* x64: Elide more uextend with extractlane
I've confirmed locally now that `pextr{b,w,d}` all zero the upper bits
of the full 64-bit register size which means that the `extractlane`
operation with a zero-extend can be elided for more cases, including
8-to-64-bit casts as well as 32-to-64.
This helps elide a few extra `mov`s in a loop I was looking at and had a
modest corresponding increase in performance (my guess was due to the
slightly decreased code size mostly as opposed to the removed `mov`s).
* Remove stray file
* x64: Refactor sextend/uextend rules
Move much of the meaty logic from these lowering rules into the
`extend_to_gpr` helper to benefit other callers of `extend_to_gpr` to
elide instructions. This additionally simplifies `sextend` and `uextend`
lowerings to rely on optimizations happening within the `extend_to_gpr`
helper.
* x64: Skip `uextend` for `pextr{b,w}` instructions
These instructions are documented as automatically zeroing the upper
bits so `uextend` operations can be skipped. This slightly improves
codegen for the wasm `i{8x16,16x8}.extract_lane_u` instructions, for
example.
* Modernize an extractor pattern
* Trim some superfluous match clauses
Additionally rejigger priorities to be "mostly default" now.
* Refactor 32-to-64 predicate to a helper
Also adjust the pattern matched in the `extend_to_gpr` helper.
* Slightly refactor pextr{b,w} case
* Review comments
* x64: Take SIGFPE signals for divide traps
Prior to this commit Wasmtime would configure `avoid_div_traps=true`
unconditionally for Cranelift. This, for the division-based
instructions, would change emitted code to explicitly trap on trap
conditions instead of letting the `div` x86 instruction trap.
There's no specific reason for Wasmtime, however, to specifically avoid
traps in the `div` instruction. This means that the extra generated
branches on x86 aren't necessary since the `div` and `idiv` instructions
already trap for similar conditions as wasm requires.
This commit instead disables the `avoid_div_traps` setting for
Wasmtime's usage of Cranelift. Subsequently the codegen rules were
updated slightly:
* When `avoid_div_traps=true`, traps are no longer emitted for `div`
instructions.
* The `udiv`/`urem` instructions now list their trap as divide-by-zero
instead of integer overflow.
* The lowering for `sdiv` was updated to still explicitly check for zero
but the integer overflow case is deferred to the instruction itself.
* The lowering of `srem` no longer checks for zero and the listed trap
for the `div` instruction is a divide-by-zero.
This means that the codegen for `udiv` and `urem` no longer have any
branches. The codegen for `sdiv` removes one branch but keeps the
zero-check to differentiate the two kinds of traps. The codegen for
`srem` removes one branch but keeps the -1 check since the semantics of
`srem` mismatch with the semantics of `idiv` with a -1 divisor
(specifically for INT_MIN).
This is unlikely to have really all that much of a speedup but was
something I noticed during #6008 which seemed like it'd be good to clean
up. Plus Wasmtime's signal handling was already set up to catch
`SIGFPE`, it was just never firing.
* Remove the `avoid_div_traps` cranelift setting
With no known users currently removing this should be possible and helps
simplify the x64 backend.
* x64: GC more support for avoid_div_traps
Remove the `validate_sdiv_divisor*` pseudo-instructions and clean up
some of the ISLE rules now that `div` is allowed to itself trap
unconditionally.
* x64: Store div trap code in instruction itself
* Keep divisors in registers, not in memory
Don't accidentally fold multiple traps together
* Handle EXC_ARITHMETIC on macos
* Update emit tests
* Update winch and tests
This commit goes through the lowerings for the CLIF `splat` instruction
and improves the support for each operator. Many of these lowerings are
mirrored from v8/SpiderMonkey and there are a number of improvements:
* AVX2 `v{p,}broadcast*` instructions are added and used when available.
* Float-based splats are much simpler and always a single-instruction
* Integer-based splats don't insert into an uninit xmm value and instead
start out with a `movd` to move into an `xmm` register. This
thoeretically breaks dependencies with prior instructions since `movd`
creates a fresh new value in the destination register.
* Loads are now sunk into all of the instructions. A new extractor,
`sinkable_load_exact`, was added to sink the i8/i16 loads.
