Looser check for overflowing_binary_op #91856

ouz-a · 2021-12-13T09:43:12Z

Fix for issue #91636 tight check resulted in ICE, this makes the check a little looser. It seems eq allows comparing of supertype and subtype if lhs = supertype and rhs = subtype but not vice versa, is this intended behavior ?

rust-highfive · 2021-12-13T09:43:15Z

Some changes occured to the CTFE / Miri engine

cc @rust-lang/miri

rust-highfive · 2021-12-13T09:43:16Z

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @cjgillot (or someone else) soon.

Please see the contribution instructions for more information.

oli-obk · 2021-12-13T11:50:03Z

Higher kinded lifetimes strike again... I think the real fix would be to use the subtyping infra, but that could be rather expensive for such a common op. Although it short circuits on type equality, so it should be cheap most of the time

ouz-a · 2021-12-13T12:03:41Z

Higher kinded lifetimes strike again... I think the real fix would be to use the subtyping infra, but that could be rather expensive for such a common op. Although it short circuits on type equality, so it should be cheap most of the time

I tried other things to fix this but they broke a lot of stuff 😸 I will look at subtyping infra, thanks!

oli-obk · 2021-12-13T13:02:39Z

I think something like tcx.at(span).sub(rhs.ty, lhs.ty).is_ok() should work

oli-obk · 2021-12-13T13:03:49Z

Or maybe check how mir validation does it, tho I fear it just bails out on such lifetimes

ouz-a · 2021-12-13T13:05:58Z

Or maybe check how mir validation does it, tho I fear it just bails out on such lifetimes

I think because of coercion mir gives up. As I posted in the issue if I use println! program compiles but there are bunch of MIR errors.

oli-obk · 2021-12-13T13:16:21Z

OK, I guess if the subtyping check works here, we should make mir validation use it, too?

ouz-a · 2021-12-13T13:37:33Z

Again for such a small case wouldn't adding an extra layer of subtyping check make it slow ? 😕

ouz-a · 2021-12-13T13:58:25Z

I think something like tcx.at(span).sub(rhs.ty, lhs.ty).is_ok() should work

Also tcx.at(span) doesn't have sub method did you meant infcx ?

oli-obk · 2021-12-13T14:07:03Z

Oh... I did... yea... spawning an infcx is serious overkill...

OK, let's keep the PR as it is, doing the real check is too expensive for CTFE

@RalfJung that ok with you?

r? @RalfJung

RalfJung · 2021-12-13T15:50:06Z

OK, I guess if the subtyping check works here, we should make mir validation use it, too?

I think MIR validation currently doesn't check much about operators (unary and binary alike), so that sounds like a separate PR to me.

RalfJung · 2021-12-13T15:50:49Z

compiler/rustc_const_eval/src/interpret/operator.rs

@@ -330,7 +330,7 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
            _ if left.layout.ty.is_any_ptr() => {
                // The RHS type must be the same *or an integer type* (for `Offset`).


Please update this comment, adding a link to #91636.

RalfJung · 2021-12-13T15:52:11Z

Yeah I am fine with this change. I proposed it. :)

@ouz-a could you add #91636 (comment) add a regression test? Other than that and the comment nit I raised, this looks good to me!

ouz-a · 2021-12-13T16:06:59Z

Going to update the comment and add a regression test. Thanks!

oli-obk · 2021-12-14T08:03:15Z

@bors r+

bors · 2021-12-14T08:03:17Z

📌 Commit b85762f has been approved by oli-obk

oli-obk · 2021-12-14T08:03:27Z

@bors rollup

RalfJung · 2021-12-14T14:20:19Z

compiler/rustc_const_eval/src/interpret/operator.rs

@@ -328,9 +328,10 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
                self.binary_int_op(bin_op, l, left.layout, r, right.layout)
            }
            _ if left.layout.ty.is_any_ptr() => {
-                // The RHS type must be the same *or an integer type* (for `Offset`).
+                // The RHS type must be a `pointer` *or an integer type* (for `Offset`).
+                // (This is workaround for the issue #91636)


Suggested change

// (This is workaround for the issue #91636)

// (Even when both sides are pointers, their type might differ, see issue #91636)

RalfJung · 2021-12-14T14:21:02Z

Comment nit
@bors r-

oli-obk · 2021-12-14T17:08:53Z

@bors r+ rollup

bors · 2021-12-14T17:08:54Z

📌 Commit a5054e3 has been approved by oli-obk

…askrgr Rollup of 7 pull requests Successful merges: - rust-lang#89825 (Make split_inclusive() on an empty slice yield an empty output) - rust-lang#91239 (regression test for issue 87490) - rust-lang#91597 (Recover on invalid operators `<>` and `<=>`) - rust-lang#91774 (Fix typo for MutVisitor) - rust-lang#91786 (Return an error when `eval_rvalue_with_identities` fails) - rust-lang#91798 (Avoid suggest adding `self` in visibility spec) - rust-lang#91856 (Looser check for overflowing_binary_op) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup

add regression test Adds a regression test for rust-lang/rust#91636 (which was fixed by rust-lang/rust#91856)

Looser check for binary_op_overflow

01948e2

rust-highfive assigned cjgillot Dec 13, 2021

rustbot added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label Dec 13, 2021

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Dec 13, 2021

This comment has been minimized.

