Straight forward extension to the recently added infrastructure which was pioneered with shl. This was originally posted as part of D99687, but split off for ease of review.
(I also decided to exclude the unknown start sign case explicitly for simplicity of understanding.)
Differential Revision: https://reviews.llvm.org/D101181
SCEV does not look through non-header PHIs inside the loop. Such phis
can be analyzed by adding separate accesses for each incoming pointer
value.
This results in 2 more loops vectorized in SPEC2000/186.crafty and
avoids regressions when sinking instructions before vectorizing.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D101286
These can be handled the same way as ule/ult, just using umax
instead of umin. This is useful in cases where the umax prevents
the upper bound from overflowing.
Differential Revision: https://reviews.llvm.org/D101196
Patch by Artem Radzikhovskyy!
Symptom: ExactSIV test produced incorrect analysis of dependencies see LIT tests
Bug: At the end of the algorithm when determining dependence direction original author forgot to divide intermediate results by gcd and round result toward zero
Although this bug can be fixed with significantly fewer changes I opted to write the code in such a way that reflects the original algorithm that Banerjee proposed, for easier reference in the future. This surprisingly results in shorter code, and fewer quotient and max/min calculations.
Changes Summary:
- fixed findGCD to return valid x and y so that they match the function description where: ax - by = gcd(a,b)
- Fixed ExactSIV test, to produce proper results
- Documented the extension of Banerjee's algorithm that the original code author introduced. Banerjee's original algorithm only tested whether Dst depends on Src, the extension also allows us to test whether Src depends on Dst, in one pass.
- ExactRDIV test worked fine. Since it uses findGCD(), it needed to be updated.Since ExactRDIV test has very few changes from the core algorithm of ExactSIV I modified the test to have consistent format as ExactSIV.
- Updated the LIT tests to be testing for correct values.
Differential Revision: https://reviews.llvm.org/D100331
As a follow-up to D95982, this patch continues unblocking optimizations that are blocked by pseudu probe instrumention.
The optimizations unblocked are:
- In-block load propagation.
- In-block dead store elimination
- Memory copy optimization that turns stores to consecutive memories into a memset.
These optimizations are local to a block, so they shouldn't affect the profile quality.
Reviewed By: wmi
Differential Revision: https://reviews.llvm.org/D100075
This patch adds the missing implementation for
TargetTransformInfo::hasActiveVectorLength() without which using
hasActiveVectorLength() causes linker error.
Patch by Vineet Kumar!
Differential Revision: https://reviews.llvm.org/D100941
ICMP_NE predicates directly overwrote the rewritten result,
instead of chaining it with previous rewrites, as was done for
ICMP_ULT and ICMP_ULE. This means that some guards were effectively
discarded, depending on their order.
The change adds support for triming and merging cold context when mergine CSSPGO profiles using llvm-profdata. This is similar to the context profile trimming in llvm-profgen, however the flexibility to trim cold context after profile is generated can be useful.
Differential Revision: https://reviews.llvm.org/D100528
CommandLine.h is indirectly included in ~50% of TUs when building
clang, and VirtualFileSystem.h is large.
(Already remarked by jhenderson on D70769.)
No behavior change.
Differential Revision: https://reviews.llvm.org/D100957
This reverts commit ea1a0d7c9a.
While this is strictly more powerful, it is also strictly slower.
InstSimplify intentionally does not perform many folds that it
is allowed to perform, if doing so requires a KnownBits calculation
that will be repeated in InstCombine.
Maybe it's worthwhile to do this here, but that needs a more
explicitly stated motivation, evaluated in a review.
It used to be that all of our intrinsics were call instructions, but over time, we've added more and more invokable intrinsics. According to the verifier, we're up to 8 right now. As IntrinsicInst is a sub-class of CallInst, this puts us in an awkward spot where the idiomatic means to check for intrinsic has a false negative if the intrinsic is invoked.
This change switches IntrinsicInst from being a sub-class of CallInst to being a subclass of CallBase. This allows invoked intrinsics to be instances of IntrinsicInst, at the cost of requiring a few more casts to CallInst in places where the intrinsic really is known to be a call, not an invoke.
After this lands and has baked for a couple days, planned cleanups:
Make GCStatepointInst a IntrinsicInst subclass.