This commit adds another case for `shuffle` lowering to the x64 backend
for the `{,v}pblendw` instruction. This instruction selects 16-bit
values from either of the inputs corresponding to an immediate 8-bit-mask where
each bit selects the corresponding lane from the inputs.
* x64: Refactor `Amode` computation in ISLE
This commit replaces the previous computation of `Amode` with a
different set of rules that are intended to achieve the same purpose but
are structured differently. The motivation for this commit is going to
become more relevant in the next commit where `lea` will be used for the
`iadd` instruction, possibly, on x64. When doing so it caused a stack
overflow in the test suite during the compilation phase of a wasm
module, namely as part of the `amode_add` function. This function is
recursively defined in terms of itself and recurses as deep as the
deepest `iadd`-chain in a program. A particular test in our test suite
has a 10k-long chain of `iadd` which ended up causing a stack overflow
in debug mode.
This stack overflow is caused because the `amode_add` helper in ISLE
unconditionally peels all the `iadd` nodes away and looks at all of
them, even if most end up in intermediate registers along the way. Given
that structure I couldn't find a way to easily abort the recursion. The
new `to_amode` helper is structured in a similar fashion but attempts to
instead only recurse far enough to fold items into the final `Amode`
instead of recursing through items which themselves don't end up in the
`Amode`. Put another way previously the `amode_add` helper might emit
`x64_add` instructions, but it no longer does that.
This goal of this commit is to preserve all the original `Amode`
optimizations, however. For some parts, though, it relies more on egraph
optimizations to run since if an `iadd` is 10k deep it doesn't try to
find a constant buried 9k levels inside there to fold into the `Amode`.
The hope, though, is that with egraphs having run already it's shuffled
constants to the right most of the time and already folded any possible
together.
* x64: Add `lea`-based lowering for `iadd`
This commit adds a rule for the lowering of `iadd` to use `lea` for 32
and 64-bit addition. The theoretical benefit of `lea` over the `add`
instruction is that the `lea` variant can emulate a 3-operand
instruction which doesn't destructively modify on of its operands.
Additionally the `lea` operation can fold in other components such as
constant additions and shifts.
In practice, however, if `lea` is unconditionally used instead of `iadd`
it ends up losing 10% performance on a local `meshoptimizer` benchmark.
My best guess as to what's going on here is that my CPU's dedicated
units for address computation are all overloaded while the ALUs are
basically idle in a memory-intensive loop. Previously when the ALU was
used for `add` and the address units for stores/loads it in theory
pipelined things better (most of this is me shooting in the dark). To
prevent the performance loss here I've updated the lowering of `iadd` to
conditionally sometimes use `lea` and sometimes use `add` depending on
how "complicated" the `Amode` is. Simple ones like `a + b` or `a + $imm`
continue to use `add` (and its subsequent hypothetical extra `mov`
necessary into the result). More complicated ones like `a + b + $imm` or
`a + b << c + $imm` use `lea` as it can remove the need for extra
instructions. Locally at least this fixes the performance loss relative
to unconditionally using `lea`.
One note is that this adds an `OperandSize` argument to the
`MInst::LoadEffectiveAddress` variant to add an encoding for 32-bit
`lea` in addition to the preexisting 64-bit encoding.
* Conditionally use `lea` based on regalloc
* x64: Add precise-output tests for div traps
This adds a suite of `*.clif` files which are intended to test the
`avoid_div_traps=true` compilation of the `{s,u}{div,rem}` instructions.
* x64: Remove conditional regalloc in `Div` instruction
Move the 8-bit `Div` logic into a dedicated `Div8` instruction to avoid
having conditionally-used registers with respect to regalloc.
* x64: Migrate non-trapping, `udiv`/`urem` to ISLE
* x64: Port checked `udiv` to ISLE
* x64: Migrate urem entirely to ISLE
* x64: Use `test` instead of `cmp` to compare-to-zero
* x64: Port `sdiv` lowering to ISLE
* x64: Port `srem` lowering to ISLE
* Tidy up regalloc behavior and fix tests
* Update docs and winch
* Review comments
* Reword again
* More refactoring test fixes
* More test fixes
* aarch64: Specialize constant vector shifts
This commit adds special lowering rules for
vector-shifts-by-constant-amounts to use dedicated instructions which
cuts down on the codegen here quite a bit for constant values.
