This is equivalent to the AMDGPUTargetMachine now, but it is the
starting point for separating R600 and GCN functionality into separate
targets.
It is recommened that users start using the gcn triple for GCN-based
GPUs, because using the r600 triple for these GPUs will be deprecated in
the future.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225277 91177308-0d34-0410-b5e6-96231b3b80d8
This patch improves the logic added at revision 224899 (see review D6728) that
teaches the backend when it is profitable to speculate calls to cttz/ctlz.
The original algorithm conservatively avoided speculating more than one
instruction from a basic block in a control flow grap modelling an if-statement.
In particular, the only allowed instruction (excluding the terminator) was a
call to cttz/ctlz. However, there are cases where we could be less conservative
and still be able to speculate a call to cttz/ctlz.
With this patch, CodeGenPrepare now tries to speculate a cttz/ctlz if the
result is zero extended/truncated in the same basic block, and the zext/trunc
instruction is "free" for the target.
Added new test cases to CodeGen/X86/cttz-ctlz.ll
Differential Revision: http://reviews.llvm.org/D6853
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225274 91177308-0d34-0410-b5e6-96231b3b80d8
In r225251, I removed an old entry from the README.txt file. While there are
several contributing factors (including pieces in Clang's ABI code), upon
further reflection, the backend part deserves a regression test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225268 91177308-0d34-0410-b5e6-96231b3b80d8
This is already handled in general when it is known the
conversion can't lose bits with smaller integer types
casted into wider floating point types.
This pattern happens somewhat often in GPU programs that cast
workitem intrinsics to float, which are often compared with 0.
Specifically handle the special case of compares with zero which
should also be known to not lose information. I had a more general
version of this which allows equality compares if the casted float is
exactly representable in the integer, but I'm not 100% confident that
is always correct.
Also fold cases that aren't integers to true / false.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225265 91177308-0d34-0410-b5e6-96231b3b80d8
Use this to test that path of invalidation. This test actually shows
redundant invalidation here that is really bad. I'm going to work on
fixing that next, but wanted to commit the test harness now that its all
working.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225257 91177308-0d34-0410-b5e6-96231b3b80d8
Requires new AsmParserOperand types that detect 16-bit and 32/64-bit mode so that we choose the right instruction based on default sizing without predicates. This is necessary since predicates mess up the disassembler table building.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225256 91177308-0d34-0410-b5e6-96231b3b80d8
Try harder to get rid of bitcast'd calls by ptrtoint/inttoptr'ing
arguments and return values when DataLayout says it is safe to do so.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225254 91177308-0d34-0410-b5e6-96231b3b80d8
remove an extra, redundant pass manager wrapping every run.
I had kept seeing these when manually testing, but it was getting really
annoying and was going to cause problems with overly eager invalidation.
The root cause was an overly complex and unnecessary pile of code for
parsing the outer layer of the pass pipeline. We can instead delegate
most of this to the recursive pipeline parsing.
I've added some somewhat more basic and precise tests to catch this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225253 91177308-0d34-0410-b5e6-96231b3b80d8
"ELF Handling for Thread-Local Storage" specifies that R_X86_64_GOTTPOFF
relocation target a movq or addq instruction.
Prohibit the truncation of such loads to movl or addl.
This fixes PR22083.
Differential Revision: http://reviews.llvm.org/D6839
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225250 91177308-0d34-0410-b5e6-96231b3b80d8
The old target DAG combine that allowed for performing int_to_fp(fp_to_int(x))
without a load/store pair is updated here with support for unsigned integers,
and to support single-precision values without a third rounding step, on newer
cores with the appropriate instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225248 91177308-0d34-0410-b5e6-96231b3b80d8
a specific analysis result.
This is quite handy to test things, and will also likely be very useful
for debugging issues. You could narrow down pass validation failures by
walking these invalidate pass runs up and down the pass pipeline, etc.
I've added support to the pass pipeline parsing to be able to create one
of these for any analysis pass desired.
Just adding this class uncovered one latent bug where the
AnalysisManager CRTP base class had a hard-coded Module type rather than
using IRUnitT.
I've also added tests for invalidation and caching of analyses in
a basic way across all the pass managers. These in turn uncovered two
more bugs where we failed to correctly invalidate an analysis -- its
results were invalidated but the key for re-running the pass was never
cleared and so it was never re-run. Quite nasty. I'm very glad to debug
this here rather than with a full system.
Also, yes, the naming here is horrid. I'm going to update some of the
names to be slightly less awful shortly. But really, I've no "good"
ideas for naming. I'll be satisfied if I can get it to "not bad".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225246 91177308-0d34-0410-b5e6-96231b3b80d8
manager tests to use them and be significantly more comprehensive.
This, naturally, uncovered a bug where the CGSCC pass manager wasn't
printing analyses when they were run.
The only remaining core manipulator is I think an invalidate pass
similar to the require pass. That'll be next. =]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225240 91177308-0d34-0410-b5e6-96231b3b80d8
is a no-op other than requiring some analysis results be available.
