When determining if an `and` should be merged into an extending load
the constant argument to the `and` is currently not checked if the
argument requires truncation. This prevents the combine happening when
the vector width is half the normal available vector width for SVE VLA
vectors.
Reviewed By: c-rhodes
Differential Revision: https://reviews.llvm.org/D129281
Unlike the name suggests this can reuse any store as a base for a
memory-based vector extract. If that store is underaligned the loads
created to extract will have an invalid alignment. Since most CPUs are
forgiving wrt alignment this is almost never an issue, on x86 this is
only reproducible by extracting a 128 bit vector out of a wider vector.
I tried making a test case in the context of
https://reviews.llvm.org/D127982 but it's really really fragile, as the
output pretty much looks like a missed optimization.
The "xor (X >> ShiftC), XorC --> (not X) >> ShiftC" fold is currently limited to the XOR mask being a shifted all-bits mask, but we can relax this to only need to match under the demanded bits.
This helps expose more bit extraction/clearing patterns and fixes the PowerPC testCompares*.ll regressions from D127115
Alive2: https://alive2.llvm.org/ce/z/fl7T7K
Differential Revision: https://reviews.llvm.org/D129933
This revision supports to scalarize a binary operation of two scalable splat vectors.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D122791
This was stored in LiveIntervals, but not actually used for anything
related to LiveIntervals. It was only used in one check for if a load
instruction is rematerializable. I also don't think this was entirely
correct, since it was implicitly assuming constant loads are also
dereferenceable.
Remove this and rely only on the invariant+dereferenceable flags in
the memory operand. Set the flag based on the AA query upfront. This
should have the same net benefit, but has the possible disadvantage of
making this AA query nonlazy.
Preserve the behavior of assuming pointsToConstantMemory implying
dereferenceable for now, but maybe this should be changed.
r175673 changed repairIntervalsInRange to find anchoring end points for
ranges automatically, but the calculation of Begin included the first
instruction found that already had an index. This patch changes it to
exclude that instruction:
1. For symmetry, so that the half open range [Begin,End) only includes
instructions that do not already have indexes.
2. As a possible performance improvement, since repairOldRegInRange
will scan fewer instructions.
3. Because repairOldRegInRange hits assertion failures in some cases
when it sees a def that already has a live interval.
(3) fixes about ten tests in the CodeGen lit test suite when
-early-live-intervals is forced on.
Differential Revision: https://reviews.llvm.org/D110182
The DAG Combiner unnecessarily restricts commutative CSE
to nodes with a single result value. This commit removes
that restriction.
Signed-off-by: Itay Bookstein <ibookstein@gmail.com>
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D129666
Undef tokens may appear in unreached code as result of RAUW of some optimization,
and it should not be considered as bad IR.
Patch by Dmitry Bakunevich!
Differential Revision: https://reviews.llvm.org/D128904
Reviewed By: mkazantsev
Added function to the ExpandVectorPredication pass to handle VP loads
and stores.
Reviewed By: simoll
Differential Revision: https://reviews.llvm.org/D109584
trunc (sign_ext_inreg X, iM) to iN --> sign_ext_inreg (trunc X to iN), iM
There are improvements on existing tests from this, and there are a pair
of large regressions in D127115 for Thumb2 caused by not folding this
pattern.
Differential Revision: https://reviews.llvm.org/D129890
D127595 added the ability to recurse up a (one-use) INSERT_VECTOR_ELT chain to create a BUILD_VECTOR before other combines manage to break the chain, something that is particularly bad in D127115.
The patch generalises this so it doesn't have to build the chain starting from the last element insertion, instead it can now start from any insertion and will recurse up the chain until it finds all elements or finds a UNDEF/BUILD_VECTOR/SCALAR_TO_VECTOR which represents that start of the chain.
Fixes several regressions in D127115
In https://reviews.llvm.org/D30114, support for mismatching address
spaces was introduced to CodeGenPrepare's optimizeMemoryInst, using
addrspacecast as it was argued that only no-op addrspacecasts would be
considered when constructing the address mode. However, by doing
inttoptr/ptrtoint, it's possible to get CGP to emit an addrspace
that's not actually no-op, introducing a miscompilation:
define void @kernel(i8* %julia_ptr) {
%intptr = ptrtoint i8* %julia_ptr to i64
%ptr = inttoptr i64 %intptr to i32 addrspace(3)*
br label %end
end:
store atomic i32 1, i32 addrspace(3)* %ptr unordered, align 4
ret void
}
Gets compiled to:
define void @kernel(i8* %julia_ptr) {
end:
%0 = addrspacecast i8* %julia_ptr to i32 addrspace(3)*
store atomic i32 1, i32 addrspace(3)* %0 unordered, align 4
ret void
}
In the case of NVPTX, this introduces a cvta.to.shared, whereas
leaving out the %end block and branch doesn't trigger this
optimization. This results in illegal memory accesses as seen in
https://github.com/JuliaGPU/CUDA.jl/issues/558
In this change, I introduced a check before doing the pointer cast
that verifies address spaces are the same. If not, it emits a
ptrtoint/inttoptr combination to get a no-op cast between address
spaces. I decided against disallowing ptrtoint/inttoptr with
non-default AS in matchOperationAddr, because now its still possible
to look through multiple sequences of them that ultimately do not
result in a address space mismatch (i.e. the second lit test).