* Fix codegen for 0-shift-rights
* Special-case zero left-shifts as well
* Remove left-shift special case
* x64: Add `shuffle` specialization for `palignr`
This commit adds specializations for the `palignr` instruction to the
x64 backend to specialize some more patterns of byte shuffles.
* Fix tests
* aarch64: Translate float and splat lowering to ISLE
I was looking into `constant_f128` and its fallback lowering into memory
and to get familiar with the code I figured it'd be good to port some
Rust logic to ISLE. This commit ports the `constant_{f128,f64,f32}`
helpers into ISLE from Rust as well as the `splat_const` helper which
ended up being closely related.
Tests reflect a number of regalloc changes that happened but also namely
one major difference is that in the lowering of `f32` a 32-bit immediate
is created now instead of a 64-bit immediate (in a GP register before
it's moved into a FP register). This semantically has no change but the
generated code is slightly different in a few minor cases.
* aarch64: Load f64/v128 constants from a pool
This commit removes the `LoadFpuConst64` and `LoadFpuConst128`
pseudo-instructions from the AArch64 backend which internally loaded a
nearby constant and then jumped over it. Constants now go through the
`VCodeConstant` infrastructure which gets placed at the end of the
function similar to how x64 works. Some minor support was added in as
well to add a new addressing mode for a `MachLabel`-relative load.
* x64: Improve memory support in `{insert,extract}lane`
This commit improves adds support to Cranelift to emit `pextr{b,w,d,q}`
with a memory destination, merging a store-of-extract operation into one
instruction. Additionally AVX support is added for the `pextr*`
instructions.
I've additionally tried to ensure that codegen tests and runtests exist
for all forms of these instructions too.
* Add missing commas
* Fix tests
* Change CLIF `shuffle` to validate lane indices
Previously the CLIF `shuffle` instruction did not perform any validation
on the lane shuffle mask and specified that out-of-bounds lanes always
returned 0 as the value. This behavior though is not required by
WebAssembly which validates that lane indices are always in-bounds.
Additionally since these are static immediates even other code
generators should be able to verify that the immediates are in-bounds.
As a result this commit updates the definition of the `shuffle`
instruction to specify that all byte immediates must be in-bounds in the
range of [0, 32). The verifier has been updated and some test cases have
been removed that were testing this functionality.
Closes#5989
* Only generate valid shuffle immediates in fuzzer
One of the cases for a splat operation, as updated in #5370, wrote to
a temp reg but then only conditionally transformed the temp into the
final destination register. In another codepath, `rd` was left
undefined. This causes a panic later when regalloc2 verifies SSA
properties of its input (here, value not def'd before use).
Fixes#5985.
* aarch64: Add `shuffle` lowerings for the `uzp{1,2}` instructions
This commit uses the same style of patterns in the x64 backend to start
adding specific lowerings of the Cranelift `shuffle` instruction to
particular AArch64 instructions.
* aarch64: Add `shuffle` lowerings to the `zip{1,2}` instructions
These instructions match the `punpck*` family of instructions on x64 and
should help provide more efficient lowerings than the current `shuffle`
fallback.
* aarch64: Add `shuffle` lowerings for `trn{1,2}`
Along the lines of prior commits adds specific patterns to lowering for
individual AArch64 instructions available.
* aarch64: Add a `shuffle` lowering for the `ext` instruction
This instruction will more-or-less concatenate two 128-bit vector
registers to create a 256-bit value, shift it right, and then take the
lower 128-bits into the destination. This can be modeled with a
`shuffle` of consecutive bytes so this adds a lowering rule to generate
this instruction.
* aarch64: Add `shuffle` special case for `dup`
This commit adds special cases for Cranelift's `shuffle` on AArch64 when
the lowering can be represented with a `dup` instruction which
broadcasts one vector's lane into all lanes of the destination.
