Commit Graph

88038 Commits

Author SHA1 Message Date
Bill Wendling
9d30e7208e Remove the getAttributesAtIndex and getNumAttrs methods in favor of using the getAttrSomewhere predicate. This prevents the uses of 'Attribute' as a collection of attributes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171271 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-31 00:49:59 +00:00
Jakub Staszak
d60b8ac64f Transform (A == C1 || A == C2) into (A & ~(C1 ^ C2)) == C1
if C1 and C2 differ only with one bit.
Fixes PR14708.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171270 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-31 00:34:55 +00:00
Hal Finkel
6eb7a4270b Support ppcf128 in SelectionDAG::getConstantFP
Fixes pr14751.

Patch by Kai; Thanks!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171261 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 19:03:32 +00:00
Nuno Lopes
98281a2050 convert a bunch of callers from DataLayout::getIndexedOffset() to GEP::accumulateConstantOffset().
The later API is nicer than the former, and is correct regarding wrap-around offsets (if anyone cares).
There are a few more places left with duplicated code, which I'll remove soon.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171259 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 16:25:48 +00:00
Dmitri Gribenko
0d80c9c8c7 Documentation: add a section to prevent spurious test failures like the one
fixed in r171243.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171258 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 14:51:03 +00:00
Bill Wendling
94e94b3506 Use the predicate methods off of AttributeSet instead of Attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171257 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 13:50:49 +00:00
Bill Wendling
8b62abdd7b Remove the Function::getRetAttributes method in favor of using the AttributeSet accessor method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171256 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 13:01:51 +00:00
Bill Wendling
39cd0c8e47 Remove Function::getParamAttributes and use the AttributeSet accessor methods instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171255 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 12:45:13 +00:00
Bill Wendling
831737d329 Remove the Function::getFnAttributes method in favor of using the AttributeSet
directly.

This is in preparation for removing the use of the 'Attribute' class as a
collection of attributes. That will shift to the AttributeSet class instead.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171253 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 10:32:01 +00:00
Bill Wendling
377660355c s/hasAttribute/contains/g to be more consistent with other method names.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171252 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 09:17:46 +00:00
Nadav Rotem
db2367512e LoopVectorizer: Fix a bug in the code that updates the loop exiting block.
LCSSA PHIs may have undef values. The vectorizer updates values that are used by outside users such as PHIs.
The bug happened because undefs are not loop values. This patch handles these PHIs.

PR14725



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171251 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 07:47:00 +00:00
Dmitri Gribenko
a6542923b8 Tests: rewrite 'opt ... %s' to 'opt ... < %s' so that opt does not emit a ModuleID
This is done to avoid odd test failures, like the one fixed in r171243.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171250 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 02:33:22 +00:00
Bill Wendling
979aff6399 Add a few more c'tors:
* One that accepts a single Attribute::AttrKind.

* One that accepts an Attribute::AttrKind plus a list of values. This is for
  attributes defined like this:

    #1 = attributes { align = 4 }

* One that accepts a string, for target-specific attributes like this:

    #2 = attributes { "cpu=cortex-a8" }


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171249 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 02:22:16 +00:00
Dmitri Gribenko
a51edf0986 Add a check to the test Analysis/ScalarEvolution/2010-09-03-RequiredTransitive.ll
This test did not test anything at all (except for opt crashing, but that was
not the reason why it was added).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171248 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 01:42:34 +00:00
Bill Wendling
529ec718ac Add a few (as yet unused) query methods to determine if the attribute that's
stored here is of a certain kind. This is in preparation for when an Attribute
object represents a single attribute, instead of a bitmask of attributes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171247 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 01:38:39 +00:00
Dmitri Gribenko
b4912b9dcc Tests: rewrite 'opt ... %s' to 'opt ... < %s' so that opt does not emit a ModuleID
This is done to avoid odd test failures, like the one fixed in r171243.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171246 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 01:28:40 +00:00
Bill Wendling
435654b181 Uniquify the AttributeImpl based on the Constant pointer, since those are
already uniquified.