This can be used in real pass pipelines to force the usually lazy
analysis running to eagerly compute something at a specific point, and
it can be used to test the pass manager infrastructure (my primary use
at the moment).
I've also added bit of pipeline parsing magic to support generating
these directly from the opt command so that you can directly use these
when debugging your analysis. The syntax is:
require<analysis-name>
This can be used at any level of the pass manager. For example:
cgscc(function(require<my-analysis>,no-op-function))
This would produce a no-op function pass requiring my-analysis, followed
by a fully no-op function pass, both of these in a function pass manager
which is nested inside of a bottom-up CGSCC pass manager which is in the
top-level (implicit) module pass manager.
I have zero attachment to the particular syntax I'm using here. Consider
it a straw man for use while I'm testing and fleshing things out.
Suggestions for better syntax welcome, and I'll update everything based
on any consensus that develops.
I've used this new functionality to more directly test the analysis
printing rather than relying on the cgscc pass manager running an
analysis for me. This is still minimally tested because I need to have
analyses to run first! ;] That patch is next, but wanted to keep this
one separate for easier review and discussion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225236 91177308-0d34-0410-b5e6-96231b3b80d8
We now produce the desired code as noted in the README.txt file (no spurious
or). Remove the README entry and improve the regression test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225214 91177308-0d34-0410-b5e6-96231b3b80d8
We now produce the desired code as noted in the README.txt file. Remove the
README entry and add a regression test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225209 91177308-0d34-0410-b5e6-96231b3b80d8
This object is meant to own the ObjectFiles and their underlying
MemoryBuffer. It is basically the equivalent of an OwningBinary
except that it efficiently handles Archives. It is optimized for
efficiently providing mappings of members of the same archive when
they are opened successively (which is standard in Darwin debug
maps, objects from the same archive will be contiguous).
Of course, the BinaryHolder will also be used by the DWARF linker
once it is commited, but for now only the debug map parser uses it.
With this change, you can run llvm-dsymutil on your Darwin debug build
of clang and get a complete debug map for it.
Differential Revision: http://reviews.llvm.org/D6690
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225207 91177308-0d34-0410-b5e6-96231b3b80d8
We now produce the desired code as noted in the README.txt file. Remove the
README entry and add a regression test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225205 91177308-0d34-0410-b5e6-96231b3b80d8
Consider this function from our README.txt file:
int foo(int a, int b) { return (a < b) << 4; }
We now explicitly track CR bits by default, so the comment in the README.txt
about not really having a SETCC is no longer accurate, but we did generate this
somewhat silly code:
cmpw 0, 3, 4
li 3, 0
li 12, 1
isel 3, 12, 3, 0
sldi 3, 3, 4
blr
which generates the zext as a select between 0 and 1, and then shifts the
result by a constant amount. Here we preprocess the DAG in order to fold the
results of operations on an extension of an i1 value into the SELECT_I[48]
pseudo instruction when the resulting constant can be materialized using one
instruction (just like the 0 and 1). This was not implemented as a DAGCombine
because the resulting code would have been anti-canonical and depends on
replacing chained user nodes, which does not fit well into the lowering
paradigm. Now we generate:
cmpw 0, 3, 4
li 3, 0
li 12, 16
isel 3, 12, 3, 0
blr
which is less silly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225203 91177308-0d34-0410-b5e6-96231b3b80d8
The 64-bit semantics of cntlzw are not special, the 32-bit population count is
stored as a 64-bit value in the range [0,32]. As a result, it is always zero
extended, and it can be added to the PPCISelDAGToDAG peephole optimization as a
frontier instruction for the removal of unnecessary zero extensions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225192 91177308-0d34-0410-b5e6-96231b3b80d8
lhbrx and lwbrx not only load their data with byte swapping, but also clear the
upper 32 bits (at least). As a result, they can be added to the PPCISelDAGToDAG
peephole optimization as frontier instructions for the removal of unnecessary
zero extensions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225189 91177308-0d34-0410-b5e6-96231b3b80d8
We used to generate code similar to:
umov.b w8, v0[2]
strb w8, [x0, x1]
because the STR*ro* patterns were preferred to ST1*.
Instead, we can avoid going through GPRs, and generate:
add x8, x0, x1
st1.b { v0 }[2], [x8]
This patch increases the ST1* AddedComplexity to achieve that.
rdar://16372710
Differential Revision: http://reviews.llvm.org/D6202
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225183 91177308-0d34-0410-b5e6-96231b3b80d8
For 0-lane stores, we used to generate code similar to:
fmov w8, s0
str w8, [x0, x1, lsl #2]
instead of:
str s0, [x0, x1, lsl #2]
To correct that: for store lane 0 patterns, directly match to STR <subreg>0.
Byte-sized instructions don't have the special case for a 0 index,
because FPR8s are defined to have untyped content.
rdar://16372710
Differential Revision: http://reviews.llvm.org/D6772
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225181 91177308-0d34-0410-b5e6-96231b3b80d8
Tag_compatibility takes two arguments, but before this patch it would
erroneously accept just one, it now produces an error in that case.
