to try to teach the logic to detect and form the good code even in
AVX2 next, but it will need to use a separate heuristic.
Finally, there is one pesky regression here where we previously would
craftily use vpermilps in AVX1 to shuffle both high and low halves at
the same time. We no longer pull that off, and not for any really good
reason. Ultimately, I think this is just another missing nuance to the
selection heuristic that I'll try to add in afterward, but this change
already seems strictly worth doing considering the magnitude of the
improvements in common matrix math shuffle patterns.
As always, please let me know if this causes a surprising regression for
you.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221861
91177308-0d34-0410-b5e6-
96231b3b80d8
Rui Ueyama [Thu, 13 Nov 2014 03:22:54 +0000 (03:22 +0000)]
llvm-readobj: Print out address table when dumping COFF delay-import table
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221855
91177308-0d34-0410-b5e6-
96231b3b80d8
Frederic Riss [Thu, 13 Nov 2014 03:20:23 +0000 (03:20 +0000)]
Add an assert and a test that verify r221709's fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221854
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 13 Nov 2014 02:42:08 +0000 (02:42 +0000)]
[x86] Don't form overly fragmented blends when splitting and
re-combining shuffles because nothing was available in the wider vector
type.
The key observation (which I've put in the comments for future
maintainers) is that at this point, no further combining is really
possible. And so even though these shuffles trivially could be combined,
we need to actually do that as we produce them when producing them this
late in the lowering.
This fixes another (huge) part of the Halide vector shuffle regressions.
As it happens, this was already well covered by the tests, but I hadn't
noticed how bad some of these got. The specific patterns that turn
directly into unpckl/h patterns were occurring *many* times in common
vector processing code.
There are still more problems here sadly, but trying to incrementally
tease them apart and it looks like this is the core of the problem in
the splitting logic.
There is some chance of regression here, you can see it in the test
changes. Specifically, where we stop forming pshufb in some cases, it is
possible that pshufb was in fact faster. Intel "says" that pshufb is
slower than the instruction sequences replacing it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221852
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Thu, 13 Nov 2014 01:44:51 +0000 (01:44 +0000)]
[CodeGenPrepare] Handle zero extensions in the TypePromotionHelper.
Prior to this patch the TypePromotionHelper was promoting only sign extensions.
Supporting zero extensions changes:
- How constants are extended.
- How sign extensions, zero extensions, and truncate are composed together.
- How the type of the extended operation is recorded. Now we need to know the
kind of the extension as well as its type.
Each change is fairly small, unlike the diff.
Most of the diff are comments/variable renaming to say "extension" instead of
"sign extension".
The performance improvements on the test suite are within the noise.
Related to <rdar://problem/
18310086>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221851
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Thu, 13 Nov 2014 00:51:12 +0000 (00:51 +0000)]
CMake: stop setting (well, #undef'ing) HOST_LINK_VERSION; it's used in Clang, not LLVM
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221849
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Thu, 13 Nov 2014 00:36:46 +0000 (00:36 +0000)]
[FastISel][AArch64] Optimize select when one of the operands is a 'true' or 'false' value.
Optimize selects of i1 in the presence of 'true' and 'false' operands to simple
logic operations.
This fixes rdar://problem/
18960150.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221848
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Thu, 13 Nov 2014 00:36:43 +0000 (00:36 +0000)]
[FastISel][AArch64] Fold the cmp into the select when possible.
This folds the compare emission into the select emission when possible, so we
can directly use the flags and don't have to emit a separate compare.
Related to rdar://problem/
18960150.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221847
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Thu, 13 Nov 2014 00:36:38 +0000 (00:36 +0000)]
[FastISel][AArch64] Extend 'select' lowering to support also i1 to i16.
Related to rdar://problem/
18960150.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221846
91177308-0d34-0410-b5e6-
96231b3b80d8
Paul Robinson [Thu, 13 Nov 2014 00:36:34 +0000 (00:36 +0000)]
Drop a few unneeded ctor calls (missed code review comment).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221845
91177308-0d34-0410-b5e6-
96231b3b80d8
Frederic Riss [Thu, 13 Nov 2014 00:15:15 +0000 (00:15 +0000)]
Revert "[dwarfdump] Add support for dumping accelerator tables."
This reverts commit r221836.