* aarch64: Add `shuffle` specializations for `rev` instructions
This commit adds shuffle mask specializations for the `rev{16,32,64}`
family of instructions on AArch64 which can be used to reverse bytes,
16-bit values, or 32-bit values within larger values.
* Fix tests
* Add doc-comments in ISLE
* x64: Optimize store-of-extract-lane-0
The `movss` and `movsd` instructions can be used to store the 0th lane
of a `t32x4` or a `t64x2` vector into memory, enabling fusing a `store`
and an `extractlane` instruction.
* Fix merge conflict with `main`
* x64: Add a smattering of lowerings for `shuffle` specializations (#5930)
* x64: Add lowerings for `punpck{h,l}wd`
Add some special cases for `shuffle` for more specialized x86
instructions.
* x64: Add `shuffle` lowerings for `pshufd`
This commit adds special-cased lowerings for the x64 `shuffle`
instruction when the `pshufd` instruction alone is necessary. This is
possible when the shuffle immediate permutes 32-bit values within one of
the vector inputs of the `shuffle` instruction, but not both.
* x64: Add shuffle lowerings for `punpck{h,l}{q,}dq`
This adds specific permutations for some x86 instructions which
specifically interleave high/low bytes for 32 and 64-bit values. This
corresponds to the preexisting specific lowerings for interleaving 8 and
16-bit values.
* x64: Add `shuffle` lowerings for `shufps`
This commit adds targeted lowerings for the `shuffle` instruction that
match the pattern that `shufps` supports. The `shufps` instruction
selects two elements from the first vector and two elements from the
second vector which means while it's not generally applicable it should
still be more useful than the catch-all lowering of `shuffle`.
* x64: Add shuffle support for `pshuf{l,h}w`
This commit adds special lowering cases for these instructions which
permute 16-bit values within a 128-bit value either within the upper or
lower half of the 128-bit value.
* x64: Specialize `shuffle` with an all-zeros immediate
Instead of loading the all-zeros immediate from a rip-relative address
at the end of the function instead generate a zero with a `pxor`
instruction and then use `pshufb` to do the broadcast.
* Review comments
* x64: Add an AVX encoding for the `pshufd` instruction
This will benefit from lack of need for alignment vs the `pshufd`
instruction if working with a memory operand and additionally, as I've
just learned, this reduces dependencies between instructions because the
`v*` instructions zero the upper bits as opposed to preserving them
which could accidentally create false dependencies in the CPU between
instructions.
* x64: Add more support for AVX loads/stores
This commit adds VEX-encoded versions of instructions such as
`mov{ss,sd,upd,ups,dqu}` for load and store operations. This also
changes some signatures so the `load` helpers specifically take a
`SyntheticAmode` argument which ended up doing a small refactoring of
the `*_regmove` variant used for `insertlane 0` into f64x2 vectors.
* x64: Enable using AVX instructions for zero regs
This commit refactors the internal ISLE helpers for creating zero'd
xmm registers to leverage the AVX support for all other instructions.
This moves away from picking opcodes to instead picking instructions
with a bit of reorganization.
* x64: Remove `XmmConstOp` as an instruction
All existing users can be replaced with usage of the `xmm_uninit_value`
helper instruction so there's no longer any need for these otherwise
constant operations. This additionally reduces manual usage of opcodes
in favor of instruction helpers.
* Review comments
* Update test expectations
* x64: Add lowerings for `punpck{h,l}wd`
Add some special cases for `shuffle` for more specialized x86
instructions.
* x64: Add `shuffle` lowerings for `pshufd`
This commit adds special-cased lowerings for the x64 `shuffle`
instruction when the `pshufd` instruction alone is necessary. This is
possible when the shuffle immediate permutes 32-bit values within one of
the vector inputs of the `shuffle` instruction, but not both.
* x64: Add shuffle lowerings for `punpck{h,l}{q,}dq`
This adds specific permutations for some x86 instructions which
specifically interleave high/low bytes for 32 and 64-bit values. This
corresponds to the preexisting specific lowerings for interleaving 8 and
16-bit values.
* x64: Add `shuffle` lowerings for `shufps`
This commit adds targeted lowerings for the `shuffle` instruction that
match the pattern that `shufps` supports. The `shufps` instruction
selects two elements from the first vector and two elements from the
second vector which means while it's not generally applicable it should
still be more useful than the catch-all lowering of `shuffle`.
