Commit Graph

40690 Commits

Author SHA1 Message Date
Ted Kremenek
865287de4a Update CMake build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111063 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-14 01:55:09 +00:00
Dan Gohman
c2f40066bb LoopSimplify shouldn't split loop backedges that use indirectbr. PR7867.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111061 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-14 00:43:09 +00:00
Dan Gohman
e2c6d131d1 Teach SimplifyCFG how to simplify indirectbr instructions.
- Eliminate redundant successors.
 - Convert an indirectbr with one successor into a direct branch.

Also, generalize SimplifyCFG to be able to be run on a function entry block.
It knows quite a few simplifications which are applicable to the entry
block, and it only needs a few checks to avoid trouble with the entry block.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111060 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-14 00:29:42 +00:00
Jim Grosbach
3d72367d30 Add a local stack object block allocation pass. This is still an
experimental pass that allocates locals relative to one another before
register allocation and then assigns them to actual stack slots as a block
later in PEI. This will eventually allow targets with limited index offset
range to allocate additional base registers (not just FP and SP) to
more efficiently reference locals, as well as handle situations where
locals cannot be referenced via SP or FP at all (dynamic stack realignment
together with variable sized objects, for example). It's currently
incomplete and almost certainly buggy. Work in progress.

Disabled by default and gated via the -enable-local-stack-alloc command
line option.

rdar://8277890



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111059 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-14 00:15:52 +00:00
Dan Gohman
61556e3a94 Add a lint check for an indirectbr destination which has not
had its address taken.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111058 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 23:56:28 +00:00
Bob Wilson
20d8e4e7aa Add a Thumb2 t2RSBrr instruction for disassembly only.
This fixes another part of PR7792.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111057 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 23:24:25 +00:00
Jakob Stoklund Olesen
0a2b2a1497 Clean up the Spiller.h interface.
The earliestStart argument is entirely specific to linear scan allocation, and
can be easily calculated by RegAllocLinearScan.

Replace std::vector with SmallVector.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111055 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 22:56:53 +00:00
Bob Wilson
703af3ab12 Temporarily disable tail calls on ARM to work around some linker problems.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111050 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 22:43:33 +00:00
Bob Wilson
38aa2871fc Move the Thumb2 SSAT and USAT optional shift operator out of the
instruction opcode.  This fixes part of PR7792.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111047 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 21:48:10 +00:00
Dan Gohman
0ad2c7ace8 Various optimizations. Don't compare two loops' depths
when they are the same loop. Don't compare two instructions'
loop depths when they are in the same block.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111045 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 21:24:58 +00:00
Jakob Stoklund Olesen
fc412d85c4 Implement splitting inside a single block.
When a live range is contained a single block, we can split it around
instruction clusters. The current approach is very primitive, splitting before
and after the largest gap between uses.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111043 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 21:18:48 +00:00
Dan Gohman
e62d58884a Fix LSR's ExtractImmediate and ExtractSymbol to avoid calling
ScalarEvolution::getAddExpr, which can be pretty expensive, when nothing
has changed, which is pretty common.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111042 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 21:17:19 +00:00
Bruno Cardoso Lopes
30baa63474 Add comments to some pattern fragments in x86
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111041 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 20:39:01 +00:00
Dan Gohman
9cba97886c When testing whether one loop contains another, test this directly
rather than testing whether the loop contains the other's header.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111039 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 20:23:25 +00:00
Dan Gohman
9f1fb42b7e Add a const.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111038 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 20:17:27 +00:00
Dan Gohman
58a85b96b8 When creating a symmetric SCEV with a constant operand, put
the constant operand on the left, as that's where ScalarEvolution
will end up canonicalizing to.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111037 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 20:17:14 +00:00
Dan Gohman
71c4144076 An add recurrence is loop-invariant in any loop inside of its
associated loop. This avoids potentially expensive traversals
of the add recurrence's operands.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111034 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 20:11:39 +00:00
Bob Wilson
6daf2a254b Refactor the code for disassembling Thumb2 saturate instructions along the
same lines as the change I made for ARM saturate instructions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111029 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 19:04:21 +00:00
Dale Johannesen
1b4051095d Revert 110491. While not wrong, it was based on a
misanalysis and is undesirable.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111028 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 18:43:45 +00:00
Bruno Cardoso Lopes
bb0a9489e0 Fix comment to reflect code, and remove an unused argument
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111022 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 17:50:47 +00:00
Bruno Cardoso Lopes
bbadd39bbb Improve comment to make explicit why not to touch this could before JIT goes MC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111021 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 17:44:10 +00:00
Jim Grosbach
c92bb50a0c tidy up whitespace a bit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111019 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 16:55:08 +00:00
Eric Christopher
63f02ac349 Revert last patch and r110954 as I meant to.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111001 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 02:37:50 +00:00
Eric Christopher
4404c00db6 Revert r110954 for now, pseudo instructions can't make it through to the JIT.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@111000 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 02:30:00 +00:00
Jakob Stoklund Olesen
2780d3ca81 Let LiveInterval::addRange extend existing ranges, it will verify that value
numbers match. The old check could accidentally leave holes in openli.

