comments for deduce_param_attrs#147873
Conversation
|
Some changes occurred to MIR optimizations cc @rust-lang/wg-mir-opt |
e2600b7 to
5f9de55
Compare
|
This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed. Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers. |
|
Could you also update the pull request description (it references code changes which are no longer there)? Thanks. r=me |
There was a problem hiding this comment.
Codegen actually uses instance_mir so using optimized_mir here seems a bit inconsistent?
Things could go quite wrong if they end up using different MIR...
5f9de55 to
82f6c60
Compare
|
Some changes occurred in compiler/rustc_codegen_ssa |
Done. I've also added comments in codegen itself. |
|
@bors r=tmiasko rollup |
It sure is, this pass is the primary reason post-mono opts are stuck. |
|
If we decide on post mono MIR passes I think it would be fine to remove attribute deduction. |
|
LLVM should often be able to infer "readonly" itself, shouldn't it?
|
Rollup merge of #147873 - RalfJung:deduce_param_attrs, r=tmiasko comments for deduce_param_attrs Cc `@saethlin` since IIRC you experimented with codegen doing post-mono MIR ops? That seems to be in conflict with this pass. Cc `@tmiasko` r? `@scottmcm`
Rollup of 4 pull requests Successful merges: - rust-lang/rust#147575 (Refactor move analysis subpath representation) - rust-lang/rust#147864 (Parse `const unsafe trait` properly) - rust-lang/rust#147868 (MirPatch: Simplify new_local.) - rust-lang/rust#147873 (comments for deduce_param_attrs) r? `@ghost` `@rustbot` modify labels: rollup
Cc @saethlin since IIRC you experimented with codegen doing post-mono MIR ops? That seems to be in conflict with this pass.
Cc @tmiasko
r? @scottmcm