Unpredicated FP-multiply of SVE vector with a vector-element given by
vector[index], for example:
fmul z0.s, z1.s, z2.s[0]
which performs an unpredicated FP-multiply of all 32-bit elements in
'z1' with the first element from 'z2'.
This patch adds restricted register classes for SVE vectors:
ZPR_3b (only z0..z7 are allowed) - for indexed vector of 16/32-bit elements.
ZPR_4b (only z0..z15 are allowed) - for indexed vector of 64-bit elements.
Reviewers: rengolin, fhahn, SjoerdMeijer, samparker, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D48823
llvm-svn: 336205
This adds the following system registers:
- RAS registers,
- MPAM registers,
- Activitiy monitor registers,
- Trace Extension registers,
- Timing insensitivity of data processing instructions,
- Enhanced Support for Nested Virtualization.
Differential Revision: https://reviews.llvm.org/D48871
llvm-svn: 336193
We currently don't any-extend vararg parameters before storing them to the stack
locations on Darwin. However, SelectionDAG however does this, and so user code
is in the wild which inadvertently relies on this extension. This can manifest
in cases where the value stored is (int)0, but the actual parameter is interpreted
by va_arg as a pointer, and so not extending to 64 bits causes the callee to
load additional undefined bits.
llvm-svn: 336120
Increment/decrement vector by multiple of predicate constraint
element count.
The variants added by this patch are:
- INCH, INCW, INC
and (saturating):
- SQINCH, SQINCW, SQINCD
- UQINCH, UQINCW, UQINCW
- SQDECH, SQINCW, SQINCD
- UQDECH, UQINCW, UQINCW
For example:
incw z0.s, all, mul #4
llvm-svn: 336090
These patches were previously reverted as they led to
buildbot time-outs caused by large switch statement in
printAliasInstr when using UBSan and O3. The issue has
been addressed with a workaround (r335525).
llvm-svn: 336079
Initial patch adding assembly support for Armv8.4-A.
Besides adding v8.4 as a supported architecture to the usual places, this also
adds target features for the different crypto algorithms. Armv8.4-A introduced
new crypto algorithms, made them optional, and allows different combinations:
- none of the v8.4 crypto functions are supported, which is independent of the
implementation of the Armv8.0 SHA1 and SHA2 instructions.
- the v8.4 SHA512 and SHA3 support is implemented, in this case the Armv8.0
SHA1 and SHA2 instructions must also be implemented.
- the v8.4 SM3 and SM4 support is implemented, which is independent of the
implementation of the Armv8.0 SHA1 and SHA2 instructions.
- all of the v8.4 crypto functions are supported, in this case the Armv8.0 SHA1
and SHA2 instructions must also be implemented.
The v8.4 crypto instructions are added to AArch64 only, and not AArch32,
and are made optional extensions to Armv8.2-A.
The user-facing Clang options will map on these new target features, their
naming will be compatible with GCC and added in follow-up patches.
The Armv8.4-A instruction sets can be downloaded here:
https://developer.arm.com/products/architecture/a-profile/exploration-tools
Differential Revision: https://reviews.llvm.org/D48625
llvm-svn: 335953
Targets should be able to define whether or not they support the outliner
without the outliner being added to the pass pipeline. Before this, the
outliner pass would be added, and ask the target whether or not it supports the
outliner.
After this, it's possible to query the target in TargetPassConfig, before the
outliner pass is created. This ensures that passing -enable-machine-outliner
will not modify the pass pipeline of any target that does not support it.
https://reviews.llvm.org/D48683
llvm-svn: 335887
Add NoTrapAfterNoreturn target option which skips emission of traps
behind noreturn calls even if TrapUnreachable is enabled.
Enable the feature on Mach-O to save code size; Comments suggest it is
not possible to enable it for the other users of TrapUnreachable.
rdar://41530228
DifferentialRevision: https://reviews.llvm.org/D48674
llvm-svn: 335877
Now that we have the ability to legalize based on MMO's. Add support for
legalizing based on AtomicOrdering and use it to correct the legalization
of the atomic instructions.
Also extend all() to be a variadic template as this ruleset now requires
3 and 4 argument versions.
llvm-svn: 335767
It isn't safe to outline sequences of instructions where x16/x17/nzcv live
across the sequence.
This teaches the outliner to check whether or not a specific canidate has
x16/x17/nzcv live across it and discard the candidate in the case that that is
true.
https://bugs.llvm.org/show_bug.cgi?id=37573https://reviews.llvm.org/D47655
llvm-svn: 335758
This patch adds a custom trunc store lowering for v4i8 vector types.
Since there is not v.4b register, the v4i8 is promoted to v4i16 (v.4h)
and default action for v4i8 is to extract each element and issue 4
byte stores.
A better strategy would be to extended the promoted v4i16 to v8i16
(with undef elements) and extract and store the word lane which
represents the v4i8 subvectores. The construction:
define void @foo(<4 x i16> %x, i8* nocapture %p) {
%0 = trunc <4 x i16> %x to <4 x i8>
%1 = bitcast i8* %p to <4 x i8>*
store <4 x i8> %0, <4 x i8>* %1, align 4, !tbaa !2
ret void
}
Can be optimized from:
umov w8, v0.h[3]
umov w9, v0.h[2]
umov w10, v0.h[1]
umov w11, v0.h[0]
strb w8, [x0, #3]
strb w9, [x0, #2]
strb w10, [x0, #1]
strb w11, [x0]
ret
To:
xtn v0.8b, v0.8h
str s0, [x0]
ret
The patch also adjust the memory cost for autovectorization, so the C
code:
void foo (const int *src, int width, unsigned char *dst)
{
for (int i = 0; i < width; i++)
*dst++ = *src++;
}
can be vectorized to:
.LBB0_4: // %vector.body
// =>This Inner Loop Header: Depth=1
ldr q0, [x0], #16
subs x12, x12, #4 // =4
xtn v0.4h, v0.4s
xtn v0.8b, v0.8h
st1 { v0.s }[0], [x2], #4
b.ne .LBB0_4
Instead of byte operations.
llvm-svn: 335735
AArch64 was only setting costs for SK_Transpose, which meant that many of the simpler shuffles (e.g. SK_Select and SK_PermuteSingleSrc for larger vector elements) was being severely overestimated by the default shuffle expansion.
This patch adds costs to help improve SLP performance and avoid a regression in reductions introduced by D48174.
I'm not very knowledgeable about AArch64 shuffle lowering so I've kept the extra costs to a minimum - someone who knows this code can add extra costs which should improve vectorization a lot more.
Differential Revision: https://reviews.llvm.org/D48172
llvm-svn: 335329
This reverts commit d8f57105010cc7e78026e511d5def873fc91e0e7.
Original Commit:
Author: Haicheng Wu <haicheng@codeaurora.org>
Date: Sun Feb 18 13:51:33 2018 +0000
[AArch64] Coalesce Copy Zero during instruction selection
Add special case for copy of zero to avoid a double copy.
Differential Revision: https://reviews.llvm.org/D36104
Author's intention is to remove a BB that has one mov instruction. In
order to do that, d8f571050 pessmizes MachineSinking by introducing a
copy, such that mov instruction is NOT moved to the BB. Optimization
downstream gets rid of the BB with only mov instruction. This works well
if we have only one fall through branch as there is only one "extra"
mov instruction.
If we have multiple fall throughs, we will have a lot of redundant movs.
In such a case, it's better to have this BB which has one mov instruction.
This is causing degradation in jpeg, fft and other codebases. I believe
if we want to remove a BB with only one branch instruction, we should not
pessimize Machine Sinking at all, and find some other solution.
llvm-svn: 335251
insertOutlinerPrologue was not used by any target, and prologue-esque code was
beginning to appear in insertOutlinerEpilogue. Refactor that into one function,
buildOutlinedFrame.
This just removes insertOutlinerPrologue and renames insertOutlinerEpilogue.
llvm-svn: 335076
This patch uses the DiagnosticPredicate for SVE predicate patterns
to improve their diagnostics, now giving a 'invalid operand' diagnostic
if the type is not an immediate or one of the expected pattern
labels.
Reviewers: samparker, SjoerdMeijer, javed.absar, fhahn
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D48220
llvm-svn: 334983
The variants added by this patch are:
- SQINC signed increment, e.g. sqinc x0, w0, all, mul #4
- SQDEC signed decrement, e.g. sqdec x0, w0, all, mul #4
- UQINC unsigned increment, e.g. uqinc w0, all, mul #4
- UQDEC unsigned decrement, e.g. uqdec w0, all, mul #4
This patch includes asmparser changes to parse a GPR64 as a GPR32 in
order to satisfy the constraint check:
x0 == GPR64(w0)
in:
sqinc x0, w0, all, mul #4
^___^ (must match)
Reviewers: rengolin, fhahn, SjoerdMeijer, samparker, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47716
llvm-svn: 334980
This patch adds instructions for comparing elements from two vectors, e.g.
cmpgt p0.s, p0/z, z0.s, z1.s
and also adds support for comparing to a 64-bit wide element vector, e.g.
cmpgt p0.s, p0/z, z0.s, z1.d
The patch also contains aliases for certain comparisons, e.g.:
cmple p0.s, p0/z, z0.s, z1.s => cmpge p0.s, p0/z, z1.s, z0.s
cmplo p0.s, p0/z, z0.s, z1.s => cmphi p0.s, p0/z, z1.s, z0.s
cmpls p0.s, p0/z, z0.s, z1.s => cmphs p0.s, p0/z, z1.s, z0.s
cmplt p0.s, p0/z, z0.s, z1.s => cmpgt p0.s, p0/z, z1.s, z0.s
llvm-svn: 334931
Support for SVE's predicated select instructions to select elements
from either vector, both in a data-vector and a predicate-vector
variant.
llvm-svn: 334905
Increment/decrement scalar register by (scaled) element count given by
predicate pattern, e.g. 'incw x0, all, mul #4'.
Reviewers: rengolin, fhahn, SjoerdMeijer, samparker, javed.absar
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D47713
llvm-svn: 334838
Some instructions require of a limited set of FP immediates as operands,
for example '#0.5 or #1.0' for SVE's FADD instruction.
This patch adds support for parsing and printing such FP immediates as
exact values (e.g. #0.499999 is not accepted for #0.5).
Reviewers: rengolin, fhahn, SjoerdMeijer, samparker, javed.absar
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D47711
llvm-svn: 334826
Summary:
For targets I'm not familiar with, I've automatically made the "default to 1 for each resource" behaviour explicit in the td files.
For more obvious cases, I've ventured a fix.
Some notes:
- Exynos is especially fishy.
- AArch64SchedThunderX2T99.td had some truncated entries. If I understand correctly, the person who wrote that interpreted the ResourceCycle as a range. I made the decision to use the upper/lower bound for consistency with the 'Latency' value. I'm sure there is a better choice.