As mentioned on D127115, this patch that attempts to recognise shuffle masks that could be simplified to a AND mask - we already have a similar transform that will fold AND -> 'clear mask' shuffle, but this patch handles cases where the referenced elements are not from the same lane indices but are known to be zero.
Differential Revision: https://reviews.llvm.org/D129150
combineShiftAnd1ToBitTest already matches "and (not (srl X, C)), 1 --> (and X, 1<<C) == 0" patterns, but we can end up with situations where the not is before the shift.
Part of some yak shaving for D127115 to generalise the "xor (X >> ShiftC), XorC --> (not X) >> ShiftC" fold.
SimplifyDemandedBits is called slightly later which allows the not(sext(x)) -> sext(not(x)) fold to occur via foldLogicOfShifts
As mentioned on D127115, we should be able to further generalise this based off the demanded bits.
This is similar to D125680, but for llvm.experimental.patchpoint
(instead of llvm.experimental.stackmap).
Differential review: https://reviews.llvm.org/D129268
Following some recent discussions, this changes the representation
of callbrs in IR. The current blockaddress arguments are replaced
with `!` label constraints that refer directly to callbr indirect
destinations:
; Before:
%res = callbr i8* asm "", "=r,r,i"(i8* %x, i8* blockaddress(@test8, %foo))
to label %asm.fallthrough [label %foo]
; After:
%res = callbr i8* asm "", "=r,r,!i"(i8* %x)
to label %asm.fallthrough [label %foo]
The benefit of this is that we can easily update the successors of
a callbr, without having to worry about also updating blockaddress
references. This should allow us to remove some limitations:
* Allow unrolling/peeling/rotation of callbr, or any other
clone-based optimizations
(https://github.com/llvm/llvm-project/issues/41834)
* Allow duplicate successors
(https://github.com/llvm/llvm-project/issues/45248)
This is just the IR representation change though, I will follow up
with patches to remove limtations in various transformation passes
that are no longer needed.
Differential Revision: https://reviews.llvm.org/D129288
If we have a variable shift amount and the demanded mask has leading
zeros, we can propagate those leading zeros to not demand those bits
from operand 0. This can allow zero_extend/sign_extend to become
any_extend. This pattern can occur due to C integer promotion rules.
This transform is already done by InstCombineSimplifyDemanded.cpp where
sign_extend can be turned into zero_extend for example.
Reviewed By: spatel, foad
Differential Revision: https://reviews.llvm.org/D121833
Widening a G_FCONSTANT by extending and then generating G_FPTRUNC doesn't produce
the same result all the time. Instead, we can just transform it to a G_CONSTANT
of the same bit pattern and truncate using a plain G_TRUNC instead.
Fixes https://github.com/llvm/llvm-project/issues/56454
Differential Revision: https://reviews.llvm.org/D129743
If an MI will not generate a target instruction, we should not compute its
latency. Then we can compute more precise instruction sequence cost, and get
better result.
Differential Revision: https://reviews.llvm.org/D129615
isSafeToExpand() for addrecs depends on whether the SCEVExpander
will be used in CanonicalMode. At least one caller currently gets
this wrong, resulting in PR50506.
Fix this by a) making the CanonicalMode argument on the freestanding
functions required and b) adding member functions on SCEVExpander
that automatically take the SCEVExpander mode into account. We can
use the latter variant nearly everywhere, and thus make sure that
there is no chance of CanonicalMode mismatch.
Fixes https://github.com/llvm/llvm-project/issues/50506.
Differential Revision: https://reviews.llvm.org/D129630
The SI machine scheduler inherits from ScheduleDAGMI.
This patch adds support for a few features that are implemented
in ScheduleDAGMI (or its base classes) that were missing so far
because their support is implemented in overridden functions.
* Support cl::opt -view-misched-dags
This option allows to open a graphical window of the scheduling DAG.
* Support cl::opt -misched-print-dags
This option allows to print the scheduling DAG in text form.
* After constructing the scheduling DAG, call postprocessDAG()
to apply any registered DAG mutations.
Note that currently there are no mutations defined in AMDGPUTargetMachine.cpp
in case SIScheduler is used.
Still add this to avoid surprises in the future in case mutations are added.
