Add stubs in IR and ABI for `f16` and `f128` #121728

tgross35 · 2024-02-28T09:09:18Z

This is the very first step toward the changes in #114607 and the f16 and f128 RFC. It adds the types to rustc_type_ir::FloatTy and rustc_abi::Primitive, and just propagates those out as unimplemented! stubs where necessary.

These types do not parse yet so there is no feature gate, and it should be okay to use unimplemented!.

The next steps will probably be AST support with parsing and the feature gate.

r? @compiler-errors
cc @Nilstrieb suggested breaking the PR up in #120645 (comment)

rustbot · 2024-02-28T09:09:27Z

This PR changes Stable MIR

cc @oli-obk, @celinval, @spastorino, @ouz-a

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

Some changes occurred in match checking

cc @Nadrieril

Some changes occurred in compiler/rustc_codegen_cranelift

cc @bjorn3

Some changes occurred to the core trait solver

cc @rust-lang/initiative-trait-system-refactor

The Miri subtree was changed

cc @rust-lang/miri

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo, @GuillaumeGomez

Some changes occurred to the CTFE / Miri engine

cc @rust-lang/miri

Some changes occurred in src/librustdoc/clean/types.rs

cc @camelid

Some changes occurred in exhaustiveness checking

cc @Nadrieril

flip1995 · 2024-02-28T09:50:11Z

src/tools/clippy/clippy_utils/src/consts.rs

@@ -778,8 +778,10 @@ pub fn mir_to_const<'tcx>(lcx: &LateContext<'tcx>, result: mir::Const<'tcx>) ->
                let range = alloc_range(offset + size * idx, size);
                let val = alloc.read_scalar(&lcx.tcx, range, /* read_provenance */ false).ok()?;
                res.push(match flt {
+                    FloatTy::F16 => unimplemented!("f16_f128"),


I'm a bit wary about adding this, especially to tools. There's definitely a risk of forgetting some of them, causing ICEs.

I can change this, what would be better?

My original change did more https://github.com/rust-lang/rust/pull/114607/files#diff-83cfffe9dbc2758ecf7b34a433525befc32ebff4525aef0de105fa5e7b659c21R808-R815, but since the surface area of this PR is intentionally much smaller, the primitives aren't available for use.

I see the benefit of small PRs and this comment isn't meant to block the PR. Just a concern, that we shouldn't forget about those unimplemented! macros.

Removing the panics could be added as a step to the tracking issue

In fairness to @tgross35, this is impossible to hit right now. I think this approach is the best for the moment.

I added a todo list that mentions these at #116909 (comment)

Nilstrieb · 2024-02-28T10:04:52Z

tests/ui/parser/f16-f128.rs

@@ -0,0 +1,37 @@
+// Make sure we don't ICE while incrementally adding f16 and f128 support


I don't think parser is quite right, the parser doesn't care :)

Yeah, this test does literally nothing lol

There didn't seem to be a good folder :) I can move it if there is a better place

I know it does nothing but I just wanted a test that can be updated as the steps move along. So as to not accidentally go from "f128 does nothing" to "typing f128 anywhere in your code causes an ICE" :)

compiler-errors

Some tiny nits, then I think this is fine

compiler/rustc_smir/src/rustc_internal/internal.rs

compiler-errors · 2024-02-28T15:18:38Z

compiler/rustc_symbol_mangling/src/typeid/typeid_itanium_cxx_abi.rs

+                FloatTy::F16 => unimplemented!("f16_f128"),
                FloatTy::F32 => 'f',
                FloatTy::F64 => 'd',
+                FloatTy::F128 => unimplemented!("f16_f128"),


@rcvalle: are there encodings for these types?

There is some more discussion for these at #114607 (comment) and rust-lang/rustc-demangle#64 for this too

I replied on the other issue, but will reply it here as well for easier reference. My recommendation is:

For f128, if it's compatible with __float128 regardless of the underlying representation (i.e., changes between IEEE 754R 128-bit floating point type or IBM extended double depending on the target), we should always encode it as 'g' as it will be how it will also be encoded by Clang. (FWIW, it seems Clang uses IEEE 754R 128-bit floating point type instead of IBM extended double for __float128 by default for ppc64le since May 2023 and __ibm128 also doesn't seem to have a distinct encoding specified in the Itanium C++ ABI mangling.)

For f16, if it's compatible with IEEE 754R 128-bit floating point type (__fp16 and _Float16 in Clang), we should always encode it as 'Dh'; if it's compatible with bfloat16 (which is unlikely?), we should always encode it as 'DF16b'.