Merge intrinsic handling in InstCombine and use idiomatic visitIntrinsicInst entry point for InstVisitor.
Do the same in SelectionDAG.
Do the same in FastISEL.
Differential Revision: https://reviews.llvm.org/D99976
We already special-cased a few interesting patterns,
but that is strictly less powerful than using KnownBits.
So instead get the known bits for the operand of `and`,
and iff all the unset bits of the `and`-mask are known to be zeros
in the operand, we can omit said `and`.
I'd reverted this in commit 3b6acb1797 due to buildbot failures. This patch contains the fix for said issue. I'd forgotten to handle the case where two phis in the same block have different operand order. We canonicalize away from this, but it's still valid IR. The tests included in this change (as opposed to simply having test output changed), crashed without the fix.
Original commit message follows...
This extends the phi handling in isKnownNonEqual with a special case based on invertible recurrences. If we can prove the recurrence is invertible (which many common ones are), we can recurse through the start operands of the recurrence skipping the phi cycle.
(Side note: Instcombine currently does not push back through these cases. I will implement that in a follow up change w/separate review.)
Differential Revision: https://reviews.llvm.org/D99912
Adding the switches to reduce diffs. I'm about to split that into an lshr part and an ashr part, doing the NFC part first makes it easier to maintain both diffs.
This extends the phi handling in isKnownNonEqual with a special case based on invertible recurrences. If we can prove the recurrence is invertible (which many common ones are), we can recurse through the start operands of the recurrence skipping the phi cycle.
(Side note: Instcombine currently does not push back through these cases. I will implement that in a follow up change w/separate review.)
Differential Revision: https://reviews.llvm.org/D99912
Previously we would use the type of the pointee to determine what to
cast the result of constant folding a load. To aid with opaque pointer
types, we should explicitly pass the type of the load rather than
looking at pointee types.
ConstantFoldLoadThroughBitcast() converts the const prop'd value to the
proper load type (e.g. [1 x i32] -> i32). Instead of calling this in
every intermediate step like bitcasts, we only call this when we
actually see the global initializer value.
In some existing uses of this API, we don't know the exact type we're
loading from immediately (e.g. first we visit a bitcast, then we visit
the load using the bitcast). In those cases we have to manually call
ConstantFoldLoadThroughBitcast() when simplifying the load to make sure
that we cast to the proper type.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D100718
Move <string> include to ImportedFunctionsInliningStatistics.cpp and add missing <memory> include as we have explicit uses of std::unique_ptr in the header.
As being discussed in https://reviews.llvm.org/D100721,
this modelling is lossy, we can't reconstruct `ash`/`ashr exact`
from it, which means that whenever we actually expand the IR,
we've just pessimized the code..
It would be good to model this pattern, after all it comes up every time
you want to compute a distance between two pointers, but not at this cost.
This reverts commit ec54867df5.
Such attributes can either be unset, or set to "true" or "false" (as string).
throughout the codebase, this led to inelegant checks ranging from
if (Fn->getFnAttribute("no-jump-tables").getValueAsString() == "true")
to
if (Fn->hasAttribute("no-jump-tables") && Fn->getFnAttribute("no-jump-tables").getValueAsString() == "true")
Introduce a getValueAsBool that normalize the check, with the following
behavior:
no attributes or attribute set to "false" => return false
attribute set to "true" => return true
Differential Revision: https://reviews.llvm.org/D99299
Use the target-independent @llvm.fptosi and @llvm.fptoui intrinsics instead.
This includes removing the instrinsics for i32x4.trunc_sat_zero_f64x2_{s,u},
which are now represented in IR as a saturating truncation to a v2i32 followed by
a concatenation with a zero vector.
Differential Revision: https://reviews.llvm.org/D100596
This is an alternative to D99759 to avoid the compile-time explosion seen in:
https://llvm.org/PR49785
Another potential solution would make the exclusion logic stronger to avoid
blowing up, but note that we reduced the complexity of the exclusion mechanism
in D16204 because it was too costly.
So I'm questioning the need for recursion/exclusion entirely - what is the
optimization value vs. cost of recursively computing known bits based on
assumptions?