Differential Revision: https://reviews.llvm.org/D128808
We shouldn't use getOpcodeDef() if we need to guarantee the def has only one
user since under the hood it may look through copies and optimization hints,
which themselves may have multiple users.
When doing scalable vectorization, the loop vectorizer uses a urem in the computation of the vector trip count. The RHS of that urem is a (possibly shifted) call to @llvm.vscale.
vscale is effectively the number of "blocks" in the vector register. (That is, types such as <vscale x 8 x i8> and <vscale x 1 x i8> both fill one 64 bit block, and vscale is essentially how many of those blocks there are in a single vector register at runtime.)
We know from the RISCV V extension specification that VLEN must be a power of two between ELEN and 2^16. Since our block size is 64 bits, the must be a power of two numbers of blocks. (For everything other than VLEN<=32, but that's already broken.)
It is worth noting that AArch64 SVE specification explicitly allows non-power-of-two sizes for the vector registers and thus can't claim that vscale is a power of two by this logic.
Differential Revision: https://reviews.llvm.org/D129609
We have the same fold in InstCombine - though implemented via OrZero flag on isKnownToBePowerOfTwo. The reasoning here is that either a) the result of the lshr is a power-of-two, or b) we have a div-by-zero triggering UB which we can ignore.
Differential Revision: https://reviews.llvm.org/D129606
Some rework of getStackGuard() based on comments in
https://reviews.llvm.org/D129505.
- getStackGuard() now creates and returns the destination
register, simplifying calls
- the pointer type is passed to getStackGuard() to avoid
recomputation
- removed PtrMemTy in emitSPDescriptorParent(), because
this type is only used here when loading the value but
not when storing the value
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D129576
To convert CTLZ to popcount we do
x = x | (x >> 1);
x = x | (x >> 2);
...
x = x | (x >>16);
x = x | (x >>32); // for 64-bit input
return popcount(~x);
This smears the most significant set bit across all of the bits
below it then inverts the remaining 0s and does a population count.
To support non-power of 2 types, the last shift amount must be
more than half of the size of the type. For i15, the last shift
was previously a shift by 4, with this patch we add another shift
of 8.
Fixes PR56457.
Differential Revision: https://reviews.llvm.org/D129431
When lowering llvm::stackprotect intrinsic, the SDAG implementation
checks useLoadStackGuardNode() to either create a LOAD_STACK_GUARD or use
the first argument of the intrinsic. This check is not present in the
IRTranslator, which results in always generating a LOAD_STACK_GUARD even
if the target does not support it.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D129505
Noticed while investigating the SystemZ regressions in D77804, prefer handling the knownbits analysis/simplification in the bitop nodes directly before falling back to SimplifyMultipleUseDemandedBits
Verify the LiveStacks analysis after a pass that claims to preserve it,
even if there are no further passes (apart from the verifier itself)
that would use the analysis.
Differential Revision: https://reviews.llvm.org/D129200
visitInlineAsm() in SDAGBuilder was duplicating a lot of the code
in ParseConstraints(), in particular all the logic to determine the
operand value and constraint VT.
Rely on the data computed by ParseConstraints() instead, and update
its ConstraintVT implementation to match getCallOperandValEVT()
more precisely.
As far as I can tell what was happening in the original code is
that the getNode call receives the same operands as the original
node with different SDNodeFlags. The logic inside getNode detects
that the node already exists and intersects the flags into the
existing node and returns it. This results in Op and NewOp for the
TLO.CombineTo call always being the same node.
We may have already called CombineTo as part of the recursive handling.
A second call to CombineTo as we unwind the recursion overwrites
the previous CombineTo. I think this means any time we updated the
poison flags that was the only change that ends up getting made
and we relied on DAGCombiner to revisit and call SimplifyDemandedBits
again. The second time the poison flags wouldn't need to be dropped
and we would keep the CombineTo call from further down the recursion.
We can instead call setFlags to drop the poison flags and remove the
call to TLO.CombineTo. This way we keep the CombineTo from deeper in
the recursion which should be more efficient.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D129511
As suggested in the post-commit feedback for D128123,
we can ease the mask constraint to ignore the MSB
(and make the code easier to read by adjusting the check).
https://alive2.llvm.org/ce/z/bbvqWv
Pointed out in Issue #56432: the current reference models may not be
quite friendly to open source projects. Their purpose is only
illustrative - the expectation is that projects would train their own.
To avoid unintentionally pulling such a model, made the URL cmake
setting require explicit user setting.
Differential Revision: https://reviews.llvm.org/D129342
Currently, an error exists when InstrRefBasedLDV observes transfers of
variables across copies, which causes it to lose track of variables
under certain circumstances, resulting in shorter lifetimes for those
variables as LDV gives up searching for live locations for them. This
patch fixes this issue by storing the currently tracked values in
the destination first, then updating them manually later without
clobbering or assigning them the wrong value.