The tests are asserting on some buildbots. This also reverts the
test part of r221837 as it relies on dwarfdump dumping the
accelerator tables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221842
91177308-0d34-0410-b5e6-
96231b3b80d8
Paul Robinson [Thu, 13 Nov 2014 00:12:14 +0000 (00:12 +0000)]
Improve long path name support on Windows.
Windows normally limits the length of an absolute path name to 260
characters; directories can have lower limits. These limits increase
to about 32K if you use absolute paths with the special '\\?\'
prefix. Teach Support\Windows\Path.inc to use that prefix as needed.
TODO: Other parts of Support could also learn to use this prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221841
91177308-0d34-0410-b5e6-
96231b3b80d8
Jordan Rose [Thu, 13 Nov 2014 00:08:41 +0000 (00:08 +0000)]
[Bitcode] AtEndOfStream should only check against the size if it's known.
This avoids an issue where AtEndOfStream mistakenly returns true at the /start/ of
a stream.
(In the rare case that the size is known and actually 0, the slow path will still
handle it correctly.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221840
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Thu, 13 Nov 2014 00:00:58 +0000 (00:00 +0000)]
Teach ScalarEvolution to sharpen range information.
If x is known to have the range [a, b), in a loop predicated by (icmp
ne x, a) its range can be sharpened to [a + 1, b). Get
ScalarEvolution and hence IndVars to exploit this fact.
This change triggers an optimization to widen-loop-comp.ll, so it had
to be edited to get it to pass.
This change was originally landed in r219834 but had a bug and broke
ASan. It was reverted in r219878, and is now being re-landed after
fixing the original bug.
phabricator: http://reviews.llvm.org/D5639
reviewed by: atrick
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221839
91177308-0d34-0410-b5e6-
96231b3b80d8
Frederic Riss [Wed, 12 Nov 2014 23:48:14 +0000 (23:48 +0000)]
Fix emission of Dwarf accelerator table when there are multiple CUs.
The DIE offset in the accel tables is an offset relative to the start
of the debug_info section, but we were encoding the offset to the
start of the containing CU.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221837
91177308-0d34-0410-b5e6-
96231b3b80d8
Frederic Riss [Wed, 12 Nov 2014 23:48:10 +0000 (23:48 +0000)]
[dwarfdump] Add support for dumping accelerator tables.
The class used for the dump only allows to dump for the moment, but
it can (and will) be easily extended to support search also.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221836
91177308-0d34-0410-b5e6-
96231b3b80d8
Frederic Riss [Wed, 12 Nov 2014 23:48:04 +0000 (23:48 +0000)]
Allow DWARFFormValue::extractValue to be called with a null CU.
Currently FormValues are only used for attributes of DIEs and thus
uers always have a CU lying around when calling into the FormValue
API.
Accelerator tables encode their information using the same Forms
as the attributes, thus it is natural to use DWARFFormValue to
extract/dump them. There is no CU in that case though. Allow the
API to be called with a null CU arguemnt by making the RelocMap
lookup conditional on the CU pointer validity. And document this
new behvior in the header. (Test coverage for this use of the API
comes in the DwarfAccelTable support patch)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221835
91177308-0d34-0410-b5e6-
96231b3b80d8
Frederic Riss [Wed, 12 Nov 2014 23:48:01 +0000 (23:48 +0000)]
Remove unsused variables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221834
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 12 Nov 2014 23:05:03 +0000 (23:05 +0000)]
[CodeGenPrepare] Replace other uses of EVT::getEVT with TL::getValueType.
r221820 fixed a problem (PR21548) where an iPTR was used in TLI legality checks,
which isn't valid and resulted in a failed assertion.
The solution was to lower pointer types into the correct target's VT, by
using TL::getValueType instead of EVT::getEVT.
This commit changes 3 other uses of EVT::getEVT, but without any tests:
- One of these non-lowered EVTs is passed to allowsMisalignedMemoryAccesses,
which goes into target's TL implementation and doesn't cause any problem (yet.)
- Two others are passed to TLI.isOperationLegalOrCustom:
- one only looks at extensions, so doesn't concern pointers.
- one only looks at binary operators, so also isn't a problem.
The latter might some day be exposed to pointers and cause the same assert as
the original PR, because there's a comment hinting at also supporting cast ops.
For consistency, update all of them and be done with it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221827
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 12 Nov 2014 22:16:55 +0000 (22:16 +0000)]
[CodeGenPrepare][AArch64] Fix a TLI legality check on iPTR to use a lowered instead.