Change-Id: I530f918587620d0d5dfebf639944d6083871ef7d
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225167 91177308-0d34-0410-b5e6-96231b3b80d8
Claim conformance to version 2.09 of the ARM ABI.
This build attribute must be emitted first amongst the build attributes when
written to an object file. This is to simplify conformance detection by
consumers.
Change-Id: If9eddcfc416bc9ad6e5cc8cdcb05d0031af7657e
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225166 91177308-0d34-0410-b5e6-96231b3b80d8
when all are being preserved.
We want to short-circuit this for a couple of reasons. One, I don't
really want passes to grow a dependency on actually receiving their
invalidate call when they've been preserved. I'm thinking about removing
this entirely. But more importantly, preserving everything is likely to
be the common case in a lot of scenarios, and it would be really good to
bypass all of the invalidation and preservation machinery there.
Avoiding calling N opaque functions to try to invalidate things that are
by definition still valid seems important. =]
This wasn't really inpsired by much other than seeing the spam in the
logging for analyses, but it seems better ot get it checked in rather
than forgetting about it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225163 91177308-0d34-0410-b5e6-96231b3b80d8
manager.
This starts to allow us to test analyses more easily, but it's really
only the beginning. Some of the code here is still untestable without
manual changes to create analysis passes, but I wanted to factor it into
a small of chunks as possible.
Next up in order to be able to test things are, in no particular order:
- No-op analyses passes so we don't have to use real ones to exercise
the pass maneger itself.
- Automatic way of generating dummy passes that require an analysis be
run, including a variant that calls a 'print' method on a pass to make
it even easier to print out the results of an analysis.
- Dummy passes that invalidate all analyses for their IR unit so we can
test invalidation and re-runs.
- Automatic way to print each analysis pass as it is re-run.
- Automatic but optional verification of analysis passes everywhere
possible.
I'm not claiming I'll get to all of these immediately, but that's what
is in the pipeline at some stage. I'm fleshing out exactly what I need
and what to prioritize by working on converting analyses and then trying
to test the conversion. =]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225162 91177308-0d34-0410-b5e6-96231b3b80d8
{code}
// loop body
... = a[i] (1)
... = a[i+1] (2)
.......
a[i+1] = .... (3)
a[i] = ... (4)
{code}
The algorithm tries to collect memory access candidates from AliasSetTracker, and then check memory dependences one another. The memory accesses are unique in AliasSetTracker, and a single memory access in AliasSetTracker may map to multiple entries in AccessAnalysis, which could cover both 'read' and 'write'. Originally the algorithm only checked 'write' entry in Accesses if only 'write' exists. This is incorrect and the consequence is it ignored all read access, and finally some RAW and WAR dependence are missed.
For the case given above, if we ignore two reads, the dependence between (1) and (3) would not be able to be captured, and finally this loop will be incorrectly vectorized.
The fix simply inserts a new loop to find all entries in Accesses. Since it will skip most of all other memory accesses by checking the Value pointer at the very beginning of the loop, it should not increase compile-time visibly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225159 91177308-0d34-0410-b5e6-96231b3b80d8
PPC has an instruction for ctlz with defined zero behavior, and our lowering of
cttz (provided by DAGCombine) is also efficient and branchless, so speculating
these makes sense.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225150 91177308-0d34-0410-b5e6-96231b3b80d8
assert out of the new pre-splitting in SROA.
This fix makes the code do what was originally intended -- when we have
a store of a load both dealing in the same alloca, we force them to both
be pre-split with identical offsets. This is really quite hard to do
because we can keep discovering problems as we go along. We have to
track every load over the current alloca which for any resaon becomes
invalid for pre-splitting, and go back to remove all stores of those
loads. I've included a couple of test cases derived from PR22093 that
cover the different ways this can happen. While that PR only really
triggered the first of these two, its the same fundamental issue.
The other challenge here is documented in a FIXME now. We end up being
quite a bit more aggressive for pre-splitting when loads and stores
don't refer to the same alloca. This aggressiveness comes at the cost of
introducing potentially redundant loads. It isn't clear that this is the
right balance. It might be considerably better to require that we only
do pre-splitting when we can presplit every load and store involved in
the entire operation. That would give more consistent if conservative
results. Unfortunately, it requires a non-trivial change to the actual
pre-splitting operation in order to correctly handle cases where we end
up pre-splitting stores out-of-order. And it isn't 100% clear that this
is the right direction, although I'm starting to suspect that it is.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225149 91177308-0d34-0410-b5e6-96231b3b80d8
r225135 added the ability to materialize i64 constants using rotations in order
to reduce the instruction count. Sometimes we can use a rotation only with some
extra masking, so that we take advantage of the fact that generating a bunch of
extra higher-order 1 bits is easy using li/lis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225147 91177308-0d34-0410-b5e6-96231b3b80d8
The required functionality has been there for some time, but I never
managed to actually wire it into the command line registry of passes.
Let's do that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225144 91177308-0d34-0410-b5e6-96231b3b80d8
Materializing full 64-bit constants on PPC64 can be expensive, requiring up to
5 instructions depending on the locations of the non-zero bits. Sometimes
materializing a rotated constant, and then applying the inverse rotation, requires
fewer instructions than the direct method. If so, do that instead.
