If the entire blocks match, we would count the branch instructions
toward the number of duplicated instructions. This doesn't match what we
do elsewhere, and was causing a bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280448 91177308-0d34-0410-b5e6-96231b3b80d8
This wasn't really well explicitly tested with a nice unittest before.
It seems good to have reasonably broken out unittests for this kind of
functionality as I'm workin go other invalidation features to make sure
none of the existing ones regress.
This still has too much duplicated code, I plan to factor that out in
a subsequent commit to use common helpers for repeated parts of this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280447 91177308-0d34-0410-b5e6-96231b3b80d8
If we failed to commit the buffer but did not die to a signal, the temp
file would remain on disk on Windows. Having an open file mapping and
file handle prevents the file from being deleted. I am choosing not to
add an assertion of success on the temp file removal, since virus
scanners and other environmental things can often cause removal to fail
in real world tools.
Also fix more temp file leaks in unit tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280445 91177308-0d34-0410-b5e6-96231b3b80d8
passes.
This simplifies the test some and makes it more focused and clear what
is being tested. It will also make it much easier to extend with further
testing of different pass behaviors.
I've also replaced a pointless module pass with running the requires
pass directly as that is all that it was really doing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280444 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This is pretty useful especially in connection with
BFI's -view-block-freq-propagation-dags. It helped me to track down the
bug that is being fixed in D24118.
While -view-block-freq-propagation-dags displays the high-level
information with static heuristics included (and block frequencies), the
new thing only shows the raw weight as presented by PGO without any of
the static estimates. This helps to distinguished what has been
measured vs. estimated.
For the sample loop in D24118, -view-block-freq-propagation-dags=integer
looks like this:
https://reviews.llvm.org/F2381352
While with -view-cfg-only you can see the underlying branch weights:
https://reviews.llvm.org/F2392296
Reviewers: dexonsmith, bogner, davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24144
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280442 91177308-0d34-0410-b5e6-96231b3b80d8
When applying our address-formation PPC64 peephole, we are reusing the @ha TOC
addis value with the low parts associated with different offsets (i.e.
different effective symbol addends). We were assuming this was okay so long as
the offsets were less than the alignment of the global variable being accessed.
This ignored the fact, however, that the TOC base pointer itself need only be
8-byte aligned. As a result, what we were doing is legal only for offsets less
than 8 regardless of the alignment of the object being accessed.
Fixes PR28727.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280441 91177308-0d34-0410-b5e6-96231b3b80d8
The logic in this function assumes that the P8 supports fusion of addis/addi,
but it does not. As a result, there is no advantage to restricting our peephole
application, merging addi instructions into dependent memory accesses, even
when the addi has multiple users, regardless of whether or not we're optimizing
for size.
We might need something like this again for the P9; I suspect we'll revisit
this code when we work on P9 tuning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280440 91177308-0d34-0410-b5e6-96231b3b80d8
Summary: This is in preparation for LoopSink pass which calls replaceDominatedUsesWith to update after sinking.
Reviewers: chandlerc, davidxl, danielcdh
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24170
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280427 91177308-0d34-0410-b5e6-96231b3b80d8
auto-brief format for doxygen comments. Most notable is switching to
that in the example doxygen comment. I've also tweaked the wording but
am happy to tweak it further if others have suggestions here.
Mostly doing this to capture something I and others have been writing
consistently and repeatedly in code reviews.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280419 91177308-0d34-0410-b5e6-96231b3b80d8
Prior to this, we could generate a vector_shuffle from an IR shuffle when the
size of the result was exactly the sum of the sizes of the input vectors.
If the output vector was narrower - e.g. a <12 x i8> being formed by a shuffle
with two <8 x i8> inputs - we would lower the shuffle to a sequence of extracts
and inserts.
Instead, we can form a larger vector_shuffle, and then extract a subvector
of the right size - e.g. shuffle the two <8 x i8> inputs into a <16 x i8>
and then extract a <12 x i8>.
This also includes a target-specific X86 combine that in the presence of
AVX2 combines:
(vector_shuffle <mask> (concat_vectors t1, undef)
(concat_vectors t2, undef))
into:
(vector_shuffle <mask> (concat_vectors t1, t2), undef)
in cases where this allows us to form VPERMD/VPERMQ.
(This is not a separate commit, as that pattern does not appear without
the DAGBuilder change.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280418 91177308-0d34-0410-b5e6-96231b3b80d8
Summary: This patch adds asm.js-style setjmp/longjmp handling support for WebAssembly. It also uses JavaScript's try and catch mechanism.
Reviewers: jpp, dschuff
Subscribers: jfb, dschuff
Differential Revision: https://reviews.llvm.org/D24121
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280415 91177308-0d34-0410-b5e6-96231b3b80d8
They're another source of generic vregs, which are going to need a type on the
definition when we remove the register width from MachineRegisterInfo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280412 91177308-0d34-0410-b5e6-96231b3b80d8
With that change, images built with 'lld-link /debug' always have a
debug directory. If no PDB filename was passed on the command line, then
the filename in the executable is empty.