This was built into the implementation from the start with 60db058,
and we have kept adding code/cost to deal with that capability.
By clearing the query's AssumptionCache inside computeKnownBitsFromAssume(),
this patch retains all existing assume functionality except refining known
bits based on even more assumptions.
We have 1 regression test that shows a difference in optimization power.
Differential Revision: https://reviews.llvm.org/D100573
Being lazy with printing the banner seems hard to reason with, we should print it
unconditionally first (it could also lead to duplicate banners if we
have multiple functions in -filter-print-funcs).
The printIR() functions were doing too many things. I separated out the
call from PrintPassInstrumentation since we were essentially doing two
completely separate things in printIR() from different callers.
There were multiple ways to generate the name of some IR. That's all
been moved to getIRName(). The printing of the IR name was also
inconsistent, now it's always "IR Dump on $foo" where "$foo" is the
name. For a function, it's the function name. For a loop, it's what's
printed by Loop::print(), which is more detailed. For an SCC, it's the
list of functions in parentheses. For a module it's "[module]", to
differentiate between a possible SCC with a function called "module".
To preserve D74814, we have to check if we're going to print anything at
all first. This is unfortunate, but I would consider this a special
case that shouldn't be handled in the core logic.
Reviewed By: jamieschmeiser
Differential Revision: https://reviews.llvm.org/D100231
SROA shifts TBAA nodes in a way that may present a problem for !tbaa but not !tbaa.struct nodes.
Differential Revision: https://reviews.llvm.org/D99851
Just like in the mul nuw case, it's sufficient that the step is
non-zero. If the step is negative, then the values will jump
between positive and negative, "crossing" zero, but the value of
the recurrence is never actually zero.
It's okay if the step is zero, we'll just stay at the same non-zero
value in that case. The valuable part of this is that the step
doesn't even need to be a constant anymore.
Stepping through callstacks in the example from D99759 reveals
this potential compile-time improvement.
The savings come from avoiding ValueTracking's computing known
bits if we have already dealt with special-case patterns.
Further improvements in this direction seem possible.
This makes a degenerate test based on PR49785 about 40x faster
(25 sec -> 0.6 sec), but it does not address the larger question
of how to limit computeKnownBitsFromAssume(). Ie, the original
test there is still infinite-time for all practical purposes.
Differential Revision: https://reviews.llvm.org/D100408
The start value can't be null for something to be a non-zero
recurrence, so hoist that common check out of the switch.
Subsequent checks may be incomplete or over-specified as noted in:
D100408
I've run into some cases where a large fraction of compile-time is
spent invalidating SCEV. One of the causes is forgetLoop(), which
walks all values that are def-use reachable from the loop header
phis. When invalidating a topmost loop, that might be close to all
values in a function. Additionally, it's fairly common for there to
not actually be anything to invalidate, but we'll still be performing
this walk again and again.
My first thought was that we don't need to continue walking the uses
if the current value doesn't have a SCEV expression. However, this
isn't quite right, because SCEV construction can skip over values
(e.g. for a chain of adds, we might only create a SCEV expression
for the final value).
What this patch does instead is to only walk the (full) def-use chain
of loop phis that have a SCEV expression. If there's no expression
for a phi, then we also don't have any dependent expressions to
invalidate.
Differential Revision: https://reviews.llvm.org/D100264
Lookup tables generate non PIC-friendly code, which requires dynamic relocation as described in:
https://bugs.llvm.org/show_bug.cgi?id=45244
This patch adds a new pass that converts lookup tables to relative lookup tables to make them PIC-friendly.
Differential Revision: https://reviews.llvm.org/D94355
SROA can handle invariant group intrinsics, let the inliner know that
for better heuristics when the intrinsics are present.
This fixes size issues in a couple files when turning on
-fstrict-vtable-pointers in Chrome.
Reviewed By: rnk, mtrofin
Differential Revision: https://reviews.llvm.org/D100249
This patch removes all uses of `std::iterator`, which was deprecated in C++17.
While this isn't currently an issue while compiling LLVM, it's useful for those using LLVM as a library.