# to view

formatting

a01b13d

rust-highfive assigned RalfJung and unassigned cjgillot Dec 13, 2021

RalfJung reviewed Dec 13, 2021

View reviewed changes

Add regression test and comment

b6c8098

This comment has been minimized.

# to view

test should pass 😅

b85762f

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 Dec 14, 2021

RalfJung reviewed Dec 14, 2021

View reviewed changes

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Dec 14, 2021

comment update

a5054e3

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 Dec 14, 2021

matthiaskrgr mentioned this pull request Dec 14, 2021

Rollup of 7 pull requests #91933

Merged

bors merged commit bae9270 into rust-lang:master Dec 15, 2021

rustbot added this to the 1.59.0 milestone Dec 15, 2021

This was referenced Dec 15, 2021

Compiler panics when comparing functions #91636

Closed

add regression test rust-lang/miri#1938

Merged

bors added a commit to rust-lang/miri that referenced this pull request Dec 15, 2021

Auto merge of #1938 - RalfJung:test, r=RalfJung

7f0e652

add regression test Adds a regression test for rust-lang/rust#91636 (which was fixed by rust-lang/rust#91856)

bors added a commit to rust-lang/miri that referenced this pull request Dec 15, 2021

Auto merge of #1938 - RalfJung:test, r=RalfJung

51e9630

add regression test Adds a regression test for rust-lang/rust#91636 (which was fixed by rust-lang/rust#91856)

bors added a commit to rust-lang/miri that referenced this pull request Dec 15, 2021

Auto merge of #1938 - RalfJung:test, r=RalfJung

2eb3b37

add regression test Adds a regression test for rust-lang/rust#91636 (which was fixed by rust-lang/rust#91856)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Looser check for overflowing_binary_op #91856

Looser check for overflowing_binary_op #91856

ouz-a commented Dec 13, 2021

rust-highfive commented Dec 13, 2021

rust-highfive commented Dec 13, 2021

This comment has been minimized.

oli-obk commented Dec 13, 2021

ouz-a commented Dec 13, 2021

oli-obk commented Dec 13, 2021

oli-obk commented Dec 13, 2021

ouz-a commented Dec 13, 2021

oli-obk commented Dec 13, 2021

ouz-a commented Dec 13, 2021 •

edited

Loading

ouz-a commented Dec 13, 2021

oli-obk commented Dec 13, 2021

RalfJung commented Dec 13, 2021

RalfJung Dec 13, 2021

RalfJung commented Dec 13, 2021

ouz-a commented Dec 13, 2021

This comment has been minimized.

oli-obk commented Dec 14, 2021

bors commented Dec 14, 2021

oli-obk commented Dec 14, 2021

RalfJung Dec 14, 2021

RalfJung commented Dec 14, 2021

oli-obk commented Dec 14, 2021

bors commented Dec 14, 2021

		@@ -330,7 +330,7 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
		_ if left.layout.ty.is_any_ptr() => {
		// The RHS type must be the same or an integer type (for `Offset`).

	// (This is workaround for the issue #91636)
	// (Even when both sides are pointers, their type might differ, see issue #91636)

Looser check for overflowing_binary_op #91856

Looser check for overflowing_binary_op #91856

Conversation

ouz-a commented Dec 13, 2021

rust-highfive commented Dec 13, 2021

rust-highfive commented Dec 13, 2021

This comment has been minimized.

oli-obk commented Dec 13, 2021

ouz-a commented Dec 13, 2021

oli-obk commented Dec 13, 2021

oli-obk commented Dec 13, 2021

ouz-a commented Dec 13, 2021

oli-obk commented Dec 13, 2021

ouz-a commented Dec 13, 2021 • edited Loading

ouz-a commented Dec 13, 2021

oli-obk commented Dec 13, 2021

RalfJung commented Dec 13, 2021

RalfJung Dec 13, 2021

Choose a reason for hiding this comment

RalfJung commented Dec 13, 2021

ouz-a commented Dec 13, 2021

This comment has been minimized.

oli-obk commented Dec 14, 2021

bors commented Dec 14, 2021

oli-obk commented Dec 14, 2021

RalfJung Dec 14, 2021

Choose a reason for hiding this comment

RalfJung commented Dec 14, 2021

oli-obk commented Dec 14, 2021

bors commented Dec 14, 2021

ouz-a commented Dec 13, 2021 •

edited

Loading