Differential Revision: https://reviews.llvm.org/D128101
This patch restores calls to has_value to make it clear that we are
checking the presence of an optional value, not the underlying value.
This patch partially reverts d08f34b592.
Differential Revision: https://reviews.llvm.org/D129454
D89489 added some logic to the interleaved access pass to attempt to
undo the folding of shuffles into binops, that instcombine performs. If
early-cse is run too, the binops may be commoned into a single operation
with multiple shuffle uses. It is still profitable reverse the transform
though, so long as all the uses are shuffles.
Differential Revision: https://reviews.llvm.org/D129419
(Reapply after revert in e9ce1a5880 due to
Fuchsia test failures. Removed changes in lib/ExecutionEngine/ other
than error categories, to be checked in more detail and reapplied
separately.)
Bulk remove many of the more trivial uses of ManagedStatic in the llvm
directory, either by defining a new getter function or, in many cases,
moving the static variable directly into the only function that uses it.
Differential Revision: https://reviews.llvm.org/D129120
Bulk remove many of the more trivial uses of ManagedStatic in the llvm
directory, either by defining a new getter function or, in many cases,
moving the static variable directly into the only function that uses it.
Differential Revision: https://reviews.llvm.org/D129120
We already handled this case for add with a constant RHS. A
similar pattern can occur for sub with a constant left hand side.
Test cases use add and a mul representing (neg (shl X, C)) because
that's what I saw in the wild. The mul will be decomposed and then
the new transform can kick in.
Tests have not been committed, but this patch shows the changes.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D128769
Originally encountered with RUST, but also there are cases with distributed LTO
where debug info dwo units contain structurally the same debug information, with
difference in DW_AT_linkage_name. This causes collision on DWO ID.
Differential Revision: https://reviews.llvm.org/D129317
SelectionDAG has a target hook, getExtendForAtomicOps, which it uses
in the computeKnownBits implementation for ATOMIC_LOAD. This is pretty
ugly (as is having a separate load opcode for atomics), so instead
allow making use of atomic zextload. Enable this for AArch64 since the
DAG path defaults in to the zext behavior.
The tablegen changes are pretty ugly, but partially helps migrate
SelectionDAG from using ISD::ATOMIC_LOAD to regular ISD::LOAD with
atomic memory operands. For now the DAG emitter will emit matchers for
patterns which the DAG will not produce.
I'm still a bit confused by the intent of the isLoad/isStore/isAtomic
bits. The DAG implementation rejects trying to use any of these in
combination. For now I've opted to make the isLoad checks also check
isAtomic, although I think having isLoad and isAtomic set on these
makes most sense.
If all the demanded bits of the AND mask covering the inserted subvector 'X' are known to be one, then the mask isn't affecting the subvector at all.
In which case, if the base vector 'C' is undef/constant, then move the AND mask up to just (constant) fold it directly.
Addresses some of the regressions from D129150, particularly the cases where we're attempting to zero the upper elements of a widened vector.
Differential Revision: https://reviews.llvm.org/D129290
After D82916 `updateAllRanges()` started to fix holes in main range with
subranges but it fails on instructions with two subregs def which are parts of
one reg. The main range constructed with //all// subranges of subregs just after
processing the first operand. So the main range gets intervals from subranges
those are not updated yet.
The patch takes into account lane mask to update the main range.
Reviewed By: rampitec, arsenm
Differential Revision: https://reviews.llvm.org/D128553
This is almost the same as the abandoned D48529, but it
allows splat vector constants too.
This replaces the x86-specific code that was added with
the alternate patch D48557 with the original generic
combine.
This transform is a less restricted form of an existing
InstCombine and the proposed SDAG equivalent for that
in D128080:
https://alive2.llvm.org/ce/z/OUm6N_
Differential Revision: https://reviews.llvm.org/D128123
This patchs adds a new metadata kind `exclude` which implies that the
global variable should be given the necessary flags during code
generation to not be included in the final executable. This is done
using the ``SHF_EXCLUDE`` flag on ELF for example. This should make it
easier to specify this flag on a variable without needing to explicitly
check the section name in the target backend.
Depends on D129053 D129052
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D129151
Currently we use the `.llvm.offloading` section to store device-side
objects inside the host, creating a fat binary. The contents of these
sections is currently determined by the name of the section while it
should ideally be determined by its type. This patch adds the new
`SHT_LLVM_OFFLOADING` section type to the ELF section types. Which
should make it easier to identify this specific data format.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D129052
This is done during type legalization since the target representation of
these nodes may not be valid until after type legalization, and after
type legalization the fact that these are dealing with i1 types may be
lost.