For some reason there're a few places that were seemingly able to use `std` functions unqualified, which no longer works after this patch. I've updated those places, but I'm not really sure why it worked in the first place.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D67586
D24453 enabled libcalls simplication for ARM PCS. This may cause
caller/callee calling conventions mismatch in some situations such as
LTO. This patch makes instcombine aware that the compatible calling
conventions differences are benign (not emitting undef idom).
Differential Revision: https://reviews.llvm.org/D99773
While, indeed, we may end up pushing less updates that we'd reserve space
for, self-dominating updates aren't often enough for that to matter.
But this should matter for normal updates.
Check the cache before calling isLoopSimplifyForm(). Otherwise we'd
always perform the check for the innermost loop and only skip it
for dominating loops.
"Does the predicate hold between two ranges?"
Not very surprisingly, some places were already doing this check,
without explicitly naming the algorithm, cleanup them all.
"Does the predicate hold between two ranges?"
Not very surprisingly, some places were already doing this check,
without explicitly naming the algorithm, cleanup them all.
Added cost estimation for switch instruction, updated costs of branches, fixed
phi cost.
Had to increase `-amdgpu-unroll-threshold-if` default value since conditional
branch cost (size) was corrected to higher value.
Test renamed to "control-flow.ll".
Removed redundant code in `X86TTIImpl::getCFInstrCost()` and
`PPCTTIImpl::getCFInstrCost()`.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D96805
When inserting a new def and renaming of uses is asked, always compute
IDF and do the renaming for the blocks with Phis in that IDF.
Resolves PR49859.
Differential Revision: https://reviews.llvm.org/D100163
In LazyValueInfoImpl::isNonNullAtEndOfBlock we populate a set of
pointers, known to be non-null at the end of a block (e.g. because we
did a load through them). We then infer that any pointer, based on an
element of this set is non-null as well ("based" here meaning a
non-null pointer is the underlying object). This is incorrect, even if
the base pointer was non-null, the value of a GEP, that lacks the
inbounds` attribute, may be null.
This issue appeared as miscompilation of the following test case:
int puts(const char *);
typedef struct iter {
int *val;
} iter_t;
static long distance(iter_t first, iter_t last) {
long r = 0;
for (; first.val != last.val; first.val++)
++r;
return r;
}
int main() {
int arr[2] = {0};
iter_t i, j;
i.val = arr;
j.val = arr + 1;
if (distance(i, j) >= 2)
puts("failed");
else
puts("passed");
}
This fixes PR49662.
Differential Revision: https://reviews.llvm.org/D99642
Add an ability to store `Offset` between partially aliased location. Use this
storage within returned `ResultAlias` instead of caching it in `AAQueryInfo`.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D98718
Main reason is preparation to transform AliasResult to class that contains
offset for PartialAlias case.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D98027
This fixes a "Cached first special instruction is wrong!" assert.
The assert fires because replacing a value with another can cause an
instruction to no longer be "special" to ICF. In this case,
devirtualization happened, turning an indirect call to a
call to a willreturn function which is no longer special.
Reviewed By: nikic, rnk
Differential Revision: https://reviews.llvm.org/D99977
clang++ uses llvm.compiler.used in certain cases to preserve
symbol which is fully inlined. D96087 has resulted in undefined
symbols in such cases. Set it to false by default to preserve
old behavior but keep the option for specific uses where we
want to ignore these (e.g. to detect a potential indirect call
to a function).
Differential Revision: https://reviews.llvm.org/D99897
A value from reachable block may come to a Phi node as its input from
unreachable block. This may confuse matchSimpleRecurrence which
has no access to DomTree and can falsely recognize something as a recurrency
because of this effect, as the attached test shows.
Patch `ae7b1e` deals with half of this problem, but it only accounts from
the case when an unreachable instruction comes to Phi as an input.
This patch provides a generalization by checking that no Phi block's
predecessor is unreachable (no matter what the input is).
Differential Revision: https://reviews.llvm.org/D99929
Reviewed By: reames
Follow up to a6d2a8d6f5. These were found by simply grepping for "::assume", and are the subset of that result which looked cleaner to me using the isa/dyn_cast patterns.
Follow up to a6d2a8d6f5. This covers all the public interfaces of the bundle related code. I tried to cleanup the internals where the changes were obvious, but there's definitely more room for improvement.
