Remove the use of -use-gnu-stack when BOLTing LLVM

[Kobzol]

This flag was (counterintuitively) removing the GNU_STACK ELF attribute, which caused the optimized libLLVM.so file to be flagged as having an executable stack on SELinux.

Removing the flag might cause issues with strip. I'm not aware that we're stripping libLLVM.so though. Does it happen anywhere?

Fixes: #105783

[rustbot]

r? @albertlarsan68

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

[Noratrieb]

This is weird..
https://github.com/llvm/llvm-project/blob/09ab1f335bce8bb4d14c28b9adad6d3bec8dfed9/bolt/lib/Rewrite/RewriteInstance.cpp#L4322-L4324
https://github.com/llvm/llvm-project/blob/09ab1f335bce8bb4d14c28b9adad6d3bec8dfed9/bolt/lib/Rewrite/RewriteInstance.cpp#L4267
When the -use-gnu-stack option is set, it appears to be replacing the PT_GNU_STACK program header with an empty PT_LOAD header, effectively deleting it. If it's not set it goes into another branch where it keeps the type.

The description of the flag is indeed very confusing.
llvm/llvm-project@7f7d4af does also talk about "issues with strip/objcopy" but it's very unclear what these issues are.

https://github.com/llvm/llvm-project/blob/09ab1f335bce8bb4d14c28b9adad6d3bec8dfed9/bolt/docs/OptimizingClang.md?plain=1#L68 Even the BOLT docs use -use-gnu-stack when optimizing clang, so something seems to be up. Probably worth asking upstream.

[Kobzol]

Yeah the name of the flag is weird. I asked about this on the BOLT Discord: https://discord.com/channels/636084430946959380/930647188944613406/1092884016627200081

[Noratrieb]

Summary from discord:
-use-gnu-stack actually means "use the program header for GNU_STACK as a slot for adding new program headers, effectively deleting it". This exists because some tools like strip and objcopy are unable to deal with the appended program headers and reusing the existing slot is fine for them. Since we aren't doing that on our libLLVM.so, it should be fine.

@bors try

[bors]

⌛ Trying commit c32953f with merge 7a2984f94b9b2cc9bcbfc178ffaec12878b7fd65...

[bors]

☀️ Try build successful - checks-actions
Build commit: 7a2984f94b9b2cc9bcbfc178ffaec12878b7fd65 (7a2984f94b9b2cc9bcbfc178ffaec12878b7fd65)

[Kobzol]

@rust-timer build 7a2984f94b9b2cc9bcbfc178ffaec12878b7fd65

[rust-timer]

Finished benchmarking commit (7a2984f94b9b2cc9bcbfc178ffaec12878b7fd65): comparison URL.

Overall result: ❌ regressions - no action needed

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -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)	1.9%	[1.9%, 1.9%]	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)	2.4%	[1.1%, 4.5%]	4
Improvements ✅ (primary)	-2.7%	[-3.2%, -2.3%]	2
Improvements ✅ (secondary)	-1.9%	[-3.2%, -1.1%]	9
All ❌✅ (primary)	-2.7%	[-3.2%, -2.3%]	2

Cycles

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

[Noratrieb]

Seems like it all still works, great!

[nikic]

@bors r+

[bors]

📌 Commit c32953f has been approved by nikic

It is now in the queue for this repository.

[bors]

⌛ Testing commit c32953f with merge 90a9f69...

[bors]

☀️ Test successful - checks-actions
Approved by: nikic
Pushing 90a9f69 to master...

[rust-timer]

Finished benchmarking commit (90a9f69): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

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

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)	2.8%	[2.4%, 3.2%]	2
Regressions ❌ (secondary)	3.0%	[1.7%, 5.1%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.4%	[-1.9%, -0.9%]	3
All ❌✅ (primary)	2.8%	[2.4%, 3.2%]	2

Cycles

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