Compute generator saved locals on MIR #101692

cjgillot · 2022-09-11T16:05:35Z

Generators are currently type-checked by introducing a witness type variable, which is unified with a GeneratorWitness(captured types) whose purpose is to ensure that the auto traits correctly migrate from the captured types to the witness type. This requires computing the captured types on HIR during type-checking, only to re-do it on MIR later.

This PR proposes to drop the HIR-based computation, and only keep the MIR one. This is done in 3 steps.

During type-checking, the witness type variable is never unified. This allows to stall all the obligations that depend on it until the end of type-checking. Then, the stalled obligations are marked as successful, and saved into the typeck results for later verification.
At type-checking writeback, witness is replaced by GeneratorWitnessMIR(def_id, substs). From this point on, all trait selection involving GeneratorWitnessMIR will fetch the MIR-computed locals, similar to what opaque types do. There is no lifetime to be preserved here: we consider all the lifetimes appearing in this witness type to be higher-ranked.
After borrowck, the stashed obligations are verified against the actually computed types, in the check_generator_obligations query. If any obligation was wrongly marked as fulfilled in step 1, it should be reported here.

There are still many issues:

~~I am not too happy having to filter out some locals from the checked bounds, I think this is MIR building that introduces raw pointers polluting the analysis;~~ solved by a check specific to static variables.
the diagnostics for captured types don't show where they are used/dropped;
I do not attempt to support chalk.

cc @eholk @jyn514 for the drop-tracking work
r? @oli-obk as you warned me of potential unsoundness

jyn514 · 2022-09-11T16:20:13Z

I don't currently have time to review a PR this large but I think this work is super exciting, thank you for tackling it ❤️

cjgillot · 2022-09-11T21:18:48Z

@bors try @rust-timer queue

rust-timer · 2022-09-11T21:18:49Z

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

bors · 2022-09-11T21:18:57Z

⌛ Trying commit 6c9ec2af2224dcbfba350ffdd69a27e5030000c1 with merge 8ea250bdb76176ecb0225baa34aa634b024ae0e2...

bors · 2022-09-11T23:09:22Z

☀️ Try build successful - checks-actions
Build commit: 8ea250bdb76176ecb0225baa34aa634b024ae0e2 (8ea250bdb76176ecb0225baa34aa634b024ae0e2)

rust-timer · 2022-09-11T23:09:24Z

Queued 8ea250bdb76176ecb0225baa34aa634b024ae0e2 with parent 0d56e34, future comparison URL.

rust-timer · 2022-09-12T00:53:35Z

Finished benchmarking commit (8ea250bdb76176ecb0225baa34aa634b024ae0e2): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: +S-waiting-on-review -S-waiting-on-perf +perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean¹	range	count²
Regressions ❌ (primary)	0.4%	[0.2%, 0.8%]	42
Regressions ❌ (secondary)	1.1%	[0.3%, 2.8%]	27
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-49.7%	[-86.8%, -0.9%]	11
All ❌✅ (primary)	0.4%	[0.2%, 0.8%]	42

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.

	mean¹	range	count²
Regressions ❌ (primary)	1.2%	[0.6%, 1.6%]	4
Regressions ❌ (secondary)	1.5%	[1.5%, 1.5%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.7%	[-2.6%, -0.7%]	2
All ❌✅ (primary)	1.2%	[0.6%, 1.6%]	4

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.

	mean¹	range	count²
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.6%	[2.4%, 2.8%]	3
Improvements ✅ (primary)	-2.5%	[-2.5%, -2.5%]	1
Improvements ✅ (secondary)	-21.1%	[-76.4%, -1.6%]	21
All ❌✅ (primary)	-2.5%	[-2.5%, -2.5%]	1

the arithmetic mean of the percent change ↩ ↩² ↩³
number of relevant changes ↩ ↩² ↩³

cjgillot · 2022-09-12T16:50:42Z

@craterbot check

craterbot · 2022-09-15T17:21:35Z

🎉 Experiment pr-101692 is completed!
📊 11235 regressed and 6 fixed (243276 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the blacklist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

rust-timer · 2023-01-28T06:30:20Z

Finished benchmarking commit (6cd6bad): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.6%	[0.4%, 0.6%]	14
Regressions ❌ (secondary)	2.2%	[0.3%, 6.0%]	22
Improvements ✅ (primary)	-0.4%	[-0.7%, -0.2%]	12
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.1%	[-0.7%, 0.6%]	26