Differential Revision: https://reviews.llvm.org/D128996
Truncates and compares require some changes to generic legalisation functions
to use ElementCount instead of getNumElements.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D129082
Prior to this change, live variable operands passed to
`llvm.experimental.stackmap` would be emitted directly to target nodes,
meaning that they don't get legalised. The upshot of this is that LLVM
may crash when encountering illegally typed target nodes.
e.g. https://github.com/llvm/llvm-project/issues/21657
This change introduces a platform independent stackmap DAG node whose
operands are legalised as per usual, thus avoiding aforementioned
crashes.
Note that some kinds of argument are still not handled properly, namely
vectors, structs, and large integers, like i128s. These will need to be
addressed in follow-up changes.
Note also that this does not change the behaviour of
`llvm.experimental.patchpoint`. A follow up change will do the same for
this intrinsic.
Differential review:
https://reviews.llvm.org/D125680
This patch adds the support for `fmax` and `fmin` operations in `atomicrmw`
instruction. For now (at least in this patch), the instruction will be expanded
to CAS loop. There are already a couple of targets supporting the feature. I'll
create another patch(es) to enable them accordingly.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D127041
As constant expressions can no longer trap, it only makes sense to
call isSafeToSpeculativelyExecute on Instructions, so limit the
API to accept only them, rather than general Operators or Values.
As integer div/rem constant expressions are no longer supported,
constants can no longer trap and are always safe to speculate.
Remove the Constant::canTrap() method and its usages.
This removes the insertvalue constant expression, as part of
https://discourse.llvm.org/t/rfc-remove-most-constant-expressions/63179.
This is very similar to the extractvalue removal from D125795.
insertvalue is also not supported in bitcode, so no auto-ugprade
is necessary.
ConstantExpr::getInsertValue() can be replaced with
IRBuilder::CreateInsertValue() or ConstantFoldInsertValueInstruction(),
depending on whether a constant result is required (with the latter
being fallible).
The ConstantExpr::hasIndices() and ConstantExpr::getIndices()
methods also go away here, because there are no longer any constant
expressions with indices.
Differential Revision: https://reviews.llvm.org/D128719
variable with its multiple aliases.
This patch handles the case where a variable has
multiple aliases.
AIX's assembly directive .set is not usable for the
aliasing purpose, and using different labels allows
AIX to emulate symbol aliases. If a value is emitted
between any two labels, meaning they are not aligned,
XCOFF will automatically calculate the offset for them.
This patch implements:
1) Emits the label of the alias just before emitting
the value of the sub-element that the alias referred to.
2) A set of aliases that refers to the same offset
should be aligned.
3) We didn't emit aliasing labels for common and
zero-initialized local symbols in
PPCAIXAsmPrinter::emitGlobalVariableHelper, but
emitted linkage for them in
AsmPrinter::emitGlobalAlias, which caused a FAILURE.
This patch fixes the bug by blocking emitting linkage
for the alias without a label.
Reviewed By: shchenz
Differential Revision: https://reviews.llvm.org/D124654
Before merging two instructions together, GISel does some sanity checks
that the folding is legal. However that check was missing that the
source of the pattern may be convergent. When the destination location
is in a different basic block, the folding is invalid.
Differential Revision: https://reviews.llvm.org/D128539
One motivation to add support for these types are the LD1Q/ST1Q
instructions in SME, for which we have defined a number of load/store
intrinsics which at the moment still take a `<vscale x 16 x i1>` predicate
regardless of their element type.
This patch adds basic support for the nxv1i1 type such that it can be passed/returned
from functions, as well as some basic support to support some existing tests that
result in a nxv1i1 type. It also adds support for splats.
Other operations (e.g. insert/extract subvector, logical ops, etc) will be
supported in follow-up patches.
Reviewed By: paulwalker-arm, efriedma
Differential Revision: https://reviews.llvm.org/D128665
In X86 we split greddy register allocation into 2 passes. The 1st pass
is to allocate tile register, and the 2nd pass is to allocate the rest
of virtual register. In most cases there is no tile register, so the 1st
pass is unnecessary. To improve the compiling time, we check if there is
any register need to be allocated by invoking callback
`ShouldAllocateClass`. If there is no register to be allocated, just
return false in the pass. This would improve the 1st greed RA pass for
normal cases.
Differential Revision: https://reviews.llvm.org/D128804
The filter clause in the landingpad may not have a GlobalVariable operand.
It may instead have a ConstantArray of operands and each operand within this
ConstantArray should also be checked to see if it is a GlobalVariable.
This patch add the check for the ConstantArray as well as a debug message that
outputs the contents of MustKeepGlobalVariables.
Reviewed By: lei, amyk, scui
Differential Revision: https://reviews.llvm.org/D128287
lowerConstant() currently accepts a number of constant expressions
which have corresponding MC expressions, but which cannot be
evaluated as a relocatable expression (unless the operands are
constant, in which case we'll just fold the expression to a constant).