PDB information would never work anyway if the PDB file name is empty,
so go ahead and try DWARF in that case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280410 91177308-0d34-0410-b5e6-96231b3b80d8
We can now maintain scalar values in VectorLoopValueMap. Thus, we no longer
have to create temporary vectors with insertelement instructions when handling
pointer induction variables. This case was mistakenly missed from r279649 when
refactoring the other scalarization code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280405 91177308-0d34-0410-b5e6-96231b3b80d8
According to spec cvtdq2pd and cvtps2pd instructions don't require memory operand to be aligned
to 16 bytes. This patch removes this requirement from the memory folding table.
Differential Revision: https://reviews.llvm.org/D23919
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280402 91177308-0d34-0410-b5e6-96231b3b80d8
My previous attempt at this connected the sub-project check targets to the test-depends target instead of to the check-all target. That resulted in the tests running multiple times on bots that built "test-depends" and "check-all" in separate build invocations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280392 91177308-0d34-0410-b5e6-96231b3b80d8
This patch moves the allocation of VectorParts for PHI nodes into the actual
PHI widening code. Previously, we allocated these VectorParts in
vectorizeBlockInLoop, and passed them by reference to widenPHIInstruction. Upon
returning, we would then map the VectorParts in VectorLoopValueMap. This
behavior is problematic for the cases where we only want to generate a scalar
version of a PHI node. For example, if in the future we only generate a scalar
version of an induction variable, we would end up inserting an empty vector
entry into the map once we return to vectorizeBlockInLoop. We now no longer
need to pass VectorParts to the various PHI widening functions, and we can keep
VectorParts allocation as close as possible to the point at which they are
actually mapped in VectorLoopValueMap.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280390 91177308-0d34-0410-b5e6-96231b3b80d8
Legalization tends to create anyext(trunc) patterns. This should always be
combined - into either a single trunc, a single ext, or nothing if the
types match exactly. But if we happen to combine the trunc first, we may pull
the trunc away from the anyext or make it implicit (e.g. the truncate(extract)
-> extract(bitcast) fold).
To prevent this, we can avoid doing the fold, similarly to how we already handle
fpround(fpextend).
Differential Revision: https://reviews.llvm.org/D23893
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280386 91177308-0d34-0410-b5e6-96231b3b80d8
Apparently nobody evaluated multiprocessing on Windows since Daniel
enabled multiprocessing on Unix in r193279. It works so far as I can
tell.
Today this is worth about an 8x speedup (631.29s to 73.25s) on my 24
core Windows machine. Hopefully this will improve Windows buildbot cycle
time, where currently it takes more time to run check-all than it does
to self-host with assertions enabled:
http://lab.llvm.org:8011/builders/clang-x86-windows-msvc2015/builds/20
build stage 2 ninja all ( 28 mins, 22 secs )
ninja check 2 stage 2 ( 37 mins, 38 secs )
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280382 91177308-0d34-0410-b5e6-96231b3b80d8
This is a partial revert of r280013. Brad King pointed out these variable names are matching CMake conventions, so we should preserve them.
I've also added a direct mapping of the LLVM_*_DIR variables which we need to make projects support building in and out of tree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280380 91177308-0d34-0410-b5e6-96231b3b80d8
Previous change broke the C API for creating an EarlyCSE pass w/
MemorySSA by adding a bool parameter to control whether MemorySSA was
used or not. This broke the OCaml bindings. Instead, change the old C
API entry point back and add a new one to request an EarlyCSE pass with
MemorySSA.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280379 91177308-0d34-0410-b5e6-96231b3b80d8
While removing a scalar shackle from an icmp fold, I noticed that I couldn't find any tests to trigger
this code path.
The 'and' shrinking transform should be handled by InstCombiner::foldCastedBitwiseLogic()
or eliminated with InstSimplify. The icmp narrowing is part of InstCombiner::foldICmpWithCastAndCast().
Differential Revision: https://reviews.llvm.org/D24031
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280370 91177308-0d34-0410-b5e6-96231b3b80d8
This was a real restriction in the original version of SinkIfThenCodeToEnd. Now it's been rewritten, the restriction can be lifted.
As part of this, we handle a very common and useful case where one of the incoming branches is actually conditional. Consider:
if (a)
x(1);
else if (b)
x(2);
This produces the following CFG:
[if]
/ \
[x(1)] [if]
| | \
| | \
| [x(2)] |
\ | /
[ end ]
[end] has two unconditional predecessor arcs and one conditional. The conditional refers to the implicit empty 'else' arc. This same pattern can also be caused by an empty default block in a switch.
We can't sink the call to x() down to end because no call to x() happens on the third incoming arc (assume that x() has sideeffects for the sake of argument; if something is safe to speculate we could indeed sink nevertheless but this cannot happen in the general case and causes many extra selects).