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)	2.5%	[2.5%, 2.5%]	1
Regressions ❌ (secondary)	2.3%	[2.3%, 2.3%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.4%	[-1.4%, -1.4%]	1
All ❌✅ (primary)	2.5%	[2.5%, 2.5%]	1

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)	3.5%	[2.3%, 4.8%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

…iler-errors Only compute mir_generator_witnesses query in drop_tracking_mir mode. Attempt to fix the perf regression in rust-lang#101692 r? `@ghost`

…ess, r=oli-obk" This reverts commit 6cd6bad, reversing changes made to 7d4df2d.

cjgillot · 2023-01-29T13:23:03Z

perf: I've managed to remove the 5% regression with #107406. Is left a ~1% regression on some incr-unchanged and incr-patched (measured in #107438).

Mark-Simulacrum · 2023-01-31T14:02:54Z

Thanks! That seems sufficient to me, marking as triaged.

cjgillot · 2023-01-31T18:29:54Z

#107443 has removed the remaining 1%.

…li-obk Compute generator saved locals on MIR Generators are currently type-checked by introducing a `witness` type variable, which is unified with a `GeneratorWitness(captured types)` whose purpose is to ensure that the auto traits correctly migrate from the captured types to the `witness` type. This requires computing the captured types on HIR during type-checking, only to re-do it on MIR later. This PR proposes to drop the HIR-based computation, and only keep the MIR one. This is done in 3 steps. 1. During type-checking, the `witness` type variable is never unified. This allows to stall all the obligations that depend on it until the end of type-checking. Then, the stalled obligations are marked as successful, and saved into the typeck results for later verification. 2. At type-checking writeback, `witness` is replaced by `GeneratorWitnessMIR(def_id, substs)`. From this point on, all trait selection involving `GeneratorWitnessMIR` will fetch the MIR-computed locals, similar to what opaque types do. There is no lifetime to be preserved here: we consider all the lifetimes appearing in this witness type to be higher-ranked. 3. After borrowck, the stashed obligations are verified against the actually computed types, in the `check_generator_obligations` query. If any obligation was wrongly marked as fulfilled in step 1, it should be reported here. There are still many issues: - ~I am not too happy having to filter out some locals from the checked bounds, I think this is MIR building that introduces raw pointers polluting the analysis;~ solved by a check specific to static variables. - the diagnostics for captured types don't show where they are used/dropped; - I do not attempt to support chalk. cc `@eholk` `@jyn514` for the drop-tracking work r? `@oli-obk` as you warned me of potential unsoundness

(GeneratorWitnessMIR) Compute generator saved locals on MIR rust-lang/rust#101692 (ParamEnv) InstCombine away intrinsic validity assertions rust-lang/rust#105582 (Primitive::Pointer) abi: add AddressSpace field to Primitive::Pointer rust-lang/rust#107248

Enable -Zdrop-tracking-mir by default This PR enables the `drop-tracking-mir` flag by default. This flag was initially implemented in rust-lang#101692. This flag computes auto-traits on generators based on their analysis MIR, instead of trying to compute on the HIR body. This removes the need for HIR-based drop-tracking, as we can now reuse the same code to compute generator witness types and to compute generator interior fields.