- The change to X86ScheduleBtVer2.td is an NFC, it just makes values more explicit.
Also see PR37310.
Reviewers: RKSimon, craig.topper, javed.absar
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D46356
llvm-svn: 334586
Register x20 is a callee-saved register which may be used for other
purposes in certain contexts, for example to hold special variables
within the kernel. This change adds support for reserving this register
both to frontend and backend to make this register usable for these
purposes.
Differential Revision: https://reviews.llvm.org/D46552
llvm-svn: 334531
On targets like Arm some relaxations may only be performed when certain
architectural features are available. As functions can be compiled with
differing levels of architectural support we must make a judgement on
whether we can relax based on the MCSubtargetInfo for the function. This
change passes through the MCSubtargetInfo for the function to
fixupNeedsRelaxation so that the decision on whether to relax can be made
per function. In this patch, only the ARM backend makes use of this
information. We must also pass the MCSubtargetInfo to applyFixup because
some fixups skip error checking on the assumption that relaxation has
occurred, to prevent code-generation errors applyFixup must see the same
MCSubtargetInfo as fixupNeedsRelaxation.
Differential Revision: https://reviews.llvm.org/D44928
llvm-svn: 334078
This is setting up to fix bug 37573 cleanly.
This moves data structures that are technically both used in some way by the
target and the general-purpose outlining algorithm into MachineOutliner.h. In
particular, the `Candidate` class is of importance.
Before, the outliner passed the locations of `Candidates` to the target, which
would then make some decisions about the prospective outlined function. This
change allows us to just pass `Candidates` along to the target. This will allow
the target to discard `Candidates` that would be considered unsafe before cost
calculation. Thus, we will be able to remove the unsafe candidates described in
the bug without resorting to torching the entire prospective function.
Also, as a side-effect, it makes the outliner a bit cleaner.
https://bugs.llvm.org/show_bug.cgi?id=37573
llvm-svn: 333952
Summary:
The new rules are straightforward. The main rules to keep in mind
are:
1. NAME is an implicit template argument of class and multiclass,
and will be substituted by the name of the instantiating def/defm.
2. The name of a def/defm in a multiclass must contain a reference
to NAME. If such a reference is not present, it is automatically
prepended.
And for some additional subtleties, consider these:
3. defm with no name generates a unique name but has no special
behavior otherwise.
4. def with no name generates an anonymous record, whose name is
unique but undefined. In particular, the name won't contain a
reference to NAME.
Keeping rules 1&2 in mind should allow a predictable behavior of
name resolution that is simple to follow.
The old "rules" were rather surprising: sometimes (but not always),
NAME would correspond to the name of the toplevel defm. They were
also plain bonkers when you pushed them to their limits, as the old
version of the TableGen test case shows.
Having NAME correspond to the name of the toplevel defm introduces
"spooky action at a distance" and breaks composability:
refactoring the upper layers of a hierarchy of nested multiclass
instantiations can cause unexpected breakage by changing the value
of NAME at a lower level of the hierarchy. The new rules don't
suffer from this problem.
Some existing .td files have to be adjusted because they ended up
depending on the details of the old implementation.
Change-Id: I694095231565b30f563e6fd0417b41ee01a12589
Reviewers: tra, simon_tatham, craig.topper, MartinO, arsenm, javed.absar
Subscribers: wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D47430
llvm-svn: 333900
For immediates used in DUP instructions that have the range
-128 to 127, or a multiple of 256 in the range -32768 to 32512,
one could argue that when the result element size is 16bits (.h),
the value can be considered both signed and unsigned.
Reviewers: rengolin, fhahn, SjoerdMeijer, samparker, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47619
llvm-svn: 333873
Print the first indexed element as a FP register, for example:
mov z0.d, z1.d[0]
Is now printed as:
mov z0.d, d1
Next to printing, this patch also adds aliases to parse 'mov z0.d, d1'.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47571
llvm-svn: 333872
Unpredicated copy of indexed SVE element to SVE vector,
along with MOV-aliases.
For example:
dup z0.h, z1.h[0]
duplicates the first 16-bit element from z1 to all elements in
the result vector z0.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D47570
llvm-svn: 333871
Predicated copy of floating-point immediate value to SVE vector,
along with MOV-aliases.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: javed.absar
Differential Revision: https://reviews.llvm.org/D47518
llvm-svn: 333869
Predicated copy of possibly shifted immediate value into SVE
vector, along with MOV-aliases.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47517
llvm-svn: 333868
Before we were relying on the any extend of the s1 to s32, but
for AAPCS we need to zero-extend it to at least s8.
Fixes PR36719
Differential Revision: https://reviews.llvm.org/D47425
llvm-svn: 333747
Unpredicated copy of floating-point immediate value into SVE vector,
along with MOV-aliases.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47482
llvm-svn: 333744
Instead of asserting when using the def_cfa directive with a register
different from fp, fallback on DWARF.
Easily triggered with:
.cfi_def_cfa x1, 32;
rdar://40249694
Differential Revision: https://reviews.llvm.org/D47593
llvm-svn: 333667
This is to make it clear what kind of bugs the LegalizerInfo::verifier
is able to catch and test its output
Reviewers: aemerson, qcolombet
Reviewed By: aemerson
Differential Revision: https://reviews.llvm.org/D46338
llvm-svn: 333597
Floating point immediate combining a negative sign and
a hexadecimal number, e.g. #-0x0 caused the compiler to crash.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: javed.absar
Differential Revision: https://reviews.llvm.org/D47483
llvm-svn: 333524
As suggested in https://bugs.llvm.org/show_bug.cgi?id=32384#c1, this change
makes the inlining of `memset()` and `memcpy()` more aggressive when
compiling for speed. The tuning remains the same when optimizing for size.
Patch by: Sebastian Pop <s.pop@samsung.com>
Evandro Menezes <e.menezes@samsung.com>
Differential revision: https://reviews.llvm.org/D45098
llvm-svn: 333429
This patch addresses the following variants:
- bitmask immediate, e.g. 'and z0.d, z0.d, #0x6'.
- unpredicated data vectors, e.g. 'and z0.d, z1.d, z2.d'.
- predicated data vectors, e.g. 'and z0.d, p0/m, z0.d, z1.d'.
And also several aliases, such as:
- ORN, alias of ORR.
- EON, alias of EOR.
- BIC, alias of AND (immediate variant)
- MOV, alias of ORR (if unpredicated and source register operands are the same)
Reviewers: rengolin, huntergr, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47363
llvm-svn: 333414
This patch adds addsub_imm8_opt_lsl_(i8|i16|i32|i64) operands
that are unsigned values in the range 0 to 255. For element widths of
16 bits or higher it may also be a signed multiple of 256 in the
range 0 to 65280.
Note: This also does some refactoring to reuse convenience function
getShiftedVal<shift>(), and now allows AArch64 scalar 'ADD #-4096' to be
accepted to be mapped to SUB #4096.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47310
llvm-svn: 333408
Unpredicated copy of optionally-shifted immediate to SVE vector,
along with MOV-aliases.
This patch contains parsing and printing support for
cpy_imm8_opt_lsl_(i8|i16|i32|i64). This operand allows a signed value in
the range -128 to +127. For element widths of 16 bits or higher it may
also be a signed multiple of 256 in the range -32768 to +32512.
For element-width of 8 bits a range of -128 to 255 is accepted, since a copy
of a byte can be considered either signed/unsigned.
Note: This patch renames tryParseAddSubImm() -> tryParseImmWithOptionalShift()
and moves the behaviour of trying to shift a plain immediate by an allowed
shift-value to its addImmWithOptionalShiftOperands() method, so that the
parsing itself is generic and allows immediates from multiple shifted operands.
This is done because an immediate can be divisible by both shifted operands.
Reviewers: rengolin, fhahn, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D47309
llvm-svn: 333263
The existing code has three different ways to try to lower a 64-bit
immediate to the sequence ORR+MOVK. The result is messy: it misses
some possible sequences, and the order of the checks means we sometimes
emit two MOVKs when we only need one.
Instead, just use a simple loop to try all possible two-instruction
ORR+MOVK sequences.
Differential Revision: https://reviews.llvm.org/D47176
llvm-svn: 333218
The AArch64 asm parser currently has custom parsing logic for .hword, .word,
and .xword. Rather than use this custom logic, we can just use
addAliasForDirective to enable the reuse of AsmParser::parseDirectiveValue.
Differential Revision: https://reviews.llvm.org/D47000
llvm-svn: 333077
When we're outlining a sequence that ends in a call, we can save up to
three instructions in the outlined function by turning the call into
a tail-call. I refer to this as thunk outlining because the resulting
outlined function looks like a thunk; suggestions welcome for a better
name.
In addition to making the outlined function shorter, thunk outlining
allows outlining calls which would otherwise be illegal to outline:
we don't need to save/restore LR, so we don't need to prove anything
about the stack access patterns of the callee.
To make this work effectively, I also added
MachineOutlinerInstrType::LegalTerminator to the generic MachineOutliner
code; this allows treating an arbitrary instruction as a terminator in
the suffix tree.
Differential Revision: https://reviews.llvm.org/D47173
llvm-svn: 333015
Summary:
This **appears** to be the last missing piece for the masked merge pattern handling in the backend.
This is [[ https://bugs.llvm.org/show_bug.cgi?id=37104 | PR37104 ]].
[[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]] will introduce an IR canonicalization that is likely bad for the end assembly.
Previously, `andps`+`andnps` / `bsl` would be generated. (see `@out`)
Now, they would no longer be generated (see `@in`), and we need to make sure that they are generated.
Differential Revision: https://reviews.llvm.org/D46528
llvm-svn: 332904
With this we gain a little flexibility in how the generic object
writer is created.
Part of PR37466.
Differential Revision: https://reviews.llvm.org/D47045
llvm-svn: 332868
To make this work I needed to add an endianness field to MCAsmBackend
so that writeNopData() implementations know which endianness to use.
Part of PR37466.
Differential Revision: https://reviews.llvm.org/D47035
llvm-svn: 332857
Provide some free functions to reduce verbosity of endian-writing
a single value, and replace the endianness template parameter with
a field.
Part of PR37466.
Differential Revision: https://reviews.llvm.org/D47032
llvm-svn: 332757
The idea is that a client that wants split dwarf would create a
specific kind of object writer that creates two files, and use it to
create the streamer.
Part of PR37466.
Differential Revision: https://reviews.llvm.org/D47050
llvm-svn: 332749
Counting the number of instructions is both unintuitive and inaccurate.
On AArch64, this only affects the generated remarks and certain rare
pseudo-instructions, but it will have a bigger impact on other targets.
Differential Revision: https://reviews.llvm.org/D46921
llvm-svn: 332685
This breaks the code which saves and restores LR, so we can't outline
without doing something more complicated for stack adjustment.
Found by inspection; we get lucky in most cases because getMemOpInfo
only handles STRWpost, not any other pre/post-increment forms. But it
hits a couple of artificial testcases in the tree.