* x64: Add shuffle support for `pshuf{l,h}w`
This commit adds special lowering cases for these instructions which
permute 16-bit values within a 128-bit value either within the upper or
lower half of the 128-bit value.
* x64: Specialize `shuffle` with an all-zeros immediate
Instead of loading the all-zeros immediate from a rip-relative address
at the end of the function instead generate a zero with a `pxor`
instruction and then use `pshufb` to do the broadcast.
* Review comments
* x64: Remove incorrect `amode_add` lowering rules
This commit removes two incorrect rules as part of the x64 backend's
computation of addressing modes. These two rules folded a zero-extended
32-bit computation into the address mode operand, but this isn't correct
as the 32-bit computation should be truncated to 32-bits but when folded
into the address mode computation it happens with 64-bit operands,
meaning truncation doesn't happen.
* Add release notes
This commit fixes an off-by-one error in the subtraction of indices when
shuffling a vector with itself. Lanes 16-and-above are mapped to select
from the first vector since the first and second element are the same,
but the subtraction was with 15 rather than 16 by accident.
* Remove the Cranelift `vselect` instruction
This instruction is documented as selecting lanes based on the "truthy"
value of the condition lane, but the current status of the
implementation of this instruction is:
* x64 - uses the high bit for `f32x4` and `f64x2` and otherwise uses the
high bit of each byte doing a byte-wise lane select rather than
whatever the controlling type is.
* AArch64 - this is the same as `bitselect` which is a bit-wise
selection rather than a lane-wise selection.
* s390x - this is the same as AArch64, a bit-wise selection rather than
lane-wise.
* interpreter - the interpreter implements the documented semantics of
selecting based on "truthy" values.
Coupled with the status of the implementation is the fact that this
instruction is not used by WebAssembly SIMD today either. The only use
of this instruction in Cranelift is the nan-canonicalization pass. By
moving nan-canonicalization to `bitselect`, since that has the desired
semantics, there's no longer any need for `vselect`.
Given this situation this commit subsqeuently removes `vselect` and all
usage of it throughout Cranelift.
Closes#5917
* Review comments
* Bring back vselect opts as bitselect opts
* Clean up vselect usage in the interpreter
* Move bitcast in nan canonicalization
* Add a comment about float optimization
This commit adds lowerings to the AArch64 backend for the `fmls`
instruction which is intended to be leveraged in the relaxed-simd
proposal for WebAssembly. This should hopefully allow for a
teeny-bit-more efficient codegen for this operator instead of using the
`fmla` instruction plus a negation instruction.
This catches a case that wasn't handled previously by #5880 to allow a
constant load to be folded into an instruction rather than forcing it to
be loaded into a temporary register.
* x64: Add `shuffle` cases for `punpck{h,l}bw`
I noticed this difference between LLVM and Cranelift for something I was
looking at recently, and while it's probably not all that common I
figured I'd add it here since it should be somewhat useful nevertheless.
* Review feedback
* Use u128 extractor instead
This instruction is only defined with i8x16 inputs and outputs so
there's no need for a type variable, so shadow the otherwise-generic `a`
result with a concrete i8x16 type.
This commit adds support for the bare lowering of the `iadd_pairwise`
instruction with `i16x8` and `i32x4` types on the x64 backend. These
lowerings are achieved with the `phaddw` and `phaddd` instructions,
respectively. Additionally AVX encodings of these instructions are added
too.
The motivation for these new lowerings comes from the relaxed-simd
proposal which will use them in the deterministic lowering of some
instructions on the x64 backend.
A number of places in the x64 backend make use of 128-bit constants for
various wasm SIMD-related instructions although most of them currently
use the `x64_xmm_load_const` helper to load the constant into a
register. Almost all xmm instructions, however, enable using a memory
operand which means that these loads can be folded into instructions to
help reduce register pressure. Automatic conversions were added for a
`VCodeConstant` into an `XmmMem` value and then explicit loads were all
removed in favor of forwarding the `XmmMem` value directly to the
underlying instruction. Note that some instances of `x64_xmm_load_const`
remain since they're used in contexts where load sinking won't work
(e.g. they're the first operand, not the second for non-commutative
instructions).