…try> Support async recursive calls (as long as they have indirection) Before rust-lang#101692, we stored coroutine witness types directly inside of the coroutine. That means that a coroutine could not contain itself (as a witness field) without creating a cycle in the type representation of the coroutine, which we detected with the `OpaqueTypeExpander`, which is used to detect cycles when expanding opaque types after that are inferred to contain themselves. After `-Zdrop-tracking-mir` was stabilized, we no longer store these generator witness fields directly, but instead behind a def-id based query. That means there is no technical obstacle in the compiler preventing coroutines from containing themselves per se, other than the fact that for a coroutine to have a non-infinite layout, it must contain itself wrapped in a layer of allocation indirection (like a `Box`). This means that it should be valid for this code to work: ``` async fn async_fibonacci(i: u32) -> u32 { if i == 0 || i == 1 { i } else { Box::pin(async_fibonacci(i - 1)).await + Box::pin(async_fibonacci(i - 2)).await } } ``` Whereas previously, you'd need to coerce the future to `Pin<Box<dyn Future<Output = ...>>` before `await`ing it, to prevent the async's desugared coroutine from containing itself across as await point. This PR does two things: 1. Remove the behavior from `OpaqueTypeExpander` where it intentionally fetches and walks through the coroutine's witness fields. This was kept around after `-Zdrop-tracking-mir` was stabilized so we would not be introducing new stable behavior, and to preserve the much better diagnostics of async recursion compared to a layout error. 2. Reworks the way we report layout errors having to do with coroutines, to make up for the diagnostic regressions introduced by (1.). We actually do even better now, pointing out the call sites of the recursion!

Support async recursive calls (as long as they have indirection) Before rust-lang#101692, we stored coroutine witness types directly inside of the coroutine. That means that a coroutine could not contain itself (as a witness field) without creating a cycle in the type representation of the coroutine, which we detected with the `OpaqueTypeExpander`, which is used to detect cycles when expanding opaque types after that are inferred to contain themselves. After `-Zdrop-tracking-mir` was stabilized, we no longer store these generator witness fields directly, but instead behind a def-id based query. That means there is no technical obstacle in the compiler preventing coroutines from containing themselves per se, other than the fact that for a coroutine to have a non-infinite layout, it must contain itself wrapped in a layer of allocation indirection (like a `Box`). This means that it should be valid for this code to work: ``` async fn async_fibonacci(i: u32) -> u32 { if i == 0 || i == 1 { i } else { Box::pin(async_fibonacci(i - 1)).await + Box::pin(async_fibonacci(i - 2)).await } } ``` Whereas previously, you'd need to coerce the future to `Pin<Box<dyn Future<Output = ...>>` before `await`ing it, to prevent the async's desugared coroutine from containing itself across as await point. This PR does two things: 1. Only report an error if an opaque expansion cycle is detected *not* through coroutine witness fields. * Instead, if we find an opaque cycle through coroutine witness fields, we compute the layout of the coroutine. If that results in a cycle error, we report it as a recursive async fn. 4. Reworks the way we report layout errors having to do with coroutines, to make up for the diagnostic regressions introduced by (1.). We actually do even better now, pointing out the call sites of the recursion!

rust-highfive assigned oli-obk Sep 11, 2022

rustbot added T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-rustdoc Relevant to the rustdoc team, which will review and decide on the PR/issue. labels Sep 11, 2022

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Sep 11, 2022

cjgillot force-pushed the generator-lazy-witness branch from ee1c144 to 2941ec7 Compare September 11, 2022 18:13

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Sep 11, 2022

This comment has been minimized.

Sign in to view

rustbot added perf-regression Performance regression. and removed S-waiting-on-perf Status: Waiting on a perf run to be completed. labels Sep 12, 2022

This comment was marked as outdated.

Sign in to view

craterbot added S-waiting-on-crater Status: Waiting on a crater run to be completed. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 12, 2022

This comment was marked as outdated.

Sign in to view

craterbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-crater Status: Waiting on a crater run to be completed. labels Sep 15, 2022

cjgillot force-pushed the generator-lazy-witness branch 2 times, most recently from 36cbe4d to f2506c0 Compare September 20, 2022 19:56

This comment has been minimized.