Add the subclass, update a few places which check for the intrinsic to use idiomatic dyn_cast, and update the public interface of AssumptionCache to use the new class. A follow up change will do the same for the newer assumption query/bundle mechanisms.
The current code does not properly handle vector indices unless they are
the first index.
At the moment LangRef gives the impression that the vector index must be
the one and only index (https://llvm.org/docs/LangRef.html#getelementptr-instruction).
But vector indices can appear at any position and according to the
verifier there may be multiple vector indices. If that's the case, the
number of elements must match.
This patch updates SimplifyGEPInst to properly handle those additional
cases.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D99961
Previously we could only vectorize FP reductions if fast math was enabled, as this allows us to
reorder FP operations. However, it may still be beneficial to vectorize the loop by moving
the reduction inside the vectorized loop and making sure that the scalar reduction value
be an input to the horizontal reduction, e.g:
%phi = phi float [ 0.0, %entry ], [ %reduction, %vector_body ]
%load = load <8 x float>
%reduction = call float @llvm.vector.reduce.fadd.v8f32(float %phi, <8 x float> %load)
This patch adds a new flag (IsOrdered) to RecurrenceDescriptor and makes use of the changes added
by D75069 as much as possible, which already teaches the vectorizer about in-loop reductions.
For now in-order reduction support is off by default and controlled with the `-enable-strict-reductions` flag.
Reviewed By: david-arm
Differential Revision: https://reviews.llvm.org/D98435
Changes getRecurrenceIdentity to always return a neutral value of -0.0 for FAdd.
Reviewed By: dmgreen, spatel
Differential Revision: https://reviews.llvm.org/D98963
This is the sibling fix to c590a9880d -
as there, we can't subsitute a vector value the equality
compare replacement that we are trying requires that the
comparison is true for the entire value. Vector select
can be partly true/false.
For use in an uncoming patch. Left out the phi case (which could otherwise fit in this framework) as it would cause infinite recursion in said patch. We can probably also leverage this in instcombine to ensure we keep the two sets of related analysis and transforms in sync.
Even if one of the operands is overdefined, we may still produce
a non-overdefined result, e.g. due to a min/max operation. This
matches our handling elsewhere, e.g. for binary operators.
The slot poisoning comment refers to a much older LVI cache
implementation.
This was prompted by D95727, which had the side-effect to break the
'release' mode build bot for ML-driven policies. The problem is that now
the pre-compiled object files don't get transitively carried through as
'source' anymore; that being said, the previous way of consuming them
was problematic, because it was only working for static builds; in
dynamic builds, the whole tf_xla_runtime was linked, which is
undesirable.
The alternative is to treat tf_xla_runtime as an archive, which then
leads to the desired effect.
Differential Revision: https://reviews.llvm.org/D99829
InstCombine performs simple forwarding from stores to loads, but
currently only handles the case where the load and store have the
same size. This extends it to also handle a store of a constant
with a larger size followed by a load with a smaller size.
This is implemented through ConstantFoldLoadThroughBitcast() which
is fairly primitive (e.g. does not allow storing a large integer
and then loading a small one), but at least can forward the first
element of a vector store. Unfortunately it seems that we currently
don't have a generic helper for "read a constant value as a different
type", it's all tangled up with other logic in either
ConstantFolding or VNCoercion.
Differential Revision: https://reviews.llvm.org/D98114
The AAMDNodes part of the MemoryLocation is not used by the BasicAA
cache, so don't store it. This reduces the size of each cache entry
from 112 bytes to 48 bytes.
BasicAA itself doesn't make use of AA metadata, but passes it
through to recursive queries and makes it part of the cache key.
Aliasing decisions that are based on AA metadata (i.e. TBAA and
ScopedAA) are based *only* on AA metadata, so checking them with
different pointer values or sizes is not useful, the result will
always be the same.
While this change is a mild compile-time improvement by itself,
the actual goal here is to reduce the size of AA cache keys in
a followup change.
Differential Revision: https://reviews.llvm.org/D90098
In order to bring up scalable vector support in LLVM incrementally,
we introduced behaviour to emit a warning, instead of an error, when
asking the wrong question of a scalable vector, like asking for the
fixed number of elements.
