Commit Graph

5323 Commits

Author SHA1 Message Date
Evan Cheng
c4b527ac06 DAGCombine's logic for forming pre- and post- indexed loads / stores were being
overly conservative. It was concerned about cases where it would prohibit
folding simple [r, c] addressing modes. e.g.
  ldr r0, [r2]
  ldr r1, [r2, #4]
=>
  ldr r0, [r2], #4
  ldr r1, [r2]
Change the logic to look for such cases which allows it to form indexed memory
ops more aggressively.

rdar://10674430


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@148086 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-13 01:37:24 +00:00
Pete Cooper
d578b905de Added FPOW, FEXP, FLOG to PromoteNode so that custom actions can be set to Promote for those operations.
Sorry, no test case yet


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@148050 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-12 21:46:18 +00:00
Evan Cheng
f0a95356d6 Allow targets to select source order pre-RA scheduler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@148033 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-12 18:27:52 +00:00
Nadav Rotem
c8d12eee12 On AVX, we can load v8i32 at a time. The bug happens when two uneven loads are used.
When we load the v12i32 type, the GenWidenVectorLoads method generates two loads: v8i32 and v4i32 
and attempts to use CONCAT_VECTORS to join them. In this fix I concat undef values to widen 
the smaller value. The test "widen_load-2.ll" also exposes this bug on AVX.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147964 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-11 20:19:17 +00:00
Chandler Carruth
f103b3d1b9 Teach the X86 instruction selection to do some heroic transforms to
detect a pattern which can be implemented with a small 'shl' embedded in
the addressing mode scale. This happens in real code as follows:

  unsigned x = my_accelerator_table[input >> 11];

Here we have some lookup table that we look into using the high bits of
'input'. Each entity in the table is 4-bytes, which means this
implicitly gets turned into (once lowered out of a GEP):

  *(unsigned*)((char*)my_accelerator_table + ((input >> 11) << 2));

The shift right followed by a shift left is canonicalized to a smaller
shift right and masking off the low bits. That hides the shift right
which x86 has an addressing mode designed to support. We now detect
masks of this form, and produce the longer shift right followed by the
proper addressing mode. In addition to saving a (rather large)
instruction, this also reduces stalls in Intel chips on benchmarks I've
measured.

In order for all of this to work, one part of the DAG needs to be
canonicalized *still further* than it currently is. This involves
removing pointless 'trunc' nodes between a zextload and a zext. Without
that, we end up generating spurious masks and hiding the pattern.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147936 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-11 08:41:08 +00:00
Chandler Carruth
732f05c41f Add 'llvm_unreachable' to passify GCC's understanding of the constraints
of several newly un-defaulted switches. This also helps optimizers
(including LLVM's) recognize that every case is covered, and we should
assume as much.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147861 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-10 18:08:01 +00:00
David Blaikie
2bd335470f Remove unnecessary default cases in switches that cover all enum values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147855 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-10 16:47:17 +00:00
Nadav Rotem
6c0366cb25 Fix a bug in the legalization of shuffle vectors. When we emulate shuffles using BUILD_VECTORS we may be using a BV of different type. Make sure to cast it back.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147851 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-10 14:28:46 +00:00
Craig Topper
704e1a0f04 Replace some uses of hasNUsesOfValue(0, X) with !hasAnyUseOfValue(X)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147733 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-07 18:31:09 +00:00
Craig Topper
cc274526ef Add some DAG combines for SUBC/SUBE. If nothing uses the carry/borrow out of subc, turn it into a sub. Turn (subc x, x) into 0 with no borrow. Turn (subc x, 0) into x with no borrow. Turn (subc -1, x) into (xor x, -1) with no borrow. Turn sube with no borrow in into subc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147728 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-07 09:06:39 +00:00
Chad Rosier
4bf76e0e28 Add comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147696 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-06 23:45:47 +00:00
Chandler Carruth
62dfc51152 Prevent a DAGCombine from firing where there are two uses of
a combined-away node and the result of the combine isn't substantially
smaller than the input, it's just canonicalized. This is the first part
of a significant (7%) performance gain for Snappy's hot decompression
loop.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147604 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-05 11:05:55 +00:00
Craig Topper
515783466c Allow vector shuffle normalizing to use concat vector even if the sources are commuted in the shuffle mask.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147527 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-04 09:23:09 +00:00
Craig Topper
481b79c950 Implement VECTOR_SHUFFLE canonicalizations during DAG combine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147525 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-04 08:07:43 +00:00
Chris Lattner
fcd7090063 Turn a few more inline asm errors into "emitErrors" instead of fatal errors.
Before we'd get:

$ clang t.c 
fatal error: error in backend: Invalid operand for inline asm constraint 'i'!

