unstably allow constants to refer to statics and read from immutable statics

[RalfJung]

I am not aware of any fundamental reason why we cannot allow constants to mention statics. What we really need is that constants do not read from statics that can change their value:

• This would break the principle that "constants behave as-if their expression was inlined everywhere and executed at runtime". This is enforced by halting const-eval interpretation when a read from a mutable global occurs.
• When building a valtree we want to be sure that the constant and everything it refers to is truly immutable. This is enforced by aborting valtree construction when a read from a mutable global occurs.

r? @oli-obk -- if you are okay with experimenting with this feature, I will create a tracking issue.
Based on and blocked on #119044; only the last commit is new.

[rustbot]

The Miri subtree was changed

cc @rust-lang/miri

Some changes occurred to the CTFE / Miri engine

cc @rust-lang/miri

Some changes occurred to the CTFE / Miri engine

cc @rust-lang/miri

[oli-obk]

fine by me, as long as the tracking issue mentions this new source of monomorphization time errors as an open question to put to T-lang

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[RalfJung]

I realized I have to make validation descend into statics, or else valtree construction can encounter unvalidated data. This also makes the feature more conservative: consts that refer to mutable memory behind a shared reference are immediately rejected, not just when they are used as a pattern.

[oli-obk]

seems fishy to allow valtrees to read from immutable statics. That means we meaningfully lose the static's identity during rountdrip.

[RalfJung]

Valtrees are about the value. How would the static's identity matter?

[RalfJung]

@bors r=oli-obk

[bors]

📌 Commit 4def373 has been approved by oli-obk

It is now in the queue for this repository.

[bors]

⌛ Testing commit 4def373 with merge 6cc4843...

[bors]

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing 6cc4843 to master...

[rust-timer]

Finished benchmarking commit (6cc4843): comparison URL.

Overall result: ❌ regressions - 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)	0.2%	[0.2%, 0.2%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
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)	-	-	0
Regressions ❌ (secondary)	3.3%	[2.3%, 4.4%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

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

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
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Bootstrap: 666.587s -> 667.259s (0.10%)
Artifact size: 308.05 MiB -> 308.02 MiB (-0.01%)