Generate more fixed_nonallocatable constraints, and add debug assertions #5132
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
elliottt
force-pushed
the
trevor/ssa-assertions
branch
2 times, most recently
from
92bc735 to
463b859
Compare
bjorn3
reviewed
Oct 26, 2022
github-actions
bot
added
cranelift
Subscribe to Label Action
This issue or pull request has been labeled: "cranelift", "cranelift:area:aarch64", "cranelift:area:machinst", "cranelift:area:x64", "isle"
Thus the following users have been cc'd because of the following labels:
To subscribe or unsubscribe from this label, edit the |
elliottt
force-pushed
the
trevor/ssa-assertions
branch
from
463b859 to
e2221c2
Compare
cfallin
reviewed
Oct 28, 2022
There was a problem hiding this comment.
One thought on s390x's EmitInfo below, but general direction seems OK.
Separately from these asserts, we were also going to add checks that we don't use any pinned vregs, right? So the vreg provided to any operand (normal use/def, fixed-constraint use/def) needs to be non-pinned, i.e. >= 128.
elliottt
force-pushed
the
trevor/ssa-assertions
branch
from
e2221c2 to
528e71e
Compare
elliottt
force-pushed
the
trevor/ssa-assertions
branch
5 times, most recently
from
f05681c to
b085ce4
Compare
elliottt
force-pushed
the
trevor/ssa-assertions
branch
5 times, most recently
from
664ffb5 to
bf21a10
Compare
elliottt
force-pushed
the
trevor/ssa-assertions
branch
from
bf21a10 to
b6aa575
Compare
elliottt
force-pushed
the
trevor/ssa-assertions
branch
from
b6aa575 to
52a14a2
Compare
elliottt
changed the title
elliottt
changed the title
jameysharp
added a commit
to jameysharp/wasmtime
that referenced
this pull request
Apr 23, 2024
When OperandCollector's reg_use/reg_late_use/reg_def/reg_early_def
methods are handed a Reg that refers to a physical ("real") register,
they all delegate to reg_fixed_nonallocatable, ignoring the constraint
kinds and positions. This behavior was introduced in bytecodealliance#5132.
In several cases, the s390x backend was calling those methods with the
result of the `gpr` or `writable_gpr` functions, which return physical
registers. In these cases we can now be more explicit that this is a
non-allocatable register.
In addition, this PR reverts bytecodealliance#4973 and bytecodealliance#5121 because they became
unecessary due, again, to bytecodealliance#5132.
github-merge-queue bot
pushed a commit
that referenced
this pull request
Apr 23, 2024
When OperandCollector's reg_use/reg_late_use/reg_def/reg_early_def
methods are handed a Reg that refers to a physical ("real") register,
they all delegate to reg_fixed_nonallocatable, ignoring the constraint
kinds and positions. This behavior was introduced in #5132.
In several cases, the s390x backend was calling those methods with the
result of the `gpr` or `writable_gpr` functions, which return physical
registers. In these cases we can now be more explicit that this is a
non-allocatable register.
In addition, this PR reverts #4973 and #5121 because they became
unecessary due, again, to #5132.
jameysharp
added a commit
to jameysharp/wasmtime
that referenced
this pull request
May 13, 2024
Consumption of non-allocatable operands was added in bytecodealliance#5253 and bytecodealliance#5132, and removed in bytecodealliance#8524 and following PRs. Now they are not only ignored by regalloc2, but the placeholders that it leaves in the allocation results are also ignored by Cranelift. So let's stop adding them to the operands list entirely.
github-merge-queue bot
pushed a commit
that referenced
this pull request
May 14, 2024
Consumption of non-allocatable operands was added in #5253 and #5132, and removed in #8524 and following PRs. Now they are not only ignored by regalloc2, but the placeholders that it leaves in the allocation results are also ignored by Cranelift. So let's stop adding them to the operands list entirely.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add assertions to the
OperandCollectorthat show we're not using pinned vregs, and usereg_fixed_nonallocatableconstraints when a real register is used with other constraint generation functions likereg_useetc.I experimented with adding debug assertions that the reused operand referenced in
reg_reuse_defconstraints was compatible with the new constraint that would be generated. This didn't end up being possible as some backends will recursively process instructions when collecting operands, making it impossible without more invasive changes to tell where that operand is in the operand vector. Given that we would need to plumb through somewhat substantial changes to support adding debug assertions, I decided that rephrasing the existing warning comment about real registers would be good enough.