Fixes PR21548. Related to PR20474.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221820
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 12 Nov 2014 21:39:01 +0000 (21:39 +0000)]
Expose the number of Newton-Raphson iterations applied to the hardware's reciprocal estimate as a parameter (x86).
This is a follow-on to r221706 and r221731 and discussed in more detail in PR21385.
This patch also loosens the testcase checking for btver2. We know that the "1.0" will be loaded, but
we can't tell exactly when, so replace the CHECK-NEXT specifiers with plain CHECKs. The CHECK-NEXT
sequence relied on a quirk of post-RA-scheduling that may change independently of anything in these tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221819
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 12 Nov 2014 21:23:34 +0000 (21:23 +0000)]
Add fortified (__*_chk) library functions to TLI (NFC)
One of them (__memcpy_chk) was already there, the others were checked
by comparing function names.
Note that the fortified libfuncs are now part of TLI, but are always
available, because they aren't generated, only optimized into the
non-checking versions.
Differential Revision: http://reviews.llvm.org/D6179
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221817
91177308-0d34-0410-b5e6-
96231b3b80d8
Timur Iskhodzhanov [Wed, 12 Nov 2014 20:21:20 +0000 (20:21 +0000)]
Temporary fix for PR21528 - use mangled C++ function names in COFF debug info to un-break ASan on Windows
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221813
91177308-0d34-0410-b5e6-
96231b3b80d8
Timur Iskhodzhanov [Wed, 12 Nov 2014 20:10:09 +0000 (20:10 +0000)]
[COFF] Make it clearer that the symbols subsection holds function display name rather than just name
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221812
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron McInally [Wed, 12 Nov 2014 19:58:54 +0000 (19:58 +0000)]
[AVX512] Add integer shift by immediate intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221811
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Wed, 12 Nov 2014 19:43:13 +0000 (19:43 +0000)]
Changing a StringRef::begin() call into StringRef::data(); NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221808
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 12 Nov 2014 18:37:00 +0000 (18:37 +0000)]
Use the return of readBytes to find out if we are at the end of the stream.
This allows the removal of isObjectEnd and opens the way for reading 64 bits
at a time.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221804
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 12 Nov 2014 18:25:47 +0000 (18:25 +0000)]
CGSCC should not treat intrinsic calls like function calls (PR21403)
Make the handling of calls to intrinsics in CGSCC consistent:
they are not treated like regular function calls because they
are never lowered to function calls.
Without this patch, we can get dangling pointer asserts from
the subsequent loop that processes callsites because it already
ignores intrinsics.
See http://llvm.org/bugs/show_bug.cgi?id=21403 for more details / discussion.
Differential Revision: http://reviews.llvm.org/D6124
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221802
91177308-0d34-0410-b5e6-
96231b3b80d8
Jingyue Wu [Wed, 12 Nov 2014 18:25:06 +0000 (18:25 +0000)]
Fix broken doxygen annotations, NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221801
91177308-0d34-0410-b5e6-
96231b3b80d8
Jingyue Wu [Wed, 12 Nov 2014 18:09:15 +0000 (18:09 +0000)]
Disable indvar widening if arithmetics on the wider type are more expensive
Summary:
Reapply r221772. The old patch breaks the bot because the @indvar_32_bit test
was run whether NVPTX was enabled or not.
IndVarSimplify should not widen an indvar if arithmetics on the wider
indvar are more expensive than those on the narrower indvar. For
instance, although NVPTX64 treats i64 as a legal type, an ADD on i64 is
twice as expensive as that on i32, because the hardware needs to
simulate a 64-bit integer using two 32-bit integers.
Split from D6188, and based on D6195 which adds NVPTXTargetTransformInfo.
Fixes PR21148.
Test Plan:
Added @indvar_32_bit that verifies we do not widen an indvar if the arithmetics
on the wider type are more expensive. This test is run only when NVPTX is
enabled.
Reviewers: jholewinski, eliben, meheff, atrick
Reviewed By: atrick
Subscribers: jholewinski, llvm-commits
Differential Revision: http://reviews.llvm.org/D6196
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221799
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 12 Nov 2014 18:07:42 +0000 (18:07 +0000)]
remove function names from comments; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221798
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 12 Nov 2014 17:11:16 +0000 (17:11 +0000)]
Return the number of read bytes in MemoryObject::readBytes.