The 'Dd, 'De', and 'Df' encodings are for when the Rust compiler and Clang support decimal floating point types.

tl;dr.: We should always encode f128 as 'g', and f16 as 'Dh'.

(Note that this independent/dissociated to Rust mangling/demangling and a different encoding can be used for it.)

Thanks, I will update these in the upcoming PR to add intrinsics.

Total miss that they didn't make __float128 mangle as q to be consistent with f and d :)

Thank you! :)

compiler-errors · 2024-02-28T15:20:19Z

src/tools/clippy/clippy_utils/src/consts.rs

@@ -778,8 +778,10 @@ pub fn mir_to_const<'tcx>(lcx: &LateContext<'tcx>, result: mir::Const<'tcx>) ->
                let range = alloc_range(offset + size * idx, size);
                let val = alloc.read_scalar(&lcx.tcx, range, /* read_provenance */ false).ok()?;
                res.push(match flt {
+                    FloatTy::F16 => unimplemented!("f16_f128"),


In fairness to @tgross35, this is impossible to hit right now. I think this approach is the best for the moment.

compiler-errors · 2024-02-28T15:20:47Z

tests/ui/parser/f16-f128.rs

@@ -0,0 +1,37 @@
+// Make sure we don't ICE while incrementally adding f16 and f128 support


Yeah, this test does literally nothing lol

compiler/stable_mir/src/ty.rs

tgross35 · 2024-02-28T17:53:00Z

@rustbot label +F-f16_and_f128

…itive` Make changes necessary to support these types in the compiler.

compiler-errors · 2024-02-28T18:26:02Z

@bors r+ rollup=never

tgross35 · 2024-02-28T19:08:45Z

Seems like bors might be taking a nap https://rust-lang.zulipchat.com/#narrow/stream/242791-t-infra/topic/Infinite.20redirect.20at.20.60bors.2Erust-lang.2Eorg.60

fmease · 2024-02-28T19:28:00Z

@bors r=compiler-errors rollup=never

bors · 2024-02-28T19:28:03Z

📌 Commit 406790e has been approved by compiler-errors

It is now in the queue for this repository.

tgross35 · 2024-02-29T19:47:51Z

@bors r=compiler-errors rollup=never

bors · 2024-02-29T19:47:55Z

@tgross35: 🔑 Insufficient privileges: Not in reviewers

bors · 2024-02-29T19:47:55Z

@tgross35: 🔑 Insufficient privileges: not in try users

tgross35 · 2024-02-29T19:48:27Z

Yup, need somebody else's help here 😆

compiler-errors · 2024-02-29T19:49:56Z

@bors r+

I believe the PR was already in the queue last time I checked this morning (until it was r-'d). This PR was not given a priority, and rollup=never PRs are merged in order of the date the PR was opened, which is why the current one is merging.

bors · 2024-02-29T19:49:58Z

📌 Commit 406790e has been approved by compiler-errors

It is now in the queue for this repository.

tgross35 · 2024-02-29T19:53:01Z

@bors r+

I believe the PR was already in the queue last time I checked this morning. This PR was not given a priority, and rollup=never PRs are merged in order of the date the PR was opened, which is why the current one is merging.

🤦‍♂️ I thought it was approval date not opening date, thanks

All the same something is weird. Every time #121242 got to the top of the queue in the past 24 hours Bors would give up and stop merging things, this one was right behind it. And it merged a stale commit of mine #119748 (comment)

tgross35 · 2024-03-01T03:18:56Z

Same PR at the top, again Bors stopped running autos. Weird (mentioned to infra on zulip)

bors · 2024-03-01T03:36:14Z

⌛ Testing commit 406790e with merge 6cbf092...

bors · 2024-03-01T06:03:35Z

☀️ Test successful - checks-actions
Approved by: compiler-errors
Pushing 6cbf092 to master...

rust-timer · 2024-03-01T07:17:37Z

Finished benchmarking commit (6cbf092): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.5%	[3.5%, 3.5%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.0%	[-5.1%, -1.5%]	3
All ❌✅ (primary)	-	-	0

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	4.3%	[2.1%, 6.3%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.7%	[-2.8%, -2.6%]	2
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 650.561s -> 653.924s (0.52%)
Artifact size: 311.48 MiB -> 311.55 MiB (0.02%)