In r225132, I added support for forming constants using bit inversion. In
effect, this reverts that commit and replaces it with rotation support. The bit
inversion is useful for turning constants that are mostly ones into ones that
are mostly zeros (thus enabling a more-efficient shift-based materialization),
but the same effect can be obtained by using negative constants and a rotate,
and that is at least as efficient, if not more.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225135 91177308-0d34-0410-b5e6-96231b3b80d8
Materializing full 64-bit constants on PPC64 can be expensive, requiring up to
5 instructions depending on the locations of the non-zero bits. Sometimes
materializing the bit-reversed constant, and then flipping the bits, requires
fewer instructions than the direct method. If so, do that instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225132 91177308-0d34-0410-b5e6-96231b3b80d8
We assumed the output of a match was a Value, this would cause us to
assert because we would fail a cast<>. Instead, use a helper in the
Operator family to hide the distinction between Value and Constant.
This fixes PR22087.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225127 91177308-0d34-0410-b5e6-96231b3b80d8
PHI nodes can have zero operands in the middle of a transform. It is
expected that utilities in Analysis don't freak out when this happens.
Note that it is considered invalid to allow these misshapen phi nodes to
make it to another pass.
This fixes PR22086.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225126 91177308-0d34-0410-b5e6-96231b3b80d8
This enhances llvm-readobj to print out the COFF export table, similar to the
-coff-import option. This is useful for testing in lld.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225120 91177308-0d34-0410-b5e6-96231b3b80d8
Weak externals are resolved statically, so we can actually generate the tail
call on PE/COFF targets without breaking the requirements. It is questionable
whether we want to propagate the current behaviour for MachO as the requirements
are part of the ARM ELF specifications, and it seems that prior to the SVN
r215890, we would have tail'ed the call. For now, be conservative and only
permit it on PE/COFF where the call will always be fully resolved.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225119 91177308-0d34-0410-b5e6-96231b3b80d8
The existing code provided for specifying a global loop alignment preference.
However, the preferred loop alignment might depend on the loop itself. For
recent POWER cores, loops between 5 and 8 instructions should have 32-byte
alignment (while the others are better with 16-byte alignment) so that the
entire loop will fit in one i-cache line.
To support this, getPrefLoopAlignment has been made virtual, and can be
provided with an optional MachineLoop* so the target can inspect the loop
before answering the query. The default behavior, as before, is to return the
value set with setPrefLoopAlignment. MachineBlockPlacement now queries the
target for each loop instead of only once per function. There should be no
functional change for other targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225117 91177308-0d34-0410-b5e6-96231b3b80d8
Most modern PowerPC cores prefer that functions and loops start on
16-byte-aligned boundaries (*), so instruct block placement, etc. to make this
happen. The branch selector has also been adjusted so account for the extra
nops that might now be inserted before loop headers.
(*) Some cores actually prefer other alignments for small loops, but that will
be addressed in a follow-up commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225115 91177308-0d34-0410-b5e6-96231b3b80d8
Newer POWER cores, and the A2, support the cmpb instruction. This instruction
compares its operands, treating each of the 8 bytes in the GPRs separately,
returning a 'mask' result of 0 (for false) or -1 (for true) in each byte.
Code generation support is added, in the form of a PPCISelDAGToDAG
DAG-preprocessing routine, that recognizes patterns close to what the
instruction computes (either exactly, or related by a constant masking
operation), and generates the cmpb instruction (along with any necessary
constant masking operation). This can be expanded if use cases arise.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225106 91177308-0d34-0410-b5e6-96231b3b80d8
We know overflow always occurs if both ~LHSKnownZero * ~RHSKnownZero
and LHSKnownOne * RHSKnownOne overflow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225077 91177308-0d34-0410-b5e6-96231b3b80d8
This is necessary to allow the disassembler to be able to handle AdSize32 instructions in 64-bit mode when address size prefix is used.
Eventually we should probably also support 'addr32' and 'addr16' in the assembler to override the address size on some of these instructions. But for now we'll just use special operand types that will lookup the current mode size to select the right instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225075 91177308-0d34-0410-b5e6-96231b3b80d8
a pre-splitting pass over loads and stores.
Historically, splitting could cause enough problems that I hamstrung the
entire process with a requirement that splittable integer loads and
stores must cover the entire alloca. All smaller loads and stores were
unsplittable to prevent chaos from ensuing. With the new pre-splitting
logic that does load/store pair splitting I introduced in r225061, we
can now very nicely handle arbitrarily splittable loads and stores. In
order to fully benefit from these smarts, we need to mark all of the
integer loads and stores as splittable.
However, we don't actually want to rewrite partitions with all integer
loads and stores marked as splittable. This will fail to extract scalar
integers from aggregates, which is kind of the point of SROA. =] In
order to resolve this, what we really want to do is only do
pre-splitting on the alloca slices with integer loads and stores fully
splittable. This allows us to uncover all non-integer uses of the alloca
that would benefit from a split in an integer load or store (and where
introducing the split is safe because it is just memory transfer from
a load to a store). Once done, we make all the non-whole-alloca integer
loads and stores unsplittable just as they have historically been,
repartition and rewrite.