This was added for the wasm SIMD proposal but I've been poking around at
this recently and the instruction can instead be represented by its
component parts with the same semantics I believe. This commit removes
the instruction and instead represents it with the existing
`iadd_pairwise` instruction (among others) and updates backends to with
new pattern matches to have the same codegen as before.
This interestingly entirely removed the codegen rule with no replacement
on the AArch64 backend as the existing rules all existed to produce the
same codegen.
This uses the `cmov`, which was previously necessary for Spectre
mitigation, to clamp the table index instead of zeroing it. By then
placing the default target as the last entry in the table, we can use
just one branch instruction in all cases.
Since there isn't a bounds-check branch any more, this sequence no
longer needs Spectre mitigation. And since we don't need to be careful
about preserving flags, half the instructions can be removed from this
pseudoinstruction and emitted as regular instructions instead.
This is a net savings of three bytes in the encoding of x64's br_table
pseudoinstruction. The generated code can sometimes be longer overall
because the blocks are emitted in a slightly different order.
My benchmark results show a very small effect on runtime performance
with this change.
The spidermonkey benchmark in Sightglass runs "1.01x faster" than main
by instructions retired, but with no significant difference in CPU
cycles. I think that means it rarely hit the default case in any
br_table instructions it executed.
The pulldown-cmark benchmark in Sightglass runs "1.01x faster" than main
by CPU cycles, but main runs "1.00x faster" by instructions retired. I
think that means this benchmark hit the default case a significant
amount of the time, so it executes a few more instructions per br_table,
but maybe the branches were predicted better.
* x64: Fill out more AVX instructions
This commit fills out more AVX instructions for SSE counterparts
currently used. Many of these instructions do not benefit from the
3-operand form that AVX uses but instead benefit from being able to use
`XmmMem` instead of `XmmMemAligned` which may be able to avoid some
extra temporary registers in some cases.
* Review comments
* Rework the blockorder module to reuse the dom tree's cfg postorder
* Update domtree tests
* Treat br_table with an empty jump table as multiple block exits
* Bless tests
* Change branch_idx to succ_idx and fix the comment
When expanding a min/max operation to a pair of icmp + select,
do not attempt to expand the input value operands twice, as
this might fail with memory operands.
Fixes https://github.com/bytecodealliance/wasmtime/issues/5859.
The relaxed-simd proposal for WebAssembly adds a fused-multiply-add
operation for `v128` types so I was poking around at Cranelift's
existing support for its `fma` instruction. I was also poking around at
the x86_64 ISA's offerings for the FMA operation and ended up with this
PR that improves the lowering of the `fma` instruction on the x64
backend in a number of ways:
* A libcall-based fallback is now provided for `f32x4` and `f64x2` types
in preparation for eventual support of the relaxed-simd proposal.
These encodings are horribly slow, but it's expected that if FMA
semantics must be guaranteed then it's the best that can be done
without the `fma` feature. Otherwise it'll be up to producers (e.g.
Wasmtime embedders) whether wasm-level FMA operations should be FMA or
multiply-then-add.
* In addition to the existing `vfmadd213*` instructions opcodes were
added for `vfmadd132*`. The `132` variant is selected based on which
argument can have a sinkable load.
* Any argument in the `fma` CLIF instruction can now have a
`sinkable_load` and it'll generate a single FMA instruction.
* All `vfnmadd*` opcodes were added as well. These are pattern-matched
where one of the arguments to the CLIF instruction is an `fneg`. I
opted to not add a new CLIF instruction here since it seemed like
pattern matching was easy enough but I'm also not intimately familiar
with the semantics here so if that's the preferred approach I can do
that too.
* x64: Enable load-coalescing for SSE/AVX instructions
This commit unlocks the ability to fold loads into operands of SSE and
AVX instructions. This is beneficial for both function size when it
happens in addition to being able to reduce register pressure.
Previously this was not done because most SSE instructions require
memory to be aligned. AVX instructions, however, do not have alignment
requirements.