…, r=compiler-errors `f16` and `f128` step 2: intrinsics Continuation of rust-lang#121728, another portion of rust-lang#114607. This PR adds `f16` and `f128` intrinsics, and hooks them up to both HIR and LLVM. This is all still unexposed to the frontend, which will probably be the next step. Also update itanium mangling per `@rcvalle's` in https://github.com/rust-lang/rust/pull/121728/files#r1506570300, and fix a typo from step 1. Once these types are usable in code, I will add the codegen tests from rust-lang#114607 (codegen is passing on that branch) This does add more `unimplemented!`s to Clippy, but I still don't think we can do better until library support is added. r? `@compiler-errors` cc `@Nilstrieb` `@rustbot` label +T-compiler +F-f16_and_f128

Rollup merge of rust-lang#121841 - tgross35:f16-f128-step2-intrinsics, r=compiler-errors `f16` and `f128` step 2: intrinsics Continuation of rust-lang#121728, another portion of rust-lang#114607. This PR adds `f16` and `f128` intrinsics, and hooks them up to both HIR and LLVM. This is all still unexposed to the frontend, which will probably be the next step. Also update itanium mangling per `@rcvalle's` in https://github.com/rust-lang/rust/pull/121728/files#r1506570300, and fix a typo from step 1. Once these types are usable in code, I will add the codegen tests from rust-lang#114607 (codegen is passing on that branch) This does add more `unimplemented!`s to Clippy, but I still don't think we can do better until library support is added. r? `@compiler-errors` cc `@Nilstrieb` `@rustbot` label +T-compiler +F-f16_and_f128

…r=compiler-errors Add stubs in IR and ABI for `f16` and `f128` This is the very first step toward the changes in rust-lang#114607 and the [`f16` and `f128` RFC](https://rust-lang.github.io/rfcs/3453-f16-and-f128.html). It adds the types to `rustc_type_ir::FloatTy` and `rustc_abi::Primitive`, and just propagates those out as `unimplemented!` stubs where necessary. These types do not parse yet so there is no feature gate, and it should be okay to use `unimplemented!`. The next steps will probably be AST support with parsing and the feature gate. r? `@compiler-errors` cc `@Nilstrieb` suggested breaking the PR up in rust-lang#120645 (comment)

…ler-errors `f16` and `f128` step 2: intrinsics Continuation of rust-lang/rust#121728, another portion of rust-lang/rust#114607. This PR adds `f16` and `f128` intrinsics, and hooks them up to both HIR and LLVM. This is all still unexposed to the frontend, which will probably be the next step. Also update itanium mangling per `@rcvalle's` in https://github.com/rust-lang/rust/pull/121728/files#r1506570300, and fix a typo from step 1. Once these types are usable in code, I will add the codegen tests from #114607 (codegen is passing on that branch) This does add more `unimplemented!`s to Clippy, but I still don't think we can do better until library support is added. r? `@compiler-errors` cc `@Nilstrieb` `@rustbot` label +T-compiler +F-f16_and_f128

…r=compiler-errors Add stubs in IR and ABI for `f16` and `f128` This is the very first step toward the changes in rust-lang#114607 and the [`f16` and `f128` RFC](https://rust-lang.github.io/rfcs/3453-f16-and-f128.html). It adds the types to `rustc_type_ir::FloatTy` and `rustc_abi::Primitive`, and just propagates those out as `unimplemented!` stubs where necessary. These types do not parse yet so there is no feature gate, and it should be okay to use `unimplemented!`. The next steps will probably be AST support with parsing and the feature gate. r? `@compiler-errors` cc `@Nilstrieb` suggested breaking the PR up in rust-lang#120645 (comment)

Relevant upstream changes: rust-lang/rust#120675: An intrinsic `Symbol` is now wrapped in a `IntrinsicDef` struct, so the relevant part of the code needed to be updated. rust-lang/rust#121464: The second argument of the `create_wrapper_file` function changed from a vector to a string. rust-lang/rust#121662: `NullOp::DebugAssertions` was renamed to `NullOp::UbCheck` and it now has data (currently unused by Kani) rust-lang/rust#121728: Introduces `F16` and `F128`, so needed to add stubs for them rust-lang/rust#121969: `parse_sess` was renamed to `psess`, so updated the relevant code. rust-lang/rust#122059: The `is_val_statically_known` intrinsic is now used in some `core::fmt` code, so had to handle it in (codegen'ed to false). rust-lang/rust#122233: This added a new `retag_box_to_raw` intrinsic. This is an operation that is primarily relevant for stacked borrows. For Kani, we just return the pointer. Resolves #3057