The result is that when there are integer loads and stores anywhere
within an alloca (such as from a memcpy of a sub-object of a larger
object), we can split them up if there are non-integer components to the
aggregate hiding beneath. I've added the challenging test cases to
demonstrate how this is able to promote to scalars even a case where we
have even *partially* overlapping loads and stores.
This restores the single-store behavior for small arrays of i8s which is
really nice. I've restored both the little endian testing and big endian
testing for these exactly as they were prior to r225061. It also forced
me to be more aggressive in an alignment test to actually defeat SROA.
=] Without the added volatiles there, we actually split up the weird i16
loads and produce nice double allocas with better alignment.
This also uncovered a number of bugs where we failed to handle
splittable load and store slices which didn't have a begininng offset of
zero. Those fixes are included, and without them the existing test cases
explode in glorious fireworks. =]
I've kept support for leaving whole-alloca integer loads and stores as
splittable even for the purpose of rewriting, but I think that's likely
no longer needed. With the new pre-splitting, we might be able to remove
all the splitting support for loads and stores from the rewriter. Not
doing that in this patch to try to isolate any performance regressions
that causes in an easy to find and revert chunk.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225074 91177308-0d34-0410-b5e6-96231b3b80d8
stores.
When there are accesses to an entire alloca with an integer
load or store as well as accesses to small pieces of the alloca, SROA
splits up the large integer accesses. In order to do that, it uses bit
math to merge the small accesses into large integers. While this is
effective, it produces insane IR that can cause significant problems in
the rest of the optimizer:
- It can cause load and store mismatches with GVN on the non-alloca side
where we end up loading an i64 (or some such) rather than loading
specific elements that are stored.
- We can't always get rid of the integer bit math, which is why we can't
always fix the loads and stores to work well with GVN.
- This is especially bad when we have operations that mix poorly with
integer bit math such as floating point operations.
- It will block things like the vectorizer which might be able to handle
the scalar stores that underly the aggregate.
At the same time, we can't just directly split up these loads and stores
in all cases. If there is actual integer arithmetic involved on the
values, then using integer bit math is actually the perfect lowering
because we can often combine it heavily with the surrounding math.
The solution this patch provides is to find places where SROA is
partitioning aggregates into small elements, and look for splittable
loads and stores that it can split all the way to some other adjacent
load and store. These are uniformly the cases where failing to split the
loads and stores hurts the optimizer that I have seen, and I've looked
extensively at the code produced both from more and less aggressive
approaches to this problem.
However, it is quite tricky to actually do this in SROA. We may have
loads and stores to the same alloca, or other complex patterns that are
hard to handle. This complexity leads to the somewhat subtle algorithm
implemented here. We have to do this entire process as a separate pass
over the partitioning of the alloca, and split up all of the loads prior
to splitting the stores so that we can handle safely the cases of
overlapping, including partially overlapping, loads and stores to the
same alloca. We also have to reconstitute the post-split slice
configuration so we can avoid iterating again over all the alloca uses
(the slow part of SROA). But we also have to ensure that when we split
up loads and stores to *other* allocas, we *do* re-iterate over them in
SROA to adapt to the more refined partitioning now required.
With this, I actually think we can fix a long-standing TODO in SROA
where I avoided splitting as many loads and stores as probably should be
splittable. This limitation historically mitigated the fallout of all
the bad things mentioned above. Now that we have more intelligent
handling, I plan to remove the FIXME and more aggressively mark integer
loads and stores as splittable. I'll do that in a follow-up patch to
help with bisecting any fallout.
The net result of this change should be more fine-grained and accurate
scalars being formed out of aggregates. At the very least, Clang now
generates perfect code for this high-level test case using
std::complex<float>:
#include <complex>
void g1(std::complex<float> &x, float a, float b) {
x += std::complex<float>(a, b);
}
void g2(std::complex<float> &x, float a, float b) {
x -= std::complex<float>(a, b);
}
void foo(const std::complex<float> &x, float a, float b,
std::complex<float> &x1, std::complex<float> &x2) {
std::complex<float> l1 = x;
g1(l1, a, b);
std::complex<float> l2 = x;
g2(l2, a, b);
x1 = l1;
x2 = l2;
}
This code isn't just hypothetical either. It was reduced out of the hot
inner loops of essentially every part of the Eigen math library when
using std::complex<float>. Those loops would consistently and
pervasively hop between the floating point unit and the integer unit due
to bit math extraction and insertion of floating point values that were
"stored" in a 64-bit integer register around the loop backedge.
