Update LLVM to 15.0.7

[alexcrichton]

This commit pulls in rust-lang/llvm-project#143 which updates the LLVM version used by rustc to 15.0.7, namely pulling in https://reviews.llvm.org/D136110 which is needed for some work I'm working on with wasm.

[rustbot]

r? @cuviper

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

[rustbot]

⚠️ Warning ⚠️

• These commits modify submodules.

[cuviper]

@bors r+ rollup=never

[bors]

📌 Commit 50a4612 has been approved by cuviper

It is now in the queue for this repository.

[alexcrichton]

Thanks for the speedy reviews! Would you be up for rubber-stamping #106644 as well perhaps? If not no worries!

[bors]

⌛ Testing commit 50a4612 with merge 645587297bd50bf6bfe30838d2f4d4e4f30f08c4...

[bors]

💔 Test failed - checks-actions

[cuviper]

CMake 3.13.4 or higher is required. You are running version 3.10.2

Hmm, indeed ubuntu:18.04 does not have new enough cmake. The command leading to that was:

RUSTDOCFLAGS=--document-private-items --document-hidden-items python3 ../x.py doc --stage 0 compiler

... which is new from #106670 (cc @albertlarsan68 @Mark-Simulacrum). I'm guessing that mingw-check got a cached LLVM when that PR landed? I can see about simply updating the mingw-check containers, but it seems a little surprising for that to start building the entire compiler if it wasn't before.

[Mark-Simulacrum]

Yeah, most likely. I think for now reverting that addition is right (the more principled fix is probably to avoid building LLVM for doc builds).

[cuviper]

Well, the upgrade proved easy enough, and is probably a good idea anyway -- #106948.

I think we can consider that revert separately, on its own merits, but it shouldn't block here after that upgrade.

[alexcrichton]

Thanks for the assistance in tracking this down! I suspected that something like that was happening when I didn't find any cmake changes in the 15.0.6 -> 15.0.7 diff. I'll wait for #106948 which I think should enable a retry of this to go through.

[a1phyr]

Could this PR also include rust-lang/llvm-project#144 ?

[alexcrichton]

Sure thing, I've pulled in that commit here too now

[cuviper]

@bors r+

[bors]

📌 Commit 3f4fdce has been approved by cuviper

It is now in the queue for this repository.

[bors]

⌛ Testing commit 3f4fdce with merge 5e37043...

[klensy]

Should rust specific patches be rebased on top, or it ok to keep them where they exist now?

[cuviper]

It's OK to leave it as-is. We'll probably be upgrading to LLVM 16 relatively soon anyway, once that branches.

[bors]

☀️ Test successful - checks-actions
Approved by: cuviper
Pushing 5e37043 to master...

[rust-timer]

Finished benchmarking commit (5e37043): 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)	-0.7%	[-0.9%, -0.6%]	6
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)	2.1%	[1.1%, 3.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-3.5%, -1.4%]	2
All ❌✅ (primary)	-	-	0

Cycles

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