The motivation here is to clarify which constant expressions are
really needed for https://discourse.llvm.org/t/rfc-remove-most-constant-expressions/63179,
and in particular clarify that we do not need to support any
division expressions, which are particularly problematic.
Differential Revision: https://reviews.llvm.org/D127972
When we fill the shape to tile configure memory, the shape is gotten
from AMX pseudo instruction. However the register for the shape may be
split or spilled by greedy RA. That cause we fill the shape to config
memory after ldtilecfg is executed, so that the shape configuration
would be wrong.
This patch is to split the tile register allocation from greedy register
allocation, so that after tile registers are allocated the shape
registers are still virtual register. The shape register only may be
redefined or multi-defined by phi elimination pass, two address pass.
That doesn't affect tile register configuration.
Differential Revision: https://reviews.llvm.org/D128584
Add a new pattern A - (B + C) ==> (A - B) - C to give machine combiner a chance
to evaluate which instruction sequence has lower latency.
Differential Revision: https://reviews.llvm.org/D124564
This is a resurrection of D106421 with the change that it keeps backward-compatibility. This means decoding the previous version of `LLVM_BB_ADDR_MAP` will work. This is required as the profile mapping tool is not released with LLVM (AutoFDO). As suggested by @jhenderson we rename the original section type value to `SHT_LLVM_BB_ADDR_MAP_V0` and assign a new value to the `SHT_LLVM_BB_ADDR_MAP` section type. The new encoding adds a version byte to each function entry to specify the encoding version for that function. This patch also adds a feature byte to be used with more flexibility in the future. An use-case example for the feature field is encoding multi-section functions more concisely using a different format.
Conceptually, the new encoding emits basic block offsets and sizes as label differences between each two consecutive basic block begin and end label. When decoding, offsets must be aggregated along with basic block sizes to calculate the final offsets of basic blocks relative to the function address.
This encoding uses smaller values compared to the existing one (offsets relative to function symbol).
Smaller values tend to occupy fewer bytes in ULEB128 encoding. As a result, we get about 17% total reduction in the size of the bb-address-map section (from about 11MB to 9MB for the clang PGO binary).
The extra two bytes (version and feature fields) incur a small 3% size overhead to the `LLVM_BB_ADDR_MAP` section size.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D121346
`commonAlignment` is a shortcut to pick the smallest of two `Align`
objects. As-is it doesn't bring much value compared to `std::min`.
Differential Revision: https://reviews.llvm.org/D128345
Information in the function `Prologue Data` is intentionally opaque.
When a function with `Prologue Data` is duplicated. The self (global
value) references inside `Prologue Data` is still pointing to the
original function. This may cause errors like `fatal error: error in backend: Cannot represent a difference across sections`.
This patch detaches the information from function `Prologue Data`
and attaches it to a function metadata node.
This and D116130 fix https://github.com/llvm/llvm-project/issues/49689.
Reviewed By: pcc
Differential Revision: https://reviews.llvm.org/D115844
This commit modifies the AsmPrinter to avoid emitting any zero-sized symbols to
the .debug_aranges table, by rounding their size up to 1. Entries with zero
length violate the DWARF 5 spec, which states:
> Each descriptor is a triple consisting of a segment selector, the beginning
> address within that segment of a range of text or data covered by some entry
> owned by the corresponding compilation unit, followed by the non-zero length
> of that range.
In practice, these zero-sized entries produce annoying warnings in lld and
cause GNU binutils to truncate the table when parsing it.
Other parts of LLVM, such as DWARFDebugARanges in the DebugInfo module
(specifically the appendRange method), already avoid emitting zero-sized
symbols to .debug_aranges, but not comprehensively in the AsmPrinter. In fact,
the AsmPrinter does try to avoid emitting such zero-sized symbols when labels
aren't involved, but doesn't when the symbol to emitted is a difference of two
labels; this patch extends that logic to handle the case in which the symbol is
defined via labels.
Furthermore, this patch fixes a bug in which `available_externally` symbols
would cause unpredictable values to be emitted into the `.debug_aranges` table
under certain circumstances. In practice I don't believe that this caused
issues up until now, but the root cause of this bug--an invalid DenseMap
lookup--triggered failures in Chromium when combined with an earlier version of
this patch. Therefore, this patch fixes that bug too.
This is a revised version of diff D126257, which was reverted due to breaking
tests. The now-reverted version of this patch didn't distinguish between
symbols that didn't have their size reported to the DwarfDebug handler and
those that had their size reported to be zero. This new version of the patch
instead restricts the special handling only to the symbols whose size is
definitively known to be zero.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D126835
These intrinsics are now fundemental for SVE code generation and have been
present for a year and a half, hence move them out of the experimental
namespace.