Differential Revision: https://reviews.llvm.org/D46920
llvm-svn: 332529
The cost computation assumes we do this correctly, but the actual
lowering was wrong.
Differential Revision: https://reviews.llvm.org/D46923
llvm-svn: 332514
For regular SVE vector operands, this patch introduces a more
sensible diagnostic when the vector has a wrong suffix (e.g. z0.s vs z0.b).
For example:
add z0.s, z1.s, z2.b -> invalid element width
^_____^
mismatch
For the vector-with-shift/extend (e.g. z0.s, uxtw #2) this patch takes
a slightly different approach and instead returns a 'invalid operand'
if the element size is not as expected. This is because the diagnostics
are more specificied to suggest using the right shift/extend suffix. This
is a trade-off not to introduce more operand classes and still provide
useful diagnostics for LD1 and PRF instructions.
For example:
ld1w z1.s, p0/z, [x0, z0.s] -> invalid shift/extend specified, expected 'z[0..31].s, (uxtw|sxtw)'
ld1w z1.d, p0/z, [x0, z0.s] -> invalid operand
^________________^
mismatch
For gather prefetches, both 'z0.s' and 'z0.d' would be allowed:
prfw #0, p0, [x0, z0.s] -> invalid shift/extend specified, expected 'z[0..31].s, (uxtw|sxtw) #2'
prfw #0, p0, [x0, z0.d] -> invalid shift/extend specified, expected 'z[0..31].d, (lsl|uxtw|sxtw) #2'
Without this change, the diagnostic would unnecessarily suggest a
different element size:
prfw #0, p0, [x0, z0.s] -> invalid shift/extend specified, expected 'z[0..31].d, (lsl|uxtw|sxtw) #2'
Reviewers: SjoerdMeijer, aemerson, fhahn, samparker, javed.absar
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D46688
llvm-svn: 332483
Keep loads and stores together (target defines how many loads
and stores to gang up), such that it will help in pairing
and vectorization.
Differential Revision https://reviews.llvm.org/D46477
llvm-svn: 332482
We currently handle all aggregates by creating one large LLT, and letting the
legalizer deal with splitting them up. However using this approach means that
we can't support big endian code correctly.
This patch changes the way that the IRTranslator deals with aggregate values,
by splitting them up into their constituent element values. To do this, parts
of the translator need to be modified to deal with multiple VRegs for a single
Value.
A new Value to VReg mapper is introduced to help keep compile time under
control, currently there is no measurable impact on CTMark despite the extra
code being generated in some cases.
Patch is based on the original work of Tim Northover.
Differential Revision: https://reviews.llvm.org/D46018
llvm-svn: 332449
This patch re-introduces the "S" inline assembler constraint. This matches
an absolute symbolic address or a label reference. The primary use case is
asm("adrp %0, %1\n\t"
"add %0, %0, :lo12:%1" : "=r"(addr) : "S"(&var));
I say re-introduces as it seems like "S" was implemented in the original
AArch64 backend, but it looks like it wasn't carried forward to the merged
backend. The original implementation had A and L modifiers that could be
used to print ":lo12:" to the string. It looks like gcc doesn't use these
and :lo12: is expected to be written in the inline assembly string so I've
not implemented A and L. Clang already supports the S modifier.
Fixes PR37180
Differential Revision: https://reviews.llvm.org/D46745
llvm-svn: 332444
When storing the 0th lane of a vector, use a simpler and usually more
efficient scalar store instead. In this case, also using the unscaled
offset.
Differential revision: https://reviews.llvm.org/D46762
llvm-svn: 332394
When storing the 0th lane of a vector, use a simpler and usually more efficient scalar store instead.
Differential revision: https://reviews.llvm.org/D46655
llvm-svn: 332251
The DEBUG() macro is very generic so it might clash with other projects.
The renaming was done as follows:
- git grep -l 'DEBUG' | xargs sed -i 's/\bDEBUG\s\?(/LLVM_DEBUG(/g'
- git diff -U0 master | ../clang/tools/clang-format/clang-format-diff.py -i -p1 -style LLVM
- Manual change to APInt
- Manually chage DOCS as regex doesn't match it.
In the transition period the DEBUG() macro is still present and aliased
to the LLVM_DEBUG() one.
Differential Revision: https://reviews.llvm.org/D43624
llvm-svn: 332240
Summary:
performPostLD1Combine in AArch64ISelLowering looks for vector
insert_vector_elt of a loaded value which it can optimize into a single
LD1LANE instruction. The code checking for the pattern was not checking
if the lane index was a constant which could cause two problems:
- an assert when lowering the LD1LANE ISD node since it assumes an
constant operand
- an assert in isel if the lane index value depends on the
post-incremented base register
Both of these issues are avoided by simply checking that the lane index
is a constant.
Fixes bug 35822.
Reviewers: t.p.northover, javed.absar
Subscribers: rengolin, kristof.beyls, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D46591
llvm-svn: 332103
Accessing the members of a large data structures needs a lot of GEPs which
usually have large offsets due to the size of the underlying data structure. If
the offsets are too large to fit into the r+i addressing mode, these GEPs cannot
be sunk to their users' blocks and many extra registers are needed then to carry
the values of these GEPs.
This patch tries to split a large data struct starting from %base like the
following.
Before:
BB0:
%base =
BB1:
%gep0 = gep %base, off0
%gep1 = gep %base, off1
%gep2 = gep %base, off2
BB2:
%load1 = load %gep0
%load2 = load %gep1
%load3 = load %gep2
After:
BB0:
%base =
%new_base = gep %base, off0
BB1:
%new_gep0 = %new_base
%new_gep1 = gep %new_base, off1 - off0
%new_gep2 = gep %new_base, off2 - off0
BB2:
%load1 = load i32, i32* %new_gep0
%load2 = load i32, i32* %new_gep1
%load3 = load i32, i32* %new_gep2
In the above example, the struct is split into two parts. The first part still
starts from %base and the second part starts from %new_base. After the
splitting, %new_gep1 and %new_gep2 have smaller offsets and then can be sunk to
BB2 and folded into their users.
The algorithm to split data structure is simple and very similar to the work of
merging SExts. First, it collects GEPs that have large offsets when iterating
the blocks. Second, it splits the underlying data structures and updates the
collected GEPs to use smaller offsets.
Differential Revision: https://reviews.llvm.org/D42759
llvm-svn: 332015
With custom lowering for vector MULLH{S,U}, it is now profitable to
vectorize a divide by constant loop for the custom types (v16i8, v8i16,
and v4i32). The cost if based on TargetLowering::Build{S,U}DIV which
uses a multiply by constant plus adjustment to express a divide by
constant.
Both {u,s}mull{2} are expressed as Instruction::Mul and shifts by
Instruction::AShr.
llvm-svn: 331873
Reverting this to see if the clang-cmake-aarch64-global-isel and
clang-cmake-aarch64-quick bots are failing because of this commit.
We know it wasn't r331819.
llvm-svn: 331846
Because we create a new kind of debug instruction, DBG_LABEL, we need to
check all passes which use isDebugValue() to check MachineInstr is debug
instruction or not. When expelling debug instructions, we should expel
both DBG_VALUE and DBG_LABEL. So, I create a new function,
isDebugInstr(), in MachineInstr to check whether the MachineInstr is
debug instruction or not.
This patch has no new test case. I have run regression test and there is
no difference in regression test.
Differential Revision: https://reviews.llvm.org/D45342
Patch by Hsiangkai Wang.
llvm-svn: 331844
Patch https://reviews.llvm.org/D41445 changed the behaviour of 'isReg()'
to also return 'true' if the parsed register operand is a vector
register. Code in the AsmMatcher checks if a register is a subclass of the
expected register class. However, even though both parsed registers map
to the same physical register, the 'v' register is of kind 'NeonVector',
where 'q' is of type Scalar, where isSubclass() does not distinguish
between the two cases.
The solution is to use an AsmOperand instead of the register directly,
and use the PredicateMethod to distinguish the two operands.
This fixes for example:
ldr v0, [x0] // 'v0' is an invalid operand for this instruction
ldr q0, [x0] // valid
Reviewers: aemerson, Gerolf, SjoerdMeijer, javed.absar
Reviewed By: aemerson
Differential Revision: https://reviews.llvm.org/D46310
llvm-svn: 331755
Summary:
Previously, a extending load was represented at (G_*EXT (G_LOAD x)).
This had a few drawbacks:
* G_LOAD had to be legal for all sizes you could extend from, even if
registers didn't naturally hold those sizes.
* All sizes you could extend from had to be allocatable just in case the
extend went missing (e.g. by optimization).
* At minimum, G_*EXT and G_TRUNC had to be legal for these sizes. As we
improve optimization of extends and truncates, this legality requirement
would spread without considerable care w.r.t when certain combines were
permitted.
* The SelectionDAG importer required some ugly and fragile pattern
rewriting to translate patterns into this style.
This patch changes the representation to:
* (G_[SZ]EXTLOAD x)
* (G_LOAD x) any-extends when MMO.getSize() * 8 < ResultTy.getSizeInBits()
which resolves these issues by allowing targets to work entirely in their
native register sizes, and by having a more direct translation from
SelectionDAG patterns.
Each extending load can be lowered by the legalizer into separate extends
and loads, however a target that supports s1 will need the any-extending
load to extend to at least s8 since LLVM does not represent memory accesses
smaller than 8 bit. The legalizer can widenScalar G_LOAD into an
any-extending load but sign/zero-extending loads need help from something
else like a combiner pass. A follow-up patch that adds combiner helpers for
for this will follow.
The new representation requires that the MMO correctly reflect the memory
access so this has been corrected in a couple tests. I've also moved the
extending loads to their own tests since they are (mostly) separate opcodes
now. Additionally, the re-write appears to have invalidated two tests from
select-with-no-legality-check.mir since the matcher table no longer contains
loads that result in s1's and they aren't legal in AArch64 anymore.
Depends on D45540
Reviewers: ab, aditya_nandakumar, bogner, rtereshin, volkan, rovka, javed.absar
Reviewed By: rtereshin
Subscribers: javed.absar, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D45541
llvm-svn: 331601
Inspired by r331508, I did a grep and found these.
Mostly just change from dyn_cast to cast. Some cases also showed a dyn_cast result being converted to bool, so those I changed to isa.
llvm-svn: 331577
Summary: Adding support for Fast flags in the SDNode to leverage fast math sub flag usage.
Reviewers: spatel, arsenm, jbhateja, hfinkel, escha, qcolombet, echristo, wristow, javed.absar
Reviewed By: spatel
Subscribers: llvm-commits, rampitec, nhaehnle, tstellar, FarhanaAleen, nemanjai, javed.absar, jbhateja, hfinkel, wdng
Differential Revision: https://reviews.llvm.org/D45710
llvm-svn: 331547
This patch adds a custom lowering for ISD::MULH{S,U} used on divide by
constant optimization (DAGCombiner::BuildSDIV and DAGCombiner::BuildUDIV).