Now we get:

$ clang t.c
t.c:16:5: error: invalid operand for inline asm constraint 'i'!
    "movq         (%4), %%mm0\n"
    ^

Which at least gets us the inline asm that is the problem.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147502 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-03 23:51:01 +00:00
Nadav Rotem
316477dd54 Fix incorrect widening of the bitcast sdnode in case the incoming operand is integer-promoted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147484 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-03 22:12:28 +00:00
Owen Anderson
517c4d7fda Remove the restriction that target intrinsics can only involve legal types. Targets can perfects well support intrinsics on illegal types, as long as they are prepared to perform custom expansion during type legalization. For example, a target where i64 is illegal might still support the i64 intrinsic operation using pairs of i32's. ARM already does some expansions like this for non-intrinsic operations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147472 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-03 20:09:02 +00:00
Elena Demikhovsky
ce58a03587 Fixed a bug in SelectionDAG.cpp.
The failure seen on win32, when i64 type is illegal.
It happens on stage of conversion VECTOR_SHUFFLE to BUILD_VECTOR.

The failure message is:
llc: SelectionDAG.cpp:784: void VerifyNodeCommon(llvm::SDNode*): Assertion `(I->getValueType() == EltVT || (EltVT.isInteger() && I->getValueType().isInteger() && EltVT.bitsLE(I->getValueType()))) && "Wrong operand type!"' failed.

I added a special test that checks vector shuffle on win32.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147445 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-03 11:59:04 +00:00
Rafael Espindola
acae2a63b9 Revert 147399. It broke CodeGen/ARM/vext.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147400 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-01 17:36:23 +00:00
Elena Demikhovsky
ac12855066 Fixed a bug in SelectionDAG.cpp.
The failure seen on win32, when i64 type is illegal.
It happens on stage of conversion VECTOR_SHUFFLE to BUILD_VECTOR.

The failure message is:
llc: SelectionDAG.cpp:784: void VerifyNodeCommon(llvm::SDNode*): Assertion `(I->getValueType() == EltVT || (EltVT.isInteger() && I->getValueType().isInteger() && EltVT.bitsLE(I->getValueType()))) && "Wrong operand type!"' failed.

I added a special test that checks vector shuffle on win32.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147399 91177308-0d34-0410-b5e6-96231b3b80d8
2012-01-01 16:22:47 +00:00
Nadav Rotem
6059b83695 PR11662.
Promotion of the mask operand needs to be done using PromoteTargetBoolean, and not padded with garbage.




git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147309 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-28 13:08:20 +00:00
Eli Friedman
d6e2560e7a Make sure DAGCombiner doesn't introduce multiple loads from the same memory location. PR10747, part 2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147283 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-26 22:49:32 +00:00
Nadav Rotem
fbb6f593c2 Fix a typo in the widening of vectors in PromoteIntRes. Patch by Shemer Anat.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147272 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-25 20:01:38 +00:00
Dylan Noblesmith
9421406aad drop unneeded config.h includes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147197 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-22 23:04:07 +00:00
Jakub Staszak
25101bb2a7 Add some constantness to BranchProbabilityInfo and BlockFrequnencyInfo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146986 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-20 20:03:10 +00:00
David Blaikie
2d24e2a396 Unweaken vtables as per http://llvm.org/docs/CodingStandards.html#ll_virtual_anch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146960 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-20 02:50:00 +00:00
Dan Gohman
e3376ecd50 Add basic generic CodeGen support for half.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146927 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-20 00:02:33 +00:00
Joerg Sonnenberger
3470693641 Allow inlining of functions with returns_twice calls, if they have the
attribute themselve.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146851 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-18 20:35:43 +00:00
Devang Patel
0508d047fe Update DebugLoc while merging nodes at -O0.
Patch by Kyriakos Georgiou!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146670 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-15 18:21:18 +00:00
Eli Friedman
ca072a3977 Don't try to form FGETSIGN after legalization; it is possible in some cases, but the existing code can't do it correctly. PR11570.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146630 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-15 02:07:20 +00:00
Owen Anderson
4e0adfa7f7 Enable synthesis of FLOG2 and FEXP2 SelectionDAG nodes from libm calls. These are already marked as illegal by default.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146623 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-15 00:54:12 +00:00
Eli Friedman
e08db65c48 Add missing cases to SDNode::getOperationName(). Patch by Micah Villmow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146548 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-14 02:28:54 +00:00
Chad Rosier
5bd83345c7 [fast-isel] Remove SelectInsertValue() as fast-isel wasn't designed to handle
instructions that define aggregate types.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146492 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-13 17:45:06 +00:00
Chandler Carruth
63974b2144 Initial CodeGen support for CTTZ/CTLZ where a zero input produces an
undefined result. This adds new ISD nodes for the new semantics,
selecting them when the LLVM intrinsic indicates that the undef behavior
is desired. The new nodes expand trivially to the old nodes, so targets
don't actually need to do anything to support these new nodes besides
indicating that they should be expanded. I've done this for all the
operand types that I could figure out for all the targets. Owners of
various targets, please review and let me know if any of these are
incorrect.