This patch puts that behaviour under a flag. The default behaviour is
that the compiler will always error, which means that all LLVM unit
tests and regression tests will now fail when a code-path is taken that
still uses the wrong interface.
The behaviour to demote an error to a warning can be individually enabled
for tools that want to support experimental use of scalable vectors.
This patch enables that behaviour when driving compilation from Clang.
This means that for users who want to try out scalable-vector support,
fixed-width codegen support, or build user-code with scalable vector
intrinsics, Clang will not crash and burn when the compiler encounters
such a case.
This allows us to do away with the following pattern in many of the SVE tests:
RUN: .... 2>%t
RUN: cat %t | FileCheck --check-prefix=WARN
WARN-NOT: warning: ...
The behaviour to emit warnings is only temporary and we expect this flag
to be removed in the future when scalable vector support is more stable.
This patch also has fixes the following tests:
unittests:
ScalableVectorMVTsTest.SizeQueries
SelectionDAGAddressAnalysisTest.unknownSizeFrameObjects
AArch64SelectionDAGTest.computeKnownBitsSVE_ZERO_EXTEND_VECTOR_INREG
regression tests:
Transforms/InstCombine/vscale_gep.ll
Reviewed By: paulwalker-arm, ctetreau
Differential Revision: https://reviews.llvm.org/D98856
Hookup TLI when inferring object size from allocation calls. This allows the analysis to prove dereferenceability for known allocation functions (such as malloc/new/etc) in addition to those marked explicitly with the allocsize attribute.
This is a follow up to 0129cd5 now that the bug fixed by e2c6621e6 is resolved.
As noted in the test, this relies on being able to prove that there is no free between allocation and context (e.g. hoist location). At the moment, this is handled conservatively. I'm working strengthening out ability to reason about no-free regions separately.
Differential Revision: https://reviews.llvm.org/D99737
We have this logic duplicated in several cases, none of which were exhaustive. Consolidate it in one place.
I don't believe this actually impacts behavior of the callers. I think they all filter their inputs such that their partial implementations were correct. If not, this might be fixing a cornercase bug.
Support deriving dereferenceability facts from allocation sites with known object sizes while correctly accounting for any possibly frees between allocation and use site. (At the moment, we're conservative and only allowing it in functions where we know we can't free.)
This is part of the work on deref-at-point semantics. I'm making the change unconditional as the miscompile in this case is way too easy to trip by accident, and the optimization was only recently added (by me).
There will be a follow up patch wiring through TLI since that should now be doable without introducing widespread miscompiles.
Differential Revision: https://reviews.llvm.org/D95815
This is a patch teaching ValueTracking that `s/u*.with.overflow` intrinsics do not
create undef/poison and they propagate poison.
I couldn't write a nice example like the one with ctpop; ValueTrackingTest.cpp were simply updated
to check these instead.
This patch helps reducing regression while fixing https://llvm.org/pr49688 .
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D99671
This fixes an issue introduced with my change d4648e, and reported in pr49768.
The root problem is that dominance collapses in unreachable code, and that LoopInfo explicitly only models reachable code. Since the recurrence matcher doesn't filter by reachability (and can't easily because not all consumers have domtree), we need to bailout before assuming that finding a recurrence implies we found a loop.
Attempts to compute savings more accurately cannot impact the set of critically important call sites.
Reviewed By: kazu
Differential Revision: https://reviews.llvm.org/D98577
We previously made a change to getUserCost to return a Invalid cost
when one of the TTI costs returned '-1' (meaning 'unknown' or
'infinitely expensive'). It makes no sense to say that:
shufflevector <2 x i8> %x, <2 x i8> %y, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
has an invalid cost. Perhaps the cost is not known, but the IR is valid
and can be code-generated. Invalid should only be used for IR that
cannot possibly be code-generated and where a cost is nonsensical.
With more passes now asserting that the cost must be valid, it is possible
that those assertions will fail for perfectly valid IR. An incomplete
cost-model probably shouldn't be a reason for the compiler to break.
It's better to consider these costs as 'very expensive' and ignore them
for other reasons. At some point, we should consider replacing -1 with
some other mechanism.