New patterns for smull and umull are added, so AArch64ISD::{S,U}MULL
can be correctly lowered to smull2 and umull2.
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D46009
llvm-svn: 331522
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Differential Revision: https://reviews.llvm.org/D46290
llvm-svn: 331272
There are two separate fixes here:
* The lowering code for non-extending loads should report UnableToLegalize instead of emitting the same instruction.
* The target should not be requesting lowering of non-extending loads.
llvm-svn: 331201
This patch extends the 'isSVEVectorRegWithShiftExtend' function to
improve diagnostics for SVE's gather load (scalar + vector) addressing
modes. Instead of always suggesting the 'unscaled' addressing mode,
the use of DiagnosticPredicate enables a more specific error message
in the context where the scaling is incorrect. For example:
ld1h z0.d, p0/z, [x0, z0.d, lsl #2]
^
shift amount should be '1'
Instead of suggesting the packed, unscaled addressing mode:
expected 'z[0..31].d, (uxtw|sxtw)'
the assembler now suggests using the proper scaling:
expected 'z[0..31].d, (lsl|uxtw|sxtw) #1'
Reviewers: fhahn, rengolin, samparker, SjoerdMeijer, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D46124
llvm-svn: 331162
Most of the add<operandname>Operands() functions are the same
and can be replaced by using a single 'RenderMethod' in
the AArch64InstrFormats.td file. Since many of the scaled
immediates (with different scaling/bits) are the same, most of
these can reuse the same AsmOperandClass.
Reviewers: fhahn, rengolin, samparker, SjoerdMeijer, javed.absar
Reviewed By: samparker
Differential Revision: https://reviews.llvm.org/D46122
llvm-svn: 331146
Summary:
Previously, a extending load was represented at (G_*EXT (G_LOAD x)).
This had a few drawbacks:
* G_LOAD had to be legal for all sizes you could extend from, even if
registers didn't naturally hold those sizes.
* All sizes you could extend from had to be allocatable just in case the
extend went missing (e.g. by optimization).
* At minimum, G_*EXT and G_TRUNC had to be legal for these sizes. As we
improve optimization of extends and truncates, this legality requirement
would spread without considerable care w.r.t when certain combines were
permitted.
* The SelectionDAG importer required some ugly and fragile pattern
rewriting to translate patterns into this style.
This patch begins changing the representation to:
* (G_[SZ]EXTLOAD x)
* (G_LOAD x) any-extends when MMO.getSize() * 8 < ResultTy.getSizeInBits()
which resolves these issues by allowing targets to work entirely in their
native register sizes, and by having a more direct translation from
SelectionDAG patterns.
This patch introduces the new generic instructions and new variation on
G_LOAD and adds lowering for them to convert back to the existing
representations.
Depends on D45466
Reviewers: ab, aditya_nandakumar, bogner, rtereshin, volkan, rovka, aemerson, javed.absar
Reviewed By: aemerson
Subscribers: aemerson, kristof.beyls, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D45540
llvm-svn: 331115
This commit makes it so that if you outline a def of some register, then the
call instruction created by the outliner actually reflects that the register
is defined by the call. It also makes it so that outlined functions don't
have the TracksLiveness property.
Outlined calls shouldn't break liveness assumptions that someone might make.
This also un-XFAILs the noredzone test, and updates the calls test.
llvm-svn: 331095
Summary:
Currently only the memory size is supported but others can be added as
needed.
narrowScalar for G_LOAD and G_STORE now correctly update the
MachineMemOperand and will refuse to legalize atomics since those need more
careful expansions to maintain atomicity.
Reviewers: ab, aditya_nandakumar, bogner, rtereshin, aemerson, javed.absar
Reviewed By: aemerson
Subscribers: aemerson, rovka, kristof.beyls, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D45466
llvm-svn: 331071
Put the first ldp at the end, so that the load-store optimizer can run
and merge the ldp and the add into a post-index ldp.
This didn't work in case no frame was needed and resulted in code size
regressions.
llvm-svn: 331044
This adds IR intrinsics for the AArch64 dot-product instructions introduced in
v8.2-A.
Differential revisioon: https://reviews.llvm.org/D46107
llvm-svn: 331036
The program might have unusual expectations for functions; for example,
the Linux kernel's build system warns if it finds references from .text
to .init.data.
I'm not sure this is something we actually want to make any guarantees
about (there isn't any explicit rule that would disallow outlining
in this case), but we might want to be conservative anyway.
Differential Revision: https://reviews.llvm.org/D46091
llvm-svn: 331007
Summary:
Use the FP for scavenged spill slot accesses to prevent corruption of
the callee-save region when the SP is re-aligned.
Based on problem and patch reported by @paulwalker-arm
This is an alternative to solution proposed in D45770
Reviewers: t.p.northover, paulwalker-arm, thegameg, javed.absar
Subscribers: qcolombet, mcrosier, paulwalker-arm, kristof.beyls, rengolin, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D46063
llvm-svn: 330976
This patch adds a new shuffle kind useful for transposing a 2xn matrix. These
transpose shuffle masks read corresponding even- or odd-numbered vector
elements from two n-dimensional source vectors and write each result into
consecutive elements of an n-dimensional destination vector. The transpose
shuffle kind is meant to model the TRN1 and TRN2 AArch64 instructions. As such,
this patch also considers transpose shuffles in the AArch64 implementation of
getShuffleCost.
Differential Revision: https://reviews.llvm.org/D45982
llvm-svn: 330941
This patch extends the PredicateMethod of AsmOperands used in SVE's
LD1 instructions with a DiagnosticPredicate. This makes them 'context
sensitive' to the operand that has been parsed and tells the user to
use the right register (with expected shift/extend), rather than telling
the immediate is out of range when it actually parsed a register.
Patch [2/2] in a series to improve assembler diagnostics for SVE:
- Patch [1/2]: https://reviews.llvm.org/D45879
- Patch [2/2]: https://reviews.llvm.org/D45880
Reviewers: olista01, stoklund, craig.topper, mcrosier, rengolin, echristo, fhahn, SjoerdMeijer, evandro, javed.absar
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D45880
llvm-svn: 330934
Before, the outliner would grab ADRPs that used LR/W30. This patch fixes
that by checking for explicit uses of those registers before the special-casing
for ADRPs.
This also adds a test that ensures that those sorts of ADRPs won't be outlined.
llvm-svn: 330783
This reland includes a check to prevent the DAG combiner from folding an
offset that is smaller than the existing one. This can cause oscillations
between two possible DAGs, which was the cause of the hang and later assertion
failure observed on the lnt-ctmark-aarch64-O3-flto bot.
http://green.lab.llvm.org/green/job/lnt-ctmark-aarch64-O3-flto/2024/
Original commit message:
> This is a code size win in code that takes offseted addresses
> frequently, such as C++ constructors that typically need to compute
> an offseted address of a vtable. This reduces the size of Chromium
> for Android's .text section by 108KB.
Differential Revision: https://reviews.llvm.org/D45199
llvm-svn: 330630
First off, this is more correct than having the B. Second off, this was making
a bot upset. This fixes that.
Update the test to include -verify-machineinstrs as well to prevent stuff like
this slipping by non debug/assert builds in the future.
llvm-svn: 330459
Summary:
In some cases the shift/extend needs to be explicitly parsed together
with the register, rather than as a separate operand. This is needed
for addressing modes where the instruction as a whole dictates the
scaling/extend, rather than specific bits in the instruction.
By parsing them as a single operand, we avoid the need to pass an
extra operand in all CodeGen patterns (because all operands need to
have an associated value), and we avoid the need to update TableGen to
accept operands that have no associated bits in the instruction.
An added benefit of parsing them together is that the assembler
can give a sensible diagnostic if the scaling is not correct.
This is patch [2/4] in a series to add assembler/disassembler support for
SVE's contiguous LD1 (scalar+scalar) instructions:
- Patch [1/4]: https://reviews.llvm.org/D45687
- Patch [2/4]: https://reviews.llvm.org/D45688
- Patch [3/4]: https://reviews.llvm.org/D45689
- Patch [4/4]: https://reviews.llvm.org/D45690
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: fhahn, SjoerdMeijer
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45688
llvm-svn: 330394
Summary:
- Renamed tryParseRegister to tryParseScalarRegister, which
now returns an OperandMatchResultTy.
- Moved matching of certain aliases into matchRegisterNameAlias.
- Changed type of most 'Reg' variables to 'unsigned'.
This is patch [1/4] in a series to add assembler/disassembler support for
SVE's contiguous LD1 (scalar+scalar) instructions:
- Patch [1/4]: https://reviews.llvm.org/D45687
- Patch [2/4]: https://reviews.llvm.org/D45688
- Patch [3/4]: https://reviews.llvm.org/D45689
- Patch [4/4]: https://reviews.llvm.org/D45690
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro, samparker
Reviewed By: samparker
Subscribers: samparker, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D45687
llvm-svn: 330311
Summary:
Added instructions for contiguous stores, ST1, with scalar+imm addressing
modes and corresponding tests. The patch also adds parsing of
'mul vl' as needed for the VL-scaled immediate.
This is patch [6/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: rengolin
Subscribers: tschuett, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D45432
llvm-svn: 330014
Summary:
Added Z_(b|h|s|d) vector list RegisterOperands along with support to
add/print the vector lists.
This is patch [5/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: fhahn
Subscribers: tschuett, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45431
llvm-svn: 330000
This is a code size win in code that takes offseted addresses
frequently, such as C++ constructors that typically need to compute
an offseted address of a vtable. This reduces the size of Chromium
for Android's .text section by 108KB.
Differential Revision: https://reviews.llvm.org/D45199
llvm-svn: 329956
AFI->setRedZone(false) was put in the wrong place before, and so it only fired
on functions that didn't have stack frames. This moves that to the top of
emitPrologue to make sure that every function without a redzone has it set
correctly.
This also adds a function representing one of the early exit cases (GHC calling
convention) to the MachineOutliner noredzone test to ensure that we can outline
from functions like these, where we never use a redzone.
llvm-svn: 329922
Summary:
Merged 'addVectorList64Operands' and 'addVectorList128Operands' into a
generic 'addVectorListOperands', which can be easily extended to work
for SVE vectors.
This is patch [4/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: rengolin
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45430
llvm-svn: 329909
Summary:
Added 'RegisterKind' to the VectorListOp structure, so that this operand
type can be reused for SVE vector lists in a later patch. It also
refactors the 'tryParseVectorList' function so it can be used directly
in the ParserMethod of an operand. The parsing can now parse multiple
kinds of vectors and recover if there is no match.
This is patch [3/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: rengolin
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45429
llvm-svn: 329900
Summary:
Place parsing of a vector index into a separate function to reduce
duplication, since the code is duplicated in both the parsing of a
Neon vector register operand and a Neon vector list.