Differential Revision: https://reviews.llvm.org/D127976
Patch was reverted in 4c5f10a due to buildbot failures, now being
reapplied with updated AArch64 and RISCV tests.
This patch adds handling for the llvm.powi.* intrinsics in
BasicTTIImplBase::getIntrinsicInstrCost() and improves vectorization.
Closes#53887.
Differential Revision: https://reviews.llvm.org/D128172
If an instruction is sunk into a loop then any kill flags on
operands declared outside the loop must be cleared as these
will be live for all loop iterations.
Fixes#46827
Reviewed By: MatzeB
Differential Revision: https://reviews.llvm.org/D126754
According to the vector spec, mf8 is not supported for i8 if ELEN
is 32. Similarily mf4 is not suported for i16/f16 or mf2 for i32/f32.
Since RVVBitsPerBlock is 64 and LMUL is calculated as
((MinNumElements * ElementSize) / RVVBitsPerBlock) this means we
need to disable any type with MinNumElements==1.
For generic IR, these types will now be widened in type legalization.
For RVV intrinsics, we'll probably hit a fatal error somewhere. I plan
to work on disabling the intrinsics in the riscv_vector.h header.
Reviewed By: arcbbb
Differential Revision: https://reviews.llvm.org/D128286
waitcnt vmcnt instructions are currently generated in loop bodies before using
values loaded outside of the loop. In some cases, it is better to flush the
vmcnt counter in a loop preheader before entering the loop body. This patch
detects these cases and generates waitcnt instructions to flush the counter.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D115747
This reverts commit 719658d078.
Breaks a few things, see comments on https://reviews.llvm.org/D128437
There's disagreement about the best fix.
So let's keep HEAD green while discussions are happening.
The base RA support infrastructure that only allow a specific register
class be allocated in RA pss. Since greedy RA, basic RA derived from
base RA, they all allow allocating specific register class. Fast RA
doesn't support allocating register for specific register class. This
patch is to enable ShouldAllocateClass in fast RA, so that it can
support allocating register for specific register class.
Differential Revision: https://reviews.llvm.org/D126771
We're slowly removing SelectionDAG::GetDemandedBits and replacing it with SimplifyMultipleUseDemandedBits, we no longer have any uses for the vector demanded elt variant.
This patch adds handling for the llvm.powi.* intrinsics in
BasicTTIImplBase::getIntrinsicInstrCost() and improves vectorization.
Closes#53887.
Differential Revision: https://reviews.llvm.org/D128172
For below case, virtual register is defined twice in the self loop. We
don't need to spill %0 after the third instruction `%0 = def (tied %0)`,
because it is defined in the second instruction `%0 = def`.
1 bb.1
2 %0 = def
3 %0 = def (tied %0)
4 ...
5 jmp bb.1
Reviewed By: MatzeB
Differential Revision: https://reviews.llvm.org/D125079
An AArch64ISD::DUP is just a splat, where the known bits for each lane
are the same as the input. This teaches that to computeKnownBitsForTargetNode.
Problems arise for constants though, as a constant BUILD_VECTOR can be
lowered to an AArch64ISD::DUP, which SimplifyDemandedBits would then
turn back into a constant BUILD_VECTOR leading to an infinite cycle.
This has been prevented by adding a isTargetCanonicalConstantNode node
to prevent the conversion back into a BUILD_VECTOR.
Differential Revision: https://reviews.llvm.org/D128144
Similar to the existing (shl (srl x, c1), c2) fold
Part of the work to fix the regressions in D77804
Differential Revision: https://reviews.llvm.org/D125836
The VT we want to shrink to may not be legal especially after type
legalization.
Fixes PR56110.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D128135
The included test hits a verifier problems as one of the instructions:
```
%113:tgpreven, %114:tgprodd = MVE_VMLSLDAVas16 %12:tgpreven(tied-def 0), %11:tgprodd(tied-def 1), %7:mqpr, %8:mqpr, 0, $noreg, $noreg
```
Has two inputs that come from different PHIs with the same base reg, but
conflicting regclasses:
```
%11:tgprodd = PHI %103:gpr, %bb.1, %16:gpr, %bb.2
%12:tgpreven = PHI %103:gpr, %bb.1, %17:gpr, %bb.2
```
The MachinePipeliner would attempt to use %103 for both the %11 and %12
operands in the prolog, constraining the register class to the common
subset of both. Unfortunately there are no registers that are both odd
and even, so the second constrainRegClass fails. Fix this situation by
inserting a COPY for the second if the call to constrainRegClass fails.
The register allocation can then fold that extra copy away. The register
allocation of Q regs changed with this test, but the R regs were the
same and no new instructions are needed in the final assembly.
Differential Revision: https://reviews.llvm.org/D127971
D125335 makes regsOverlap skip following control flow, which is not entended
in the original code.
