Allow MIR debuginfo to point to a variable's address

[cjgillot]

MIR optimizations currently do not to operate on borrowed locals.

When enabling #106285, many borrows will be left as-is because they are used in debuginfo. This pass allows to replace this pattern directly in MIR debuginfo:

a => _1
_1 = &raw? mut? _2

becomes

a => &_2
// No statement to borrow _2.

This pass is implemented as a drive-by in ReferencePropagation MIR pass.

This transformation allows following following MIR opts to treat _2 as an unborrowed local, and optimize it as such, even in builds with debuginfo.

In codegen, when encountering a => &..&_2, we create a list of allocas:

store ptr %_2.dbg.spill, ptr %a.ref0.dbg.spill
store ptr %a.ref0.dbg.spill, ptr %a.ref1.dbg.spill
...
call void @llvm.dbg.declare(metadata ptr %a.ref{n}.dbg.spill, /* ... */)

Caveat: this transformation looses the exact type, we do not differentiate a as a immutable, mutable reference or a raw pointer. Everything is declared to *mut to codegen. I'm not convinced this is a blocker.

[rustbot]

r? @jackh726

(rustbot has picked a reviewer for you, use r? to override)

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[jackh726]

r? mir-opt

[oli-obk]

r=me with the debuginfo test extended to have a case where references is 2 or more.

[bors]

☔ The latest upstream changes (presumably #111493) made this pull request unmergeable. Please resolve the merge conflicts.

[cjgillot]

@oli-obk to simplify triggering the case with 2 or more references, I added 1965237 too. Is the debuginfo test sufficient, or do you need me to check more than 2?

[oli-obk]

That's perfect, thanks!

@bors r+

[bors]

📌 Commit 8fb888d has been approved by oli-obk

It is now in the queue for this repository.

[bors]

⌛ Testing commit 8fb888d with merge bc88895...

[bors]

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing bc88895 to master...

[rust-timer]

Finished benchmarking commit (bc88895): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

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
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.6%	[-2.6%, -2.6%]	1
All ❌✅ (primary)	-	-	0

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)	4.7%	[4.2%, 5.5%]	3
Regressions ❌ (secondary)	2.3%	[2.1%, 2.5%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.0%	[-2.0%, -2.0%]	1
All ❌✅ (primary)	4.7%	[4.2%, 5.5%]	3

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.2%	[-3.2%, -3.2%]	1
All ❌✅ (primary)	-	-	0

Binary size

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)	0.1%	[0.0%, 0.2%]	59
Regressions ❌ (secondary)	0.1%	[0.0%, 0.2%]	32
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.1%	[0.0%, 0.2%]	59

Bootstrap: 661.349s -> 660.613s (-0.11%)