This is patch [2/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: rengolin
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45428
llvm-svn: 329809
In r329691, we would choose FP even if the offset wouldn't fit, just
because the offset is smaller than the one from BP. This made many
accesses through FP need to scavenge a register, which resulted in
slower and bigger code for no good reason.
This patch now always picks the offset that fits first, even if FP is
preferred.
llvm-svn: 329797
Summary:
Merged 'tryMatchVectorRegister' (specific to Neon) and
'tryParseSVERegister' into a single 'tryParseVectorRegister' function, and
created a generic 'parseVectorKind()' function that returns the #Elements
and ElementWidth of a vector suffix. This reduces the duplication of
this functionality between two the vector implementations.
This is patch [1/6] in a series to add assembler/disassembler support for
SVE's contiguous ST1 (scalar+imm) instructions.
Reviewers: fhahn, rengolin, javed.absar, huntergr, SjoerdMeijer, t.p.northover, echristo, evandro
Reviewed By: fhahn
Subscribers: tschuett, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D45427
llvm-svn: 329782
Summary:
When inserting MOVs to avoid Falkor HWPF collisions, the non-base
register operand of load instructions (e.g. a register offset) was not
being considered live, so it could potentially have been used as a
scratch register, clobbering the actual offset value.
Reviewers: mcrosier
Subscribers: rengolin, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45502
llvm-svn: 329761
This commit fixes the bot failures that were coming up before with r329716.
The fix was to move the check for "isInSection()" inside of the if condition
and emit the error there instead of waiting to get past the unreachable statement.
This should work in debug and release builds now.
llvm-svn: 329746
There was missing nullptr check before a call to getSection() in
recordRelocation. This would result in a segfault in code like the attached
test.
This adds the missing check and a test which makes sure we get the expected
error output.
llvm-svn: 329716
In the presence of variable-sized stack objects, we always picked the
base pointer when resolving frame indices if it was available.
This makes us hit an assert where we can't reach the emergency spill
slot if it's too far away from the base pointer. Since on AArch64 we
decide to place the emergency spill slot at the top of the frame, it
makes more sense to use FP to access it.
The changes here don't affect only emergency spill slots but all the
frame indices. The goal here is to try to choose between FP, BP and SP
so that we minimize the offset and avoid scavenging, or worse, asserting
when trying to access a slot allocated by the scavenger.
Previously discussed here: https://reviews.llvm.org/D40876.
Differential Revision: https://reviews.llvm.org/D45358
llvm-svn: 329691
Much like any written register in load/store instructions, the status register
is not allowed to overlap with any others. So diagnose it like we already do
with the other cases.
llvm-svn: 329687
Lower is slightly odd. It often doesn't change the type but the lowerings
do use the new type to decide what code to create. Treat it like a mutation
but provide convenience functions that re-use the existing type.
Re-uses the existing tests:
test/CodeGen/AArch64/GlobalISel/legalize-rem.mir
test/CodeGen/AArch64/GlobalISel//legalize-mul.mir
test/CodeGen/AArch64/GlobalISel//legalize-cmpxchg-with-success.mir
llvm-svn: 329623
This is a code size win in code that takes offseted addresses
frequently, such as C++ constructors that typically need to compute
an offseted address of a vtable. It reduces the size of Chromium for
Android's .text section by 46KB, or 56KB with ThinLTO (which exposes
more opportunities to use a direct access rather than a GOT access).
Because the addend range is limited in COFF and Mach-O, this is
enabled for ELF only.
Differential Revision: https://reviews.llvm.org/D45199
llvm-svn: 329611
The TargetSchedModel is always initialized using the TargetSubtargetInfo's
MCSchedModel and TargetInstrInfo, so we don't need to extract those and
pass 3 parameters to init().
Differential Revision: https://reviews.llvm.org/D44789
llvm-svn: 329540
The implementation of shadow call stack on aarch64 is quite different to
the implementation on x86_64. Instead of reserving a segment register for
the shadow call stack, we reserve the platform register, x18. Any function
that spills lr to sp also spills it to the shadow call stack, a pointer to
which is stored in x18.
Differential Revision: https://reviews.llvm.org/D45239
llvm-svn: 329236
The MachineOutliner has a bunch of target hooks that will call llvm_unreachable
if the target doesn't implement them. Therefore, if you enable the outliner on
such a target, it'll just crash. It'd be much better if it'd just *not* run
the outliner at all in this case.
This commit adds a hook to TargetInstrInfo that returns false by default.
Targets that implement the hook make it return true. The outliner checks the
return value of this hook to decide whether or not to continue.
llvm-svn: 329220
Summary:
r327219 added wrappers to std::sort which randomly shuffle the container before sorting.
This will help in uncovering non-determinism caused due to undefined sorting
order of objects having the same key.
To make use of that infrastructure we need to invoke llvm::sort instead of std::sort.
Note: This patch is one of a series of patches to replace *all* std::sort to llvm::sort. Refer the comments section in D44363 for a list of all the required patches.
Reviewers: t.p.northover, jmolloy, RKSimon, rengolin
Reviewed By: rengolin
Subscribers: dexonsmith, rengolin, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D44853
llvm-svn: 329216
Makes it easier to see mistakes such as the one fixed in r329178 and makes
the different target CMakeLists more consistent.
Also remove some stale-looking comments from the Nios2 target cmakefile.
No intended behavior change.
llvm-svn: 329181
This patch adds a hasRedZone() function to AArch64MachineFunctionInfo. It
returns true if the function is known to use a redzone, false if it is known
to not use a redzone, and no value otherwise.
This removes the requirement to pass -mno-red-zone when outlining for AArch64.
https://reviews.llvm.org/D45189
llvm-svn: 329120
Currently EVT is in the IR layer only because of Function.cpp needing a very small piece of the functionality of EVT::getEVTString(). The rest of EVT is used in codegen making CodeGen a better place for it.
The previous code converted a Type* to EVT and then called getEVTString. This was only expected to handle the primitive types from Type*. Since there only a few primitive types, we can just print them as strings directly.
Differential Revision: https://reviews.llvm.org/D45017
llvm-svn: 328806
This commit simplifies the call outlining logic by removing references to the
Function associated with the callee. To do this, it requires that valid
callee save info is available to the outliner.
llvm-svn: 328719
If an ADRP appears with, say, a CPI operand, we shouldn't outline it.
This moves the check for unsafe operands so that it occurs before the special-case
for ADRPs. Also add a test for outlining ADRPs.
llvm-svn: 328674
Summary:
This is a canonical way to teach objdump to print the target
symbols for branches when disassembling AArch64 code.
Reviewers: evandro, t.p.northover, espindola
Reviewed By: t.p.northover
Differential Revision: https://reviews.llvm.org/D44851
llvm-svn: 328638
This is used by llvm tblgen as well as by LLVM Targets, so the only
common place is Support for now. (maybe we need another target for these
sorts of things - but for now I'm at least making them correct & we can
make them better if/when people have strong feelings)
llvm-svn: 328395
Loads and stores can only shift the offset register by the size of the value
being loaded, but currently the DAGCombiner will reduce the width of the load
if it's followed by a trunc making it impossible to later combine the shift.
Solve this by implementing shouldReduceLoadWidth for the AArch64 backend and
make it prevent the width reduction if this is what would happen, though do
allow it if reducing the load width will let us eliminate a later sign or zero
extend.
Differential Revision: https://reviews.llvm.org/D44794
llvm-svn: 328321
Summary:
This pass sinks COPY instructions into a successor block, if the COPY is not
used in the current block and the COPY is live-in to a single successor
(i.e., doesn't require the COPY to be duplicated). This avoids executing the
the copy on paths where their results aren't needed. This also exposes
additional opportunites for dead copy elimination and shrink wrapping.
These copies were either not handled by or are inserted after the MachineSink
pass. As an example of the former case, the MachineSink pass cannot sink
COPY instructions with allocatable source registers; for AArch64 these type
of copy instructions are frequently used to move function parameters (PhyReg)
into virtual registers in the entry block..
For the machine IR below, this pass will sink %w19 in the entry into its
successor (%bb.1) because %w19 is only live-in in %bb.1.
```
%bb.0:
%wzr = SUBSWri %w1, 1
%w19 = COPY %w0
Bcc 11, %bb.2
%bb.1:
Live Ins: %w19
BL @fun
%w0 = ADDWrr %w0, %w19
RET %w0
%bb.2:
%w0 = COPY %wzr
RET %w0
```
As we sink %w19 (CSR in AArch64) into %bb.1, the shrink-wrapping pass will be
able to see %bb.0 as a candidate.
With this change I observed 12% more shrink-wrapping candidate and 13% more dead copies deleted in spec2000/2006/2017 on AArch64.
Reviewers: qcolombet, MatzeB, thegameg, mcrosier, gberry, hfinkel, john.brawn, twoh, RKSimon, sebpop, kparzysz
Reviewed By: sebpop
Subscribers: evandro, sebpop, sfertile, aemerson, mgorny, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D41463
llvm-svn: 328237
When outlining calls, the outliner needs to update CFI to ensure that, say,
exception handling works. This commit adds that functionality and adds a test
just for call outlining.
Call outlining stuff in machine-outliner.mir should be moved into
machine-outliner-calls.mir in a later commit.
llvm-svn: 327917
This extends the use of this attribute on ARM and AArch64 from
SVN r325900 (where it was only checked for fixed stack
allocations on ARM/AArch64, but for all stack allocations on X86).
This also adds a testcase for the existing use of disabling the
fixed stack probe with the attribute on ARM and AArch64.
Differential Revision: https://reviews.llvm.org/D44291
llvm-svn: 327897
Summary:
The docs already claim that this happens, but so far it hasn't. As a
consequence, existing TableGen files get this wrong a lot, but luckily
the fixes are all reasonably straightforward.
To make this work with all the existing forms of self-references (since
the true type of a record is only built up over time), the lookup of
self-references in !cast is delayed until the final resolving step.
Change-Id: If5923a72a252ba2fbc81a889d59775df0ef31164
Reviewers: arsenm, craig.topper, tra, MartinO
Subscribers: wdng, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D44475
llvm-svn: 327849
This is similar to the check later when we remap some of the instructions from one class to a new one. But if we reuse the class we don't get to do that check.
So many CPUs have violations of this check that I had to add a flag to the SchedMachineModel to allow it to be disabled. Hopefully we can get those cleaned up quickly and remove this flag.
A lot of the violations are due to overlapping regular expressions, but that's not the only kind of issue it found.
llvm-svn: 327808
At the point the outliner runs, KILLs don't impact anything, but they're still
considered unique instructions. This commit makes them invisible like
DebugValues so that they can still be outlined without impacting outlining
decisions.
llvm-svn: 327760
This patch provides an implementation of getArithmeticReductionCost for
AArch64. We can specialize the cost of add reductions since they are computed
using the 'addv' instruction.