Differential Revision: https://reviews.llvm.org/D128039
WidenVecOp_INSERT_SUBVECTOR only supported cases where widening
effectively converts the insert into a copy. However, when the
widened subvector is no bigger than the vector being inserted into
and we can be sure there's no loss of data, we can simply emit
another INSERT_SUBVECTOR.
Fixes: #54982
Differential Revision: https://reviews.llvm.org/D127508
MinRCSize is 4 and prevents constrainRegClass from changing the
register class if the new class has size less than 4.
IMPLICIT_DEF gets a unique vreg for each use and will be removed
by the ProcessImplicitDef pass before register allocation. I don't
think there is any reason to prevent constraining the virtual register
to whatever register class the use needs.
The attached test case was previously creating a copy of IMPLICIT_DEF
because vrm8nov0 has 3 registers in it.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D128005
This helps handling a case where the BUILD_VECTOR has i16 element type
and i32 constant operands
t2: v8i16 = setcc t8, t17, setult:ch
t3: v8i16 = BUILD_VECTOR Constant:i32<1>, ...
t4: v8i16 = and t2, t3
t5: v8i16 = add t8, t4
This can be turned into t5: v8i16 = sub t8, t2, and allows us to remove
t3 and t4 from the DAG.
Differential Revision: https://reviews.llvm.org/D127354
This reverts commit 7207373e1e.
We found another RISC-V bug when landing D126048, and it has been fixed
by D127642 now.
Differential Revision: https://reviews.llvm.org/D126048
This is needed by our downstream and makes bf16 and f16 have the
same set of scalable vector types.
Reviewed By: rui.zhang
Differential Revision: https://reviews.llvm.org/D127877
The use operand may be undefined. In that case we can just continue to
check the next operand since it won't increase register pressure.
Differential Revision: https://reviews.llvm.org/D127848
This is modeled after the half-precision fp support. Two new nodes are
introduced for casting from and to bf16. Since casting from bf16 is a
simple operation I opted to always directly lower it to integer
arithmetic. The other way round is more complicated if you want to
preserve IEEE semantics, so it's handled by a new __truncsfbf2
compiler-rt builtin.
This is of course very bare bones, but sufficient to get a semi-softened
fadd on x86.
Possible future improvements:
- Targets with bf16 conversion instructions can now make fp_to_bf16 legal
- The software conversion to bf16 can be replaced by a trivial
implementation under fast math.
Differential Revision: https://reviews.llvm.org/D126953
When compiling for the RWPI relocation model [1], the debug information
is wrong for readonly global variables.
Writable global variables are accessed by the static base register (R9
on ARM) in the RWPI relocation model. This is being correctly generated
Readonly global variables are not accessed by the static base register
in the RWPI relocation model. This case is incorrectly generating the
same debugging information as for writable global variables.
References:
[1] ARM Read-Write Position Independence: https://github.com/ARM-software/abi-aa/blob/main/aapcs32/aapcs32.rst#read-write-position-independence-rwpi
Differential Revision: https://reviews.llvm.org/D126361
GetValueInMiddleOfBlock uses result of GetValueAtEndOfBlockInternal if there is no value
defined for current basic block.
If there is already a value it tries (in this order):
to find single register coming from all predecessors
find existing phi node which matches our incoming registers
build new phi.
The compile time improvement is to use current available value if
it is defined out of current BB or it is a PHI register.
This is due to it can be used in the middle basic block.
Reviewed By: sameerds
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D126523
When spilling CSRs, FixupStatepoint pass does simple copy propagation,
trying to find COPY instruction which defines register being spilled
and spill COPY source instead. I.e., if we have CSR $x and found
$x = COPY $y
we will spill $y instead.
But we may be unable to delete COPY instruction for some reason.
Then, spill will be inserted after it, adding another use of $y.
If COPY instruction was last use of $y (killed it), after insertion of
the spill it is not, so `isKill` flag must be cleared. We failed to do
so and this patch fixes this issue.
Reviewed By: skatkov
Differential Revision: https://reviews.llvm.org/D127308
This is a fix for https://github.com/llvm/llvm-project/issues/55827.
When register we are trying to re-color is split the original register (we tried to recover)
has no uses after the split. However in rollback actions we assign back physical register to it.
Later it causes different assertions. One of them is in attached test.
This CL fixes this by avoiding assigning physical register back to register which has no usage
or its live interval now is empty.
Reviewed By: arsenm, qcolombet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D127281
The last of getEvictor use was removed on Jun 5, 2022 in commit
5c06f7168f, which was itself a patch to
remove unused code.
Once we remove getEvictor, EvictionTrack becomes a write-only data
structure. The data in it won't affect compilation, so the entire
class is essentially dead.
Kazu Hirata