So far, this change has passed a bootstrap and I have done some other
testing and so far, no issues. That doesn't mean there won't be though,
so I'll be prepared to help with any fallout. If you performance swings
in particular, please let me know. I'm very curious what all the impact
of this change will be. Stay tuned for the follow-up to also split more
integer loads and stores.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225061 91177308-0d34-0410-b5e6-96231b3b80d8
This is the second installment of improvements to instruction selection for "bit
permutation" instruction sequences. r224318 added logic for instruction
selection for 32-bit bit permutation sequences, and this adds lowering for
64-bit sequences. The 64-bit sequences are more complicated than the 32-bit
ones because:
a) the 64-bit versions of the 32-bit rotate-and-mask instructions
work by replicating the lower 32-bits of the value-to-be-rotated into the
upper 32 bits -- and integrating this into the cost modeling for the various
bit group operations is non-trivial
b) unlike the 32-bit instructions in 32-bit mode, the rotate-and-mask instructions
cannot, in one instruction, specify the
mask starting index, the mask ending index, and the rotation factor. Also,
forming arbitrary 64-bit constants is more complicated than in 32-bit mode
because the number of instructions necessary is value dependent.
Plus, support for 'late masking' was added: it is sometimes more efficient to
treat the overall value as if it had no mandatory zero bits when planning the
bit-group insertions, and then mask them in at the very end. Unfortunately, as
the structure of the bit groups is different in the two cases, the more
feasible implementation technique was to generate both instruction sequences,
and then pick the shorter one.
And finally, we now generate reasonable code for i64 bswap:
rldicl 5, 3, 16, 0
rldicl 4, 3, 8, 0
rldicl 6, 3, 24, 0
rldimi 4, 5, 8, 48
rldicl 5, 3, 32, 0
rldimi 4, 6, 16, 40
rldicl 6, 3, 48, 0
rldimi 4, 5, 24, 32
rldicl 5, 3, 56, 0
rldimi 4, 6, 40, 16
rldimi 4, 5, 48, 8
rldimi 4, 3, 56, 0
vs. what we used to produce:
li 4, 255
rldicl 5, 3, 24, 40
rldicl 6, 3, 40, 24
rldicl 7, 3, 56, 8
sldi 8, 3, 8
sldi 10, 3, 24
sldi 12, 3, 40
rldicl 0, 3, 8, 56
sldi 9, 4, 32
sldi 11, 4, 40
sldi 4, 4, 48
andi. 5, 5, 65280
andis. 6, 6, 255
andis. 7, 7, 65280
sldi 3, 3, 56
and 8, 8, 9
and 4, 12, 4
and 9, 10, 11
or 6, 7, 6
or 5, 5, 0
or 3, 3, 4
or 7, 9, 8
or 4, 6, 5
or 3, 3, 7
or 3, 3, 4
which is 12 instructions, instead of 25, and seems optimal (at least in terms
of code size).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225056 91177308-0d34-0410-b5e6-96231b3b80d8
The issues was that AArch64 has additional restrictions on when local
relocations can be used. We have to take those into consideration when
deciding to put a L symbol in the symbol table or not.
Original message:
Remove doesSectionRequireSymbols.
In an assembly expression like
bar:
.long L0 + 1
the intended semantics is that bar will contain a pointer one byte past L0.
In sections that are merged by content (strings, 4 byte constants, etc), a
single position in the section doesn't give the linker enough information.
For example, it would not be able to tell a relocation must point to the
end of a string, since that would look just like the start of the next.
The solution used in ELF to use relocation with symbols if there is a non-zero
addend.
In MachO before this patch we would just keep all symbols in some sections.
This would miss some cases (only cstrings on x86_64 were implemented) and was
inefficient since most relocations have an addend of 0 and can be represented
without the symbol.
This patch implements the non-zero addend logic for MachO too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225048 91177308-0d34-0410-b5e6-96231b3b80d8
We are allowed to move the 'B' to the right hand side if we an prove
there is no signed overflow and if the comparison itself is signed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225034 91177308-0d34-0410-b5e6-96231b3b80d8
Too many different comment characters - instead of trying to account for
them all, instead disable the comments and just check for end-of-line
instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225020 91177308-0d34-0410-b5e6-96231b3b80d8
GCC does this for non-zero discriminators and since GCC doesn't produce
column info, that was the only place it comes up there. For LLVM, since
we can emit discriminators and/or column info, it makes more sense to
invert the condition and just test for changes in line number.
This should resolve at least some of the GDB 7.5 test suite failures
created by recent Clang changes that increase the location fidelity
(which, since Clang defaults to including column info on Linux by
default created a bunch of cases that confused GDB).
In theory we could do this better/differently by grouping actual source
statements together in a similar manner to the way lexical scopes are
handled but given that GDB isn't really in a position to consume that (&
users are probably somewhat used to different lines being different
'statements') this seems the safest and cheapest change. (I'm concerned
that doing this 'right' would bloat the debugloc data even further -
something Duncan's working hard to address)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225011 91177308-0d34-0410-b5e6-96231b3b80d8
Under the large code model, we cannot assume that __morestack lives within
2^31 bytes of the call site, so we cannot use pc-relative addressing. We
cannot perform the call via a temporary register, as the rax register may
be used to store the static chain, and all other suitable registers may be
either callee-save or used for parameter passing. We cannot use the stack
at this point either because __morestack manipulates the stack directly.
To avoid these issues, perform an indirect call via a read-only memory
location containing the address.