The solution implemented here is one recommended by Chris which is to
add a new `XmmMemAligned` newtype wrapper around `XmmMem`. All SSE
instructions are now annotated as requiring an `XmmMemAligned` operand
except for a new new instruction styles used specifically for
instructions that don't require alignment (e.g. `movdqu`, `*sd`, and
`*ss` instructions). All existing instruction helpers continue to take
`XmmMem`, however. This way if an AVX lowering is chosen it can be used
as-is. If an SSE lowering is chosen, however, then an automatic
conversion from `XmmMem` to `XmmMemAligned` kicks in. This automatic
conversion only fails for unaligned addresses in which case a load
instruction is emitted and the operand becomes a temporary register
instead. A number of prior `Xmm` arguments have now been converted to
`XmmMem` as well.
One change from this commit is that loading an unaligned operand for an
SSE instruction previously would use the "correct type" of load, e.g.
`movups` for f32x4 or `movup` for f64x2, but now the loading happens in
a context without type information so the `movdqu` instruction is
generated. According to [this stack overflow question][question] it
looks like modern processors won't penalize this "wrong" choice of type
when the operand is then used for f32 or f64 oriented instructions.
Finally this commit improves some reuse of logic in the `put_in_*_mem*`
helper to share code with `sinkable_load` and avoid duplication. With
this in place some various ISLE rules have been updated as well.
In the tests it can be seen that AVX-instructions are now automatically
load-coalesced and use memory operands in a few cases.
[question]: https://stackoverflow.com/questions/40854819/is-there-any-situation-where-using-movdqu-and-movupd-is-better-than-movups
* Fix tests
* Fix move-and-extend to be unaligned
These don't have alignment requirements like other xmm instructions as
well. Additionally add some ISA tests to ensure that their output is
tested.
* Review comments
This is a follow-up to comments in #5795 to remove some cruft in the x64
instruction model to ensure that the shape of an `Inst` reflects what's
going to happen in regalloc and encoding. This accessor was used to
handle `round*`, `pextr*`, and `pshufb` instructions. The `round*` ones
had already moved to the appropriate `XmmUnary*` variant and `pshufb`
was additionally moved over to that variant as well.
The `pextr*` instructions got a new `Inst` variant and additionally had
their constructors slightly modified to no longer require the type as
input. The encoding for these instructions now automatically handles the
various type-related operands through a new `SseOpcode::Pextrq` operand
to represent 64-bit movements.
* x64: Add most remaining AVX lowerings
This commit goes through `inst.isle` and adds a corresponding AVX
lowering for most SSE lowerings. I opted to skip instructions where the
SSE lowering didn't read/modify a register, such as `roundps`. I think
that AVX will benefit these instructions when there's load-merging since
AVX doesn't require alignment, but I've deferred that work to a future
PR.
Otherwise though in this PR I think all (or almost all) of the 3-operand
forms of AVX instructions are supported with their SSE counterparts.
This should ideally improve codegen slightly by removing register
pressure and the need for `movdqa` between registers. I've attempted to
ensure that there's at least one codegen test for all the new instructions.
As a side note, the recent capstone integration into `precise-output`
tests helped me catch a number of encoding bugs much earlier than
otherwise, so I've found that incredibly useful in tests!
* Move `vpinsr*` instructions to their own variant
Use true `XmmMem` and `GprMem` types in the instruction as well to get
more type-level safety for what goes where.
* Remove `Inst::produces_const` accessor
Instead of conditionally defining regalloc and various other operations
instead add dedicated `MInst` variants for operations which are intended
to produce a constant to have more clear interactions with regalloc and
printing and such.
* Fix tests
* Register traps in `MachBuffer` for load-folding ops
This adds a missing `add_trap` to encoding of VEX instructions with
memory operands to ensure that if they cause a segfault that there's
appropriate metadata for Wasmtime to understand that the instruction
could in fact trap. This fixes a fuzz test case found locally where v8
trapped and Wasmtime didn't catch the signal and crashed the fuzzer.