Also let useIntv add all ranges for the phi-def value inserted by
enterIntvAtEnd. This works as long at the value mapping is established in
enterIntvAtEnd.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110995 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 01:05:26 +00:00
Jakob Stoklund Olesen
00667a54f2 Remember to actually update SplitAnalysis statistics now that we have a fancy
function to do it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110994 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 01:05:23 +00:00
Nate Begeman
7f1f4089a1 Reapply this transformation now that it is passing the external test which it previously failed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110987 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 00:17:53 +00:00
Nate Begeman
a966af2f4c Move some code from Verifier into SVI::isValidOperands. This allows us to catch bad shufflevector operations when they are created, rather than waiting for someone to notice later on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110986 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-13 00:16:46 +00:00
Dan Gohman
ca2835411f Trim #includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110983 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 23:56:03 +00:00
Dan Gohman
95df619a40 Tidy up whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110982 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 23:50:08 +00:00
Dan Gohman
9e2f6284bc Use .empty() instead of .size().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110981 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 23:46:28 +00:00
Jakob Stoklund Olesen
09c45d2fe8 Handle an empty dupli.
This can happen if the original interval has been broken into two disconnected
parts. Ideally, we should be able to detect when the graph is disconnected and
create separate intervals, but that code is not implemented yet.

Example:

Two basic blocks are both branching to a loop header. Our interval is defined in
both basic blocks, and live into the loop along both edges.

We decide to split the interval around the loop. The interval is split into an
inside part and an outside part. The outside part now has two disconnected
segments, one in each basic block.

If we later decide to split the outside interval into single blocks, we get one
interval per basic block and an empty dupli for the remainder.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110976 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 23:02:57 +00:00
Jakob Stoklund Olesen
2dee7a527b Update the SplitAnalysis statistics as uses are moved from curli to the new
split intervals. THis means the analysis can be used for multiple splits as long
as curli doesn't shrink.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110975 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 23:02:55 +00:00
Chris Lattner
1522ce9770 fix PR7876: If ipsccp decides that a function's address is taken
before it rewrites the code, we need to use that in the post-rewrite pass.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110962 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 22:25:23 +00:00
Bruno Cardoso Lopes
64baddc0f2 Some small clean-up: use of pseudo instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110954 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 20:55:18 +00:00
Johnny Chen
1adc40cac3 Cleaned up the for-disassembly-only entries in the arm instruction table so that
the memory barrier variants (other than 'SY' full system domain read and write)
are treated as one instruction with option operand.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110951 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 20:46:17 +00:00
Jakob Stoklund Olesen
fff2c4726b Also recompute HasPHIKill flags in LiveInterval::RenumberValues.
If a phi-def value were removed from the interval, the phi-kill flags are no
longer valid.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110949 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 20:38:03 +00:00
Evan Cheng
719510a178 Make sure ARM constant island pass does not break up an IT block. If the split point is in the middle of an IT block, it should move it up to just above the IT instruction. rdar://8302637
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110947 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 20:30:05 +00:00
Bruno Cardoso Lopes
642eb02045 - Teach SSEDomainFix to switch between different levels of AVX instructions. Here we guess that AVX will have domain issues, so just implement them for consistency and in the future we remove if it's unnecessary.
- Make foldMemoryOperandImpl aware of 256-bit zero vectors folding and support the 128-bit counterparts of AVX too.
- Make sure MOV[AU]PS instructions are only selected when SSE1 is enabled, and duplicate the patterns to match AVX.
- Add a testcase for a simple 128-bit zero vector creation.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110946 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 20:20:53 +00:00
Jakob Stoklund Olesen
1b2932024f Remove trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110944 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 20:01:23 +00:00
Jakob Stoklund Olesen
e1f543fbb3 Clean up debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110940 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 18:50:55 +00:00
Bruno Cardoso Lopes
6da9cee0f1 Define AVX 128-bit pattern versions of SET0PS/PD.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110937 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 18:20:59 +00:00
Jakob Stoklund Olesen
f1b05f2b0e Implement single block splitting.
Before spilling a live range, we split it into a separate range for each basic
block where it is used. That way we only get one reload per basic block if the
new smaller ranges can allocate to a register.

This type of splitting is already present in the standard spiller.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110934 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 17:07:14 +00:00
Dan Gohman
727356fc7d Optimize ScalarEvolution::getAddExpr's operand factoring code by
having it finish processing all of the muliply operands before
starting the whole getAddExpr process over again, instead of
immediately after the first simplification.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110916 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 15:00:23 +00:00
Dan Gohman
918e76b8b5 Hoist some loop-invariant code out of a hot loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110915 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 14:52:55 +00:00
Dan Gohman
dc7692b1ad Optimize ScalarEvolution::getAddExpr's duplicate operand detection
by having it finish processing the whole operand list before
starting the whole getAddExpr process over again, instead of
immediately after the first duplicate is found.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110914 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 14:46:54 +00:00
Duncan Sands
335db22392 Add a 'normalize' method to the Triple class, which takes a mucked up
target triple and straightens it out.  This does less than gcc's script
config.sub, for example it turns i386-mingw32 into i386--mingw32 not
i386-pc-mingw32, but it does a decent job of turning funky triples into
something that the rest of the Triple class can understand.  The plan
is to use this to canonicalize triple's when they are first provided
by users, and have the rest of LLVM only deal with canonical triples.
Once this is done the special case workarounds in the Triple constructor
can be removed, making the class more regular and easier to use.  The
comments and unittests for the Triple class are already adjusted in this
patch appropriately for this brave new world of increased uniformity.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110909 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 11:31:39 +00:00
Eric Christopher
7486278900 Temporarily revert 110737 and 110734, they were causing failures
in an external testsuite.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110905 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 07:01:22 +00:00
Bruno Cardoso Lopes
4d04362813 Fix comment order
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110898 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 02:08:52 +00:00
Bruno Cardoso Lopes
8c05a850f4 Begin to support some vector operations for AVX 256-bit intructions. The long
term goal here is to be able to match enough of vector_shuffle and build_vector
so all avx intrinsics which aren't mapped to their own built-ins but to
shufflevector calls can be codegen'd. This is the first (baby) step, support
building zeroed vectors.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@110897 91177308-0d34-0410-b5e6-96231b3b80d8
2010-08-12 02:06:36 +00:00