Note: This will be expanded in the future to add more than just one pointer
value.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171245 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 01:23:08 +00:00
Bill Wendling
c966e08a98 s/Raw/getBitMask/g to be more in line with current naming conventions. This method won't be sticking around.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171244 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 01:05:42 +00:00
NAKAMURA Takumi
f35cb76a70 llvm/test/Transforms/GVN/null-aliases-nothing.ll: Fix a RUN line not to emit ModuleID.
Larry Evans reported it fails if source tree contains "load", like "download".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171243 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-30 00:33:26 +00:00
Craig Topper
22d8f0d685 Remove intrinsic specific instructions for (V)SQRTPS/PD. Instead lower to target-independent ISD nodes and use the existing patterns for those.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171237 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 18:18:20 +00:00
Craig Topper
6f57f39e4a Merge similar functionality using a nested switch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171229 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 17:19:06 +00:00
Craig Topper
6d183e4007 Remove intrinsic specific instructions for SSE/SSE2/AVX floating point max/min instructions. Lower them to target specific nodes and use those patterns instead. This also allows them to be commuted if UnsafeFPMath is enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171227 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 16:44:25 +00:00
Jakub Staszak
c20323a758 Simplify code, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171226 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 15:57:26 +00:00
Jakub Staszak
3c0307d760 Delete executive bit on ./lib/Target/Hexagon/HexagonAsmPrinter.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171225 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 15:23:06 +00:00
Bill Wendling
7c1683d47c Use a 'Constant' object instead of a bit field to store the attribute data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171221 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 12:29:38 +00:00
Bill Wendling
4d5e5e5d61 Use the accessor method instead of the raw ivar to get the bits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171220 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-29 12:10:46 +00:00
Chandler Carruth
a84e7508eb Nuke some dead code that snuck in some how. I thought I had already
deleted this, but apparantly not. Charmingly, Clang didn't warn on it
but GCC did.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171197 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 14:50:51 +00:00
Chandler Carruth
73527d30cd Fix a stunning oversight in the inline cost analysis. It was never
propagating one of the values it simplified to a constant across
a myriad of instructions. Notably, ptrtoint instructions when we had
a constant pointer (say, 0) didn't propagate that, blocking a massive
number of down-stream optimizations.

This was uncovered when investigating why we fail to inline and delete
the boilerplate in:

  void f() {
    std::vector<int> v;
    v.push_back(1);
  }

It turns out most of the efforts I've made thus far to improve the
analysis weren't making it far purely because of this. After this is
fixed, the store-to-load forwarding patch enables LLVM to optimize the
above to an empty function. We still can't nuke a second push_back, but
for different reasons.

There is a very real chance this will cause somewhat noticable changes
in inlining behavior, so please let me know if you see regressions (or
improvements!) because of this patch.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 14:43:42 +00:00
Chandler Carruth
ba94204e94 Teach the inline cost analysis about calls that can be simplified and
how to propagate constants through insert and extract value
instructions.

With the recent improvements to instsimplify, this allows inline cost
analysis to constant fold through intrinsic functions, including notably
the with.overflow intrinsic math routines which often show up inside of
STL abstractions. This is yet another piece in the puzzle of breaking
down the code for:

  void f() {
    std::vector<int> v;
    v.push_back(1);
  }

But it still isn't enough. There are a pile of bugs in inline cost still
blocking this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171195 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 14:23:32 +00:00
Chandler Carruth
e949aa1c91 Teach instsimplify to use the constant folder where appropriate for
constant folding calls. Add the initial tests for this which show that
now instsimplify can simplify blindingly obvious code patterns expressed
with both intrinsics and library calls.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171194 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 14:23:29 +00:00
Chandler Carruth
c98bd9f1a7 Add entry points to instsimplify for simplifying calls. The entry points
are nice and decomposed so that we can simplify synthesized calls as
easily as actually call instructions. The internal utility still has the
same behavior, it just now operates on a more generic interface so that
I can extend the set of call simplifications that instsimplify knows
about.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171189 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 11:30:55 +00:00
Alexey Samsonov
f045df1b8b Add proper support for -fsanitize-blacklist= flag for TSan and MSan. LLVM part.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171183 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 09:30:44 +00:00
Nadav Rotem
ae34b4280e CostModel: initial checkin for code that estimates the cost of special shuffles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171180 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 08:19:03 +00:00
Nadav Rotem
40ef8b7548 wrap 80-col lines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171179 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 07:28:43 +00:00
Nadav Rotem
0509db2738 AVX: Move the ZEXT/ANYEXT DAGCo optimizations to the lowering of these optimizations. The old test cases still cover all of these lowering/optimizations. The single change that we have is that now anyext does not need to zero a register, because it does not use the exact code path as the zero_extend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171178 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-28 05:45:24 +00:00
Nadav Rotem
587fb1dd30 Reverse the 'if' condition and reduce the indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171172 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 23:08:05 +00:00
Craig Topper
cccccabd07 Merge basic_sse12_fp_binop_p_int and basic_sse12_fp_binop_p_y_int multiclasses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171171 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 22:53:47 +00:00
Nadav Rotem
1a330af3b5 AVX/AVX2: Move the SEXT lowering code from a target specific DAGco to a lowering function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171170 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 22:47:16 +00:00
Craig Topper
d5fc507ff1 Merge basic_sse12_fp_binop_p and basic_sse12_fp_binop_p_y multiclasses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171166 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 18:51:50 +00:00
Chandler Carruth
a455fdd7e1 Add support to BasicBlocks for iterating backwards over the
instructions. This just exposes the already present reverse iterators of
the instruction ilist.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171159 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 12:00:56 +00:00
Chandler Carruth
edf315cd71 Provide a common half-open interval map info implementation, and just
re-use that for SlotIndexes. This way other users who want half-open
semantics can share the implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171158 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 11:29:17 +00:00
Chandler Carruth
7ccc2f7a78 Make this parameter be named consistently with most other
getAnalysisUsage implementations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171157 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 11:17:15 +00:00
Sean Silva
bdb0c0aaf3 docs: Add FAQ about "storing to a virtual register".
This came up for the N+1'st time today in IRC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171155 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 10:23:04 +00:00
Sean Silva
6fa16e192f docs: Move link to the new "external tutorials" area.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171154 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 08:57:08 +00:00
Alexey Samsonov
1c8b825c43 [ASan] Fix lifetime intrinsics handling. Now for each intrinsic we check if it describes one of 'interesting' allocas. Assume that allocas can go through casts and phi-nodes before apperaring as llvm.lifetime arguments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171153 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 08:50:58 +00:00
Nadav Rotem
3190be9c9a DAGCombinerInformation: add a getter that exposes the dagcombine level.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171152 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 08:44:35 +00:00
Alexey Samsonov
44185d4908 Fix new[]/delete mismatch in FullDependence spotted by AddressSanitizer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171150 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 08:40:37 +00:00
Nadav Rotem
898c5e86fb docs: Update the benchmark with updated perf numbers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171149 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 08:32:44 +00:00
Nadav Rotem
d6fb53adb1 On AVX/AVX2 the type v8i1 is legalized to v8i16, which is an XMM sized
register. In most cases we actually compare or select YMM-sized registers
and mixing the two types creates horrible code. This commit optimizes
some of the transition sequences.

PR14657.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171148 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 08:15:45 +00:00
Nadav Rotem
3c22a44400 AVX/AVX2: Move the code that lowers vector-trunc from a DAGCo-hook to custom lowering hook.
The vector truncs were scalarized during LegalizeVectorOps, later vectorized again by some DAGCombine optimization
and finally, lowered by a dagcombing optimization. Now, they are properly lowered during LegalizeVectorOps.
No new testcase because the original testcases still work.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171146 91177308-0d34-0410-b5e6-96231b3b80d8
2012-12-27 07:45:10 +00:00