Note that the expand behavior is *conservatively correct*, and exactly
matches LLVM's current behavior with these operations. Ideally this
patch will not change behavior in any way. For example the regtest suite
finds the exact same instruction sequences coming out of the code
generator. That's why there are no new tests here -- all of this is
being exercised by the existing test suite.

Thanks to Duncan Sands for reviewing the various bits of this patch and
helping me get the wrinkles ironed out with expanding for each target.
Also thanks to Chris for clarifying through all the discussions that
this is indeed the approach he was looking for. That said, there are
likely still rough spots. Further review much appreciated.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146466 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-13 01:56:10 +00:00
Chad Rosier
f2a745efe5 [fast-isel] Guard "exhastive" fast-isel output with -fast-isel-verbose2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146453 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-13 00:05:11 +00:00
Daniel Dunbar
4ab406d7fc LLVMBuild: Remove trailing newline, which irked me.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146409 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-12 19:48:00 +00:00
Chad Rosier
4552d3e22a [fast-isel] SelectInsertValue seems to be causing miscompiles for ARM. Disable while I investigate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146331 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-10 21:27:40 +00:00
Chad Rosier
b435aa2c1d Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146327 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-10 19:48:51 +00:00
Chad Rosier
cd462d055f [fast-isel] Add support for selecting insertvalue.
rdar://10530851

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146276 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-09 20:09:54 +00:00
Eli Friedman
2dd0353fec Fix a couple of logic bugs in TargetLowering::SimplifyDemandedBits. PR11514.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146219 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-09 01:16:26 +00:00
Owen Anderson
243eb9ecbb Enhance both TargetLibraryInfo and SelectionDAGBuilder so that the latter can use the former to prevent the formation of libm SDNode's when -fno-builtin is passed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146193 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-08 22:15:21 +00:00
Chad Rosier
73e08d3507 Add rather verbose stats for fast-isel failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146186 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-08 21:37:10 +00:00
Owen Anderson
4a4fdf3476 Teach SelectionDAG to match more calls to libm functions onto existing SDNodes. Mark these nodes as illegal by default, unless the target declares otherwise.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146171 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-08 19:32:14 +00:00
Nadav Rotem
44bac7cd65 Fix a bug in the integer-promotion of bitcast operations on vector types.
We must not issue a bitcast operation for integer-promotion of vector types, because the
location of the values in the vector may be different.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146150 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-08 13:10:01 +00:00
Eli Friedman
0e6307f642 Make sure we correctly set LiveRegGens when a call is unscheduled. <rdar://problem/10460321>. No testcase because this is very sensitive to scheduling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146087 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-07 22:24:28 +00:00
Eli Friedman
30c44e18bf Fix an assertion in the scheduler. PR11386. No testcase included because it's rather delicate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146083 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-07 22:06:02 +00:00
Nick Lewycky
e77ae2d692 These global variables aren't thread-safe, STATISTIC is. Andy Trick tells me
that he isn't using these any more, so just delete them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146076 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-07 21:35:59 +00:00
Evan Cheng
5a96b3dad2 Add bundle aware API for querying instruction properties and switch the code
generator to it. For non-bundle instructions, these behave exactly the same
as the MC layer API.

For properties like mayLoad / mayStore, look into the bundle and if any of the
bundled instructions has the property it would return true.
For properties like isPredicable, only return true if *all* of the bundled
instructions have the property.
For properties like canFoldAsLoad, isCompare, conservatively return false for
bundles.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146026 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-07 07:15:52 +00:00
Eli Friedman
1c663fee56 Zap unnecessary isIntDivCheap() check. PR11485. No testcase because this doesn't affect any in-tree target.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@146015 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-07 03:55:52 +00:00
Eli Friedman
26323442d5 Fix an optimization involving EXTRACT_SUBVECTOR in DAGCombine so it behaves correctly. PR11494.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@145996 91177308-0d34-0410-b5e6-96231b3b80d8
2011-12-07 00:11:56 +00:00