Reviewed By: paulwalker-arm, dmgreen
Differential Revision: https://reviews.llvm.org/D99502
Lookup tables generate non PIC-friendly code, which requires dynamic relocation as described in:
https://bugs.llvm.org/show_bug.cgi?id=45244
This patch adds a new pass that converts lookup tables to relative lookup tables to make them PIC-friendly.
Differential Revision: https://reviews.llvm.org/D94355
This can only happen if offset types that are larger than the
pointer size are involved. The previous implementation did not
assert in this case because it initialized the APInts to the
width of one of the variables -- though I strongly suspect it
did not compute correct results in this case.
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=32621
reported by fhahn.
If the sizes of both memory locations are unknown, we can only
perform a check on the underlying objects. There's no point in
going through GEP decomposition in this case.
The current linear expression decomposition handles zext/sext by
decomposing the casted operand, and then checking NUW/NSW flags
to determine whether the extension can be distributed. This has
some disadvantages:
First, it is not possible to perform a partial decomposition. If
we have zext((x + C1) +<nuw> C2) then we will fail to decompose
the expression entirely, even though it would be safe and
profitable to decompose it to zext(x + C1) +<nuw> zext(C2)
Second, we may end up performing unnecessary decompositions,
which will later be discarded because they lack nowrap flags
necessary for extensions.
Third, correctness of the code is not entirely obvious: At a high
level, we encounter zext(x -<nuw> C) in the form of a zext on the
linear expression x + (-C) with nuw flag set. Notably, this case
must be treated as zext(x) + -zext(C) rather than zext(x) + zext(-C).
The code handles this correctly by speculatively zexting constants
to the final bitwidth, and performing additional fixup if the
actual extension turns out to be an sext. This was not immediately
obvious to me.
This patch inverts the approach: An ExtendedValue represents a
zext(sext(V)), and linear expression decomposition will try to
decompose V further, either by absorbing another sext/zext into the
ExtendedValue, or by distributing zext(sext(x op C)) over a binary
operator with appropriate nsw/nuw flags. At each step we can
determine whether distribution is legal and abort with a partial
decomposition if not. We also know which extensions we need to
apply to constants, and don't need to speculate or fixup.
While explicit sext instructions were handled correctly, the
implicit sext that occurs if the offset is smaller than the
pointer size blindly assumed that sext(X * Scale + Offset) is the
same as sext(X) * Scale + Offset, which is obviously not correct.
Fix this by extracting the code that handles linear expression
extension and reusing it for the implicit sext as well.
A number of variables need to be correctly initialized on entry
to GetLinearExpression() for the implementation to behave reasonably.
The fact that SExtBits can currenlty be non-zero on entry is a bug,
as demonstrated by the added test: For implicit sexts by the GEP,
we do currently skip legality checks.
Currently, we'd produce an incorrect decomposition, because we
already recursively called GetLinearExpression(), so the Scale=1,
Offset=0 will not necessarily be relative to the shl itself.
Now, this doesn't actually matter for functional correctness,
because such a shift is poison anyway, so its okay to return
an incorrect decomposition. It's still unnecessarily confusing
though, and we can easily avoid this by checking the bitwidth
earlier.
Nowrap flags between mul and shl differ in that mul nsw allows
multiplication of 1 * INT_MIN, while shl nsw does not. This means
that it is always fine to transfer shl nowrap flags to muls, but
not necessarily the other way around. In this case the NUW/NSW
results refer to mul/add operations, so it's fine to retain the
flags from the shl.
Handle (x << s) != (y << s) where x != y and the shifts are
non-wrapping. Once again, this establishes parity with the
corresponing mul fold that already exists. The shift case is
more powerful because we don't need to guard against multiplies
by zero.
This handles the pattern X != X << C for non-zero X and C and a
non-overflowing shift. This establishes parity with the corresponing
fold for multiplies.
This is mainly for clarity: It doesn't make sense to do any
negative/positive checks when dealing with a nuw add/mul. These
only make sense to nsw add/mul.
This patch enables the cost-benefit-analysis-based inliner by default
if we have instrumentation profile.
- SPEC CPU 2017 shows a 0.4% improvement.
- An internal large benchmark shows a 0.9% reduction in the cycle
count along with 14.6% reduction in the number of call instructions
executed.
