rust/compiler/rustc_codegen_ssa/src/mir/analyze.rs, branch automation/bors/tryhttps://github.com/rust-lang/rust
http://git.dreamy.place/mirrors/rust/atom?h=automation/bors/try2025-10-02T06:55:50+00:00mir-opt: Eliminate dead ref statements2025-10-02T06:55:50+00:00dianqkdianqk@dianqk.net2025-06-08T07:30:09+00:00urn:sha1:571412f8190089c36758031fe09fc0ece59be6b7ProjectionElem::Subtype -> CastKind::Subtype2025-09-26T08:25:26+00:00beepster409619316085+beepster4096@users.noreply.github.com2025-09-26T03:54:54+00:00urn:sha1:aa5a21450a070fdea66a07d3cab3b69e6735c328Auto merge of #144347 - scottmcm:ssa-enums-v0, r=WaffleLapkin2025-07-27T06:37:55+00:00borsbors@rust-lang.org2025-07-27T06:37:55+00:00urn:sha1:86ef32029427cfc4161a3fd7a51992302f7c5552
No longer need `alloca`s for consuming `Result<!, i32>` and similar
In optimized builds GVN gets rid of these already, but in `opt-level=0` we actually make `alloca`s for this, which particularly impacts `?`-style things that use actually-only-one-variant types like this.
While doing so, rewrite `LocalAnalyzer::process_place` to be non-recursive, solving a 6+ year old FIXME.
r? codegen
Remove `rvalue_creates_operand` entirely2025-07-23T15:40:27+00:00Scott McMurrayscottmcm@users.noreply.github.com2025-07-23T15:40:27+00:00urn:sha1:b7e025cfb64094d8672c752a9fffb12e9bc79567
Split to a separate commit to it could be reverted later if necessary, should we get new `Rvalue`s where we can't handle it this way.
No longer need `alloca`s for consuming `Result<!, i32>` and similar2025-07-23T07:09:36+00:00Scott McMurrayscottmcm@users.noreply.github.com2025-07-23T06:23:02+00:00urn:sha1:6a5c7e04156125cb8c1ff78eae95e6ff076f51ed
In optimized builds GVN gets rid of these already, but in `opt-level=0` we actually make `alloca`s for this, which particularly impacts `?`-style things that use actually-only-one-variant types like this.
Let `rvalue_creates_operand` return true for *all* `Rvalue::Aggregate`s2025-07-08T06:08:10+00:00Scott McMurrayscottmcm@users.noreply.github.com2025-07-05T06:16:41+00:00urn:sha1:8cf2c71243f8c093e7728370e6653aa940dbaf22
Inspired by <https://github.com/rust-lang/rust/pull/138759#discussion_r2156375342> where I noticed that we were nearly at this point, plus the comments I was writing in 143410 that reminded me a type-dependent `true` is fine.
This PR splits the `OperandRef::builder` logic out to a separate type, with the updates needed to handle SIMD as well. In doing so, that makes the existing `Aggregate` path in `codegen_rvalue_operand` capable of handing SIMD values just fine.
As a result, we no longer need to do layout calculations for aggregate result types when running the analysis to determine which things can be SSA in codegen.
Don't `alloca` just to look at a discriminant2025-03-12T07:56:43+00:00Scott McMurrayscottmcm@users.noreply.github.com2025-03-12T07:38:14+00:00urn:sha1:143f39362aa3fe30e19de5d2a29bf6535e8f975f
Today we're making LLVM do a bunch of extra work for every enum you match on, even trivial stuff like `Option<bool>`. Let's not.
rename `BitSet` to `DenseBitSet`2025-01-11T11:34:01+00:00Rémy Rakicremy.rakic+github@gmail.com2025-01-07T15:19:05+00:00urn:sha1:a13354bea05968799a5be5521691322274fa6a9e
This should make it clearer that this bitset is dense, with the
advantages and disadvantages that it entails.
Reformat using the new identifier sorting from rustfmt2024-09-22T23:11:29+00:00Michael Gouletmichael@errs.io2024-09-22T23:05:04+00:00urn:sha1:c682aa162b0d41e21cc6748f4fecfe01efb69d1fCompute reachable locals as part of non_ssa_locals2024-09-21T05:07:00+00:00Ben Kimockkimockb@gmail.com2024-08-25T01:03:38+00:00urn:sha1:523f8f8398f84d13c7d38749de53b701ccfb4f50