This solution is not perfect, as it assumes that the .rodata section
is laid out within 2^31 bytes of each function body, but this seems to
be sufficient for JIT.
Differential Revision: http://reviews.llvm.org/D6787
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225003 91177308-0d34-0410-b5e6-96231b3b80d8
If a linker directive is already quoted, don't try to quote it again, otherwise it creates a mess.
This pops up in places like:
#pragma comment(linker,"\"/foo bar'\"")
Differential Revision: http://reviews.llvm.org/D6792
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224998 91177308-0d34-0410-b5e6-96231b3b80d8
In an assembly expression like
bar:
.long L0 + 1
the intended semantics is that bar will contain a pointer one byte past L0.
In sections that are merged by content (strings, 4 byte constants, etc), a
single position in the section doesn't give the linker enough information.
For example, it would not be able to tell a relocation must point to the
end of a string, since that would look just like the start of the next.
The solution used in ELF to use relocation with symbols if there is a non-zero
addend.
In MachO before this patch we would just keep all symbols in some sections.
This would miss some cases (only cstrings on x86_64 were implemented) and was
inefficient since most relocations have an addend of 0 and can be represented
without the symbol.
This patch implements the non-zero addend logic for MachO too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224985 91177308-0d34-0410-b5e6-96231b3b80d8
It looks like the original intent was to check which symbols were created.
With macho-dump the sections were being checked just to match which symbol
was in which section.
llvm-objdump prints the section a symbol is in.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224980 91177308-0d34-0410-b5e6-96231b3b80d8
These are simply a collection of tests intended to show that information about the contents of gc references in the heap is lost at a statepoint. I've tried to write them so that they don't disallow correct transformations, while still being fairly easy to understand.
p.s. Ideas for additional tests are welcome.
Differential Revision: http://reviews.llvm.org/D6491
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224971 91177308-0d34-0410-b5e6-96231b3b80d8
This change implements four basic optimizations:
If a relocated value isn't used, it doesn't need to be relocated.
If the value being relocated is null, relocation doesn't change that. (Technically, this might be collector specific. I don't know of one which it doesn't work for though.)
If the value being relocated is undef, the relocation is meaningless.
If the value being relocated was known nonnull, the relocated pointer also isn't null. (Since it points to the same source language object.)
I outlined other planned work in comments.
Differential Revision: http://reviews.llvm.org/D6600
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224968 91177308-0d34-0410-b5e6-96231b3b80d8
In LICM, we have a check for an instruction which is guaranteed to execute and thus can't introduce any new faults if moved to the preheader. To handle a function which might unconditionally throw when first called, we check for any potentially throwing call in the loop and give up.
This is unfortunate when the potentially throwing condition is down a rare path. It prevents essentially all LICM of potentially faulting instructions where the faulting condition is checked outside the loop. It also greatly diminishes the utility of loop unswitching since control dependent instructions - which are now likely in the loops header block - will not be lifted by subsequent LICM runs.
define void @nothrow_header(i64 %x, i64 %y, i1 %cond) {
; CHECK-LABEL: nothrow_header
; CHECK-LABEL: entry
; CHECK: %div = udiv i64 %x, %y
; CHECK-LABEL: loop
; CHECK: call void @use(i64 %div)
entry:
br label %loop
loop: ; preds = %entry, %for.inc
%div = udiv i64 %x, %y
br i1 %cond, label %loop-if, label %exit
loop-if:
call void @use(i64 %div)
br label %loop
exit:
ret void
}
The current patch really only helps with non-memory instructions (i.e. divs, etc..) since the maythrow call down the rare path will be considered to alias an otherwise hoistable load. The one exception is that it does kick in for loads which are known to be invariant without regard to other possible stores, i.e. those marked with either !invarant.load metadata of tbaa 'is constant memory' metadata.
Differential Revision: http://reviews.llvm.org/D6725
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224965 91177308-0d34-0410-b5e6-96231b3b80d8
This patches fixes a miscompile where we were assuming that loading from null is undefined and thus we could assume it doesn't happen. This transform is perfectly legal in address space 0, but is not neccessarily legal in other address spaces.
We really should introduce a hook to control this property on a per target per address space basis. We may be loosing valuable optimizations in some address spaces by being too conservative.
Original patch by Thomas P Raoux (submitted to llvm-commits), tests and formatting fixes by me.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224961 91177308-0d34-0410-b5e6-96231b3b80d8
The else case ResultReg was not checked for validity.
To my surprise, this case was not hit in any of the
existing test cases. This includes a new test cases
that tests this path.
Also drop the `target triple` declaration from the
original test as suggested by H.J. Lu, because
apparently with it the test won't be run on Linux
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224901 91177308-0d34-0410-b5e6-96231b3b80d8
If the control flow is modelling an if-statement where the only instruction in
the 'then' basic block (excluding the terminator) is a call to cttz/ctlz,
CodeGenPrepare can try to speculate the cttz/ctlz call and simplify the control
flow graph.
