Avoid creating overlapping assignments in MatchBranchSimplification

[saethlin]

MatchBranchSimplification can turn a match into an assignment, but doesn't make sure the LHS and RHS do not overlap. For example, the enum version of the mir-opt test in this PR is currently compiled with -Zmir-opt-level=0 -Zmir-enable-passes=+MatchBranchSimplification to

fn g(_1: Foo) -> Foo {
    let mut _0: Foo;
    let mut _2: Foo;
    let mut _3: *const Foo;
    let mut _4: u8;

    bb0: {
        _2 = copy _1;
        _3 = &raw const _2;
        _4 = discriminant((*_3));
        _2 = no_retag copy (*_3); // Since _3 = &raw const 2 above, this overlaps!
        _0 = copy _2;
        return;
    }
}

While this is technically a miscompile, I haven't found any way to turn this into something exciting, in part because the only thing that can be made to overlap here is a by-value enum discriminant which is always register-sized so backends lower this in a way that is okay with overlapping.

[saethlin]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 511c733 (511c733988d791a6dfd9fbc8c8b3d608dc4f84c8)
Base parent: 574ff7d (574ff7d98bd6d037e5236a8453029173b32631fd)

[rust-timer]

Finished benchmarking commit (511c733): comparison URL.

Overall result: ❌✅ regressions and improvements - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

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

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.3%	[0.3%, 0.3%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.4%	[-0.4%, -0.4%]	1
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (secondary 1.7%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	4.7%	[3.4%, 5.1%]	5
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-5.8%	[-7.9%, -3.7%]	2
All ❌✅ (primary)	-	-	0

Cycles

Results (primary -0.7%, secondary 0.5%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	1.3%	[1.3%, 1.3%]	1
Regressions ❌ (secondary)	13.9%	[8.4%, 24.5%]	3
Improvements ✅ (primary)	-2.6%	[-2.6%, -2.6%]	1
Improvements ✅ (secondary)	-5.2%	[-7.4%, -3.3%]	7
All ❌✅ (primary)	-0.7%	[-2.6%, 1.3%]	2

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 496.664s -> 494.065s (-0.52%)
Artifact size: 406.89 MiB -> 406.92 MiB (0.01%)

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[rustbot]

r? @TaKO8Ki

rustbot has assigned @TaKO8Ki.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 76 candidates
• Random selection from 19 candidates

[saethlin]

r? mir-opt