Differential Revision: https://reviews.llvm.org/D44490
llvm-svn: 327702
Optionally allow the order of restoring the callee-saved registers in the
epilogue to be reversed.
The flag -reverse-csr-restore-seq generates the following code:
```
stp x26, x25, [sp, #-64]!
stp x24, x23, [sp, #16]
stp x22, x21, [sp, #32]
stp x20, x19, [sp, #48]
; [..]
ldp x24, x23, [sp, #16]
ldp x22, x21, [sp, #32]
ldp x20, x19, [sp, #48]
ldp x26, x25, [sp], #64
ret
```
Note how the CSRs are restored in the same order as they are saved.
One exception to this rule is the last `ldp`, which allows us to merge
the stack adjustment and the ldp into a post-index ldp. This is done by
first generating:
ldp x26, x27, [sp]
add sp, sp, #64
which gets merged by the arm64 load store optimizer into
ldp x26, x25, [sp], #64
The flag is disabled by default.
llvm-svn: 327569
Summary:
Fixes an UB caught by sanitizer. The shift amount might be larger than 32 so the operand should be 1ULL.
In this patch, we replace the original expression with existing API with uint64_t type.
Reviewers: eli.friedman, rengolin
Reviewed By: rengolin
Subscribers: rengolin, javed.absar, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D44234
llvm-svn: 326969
Since there is no instruction for integer vector division, factor in the
cost of singling out each element to be used with the scalar division
instruction.
Differential revision: https://reviews.llvm.org/D43974
llvm-svn: 326955
The attached testcase started failing after the patch to define
isExtractSubvectorCheap with the following pattern mismatch:
ISEL: Starting pattern match
Initial Opcode index to 85068
Match failed at index 85076
LLVM ERROR: Cannot select: t47: v8i16 = insert_subvector undef:v8i16, t43, Constant:i64<0>
The code generated from llvm/lib/Target/AArch64/AArch64InstrInfo.td
def : Pat<(insert_subvector undef, (v4i16 FPR64:$src), (i32 0)),
(INSERT_SUBREG (v8i16 (IMPLICIT_DEF)), FPR64:$src, dsub)>;
is in ninja/lib/Target/AArch64/AArch64GenDAGISel.inc
At the location of the error it is:
/* 85076*/ OPC_CheckChild2Type, MVT::i32,
And it failed to match the type of operand 2.
Adding another def-pat for i64 fixes the failed def-pat error:
def : Pat<(insert_subvector undef, (v4i16 FPR64:$src), (i64 0)),
(INSERT_SUBREG (v8i16 (IMPLICIT_DEF)), FPR64:$src, dsub)>;
llvm-svn: 326949
Following the ARM-neon backend, define isExtractSubvectorCheap to return true
when extracting low and high part of a neon register.
The patch disables a test in llvm/test/CodeGen/AArch64/arm64-ext.ll This
testcase is fragile in the sense that it requires a BUILD_VECTOR to "survive"
all DAG transforms until ISelLowering. The testcase is supposed to check that
AArch64TargetLowering::ReconstructShuffle() works, and for that we need a
BUILD_VECTOR in ISelLowering. As we now transform the BUILD_VECTOR earlier into
an VEXT + vector_shuffle, we don't have the BUILD_VECTOR pattern when we get to
ISelLowering. As there is no way to disable the combiner to only exercise the
code in ISelLowering, the patch disables the testcase.
Differential revision: https://reviews.llvm.org/D43973
llvm-svn: 326811
The error occurs when reading i16 elements (as in the testcase) from a v8i8
with a pattern of <0,2,4,6>. As all the data in the vector is accessed, the
operation is not a VUZP. The patch stops the pattern recognition of VUZP when
EXTRACT_VECTOR_ELT has a different element type than BUILD_VECTOR.
llvm-svn: 326722
Use the whole gammut of constant immediates available to set up a vector.
Instead of using, for example, `mov w0, #0xffff; dup v0.4s, w0`, which
transfers between register files, use the more efficient `movi v0.4s, #-1`
instead. Not limited to just a few values, but any immediate value that can
be encoded by all the variants of `FMOV`, `MOVI`, `MVNI`, thus eliminating
the need to there be patterns to optimize special cases.
Differential revision: https://reviews.llvm.org/D42133
llvm-svn: 326718
when a BUILD_VECTOR is created out of a sequence of EXTRACT_VECTOR_ELT with a
specific pattern sequence, either <0, 2, 4, ...> or <1, 3, 5, ...>, replace the
BUILD_VECTOR with either vuzp1 or vuzp2.
With this patch LLVM generates the following code for the first function fun1 in the testcase:
adrp x8, .LCPI0_0
ldr q0, [x8, :lo12:.LCPI0_0]
tbl v0.16b, { v0.16b }, v0.16b
ext v1.16b, v0.16b, v0.16b, #8
uzp1 v0.8b, v0.8b, v1.8b
str d0, [x8]
ret
Without this patch LLVM currently generates this code:
adrp x8, .LCPI0_0
ldr q0, [x8, :lo12:.LCPI0_0]
tbl v0.16b, { v0.16b }, v0.16b
mov v1.16b, v0.16b
mov v1.b[1], v0.b[2]
mov v1.b[2], v0.b[4]
mov v1.b[3], v0.b[6]
mov v1.b[4], v0.b[8]
mov v1.b[5], v0.b[10]
mov v1.b[6], v0.b[12]
mov v1.b[7], v0.b[14]
str d1, [x8]
ret
llvm-svn: 326443
Emulated TLS is enabled by llc flag -emulated-tls,
which is passed by clang driver.
When llc is called explicitly or from other drivers like LTO,
missing -emulated-tls flag would generate wrong TLS code for targets
that supports only this mode.
Now use useEmulatedTLS() instead of Options.EmulatedTLS to decide whether
emulated TLS code should be generated.
Unit tests are modified to run with and without the -emulated-tls flag.
Differential Revision: https://reviews.llvm.org/D42999
llvm-svn: 326341
Currently we assert that only non target specific opcodes can have
missing RegisterClass constraints in the MCDesc. The backend can have
instructions with register operands but don't have RegisterClass
constraints (say using unknown_class) in which case the instruction
defining the register will constrain it.
Change the assert to only fire if a def has no regclass.
https://reviews.llvm.org/D43409
llvm-svn: 326142
This feature enables the fusion of the comparison and the conditional select
instructions together.
Differential revision: https://reviews.llvm.org/D42392
llvm-svn: 325939
Summary:
Add a target option AllowRegisterRenaming that is used to opt in to
post-register-allocation renaming of registers. This is set to 0 by
default, which causes the hasExtraSrcRegAllocReq/hasExtraDstRegAllocReq
fields of all opcodes to be set to 1, causing
MachineOperand::isRenamable to always return false.
Set the AllowRegisterRenaming flag to 1 for all in-tree targets that
have lit tests that were effected by enabling COPY forwarding in
MachineCopyPropagation (AArch64, AMDGPU, ARM, Hexagon, Mips, PowerPC,
RISCV, Sparc, SystemZ and X86).
Add some more comments describing the semantics of the
MachineOperand::isRenamable function and how it is set and maintained.
Change isRenamable to check the operand's opcode
hasExtraSrcRegAllocReq/hasExtraDstRegAllocReq bit directly instead of
relying on it being consistently reflected in the IsRenamable bit
setting.
Clear the IsRenamable bit when changing an operand's register value.
Remove target code that was clearing the IsRenamable bit when changing
registers/opcodes now that this is done conservatively by default.
Change setting of hasExtraSrcRegAllocReq in AMDGPU target to be done in
one place covering all opcodes that have constant pipe read limit
restrictions.
Reviewers: qcolombet, MatzeB
Subscribers: aemerson, arsenm, jyknight, mcrosier, sdardis, nhaehnle, javed.absar, tpr, arichardson, kristof.beyls, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, jordy.potman.lists, apazos, sabuasal, niosHD, escha, nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D43042
llvm-svn: 325931
Move checks for each fusion case into separate functions for better
legibility and maintainability.
Differential revision: https://reviews.llvm.org/D43649
llvm-svn: 325844
Get rid of icky goto loops and make the code easier to maintain. Otherwise,
NFC.
Restore r324903 and fix PR36369.
Differentail revision: https://reviews.llvm.org/D43364
llvm-svn: 325621
This is a follow on commit to r[x] where we fix the other direction of copy.
For this case, after converting the source from gpr32 -> fpr32, we use a
subregister copy, which is essentially what EXTRACT_SUBREG does in SDAG land.
https://reviews.llvm.org/D43444
llvm-svn: 325550
This makes sure that alloca() function calls properly probe the
stack as needed.
Differential Revision: https://reviews.llvm.org/D42356
llvm-svn: 325433
The data type is assumed to be a vector, but sometimes it is not, leading
to an assertion.
Add simple test-case to verify this.
Differential revision: https://reviews.llvm.org/D42599
llvm-svn: 325378
* Document most API's
* Delete a useless function call
* Fix a discrepancy between the single and multi-opcode variants of
getActionDefinitions().
The multi-opcode variant now requires that more than one opcode is requested.
Previously it acted much like the single-opcode form but unnecessarily
enforced the requirements of the multi-opcode form.
llvm-svn: 325067
It caused "Cannot select: t33: f64 = AArch64ISD::FMOV Constant:i32<0>"
in Chromium builds. See PR36369.
> Get rid of icky goto loops and make the code easier to maintain (NFC).
>
> Differential revision: https://reviews.llvm.org/D42723
llvm-svn: 325034
Armv8.1-A added an atomic load-clear instruction (which performs bitwise
and with the complement of it's operand), but not a load-and
instruction. Our current code-generation for atomic load-and always
inserts an MVN instruction to invert its argument, even if it could be
folded into a constant or another instruction.
This adds lowering early in selection DAG to convert a load-and
operation into an xor with -1 and a load-clear, allowing the normal DAG
optimisations to work on it.
To do this, I've had to add a new ISD opcode, ATOMIC_LOAD_CLR. I don't
see any easy way to do this with an AArch64-specific ISD node, because
the code-generation for atomic operations assumes the SDNodes are of
type AtomicSDNode.
I've left the old tablegen patterns in because they are still needed for
global isel.
Differential revision: https://reviews.llvm.org/D42478
llvm-svn: 324908
Armv8.1-A added an atomic load-add instruction, but not a load-subtract
instruction. Our current code-generation for atomic load-subtract always
inserts a NEG instruction to negate it's argument, even if it could be
folded into a constant or another instruction.
This adds lowering early in selection DAG to convert a load-subtract
operation into a subtract and a load-add, allowing the normal DAG
optimisations to work on it.
I've left the old tablegen patterns in because they are still needed for
global isel.
Some of the tests in this patch are copied from D35375 by Chad Rosier (which
was abandoned).