We are now able to detect this case and split off the unconditional arcs to a common successor:
[if]
/ \
[x(1)] [if]
| | \
| | \
| [x(2)] |
\ / |
[sink.split] |
\ /
[ end ]
Now we can sink the call to x() into %sink.split. This can cause significant code simplification in many testcases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280364 91177308-0d34-0410-b5e6-96231b3b80d8
If an attribute name has special characters such as '\01', it is not
properly printed in LLVM assembly language format. Since the format
expects the special characters are printed as it is, it has to contain
escape characters to make it printable.
Before:
attributes #0 = { ... "counting-function"="^A__gnu_mcount_nc" ...
After:
attributes #0 = { ... "counting-function"="\01__gnu_mcount_nc" ...
Reviewers: hfinkel, rengolin, rjmccall, compnerd
Subscribers: nemanjai, mcrosier, hans, shenhan, majnemer, llvm-commits
Differential Revision: https://reviews.llvm.org/D23792
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280357 91177308-0d34-0410-b5e6-96231b3b80d8
r279460 rewrote this function to be able to handle more than two incoming edges and took pains to ensure this didn't regress anything.
This time we change the logic for determining if an instruction should be sunk. Previously we used a single pass greedy algorithm - sink instructions until one requires more than one PHI node or we run out of instructions to sink.
This had the problem that sinking instructions that had non-identical but trivially the same operands needed extra logic so we sunk them aggressively. For example:
%a = load i32* %b %d = load i32* %b
%c = gep i32* %a, i32 0 %e = gep i32* %d, i32 1
Sinking %c and %e would naively require two PHI merges as %a != %d. But the loads are obviously equivalent (and maybe can't be hoisted because there is no common predecessor).
This is why we implemented the fairly complex function areValuesTriviallySame(), to look through trivial differences like this. However it's just not clever enough.
Instead, throw areValuesTriviallySame away, use pointer equality to check equivalence of operands and switch to a two-stage algorithm.
In the "scan" stage, we look at every sinkable instruction in isolation from end of block to front. If it's sinkable, we keep track of all operands that required PHI merging.
In the "sink" stage, we iteratively sink the last non-terminator in the source blocks. But when calculating how many PHIs are actually required to be inserted (to work out if we should stop or not) we remove any values that have already been sunk from the set of PHI-merges required, which allows us to be more aggressive.
This turns an algorithm with potentially recursive lookahead (looking through GEPs, casts, loads and any other instruction potentially not CSE'd) to two linear scans.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280351 91177308-0d34-0410-b5e6-96231b3b80d8
LLVM has an @llvm.eh.dwarf.cfa intrinsic, used to lower the GCC-compatible
__builtin_dwarf_cfa() builtin. As pointed out in PR26761, this is currently
broken on PowerPC (and likely on ARM as well). Currently, @llvm.eh.dwarf.cfa is
lowered using:
ADD(FRAMEADDR, FRAME_TO_ARGS_OFFSET)
where FRAME_TO_ARGS_OFFSET defaults to the constant zero. On x86,
FRAME_TO_ARGS_OFFSET is lowered to 2*SlotSize. This setup, however, does not
work for PowerPC. Because of the way that the stack layout works, the canonical
frame address is not exactly (FRAMEADDR + FRAME_TO_ARGS_OFFSET) on PowerPC
(there is a lower save-area offset as well), so it is not just a matter of
implementing FRAME_TO_ARGS_OFFSET for PowerPC (unless we redefine its
semantics -- We can do that, since it is currently used only for
@llvm.eh.dwarf.cfa lowering, but the better to directly lower the CFA construct
itself (since it can be easily represented as a fixed-offset FrameIndex)). Mips
currently does this, but by using a custom lowering for ADD that specifically
recognizes the (FRAMEADDR, FRAME_TO_ARGS_OFFSET) pattern.
This change introduces a ISD::EH_DWARF_CFA node, which by default expands using
the existing logic, but can be directly lowered by the target. Mips is updated
to use this method (which simplifies its implementation, and I suspect makes it
more robust), and updates PowerPC to do the same.
Fixes PR26761.
Differential Revision: https://reviews.llvm.org/D24038
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280350 91177308-0d34-0410-b5e6-96231b3b80d8
As discussed in https://reviews.llvm.org/D22666, our current mechanism to
support -pg profiling, where we insert calls to mcount(), or some similar
function, is fundamentally broken. We insert these calls in the frontend, which
means they get duplicated when inlining, and so the accumulated execution
counts for the inlined-into functions are wrong.
Because we don't want the presence of these functions to affect optimizaton,
they should be inserted in the backend. Here's a pass which would do just that.
The knowledge of the name of the counting function lives in the frontend, so
we're passing it here as a function attribute. Clang will be updated to use
this mechanism.
Differential Revision: https://reviews.llvm.org/D22825
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@280347 91177308-0d34-0410-b5e6-96231b3b80d8