rustbot assigned compiler-errors Feb 28, 2024

tgross35 force-pushed the f16-f128-step1-ty-updates branch 2 times, most recently from c126076 to 44c1336 Compare February 28, 2024 09:25

flip1995 reviewed Feb 28, 2024

View reviewed changes

Nilstrieb reviewed Feb 28, 2024

View reviewed changes

compiler-errors approved these changes Feb 28, 2024

View reviewed changes

tgross35 force-pushed the f16-f128-step1-ty-updates branch from 44c1336 to 2dc026a Compare February 28, 2024 17:50

compiler-errors reviewed Feb 28, 2024

View reviewed changes

compiler/stable_mir/src/ty.rs Outdated Show resolved Hide resolved

rustbot added the F-f16_and_f128 `#![feature(f16_and_f128)]` label Feb 28, 2024

tgross35 added 5 commits February 28, 2024 12:58

Add f16 and f128 to rustc_type_ir::FloatTy and `rustc_abi::Prim…

e3f63d9

…itive` Make changes necessary to support these types in the compiler.

Add basic support for f16 and f128 to rustdoc

98e1728

Add stubs for f16 and f128 to clippy

ae70502

Add stubs for f16 and f128 to miri

c17a516

Add a basic test for f16 and f128

406790e

tgross35 force-pushed the f16-f128-step1-ty-updates branch from 2dc026a to 406790e Compare February 28, 2024 17:58

tgross35 mentioned this pull request Feb 28, 2024

Tracking Issue for f16 and f128 float types #116909

Open

3 tasks

fmease closed this Feb 28, 2024

fmease reopened this Feb 28, 2024

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Feb 28, 2024

bors added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Feb 29, 2024

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 29, 2024

bors added the merged-by-bors This PR was explicitly merged by bors. label Mar 1, 2024

bors merged commit 6cbf092 into rust-lang:master Mar 1, 2024
23 checks passed

rustbot added this to the 1.78.0 milestone Mar 1, 2024

tgross35 deleted the f16-f128-step1-ty-updates branch March 1, 2024 06:12

This was referenced Mar 1, 2024

Support f16 and f128 rust-lang/rustc_codegen_gcc#461

Open

Support f16 and f128 rust-lang/rustc_codegen_cranelift#1461

Open

f16 and f128 step 2: intrinsics #121841

Merged

tgross35 mentioned this pull request Mar 3, 2024

f16 and f128 step 3: compiler support & feature gate #121926

Merged

rcvalle mentioned this pull request Mar 6, 2024

Add demangling for f16 and f128 rust-lang/rustc-demangle#64

Open

zhassan-aws mentioned this pull request Mar 11, 2024

Upgrade toolchain to 2024-03-11 model-checking/kani#3071

Merged

		@@ -0,0 +1,37 @@
		// Make sure we don't ICE while incrementally adding f16 and f128 support

Add stubs in IR and ABI for f16 and f128 #121728

Add stubs in IR and ABI for f16 and f128 #121728

Conversation

tgross35 commented Feb 28, 2024 • edited

rustbot commented Feb 28, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tgross35 Feb 28, 2024 • edited

Choose a reason for hiding this comment

compiler-errors left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tgross35 Feb 28, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tgross35 commented Feb 28, 2024

compiler-errors commented Feb 28, 2024

tgross35 commented Feb 28, 2024

fmease commented Feb 28, 2024

bors commented Feb 28, 2024

tgross35 commented Feb 29, 2024

bors commented Feb 29, 2024

bors commented Feb 29, 2024

tgross35 commented Feb 29, 2024

compiler-errors commented Feb 29, 2024 • edited

bors commented Feb 29, 2024

tgross35 commented Feb 29, 2024 • edited

tgross35 commented Mar 1, 2024 • edited

bors commented Mar 1, 2024

bors commented Mar 1, 2024

rust-timer commented Mar 1, 2024

Overall result: no relevant changes - no action needed

Add stubs in IR and ABI for `f16` and `f128` #121728

Add stubs in IR and ABI for `f16` and `f128` #121728

tgross35 commented Feb 28, 2024 •

edited

tgross35 Feb 28, 2024 •

edited

tgross35 Feb 28, 2024 •

edited

compiler-errors commented Feb 29, 2024 •

edited

tgross35 commented Feb 29, 2024 •

edited

tgross35 commented Mar 1, 2024 •

edited