Returning more information will allow BitstreamReader to be simplified a bit
and changed to read 64 bits at a time.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221794
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Hibbits [Wed, 12 Nov 2014 16:50:15 +0000 (16:50 +0000)]
Revert part of the PIC tests (TLS part)
This change actually wasn't warranted for -O0, and the new changes prove it and
break the build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221793
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Hibbits [Wed, 12 Nov 2014 16:40:00 +0000 (16:40 +0000)]
Fix thet tests.
I seem to have missed the update I made for changing 'flag_pic' to "PIC Level".
Mea culpa.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221792
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Hibbits [Wed, 12 Nov 2014 15:16:30 +0000 (15:16 +0000)]
Add support for small-model PIC for PowerPC.
Summary:
Large-model was added first. With the addition of support for multiple PIC
models in LLVM, now add small-model PIC for 32-bit PowerPC, SysV4 ABI. This
generates more optimal code, for shared libraries with less than about 16380
data objects.
Test Plan: Test cases added or updated
Reviewers: joerg, hfinkel
Reviewed By: hfinkel
Subscribers: jholewinski, mcrosier, emaste, llvm-commits
Differential Revision: http://reviews.llvm.org/D5399
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221791
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 12 Nov 2014 14:48:38 +0000 (14:48 +0000)]
Reduce code duplication a bit. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221785
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 12 Nov 2014 14:23:04 +0000 (14:23 +0000)]
Fix the test.
It was broken since r221708.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221783
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Wed, 12 Nov 2014 14:01:17 +0000 (14:01 +0000)]
Fixing more -Wcast-qual warnings; NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221782
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Wed, 12 Nov 2014 13:55:27 +0000 (13:55 +0000)]
Fixing a -Wcast-qual warning; NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221781
91177308-0d34-0410-b5e6-
96231b3b80d8
Zoran Jovanovic [Wed, 12 Nov 2014 13:30:10 +0000 (13:30 +0000)]
[mips][micromips] Add predicate 'InMicroMips' at CodeGen patterns for microMIPS instructions
Differential Revision: http://reviews.llvm.org/D6198
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221780
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 12 Nov 2014 10:05:18 +0000 (10:05 +0000)]
[x86] Start improving the matching of unpck instructions based on test
cases from Halide folks. This initial step was extracted from
a prototype change by Clay Wood to try and address regressions found
with Halide and the new vector shuffle lowering.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221779
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 12 Nov 2014 09:17:15 +0000 (09:17 +0000)]
[x86] Clean up a bunch of vector shuffle tests with my script. Notably,
removes windows line endings and other noise. This is in prelude to
making substantive changes to these tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221776
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 12 Nov 2014 07:42:26 +0000 (07:42 +0000)]
MCDisassembler::getInstruction():: Prune also "\param Region", since it was removed in r221751. [-Wdocumentation]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221775
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Wed, 12 Nov 2014 07:31:03 +0000 (07:31 +0000)]
AVX-512: Intrinsics for ERI
3 instructions: vrcp28, vrsqrt28, vexp2, only vector forms.
Intrinsics include SAE (Suppres All Exceptions) parameter.
http://reviews.llvm.org/D6214
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221774
91177308-0d34-0410-b5e6-
96231b3b80d8
Jingyue Wu [Wed, 12 Nov 2014 07:19:25 +0000 (07:19 +0000)]
Reverts r221772 which fails tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221773
91177308-0d34-0410-b5e6-
96231b3b80d8
Jingyue Wu [Wed, 12 Nov 2014 06:58:45 +0000 (06:58 +0000)]
Disable indvar widening if arithmetics on the wider type are more expensive
Summary:
IndVarSimplify should not widen an indvar if arithmetics on the wider
indvar are more expensive than those on the narrower indvar. For
instance, although NVPTX64 treats i64 as a legal type, an ADD on i64 is
twice as expensive as that on i32, because the hardware needs to
simulate a 64-bit integer using two 32-bit integers.
Split from D6188, and based on D6195 which adds NVPTXTargetTransformInfo.
Fixes PR21148.
Test Plan:
Added @indvar_32_bit that verifies we do not widen an indvar if the arithmetics
on the wider type are more expensive.
Reviewers: jholewinski, eliben, meheff, atrick
Reviewed By: atrick
Subscribers: jholewinski, llvm-commits
Differential Revision: http://reviews.llvm.org/D6196
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221772
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 12 Nov 2014 04:31:19 +0000 (04:31 +0000)]
Delete dead code. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221770
91177308-0d34-0410-b5e6-
96231b3b80d8