* x64: Add rudimentary support for some AVX instructions
I was poking around Spidermonkey's wasm backend and saw that the various
assembler functions used are all `v*`-prefixed which look like they're
intended for use with AVX instructions. I looked at Cranelift and it
currently doesn't have support for many AVX-based instructions, so I
figured I'd take a crack at it!
The support added here is a bit of a mishmash when viewed alone, but my
general goal was to take a single instruction from the SIMD proposal for
WebAssembly and migrate all of its component instructions to AVX. I, by
random chance, picked a pretty complicated instruction of `f32x4.min`.
This wasm instruction is implemented on x64 with 4 unique SSE
instructions and ended up being a pretty good candidate.
Further digging about AVX-vs-SSE shows that there should be two major
benefits to using AVX over SSE:
* Primarily AVX instructions largely use a three-operand form where two
input registers are operated with and an output register is also
specified. This is in contrast to SSE's predominant
one-register-is-input-but-also-output pattern. This should help free
up the register allocator a bit and additionally remove the need for
movement between registers.
* As #4767 notes the memory-based operations of VEX-encoded instructions
(aka AVX instructions) do not have strict alignment requirements which
means we would be able to sink loads and stores into individual
instructions instead of having separate instructions.
So I set out on my journey to implement the instructions used by
`f32x4.min`. The first few were fairly easy. The machinst backends are
already of the shape "take these inputs and compute the output" where
the x86 requirement of a register being both input and output is
postprocessed in. This means that the `inst.isle` creation helpers for
SSE instructions were already of the correct form to use AVX. I chose to
add new `rule` branches for the instruction creation helpers, for
example `x64_andnps`. The new `rule` conditionally only runs if AVX is
enabled and emits an AVX instruction instead of an SSE instruction for
achieving the same goal. This means that no lowerings of clif
instructions were modified, instead just new instructions are being
generated.
The VEX encoding was previously not heavily used in Cranelift. The only
current user are the FMA-style instructions that Cranelift has at this
time. These FMA instructions have one extra operand than `vandnps`, for
example, so I split the existing `XmmRmRVex` into a few more variants to
fit the shape of the instructions that needed generating for
`f32x4.min`. This was accompanied then with more AVX opcode definitions,
more emission support, etc.
Upon implementing all of this it turned out that the test suite was
failing on my machine due to the memory-operand encodings of VEX
instructions not being supported. I didn't explicitly add those in
myself but some preexisting RIP-relative addressing was leaking into the
new instructions with existing tests. I opted to go ahead and fill out
the memory addressing modes of VEX encoding to get the tests passing
again.
All-in-all this PR adds new instructions to the x64 backend for a number
of AVX instructions, updates 5 existing instruction producers to use AVX
instructions conditionally, implements VEX memory operands, and adds
some simple tests for the new output of `f32x4.min`. The existing
runtest for `f32x.min` caught a few intermediate bugs along the way and
I additionally added a plain `target x86_64` to that runtest to ensure
that it executes with and without AVX to test the various lowerings.
I'll also note that this, and future support, should be well-fuzzed
through Wasmtime's fuzzing which may explicitly disable AVX support
despite the machine having access to AVX, so non-AVX lowerings should be
well-tested into the future.
It's also worth mentioning that I am not an AVX or VEX or x64 expert.
Implementing the memory operand part for VEX was the hardest part of
this PR and while I think it should be good someone else should
definitely double-check me. Additionally I haven't added many
instructions to the x64 backend yet so I may have missed obvious places
to tests or such, so am happy to follow-up with anything to be more
thorough if necessary.
Finally I should note that this is just the tip of the iceberg when it
comes to AVX. My hope is to get some of the idioms sorted out to make it
easier for future PRs to add one-off instruction lowerings or such.
* Review feedback
Rework br_table to use BlockCall, allowing us to avoid adding new nodes during ssa construction to hold block arguments. Additionally, many places where we previously matched on InstructionData to extract branch destinations can be replaced with a use of branch_destination or branch_destination_mut.
As a follow-up to #5780, disassemble the regions identified by bb_starts, falling back on disassembling the whole buffer. This ensures that instructions like br_table that introduce a lot of constants don't throw off capstone for the remainder of the function.
---------
Co-authored-by: Jamey Sharp <jamey@minilop.net>