Sign in to view

cjgillot force-pushed the generator-lazy-witness branch from d39d4d0 to b134e71 Compare September 21, 2022 17:26

bors mentioned this pull request Jan 28, 2023

Rollup of 6 pull requests #107356

Closed

cjgillot mentioned this pull request Jan 28, 2023

Only compute mir_generator_witnesses query in drop_tracking_mir mode. #107406

Merged

cjgillot deleted the generator-lazy-witness branch January 28, 2023 10:54

cjgillot added a commit to cjgillot/rust that referenced this pull request Jan 29, 2023

Revert "Auto merge of rust-lang#101692 - cjgillot:generator-lazy-witn…

16f39c5

…ess, r=oli-obk" This reverts commit 6cd6bad, reversing changes made to 7d4df2d.

Mark-Simulacrum added the perf-regression-triaged The performance regression has been triaged. label Jan 31, 2023

calleum mentioned this pull request Feb 2, 2023

Box syntax and generator_clone can lead to double free #105084

Closed

dtolnay mentioned this pull request Feb 3, 2023

Autotrait bounds on dyn-safe trait methods #107082

Merged

tedinski mentioned this pull request Feb 3, 2023

Update Rust toolchain to nightly-2023-02-03 model-checking/kani#2184

Closed

4 tasks

eholk mentioned this pull request Feb 23, 2023

DO NOT MERGE (yet): Enable -Zdrop-tracking by default #97334

Closed

qinheping mentioned this pull request Mar 9, 2023

Upgrade Rust toolchain to nightly-2023-02-03 model-checking/kani#2291

Merged

4 tasks

ehuss mentioned this pull request Mar 20, 2023

CI random failure on x86_64-apple-2 InvalidInput in create_dir_all while running test/rustdoc/primitive/no_std.rs #109397

Closed

This was referenced Apr 28, 2023

ICE: ast lowering: no entry for node id: NodeId(53) #110899

Closed

ICE: expected ConstKind::Value, got Unevaluated(UnevaluatedConst... #110892

Closed

cjgillot mentioned this pull request Jun 8, 2023

Enable -Zdrop-tracking-mir by default #107421

Merged

compiler-errors mentioned this pull request Nov 12, 2023

Support async recursive calls (as long as they have indirection) #117703

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Compute generator saved locals on MIR #101692

Compute generator saved locals on MIR #101692

cjgillot commented Sep 11, 2022 •

edited

Loading

jyn514 commented Sep 11, 2022

This comment has been minimized.

cjgillot commented Sep 11, 2022

rust-timer commented Sep 11, 2022

bors commented Sep 11, 2022

This comment has been minimized.

bors commented Sep 11, 2022

rust-timer commented Sep 11, 2022

rust-timer commented Sep 12, 2022

cjgillot commented Sep 12, 2022

This comment was marked as outdated.

This comment was marked as outdated.

craterbot commented Sep 15, 2022

This comment has been minimized.

rust-timer commented Jan 28, 2023

cjgillot commented Jan 29, 2023

Mark-Simulacrum commented Jan 31, 2023

cjgillot commented Jan 31, 2023

Compute generator saved locals on MIR #101692

Compute generator saved locals on MIR #101692

Conversation

cjgillot commented Sep 11, 2022 • edited Loading

jyn514 commented Sep 11, 2022

This comment has been minimized.

cjgillot commented Sep 11, 2022

rust-timer commented Sep 11, 2022

bors commented Sep 11, 2022

This comment has been minimized.

bors commented Sep 11, 2022

rust-timer commented Sep 11, 2022

rust-timer commented Sep 12, 2022

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Instruction count

Max RSS (memory usage)

Cycles

Footnotes

cjgillot commented Sep 12, 2022

This comment was marked as outdated.

This comment was marked as outdated.

craterbot commented Sep 15, 2022

This comment has been minimized.

rust-timer commented Jan 28, 2023

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Instruction count

Max RSS (memory usage)

Cycles

cjgillot commented Jan 29, 2023

Mark-Simulacrum commented Jan 31, 2023

cjgillot commented Jan 31, 2023

cjgillot commented Sep 11, 2022 •

edited

Loading