Differential revision: https://reviews.llvm.org/D42477
llvm-svn: 324892
Summary:
This change is part of step five in the series of changes to remove alignment argument from
memcpy/memmove/memset in favour of alignment attributes. In particular, this changes
AArch64FastISel to cease using the old getAlignment() API of MemoryIntrinsic in favour of getting
source & dest specific alignments through the new API.
Steps:
Step 1) Remove alignment parameter and create alignment parameter attributes for
memcpy/memmove/memset. ( rL322965, rC322964, rL322963 )
Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing
source and dest alignments. ( rL323597 )
Step 3) Update Clang to use the new IRBuilder API. ( rC323617 )
Step 4) Update Polly to use the new IRBuilder API. ( rL323618 )
Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API,
and those that use use MemIntrinsicInst::[get|set]Alignment() to use [get|set]DestAlignment()
and [get|set]SourceAlignment() instead. ( rL323886, r323891, rL324148, rL324273, rL324278,
rL324384, rL324395, rL324402, rL324626, rL324642, rL324653, rL324654 )
Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the
MemIntrinsicInst::[get|set]Alignment() methods.
Reference
http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.htmlhttp://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
llvm-svn: 324773
Enable multiple COPY hints to eliminate more COPYs during register allocation.
Note that this is something all targets should do, see
https://reviews.llvm.org/D38128.
Review: Martin Storsjö
llvm-svn: 324720
We were generating "fmov h0, wzr" instructions when FullFP16 is not enabled.
I've not added any tests, because the problem was visible in:
test/CodeGen/AArch64/arm64-zero-cycle-zeroing.ll,
which I had to change: I don't think Cyclone has FullFP16 enabled
by default, so it shouldn't be using this v8.2a instruction.
I've also removed these rdar tags, please shout if there are any objections.
Differential Revision: https://reviews.llvm.org/D43020
llvm-svn: 324581
Summary: Adds support for the SVE AND instruction with vector and logical-immediate operands, and their corresponding aliases.
Reviewers: fhahn, rengolin, samparker, echristo, aadg, kristof.beyls
Reviewed By: fhahn
Subscribers: aemerson, javed.absar, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D42295
llvm-svn: 324343
This fixes a crash where the user is a COPY, which deliberately does not
constrain its source operands, resulting in a vreg without a reg class escaping
selection.
Differential Revision: https://reviews.llvm.org/D42697
llvm-svn: 324047
I added this comment with D42323, but as discussed in D42806, the architecture
does the right thing for denorms. We don't even need the select on 0.0 here?
llvm-svn: 323996
As shown in the example in PR34994:
https://bugs.llvm.org/show_bug.cgi?id=34994
...we can return a very wrong answer (inf instead of 0.0) for square root when
using a reciprocal square root estimate instruction.
Here, I've conditionalized the filtering out of denorms based on the function
having "denormal-fp-math"="ieee" in its attributes. The other options for this
attribute are 'preserve-sign' and 'positive-zero'.
So we don't generate this extra code by default with just '-ffast-math' (because
then there's no denormal attribute string at all), but it works if you specify
'-ffast-math -fdenormal-fp-math=ieee' from clang.
As noted in the review, there may be other problems in clang that affect the
results depending on platform (Linux x86 at least), but this should allow
creating the desired codegen.
Differential Revision: https://reviews.llvm.org/D42323
llvm-svn: 323981
This feature enables the fusion of the address generation and a
corresponding load or store together.
Differential revision: https://reviews.llvm.org/D42393
llvm-svn: 323782
This feature enables special handling of cheap as move in the existing
custom handling specifically for Exynos processors.
Differential revision: https://reviews.llvm.org/D42387
llvm-svn: 323774
This reverts commit r322917 due to multiple performance regressions in spec2006
and spec2017. XFAILed llvm/test/CodeGen/AArch64/big-callframe.ll which initially
motivated this change.
llvm-svn: 323683
Summary:
As discussed in D42244, we have difficulty describing the legality of some
operations. We're not able to specify relationships between types.
For example, declaring the following
setAction({..., 0, s32}, Legal)
setAction({..., 0, s64}, Legal)
setAction({..., 1, s32}, Legal)
setAction({..., 1, s64}, Legal)
currently declares these type combinations as legal:
{s32, s32}
{s64, s32}
{s32, s64}
{s64, s64}
but we currently have no means to say that, for example, {s64, s32} is
not legal. Some operations such as G_INSERT/G_EXTRACT/G_MERGE_VALUES/
G_UNMERGE_VALUES have relationships between the types that are currently
described incorrectly.
Additionally, G_LOAD/G_STORE currently have no means to legalize non-atomics
differently to atomics. The necessary information is in the MMO but we have no
way to use this in the legalizer. Similarly, there is currently no way for the
register type and the memory type to differ so there is no way to cleanly
represent extending-load/truncating-store in a way that can't be broken by
optimizers (resulting in illegal MIR).
It's also difficult to control the legalization strategy. We've added support
for legalizing non-power of 2 types but there's still some hardcoded assumptions
about the strategy. The main one I've noticed is that type0 is always legalized
before type1 which is not a good strategy for `type0 = G_EXTRACT type1, ...` if
you need to widen the container. It will converge on the same result eventually
but it will take a much longer route when legalizing type0 than if you legalize
type1 first.
Lastly, the definition of legality and the legalization strategy is kept
separate which is not ideal. It's helpful to be able to look at a one piece of
code and see both what is legal and the method the legalizer will use to make
illegal MIR more legal.
This patch adds a layer onto the LegalizerInfo (to be removed when all targets
have been migrated) which resolves all these issues.
Here are the rules for shift and division:
for (unsigned BinOp : {G_LSHR, G_ASHR, G_SDIV, G_UDIV})
getActionDefinitions(BinOp)
.legalFor({s32, s64}) // If type0 is s32/s64 then it's Legal
.clampScalar(0, s32, s64) // If type0 is <s32 then WidenScalar to s32
// If type0 is >s64 then NarrowScalar to s64
.widenScalarToPow2(0) // Round type0 scalars up to powers of 2
.unsupported(); // Otherwise, it's unsupported
This describes everything needed to both define legality and describe how to
make illegal things legal.
Here's an example of a complex rule:
getActionDefinitions(G_INSERT)
.unsupportedIf([=](const LegalityQuery &Query) {
// If type0 is smaller than type1 then it's unsupported
return Query.Types[0].getSizeInBits() <= Query.Types[1].getSizeInBits();
})
.legalIf([=](const LegalityQuery &Query) {
// If type0 is s32/s64/p0 and type1 is a power of 2 other than 2 or 4 then it's legal
// We don't need to worry about large type1's because unsupportedIf caught that.
const LLT &Ty0 = Query.Types[0];
const LLT &Ty1 = Query.Types[1];
if (Ty0 != s32 && Ty0 != s64 && Ty0 != p0)
return false;
return isPowerOf2_32(Ty1.getSizeInBits()) &&
(Ty1.getSizeInBits() == 1 || Ty1.getSizeInBits() >= 8);
})
.clampScalar(0, s32, s64)
.widenScalarToPow2(0)
.maxScalarIf(typeInSet(0, {s32}), 1, s16) // If type0 is s32 and type1 is bigger than s16 then NarrowScalar type1 to s16
.maxScalarIf(typeInSet(0, {s64}), 1, s32) // If type0 is s64 and type1 is bigger than s32 then NarrowScalar type1 to s32
.widenScalarToPow2(1) // Round type1 scalars up to powers of 2
.unsupported();
This uses a lambda to say that G_INSERT is unsupported when type0 is bigger than
type1 (in practice, this would be a default rule for G_INSERT). It also uses one
to describe the legal cases. This particular predicate is equivalent to:
.legalFor({{s32, s1}, {s32, s8}, {s32, s16}, {s64, s1}, {s64, s8}, {s64, s16}, {s64, s32}})
In terms of performance, I saw a slight (~6%) performance improvement when
AArch64 was around 30% ported but it's pretty much break even right now.
I'm going to take a look at constexpr as a means to reduce the initialization
cost.
Future work:
* Make it possible for opcodes to share rulesets. There's no need for
G_LSHR/G_ASHR/G_SDIV/G_UDIV to have separate rule and ruleset objects. There's
no technical barrier to this, it just hasn't been done yet.
* Replace the type-index numbers with an enum to get .clampScalar(Type0, s32, s64)
* Better names for things like .maxScalarIf() (clampMaxScalar?) and the vector rules.
* Improve initialization cost using constexpr
Possible future work:
* It's possible to make these rulesets change the MIR directly instead of
returning a description of how to change the MIR. This should remove a little
overhead caused by parsing the description and routing to the right code, but
the real motivation is that it removes the need for LegalizeAction::Custom.
With Custom removed, there's no longer a requirement that Custom legalization
change the opcode to something that's considered legal.
Reviewers: ab, t.p.northover, qcolombet, rovka, aditya_nandakumar, volkan, reames, bogner
Reviewed By: bogner
Subscribers: hintonda, bogner, aemerson, mgorny, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D42251
llvm-svn: 323681
Summary:
The improvements to the LegalizerInfo discussed in D42244 require that
LegalizerInfo::LegalizeAction be available for use in other classes. As such,
it needs to be moved out of LegalizerInfo. This has been done separately to the
next patch to minimize the noise in that patch.
llvm-svn: 323669
Summary:
All variants of isLogicalImm[Not](32|64) can be combined into a single templated function, same for printLogicalImm(32|64).
By making it use a template instead, further SVE patches can use it for other data types as well (e.g. 8, 16 bits).
Reviewers: fhahn, rengolin, aadg, echristo, kristof.beyls, samparker
Reviewed By: samparker
Subscribers: aemerson, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D42294
llvm-svn: 323646
The Large System Extension added an atomic compare-and-swap instruction
that operates on a pair of 64-bit registers, which we can use to
implement a 128-bit cmpxchg.
Because i128 is not a legal type for AArch64 we have to do all of the
instruction selection in C++, and the instruction requires even/odd
register pairs, so we have to wrap it in REG_SEQUENCE and EXTRACT_SUBREG
nodes. This is very similar to what we do for 64-bit cmpxchg in the ARM
backend.
Differential revision: https://reviews.llvm.org/D42104
llvm-svn: 323634
Previously some targets printed their own message at the start of Select to indicate what they were selecting. For the targets that didn't, it means there was no print of the root node before any custom handling in the target executed. So if the target did something custom and never called SelectNodeCommon, no print would be made. For the targets that did print a message in Select, if they didn't custom handle a node SelectNodeCommon would reprint the root node before walking the isel table.
It seems better to just print the message before the call to Select so all targets behave the same. And then remove the root node printing from SelectNodeCommon and just leave a message that says we're starting the table search.
There were also some oddities in blank line behavior. Usually due to a \n after a call to SelectionDAGNode::dump which already inserted a new line.
llvm-svn: 323551
This patch enables aggressive FMA by default on T99, and provides a -mllvm
option to enable the same on other AArch64 micro-arch's (-mllvm
-aarch64-enable-aggressive-fma).