Example:
\code
entry:
%cmp = icmp eq i64 %val, 0
br i1 %cmp, label %end.bb, label %then.bb
then.bb:
%c = tail call i64 @llvm.cttz.i64(i64 %val, i1 true)
br label %end.bb
end.bb:
%cond = phi i64 [ %c, %then.bb ], [ 64, %entry]
\code
In this example, basic block %then.bb is taken if value %val is not zero.
Also, the phi node in %end.bb would propagate the size-of in bits of %val
only if %val is equal to zero.
With this patch, CodeGenPrepare will try to hoist the call to cttz from %then.bb
into basic block %entry only if cttz is cheap to speculate for the target.
Added two new hooks in TargetLowering.h to let targets customize the behavior
(i.e. decide whether it is cheap or not to speculate calls to cttz/ctlz). The
two new methods are 'isCheapToSpeculateCtlz' and 'isCheapToSpeculateCttz'.
By default, both methods return 'false'.
On X86, method 'isCheapToSpeculateCtlz' returns true only if the target has
LZCNT. Method 'isCheapToSpeculateCttz' only returns true if the target has BMI.
Differential Revision: http://reviews.llvm.org/D6728
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224899 91177308-0d34-0410-b5e6-96231b3b80d8
Masked vector intrinsics are a part of common LLVM IR, but they are really supported on AVX2 and AVX-512 targets. I added a code that translates masked intrinsic for all other targets. The masked vector intrinsic is converted to a chain of scalar operations inside conditional basic blocks.
http://reviews.llvm.org/D6436
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224897 91177308-0d34-0410-b5e6-96231b3b80d8
Determining the address of a TLS variable results in a function call in
certain TLS models. This means that a simple ICmpInst might actually
result in invalidating the CTR register.
In such cases, do not attempt to rely on the CTR register for loop
optimization purposes.
This fixes PR22034.
Differential Revision: http://reviews.llvm.org/D6786
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224890 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
Consider the following IR:
%3 = load i8* undef
%4 = trunc i8 %3 to i1
%5 = call %jl_value_t.0* @foo(..., i1 %4, ...)
ret %jl_value_t.0* %5
Bools (that are the result of direct truncs) are lowered as whatever
the argument to the trunc was and a "and 1", causing the part of the
MBB responsible for this argument to look something like this:
%vreg8<def,tied1> = AND8ri %vreg7<kill,tied0>, 1, %EFLAGS<imp-def>; GR8:%vreg8,%vreg7
Later, when the load is lowered, it will insert
%vreg15<def> = MOV8rm %vreg14, 1, %noreg, 0, %noreg; mem:LD1[undef] GR8:%vreg15 GR64:%vreg14
but remember to (at the end of isel) replace vreg7 by vreg15. Now for
the bug. In fast isel lowering, we mistakenly mark vreg8 as the result
of the load instead of the trunc. This adds a fixup to have
vreg8 replaced by whatever the result of the load is as well, so
we end up with
%vreg15<def,tied1> = AND8ri %vreg15<kill,tied0>, 1, %EFLAGS<imp-def>; GR8:%vreg15
which is an SSA violation and causes problems later down the road.
This fixes PR21557.
Test Plan: Test test case from PR21557 is added to the test suite.
Reviewers: ributzka
Reviewed By: ributzka
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6245
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224884 91177308-0d34-0410-b5e6-96231b3b80d8
A multiply cannot unsigned wrap if there are bitwidth, or more, leading
zero bits between the two operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224849 91177308-0d34-0410-b5e6-96231b3b80d8
We already utilize this logic for reducing overflow intrinsics, it makes
sense to reuse it for normal multiplies as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224847 91177308-0d34-0410-b5e6-96231b3b80d8
When materializing constant i1 values, they must be zero extended. We represent
i1 values as [0, 1], not [0, -1], in i32 registers. As it turns out, this code
path was dead for i1 values prior to r216006 (which is why this did not manifest in
miscompiles until recently).
Fixes -O0 self-hosting on PPC64/Linux.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224842 91177308-0d34-0410-b5e6-96231b3b80d8
It's possible to have a prior definition of a symbol in module asm.
Raise an error instead of crashing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224828 91177308-0d34-0410-b5e6-96231b3b80d8
.set directives may be overridden by other .set directives as well as
label definitions.
This fixes PR22019.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224811 91177308-0d34-0410-b5e6-96231b3b80d8
Correct the line information generation for preprocessed assembly. Although we
tracked the source information for the macro instantiation, we failed to account
for the fact that we were instantiating a macro, which is populated into a new
buffer and that the line information would be relative to the definition rather
than the actual instantiation location. This could cause the line number
associated with the statement to be very high due to wrapping of the difference
calculated for the preprocessor line information emitted into the stream.
Properly calculate the line for the macro instantiation, referencing the line
where the macro is actually used as GCC/gas do.
The test case uses x86, though the same problem exists on any other target using
the LLVM IAS.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224810 91177308-0d34-0410-b5e6-96231b3b80d8
Patch by Ramkumar Ramachandra <artagnon@gmail.com>.
Also remove Llvm_executionengine.get_pointer_to_global, as it
is actually deprecated and didn't appear in a stable release.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224801 91177308-0d34-0410-b5e6-96231b3b80d8