Differential Revision: https://reviews.llvm.org/D98213
getPointersDiff would previously round down the difference between two
pointers to a multiple of the element size of the pointee, which could
result in a pointer value being decreased a little.
Alexey Bataev has graciously agreed to add a testcase for this;
submitting the bugfix now to unblock.
loop:
%cmp.0 = phi i32 [ 3, %entry ], [ %inc, %loop ]
%pos.0 = phi i32 [ 1, %entry ], [ %cmp.0, %loop ]
...
%inc = add i32 %cmp.0, 1
br label %loop
On above example, %pos.0 uses previous iteration's %cmp.0 with backedge
according to PHI's instruction's defintion. If the %inc is not same among
iterations, we can say the two PHIs are not same.
Differential Revision: https://reviews.llvm.org/D98422
Rather than special-casing assume in BasicAA getModRefBehavior(),
do this one level higher, in the attribute handling of CallBase.
For assumes with operand bundles, the inaccessiblememonly attribute
applies regardless of operand bundles.
With cost-benefit analysis for inlining, we bypass the cost-threshold by returning inline result from call analyzer early.
However the cost and threshold are still available from call analyzer, and when cost is actually higher than threshold, we incorrect set the reason.
The change makes the decision from cost-benefit analysis explicit. It's mostly NFC, except that it allows the priority-based sample loader inliner used by CSSPGO to use cost-benefit heuristic.
Differential Revision: https://reviews.llvm.org/D99302
This patch enables the cost-benefit-analysis-based inliner by default
if we have instrumentation profile.
- SPEC CPU 2017 shows a 0.4% improvement.
- An internal large benchmark shows a 0.9% reduction in the cycle
count along with 14.6% reduction in the number of call instructions
executed.
Differential Revision: https://reviews.llvm.org/D98213
This is similar to the select logic just ahead of the new code.
Min/max choose exactly one value from the inputs, so if both of
those are a power-of-2, then the result must be a power-of-2.
This might help with D98152, but we likely still need other
pieces of the puzzle to avoid regressions.
The change in PatternMatch.h is needed to build with clang.
It's possible there is a better way to deal with the 'const'
incompatibities.
Differential Revision: https://reviews.llvm.org/D99276
SCEV currently tries to prove implications of x pred y by also
trying to imply ~y pred ~x. This is expensive in terms of
compile-time (in fact, the majority of isImpliedCond compile-time
is spent here) and generally not fruitful. The issue is that this
also swaps the operands and thus breaks canonical ordering. If
originally we were trying to prove an implication like
X > C1 -> Y > C2, then we'll now try to prove X > C1 -> C3 > ~Y,
which will not work.
The only real case where we can get some use out of this transform
is if the original conditions were in the form X > C1 -> Y < C2, were
then swapped to X > C1 -> C2 > Y and are then swapped again here to
X > C1 -> ~Y > C3.
As such, handle this at a higher level, where we are doing the
swapping in the first place. There's four different ways that we
can line up a predicate and a swapped predicate, so we use some
heuristics to pick some profitable way.
Because we now try this transform at a higher level
(isImpliedCondOperands rather than isImpliedCondOperandsHelper),
we can also prove additional facts. Of the added tests, one was
proven previously while the other wasn't.
Differential Revision: https://reviews.llvm.org/D90926
Lookup tables generate non PIC-friendly code, which requires dynamic relocation as described in:
https://bugs.llvm.org/show_bug.cgi?id=45244
This patch adds a new pass that converts lookup tables to relative lookup tables to make them PIC-friendly.
Differential Revision: https://reviews.llvm.org/D94355
FindAvailableLoadedValue() relies on FindAvailablePtrLoadStore() to run
the alias analysis when searching for an equivalent value. However,
FindAvailablePtrLoadStore() calls the alias analysis framework with a
memory location for the load constructed from an address and a size,
which thus lacks TBAA metadata info. This commit modifies
FindAvailablePtrLoadStore() to accept an optional memory location as
parameter to allow FindAvailableLoadedValue() to create it based on the
load instruction, which would then have TBAA metadata info attached.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D99206
This patch changes the interface to take a RegisterKind, to indicate
whether the register bitwidth of a scalar register, fixed-width vector
register, or scalable vector register must be returned.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D98874
Philip Reames