Test case demonstrating the effects on T99 is included.
Patch by: steleman (Stefan Teleman)
Differential Revision: https://reviews.llvm.org/D40696
llvm-svn: 323474
The tablegen imported patterns for sext(load(a)) don't check for single uses
of the load or delete the original after matching. As a result two loads are
left in the generated code. This particular issue will be fixed by adding
support for a G_SEXTLOAD opcode in future.
There are however other potential issues around this that wouldn't be fixed by
a G_SEXTLOAD, so until we have a proper solution we don't try to handle volatile
loads at all in the AArch64 selector.
Fixes/works around PR36018.
llvm-svn: 323371
Summary:
Loads/stores of some NEON vector types are promoted to other vector
types with different lane sizes but same vector size. This is not a
problem in little-endian but, when in big-endian, it requires
additional byte reversals required to preserve the lane ordering
while keeping the right endianness of the data inside each lane.
For example:
%1 = load <4 x half>, <4 x half>* %p
results in the following assembly:
ld1 { v0.2s }, [x1]
rev32 v0.4h, v0.4h
This patch changes the promotion of these loads/stores so that the
actual vector load/store (LD1/ST1) takes care of the endianness
correctly and there is no need for further byte reversals. The
previous code now results in the following assembly:
ld1 { v0.4h }, [x1]
Reviewers: olista01, SjoerdMeijer, efriedma
Reviewed By: efriedma
Subscribers: aemerson, rengolin, javed.absar, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D42235
llvm-svn: 323325
Remove FeatureSlowMisaligned128Store from cyclone flags.
This flag causes splitting of 16 byte wide stores into 2 stored of 8
bytes. This was useful on older apple CPUs which were slow for 16byte
stores that were not aligned on 16byte. As the compiler often cannot
predict the actual alignment, the splitting was choosen.
This has been a topic for a lot of debate as the splitting also
decreases performance for some benchmarks. Measuring the effects on
newer apple chips (rdar://35525421) shows that it harms more cases than
it helps. So it is time to retire this workaround.
llvm-svn: 323289
Some nodes produce multiple values so when obtaining the type of an ISD::OR we
need to make sure we ask for the correct one. Hopefully that's all of them.
llvm-svn: 323205
Improves the code generation for v4f16 FCMP instructions when FullFP16 is not supported.
Generating FCTVL(s) rather than a longer series of FCVTs.
Differential Revision: https://reviews.llvm.org/D41772
llvm-svn: 323118
Summary:
This patch adds support for parsing/printing of named or unnamed
patterns that are used in SVE's PTRUE instruction, amongst others.
The pattern can be specified as a named pattern to initialize the predicate
vector or it can be specified as an immediate in the range 0-31.
Reviewers: fhahn, rengolin, evandro, mcrosier, t.p.northover
Reviewed By: fhahn
Subscribers: aemerson, javed.absar, tschuett, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D41818
llvm-svn: 323098
Re-commit of r322200: The testcase shouldn't hit machineverifiers
anymore with r322917 in place.
Large callframes (calls with several hundreds or thousands or
parameters) could lead to situations in which the emergency spillslot is
out of range to be addressed relative to the stack pointer.
This commit forces the use of a frame pointer in the presence of large
callframes.
This commit does several things:
- Compute max callframe size at the end of instruction selection.
- Add mirFileLoaded target callback. Use it to compute the max callframe size
after loading a .mir file when the size wasn't specified in the file.
- Let TargetFrameLowering::hasFP() return true if there exists a
callframe > 255 bytes.
- Always place the emergency spillslot close to FP if we have a frame
pointer.
- Note that `useFPForScavengingIndex()` would previously return false
when a base pointer was available leading to the emergency spillslot
getting allocated late (that's the whole effect of this callback).
Which made no sense to me so I took this case out: Even though the
emergency spillslot is technically not referenced by FP in this case
we still want it allocated early.
Differential Revision: https://reviews.llvm.org/D40876
llvm-svn: 322919
Do not create CALLSEQ_START/CALLSEQ_END when there is no callframe to
setup and the callframe size is 0.
- Fixes an invalid callframe nesting for byval arguments, which would
look like this before this patch (as in `big-byval.ll`):
...
ADJCALLSTACKDOWN 32768, 0, ... # Setup for extfunc
...
ADJCALLSTACKDOWN 0, 0, ... # setup for memcpy
...
BL &memcpy ...
ADJCALLSTACKUP 0, 0, ... # destroy for memcpy
...
BL &extfunc
ADJCALLSTACKUP 32768, 0, ... # destroy for extfunc
- Saves us two instructions in the common case of zero-sized stackframes.
- Remove an unnecessary scheduling barrier (hence the small unittest
changes).
Differential Revision: https://reviews.llvm.org/D42006
llvm-svn: 322917
Every known PE COFF target emits /EXPORT: linker flags into a .drective
section. The AsmPrinter should handle this.
While we're at it, use global_values() and emit each export flag with
its own .ascii directive. This should make the .s file output more
readable.
llvm-svn: 322788
Summary:
This patch adds a new target option in order to control GlobalISel.
This will allow the users to enable/disable GlobalISel prior to the
backend by calling `TargetMachine::setGlobalISel(bool Enable)`.
No test case as there is already a test to check GlobalISel
command line options.
See: CodeGen/AArch64/GlobalISel/gisel-commandline-option.ll.
Reviewers: qcolombet, aemerson, ab, dsanders
Reviewed By: qcolombet
Subscribers: rovka, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D42137
llvm-svn: 322773
Summary:
Loading a vector of 4 half-precision FP sometimes results in an LD1
of 2 single-precision FP + a reversal. This results in an incorrect
byte swap due to the conversion from little endian to big endian.
In order to generate the correct byte swap, it is easier to
generate the correct LD1 of 4 half-precision FP, thus avoiding the
subsequent reversal.
Reviewers: craig.topper, jmolloy, olista01
Reviewed By: olista01
Subscribers: efriedma, samparker, SjoerdMeijer, rogfer01, aemerson, rengolin, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D41863
llvm-svn: 322663
Summary:
This patch adds CustomRenderer which renders the matched
operands to the specified instruction.
Targets can enable the matching of SDNodeXForm by adding
a definition that inherits from GICustomOperandRenderer and
GISDNodeXFormEquiv as follows.
def gi_imm8 : GICustomOperandRenderer<"renderImm8”>,
GISDNodeXFormEquiv<imm8_xform>;
Custom renderer functions should be of the form:
void render(MachineInstrBuilder &MIB, const MachineInstr &I);
Reviewers: dsanders, ab, rovka
Reviewed By: dsanders
Subscribers: kristof.beyls, javed.absar, llvm-commits, mgrang, qcolombet
Differential Revision: https://reviews.llvm.org/D42012
llvm-svn: 322582
*Mostly* NFC. Still updating the test though just for completeness.
This moves the hasAddressTaken check to MachineOutliner.cpp and replaces it
with a per-basic block test rather than a per-function test. The old test was
too conservative and was preventing functions in C programs from being
outlined even though they were safe to outline.
This was mostly a problem in C sources.
llvm-svn: 322425
Fix typos in the default scheduling resources when using the post indexed
addressing modes.
Differential revision: https://reviews.llvm.org/D40511
llvm-svn: 322392
Summary:
Very basic stack instrumentation using tagged pointers.
Tag for N'th alloca in a function is built as XOR of:
* base tag for the function, which is just some bits of SP (poor
man's random)
* small constant which is a function of N.
Allocas are aligned to 16 bytes. On every ReturnInst allocas are
re-tagged to catch use-after-return.
This implementation has a bunch of issues that will be taken care of
later:
1. lifetime intrinsics referring to tagged pointers are not
recognized in SDAG. This effectively disables stack coloring.
2. Generated code is quite inefficient. There is one extra
instruction at each memory access that adds the base tag to the
untagged alloca address. It would be better to keep tagged SP in a
callee-saved register and address allocas as an offset of that XOR
retag, but that needs better coordination between hwasan
instrumentation pass and prologue/epilogue insertion.
3. Lifetime instrinsics are ignored and use-after-scope is not
implemented. This would be harder to do than in ASan, because we
need to use a differently tagged pointer depending on which
lifetime.start / lifetime.end the current instruction is dominated
/ post-dominated.
Reviewers: kcc, alekseyshl
Subscribers: srhines, kubamracek, javed.absar, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D41602
llvm-svn: 322324
In practice, this patch has no effect on scheduling.
There is no test case as there already exists a comprehensive test case for
LSE Atomics.
Patch by Stefan Teleman
Differential Revision: https://reviews.llvm.org/D40694
llvm-svn: 322291
Revert for now as the testcase is hitting a pre-existing verifier error
that manifest as a failure when expensive checks are enabled (or
-verify-machineinstrs) is used.
This reverts commit r322200.
llvm-svn: 322231
ADRP instructions weren't being outlined because they're PC-relative and thus
fail the LR checks. This patch adds a special case for ADRPs to
getOutliningType to make sure that ADRPs can be outlined and updates the MIR
test.
llvm-svn: 322207
Large callframes (calls with several hundreds or thousands or
parameters) could lead to situations in which the emergency spillslot is
out of range to be addressed relative to the stack pointer.
This commit forces the use of a frame pointer in the presence of large
callframes.
This commit does several things:
- Compute max callframe size at the end of instruction selection.
- Add mirFileLoaded target callback. Use it to compute the max callframe size
after loading a .mir file when the size wasn't specified in the file.
- Let TargetFrameLowering::hasFP() return true if there exists a
callframe > 255 bytes.
- Always place the emergency spillslot close to FP if we have a frame
pointer.
- Note that `useFPForScavengingIndex()` would previously return false
when a base pointer was available leading to the emergency spillslot
getting allocated late (that's the whole effect of this callback).
Which made no sense to me so I took this case out: Even though the
emergency spillslot is technically not referenced by FP in this case
we still want it allocated early.
Differential Revision: https://reviews.llvm.org/D40876
llvm-svn: 322200
Summary:
This extends TableGen's AsmMatcherEmitter with code that generates
a table with tied-operand constraints. The constraints are checked
when parsing the instruction. If an operand is not equal to its tied operand,
the assembler will give an error.
Patch [2/3] in a series to add operand constraint checks for SVE's predicated ADD/SUB.
Reviewers: olista01, rengolin, mcrosier, fhahn, craig.topper, evandro, echristo
Reviewed By: fhahn
Subscribers: javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D41446
llvm-svn: 322166
Fixed issue that was found on sanitizer-x86_64-linux-fast.
I changed the result type of 'Parser.getTok().getString().lower()'
in AArch64AsmParser::tryParseSVEPredicateVector() from 'StringRef' to
'auto', since StringRef::lower() returns a std::string.
llvm-svn: 322092
Sander de Smalen