Support s390x z13 vector ABI #131586

taiki-e · 2024-10-12T07:09:58Z

This resolves the following fixmes:

rust/compiler/rustc_target/src/abi/call/s390x.rs

Lines 1 to 2 in 58420a0

    
           // FIXME: The assumes we're using the non-vector ABI, i.e., compiling 
        
           // for a pre-z13 machine or using -mno-vx.

rust/compiler/rustc_target/src/spec/targets/s390x_unknown_linux_gnu.rs

Lines 9 to 11 in 58420a0

    
           // FIXME: The ABI implementation in abi/call/s390x.rs is for now hard-coded to assume the no-vector 
        
           // ABI. Pass the -vector feature string to LLVM to respect this assumption. 
        
           base.features = "-vector".into();

Refs: Section 1.2.3 "Parameter Passing" and section 1.2.5 "Return Values" in ELF Application Binary Interface s390x Supplement, Version 1.6.1 (lzsabi_s390x.pdf in https://github.com/IBM/s390x-abi/releases/tag/v1.6.1)

This PR extends #127731 (merged) 's ABI check to handle cases where vector target feature is disabled.
If we do not do ABI check, we run into the ABI problems as described in #116558 and #130869 (comment), and the problem of the compiler generating strange code (#131586 (comment)).

cc @uweigand

@rustbot label +O-SystemZ +A-ABI

rustbot · 2024-10-12T07:10:06Z

r? @jieyouxu

rustbot has assigned @jieyouxu.
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

jieyouxu · 2024-10-12T07:18:33Z

r? compiler

compiler/rustc_target/src/abi/call/s390x.rs

uweigand

In addition to the problem mentioned inline, we should also verify that building with soft-float works as expected. This option implicitly disables any use of floating-point or vector registers, and therefore also disables use of the vector facility. I believe the LLVM back-end should handle that logic correctly, but it would be good to have the test check.

uweigand · 2024-10-14T08:42:50Z

tests/assembly/s390x-vector-abi.rs

+    *x
+}
+
+// FIXME: should check output for z10, but it is very long...


It's also incorrect. Without the vector feature, vector types of any length are supposed to be passed via implicit reference, which in the case means a pointer in %r2. The body of the function should simply be:

lg %r2, 0(%r2)

Not sure what exactly the compiler is doing here, but it appears to have scalarized the input and pass it as 8 distinct one-byte arguments ...

This case and soft-float case should eventually result in a compilation error once the ABI check is implemented, and I will add UI tests for these cases once the ABI check is implemented.

Fair enough. I just wanted to point out the incorrect code, in case we do end up going back to supporting the no-vector case.

bors · 2024-10-16T00:02:38Z

☔ The latest upstream changes (presumably #131747) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2024-10-25T18:02:56Z

☔ The latest upstream changes (presumably #127731) made this pull request unmergeable. Please resolve the merge conflicts.

taiki-e · 2024-10-25T19:31:32Z

tests/ui/simd-abi-checks-s390x.rs

+// FIXME: +soft-float itself doesn't set -vector
+//@[z13_soft_float] compile-flags: --target s390x-unknown-linux-gnu -C target-cpu=z13 -C target-feature=-vector,+soft-float


I added a soft-float case per #131586 (review), but the soft-float target feature affects the float ABI, so its use with the -C target-feature flag will likely be deprecated (#116344).

Also, the soft-float target feature is recognized by LLVM but not by rustc, so I get the following warning:

warning: unknown and unstable feature specified for `-Ctarget-feature`: `soft-float` | = note: it is still passed through to the codegen backend, but use of this feature might be unsound and the behavior of this feature can change in the future

https://godbolt.org/z/x75bPveh5

(Btw, it seems other architecture's soft-float targets also explicitly disable SIMD-related features)

rust/compiler/rustc_target/src/spec/targets/x86_64_unknown_none.rs

Line 23 in 6faf0bd

features: "-mmx,-sse,-sse2,-sse3,-ssse3,-sse4.1,-sse4.2,-avx,-avx2,+soft-float".into(),

rustbot · 2024-10-25T19:32:19Z

These commits modify compiler targets.
(See the Target Tier Policy.)

taiki-e · 2024-10-25T19:32:26Z

Rewritten based on the merged #127731.

@rustbot label +S-waiting-on-review, -S-blocked

taiki-e · 2024-10-25T21:10:00Z

Hmm, #127731 will likely be reverted for now: #132152

@rustbot label +S-blocked

compiler-errors · 2024-10-25T21:33:18Z

Sorry for the churn @taiki-e 😭

bors · 2024-10-26T07:00:27Z

☔ The latest upstream changes (presumably #132152) made this pull request unmergeable. Please resolve the merge conflicts.

rustbot assigned jieyouxu Oct 12, 2024

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. O-SystemZ Target: SystemZ processors (s390x) labels Oct 12, 2024

taiki-e mentioned this pull request Oct 12, 2024

s390x vector facilities support #130869

Open

10 tasks

This comment has been minimized.

Sign in to view

rustbot assigned compiler-errors and unassigned jieyouxu Oct 12, 2024

taiki-e force-pushed the s390x-vector-abi branch from 7656805 to 07ca922 Compare October 12, 2024 07:19

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch 3 times, most recently from c14d9b2 to d39e822 Compare October 12, 2024 13:40

This comment was marked as outdated.

Sign in to view

taiki-e commented Oct 12, 2024

View reviewed changes

compiler/rustc_target/src/abi/call/s390x.rs Outdated Show resolved Hide resolved

This comment was marked as outdated.

Sign in to view

taiki-e mentioned this pull request Oct 13, 2024

Support #[repr(simd)] types in input/output of s390x inline assembly #131664

Draft

uweigand reviewed Oct 14, 2024

View reviewed changes

taiki-e force-pushed the s390x-vector-abi branch 2 times, most recently from 0a6b7ad to c1835bd Compare October 14, 2024 11:31

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch from c1835bd to f4f0361 Compare October 14, 2024 12:19

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch from f4f0361 to e18682e Compare October 14, 2024 13:30

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch from ec20e94 to 8c4ec03 Compare October 14, 2024 14:37

rustbot added A-testsuite Area: The testsuite used to check the correctness of rustc T-bootstrap Relevant to the bootstrap subteam: Rust's build system (x.py and src/bootstrap) labels Oct 14, 2024

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch 2 times, most recently from 546fd91 to 857dbc6 Compare October 14, 2024 15:10

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch from 857dbc6 to 42a4984 Compare October 15, 2024 14:48

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch from 42a4984 to 1a8d1aa Compare October 15, 2024 15:19

This comment has been minimized.

Sign in to view

taiki-e force-pushed the s390x-vector-abi branch 2 times, most recently from a9242d0 to 524c9f6 Compare October 15, 2024 15:54

taiki-e force-pushed the s390x-vector-abi branch from 524c9f6 to c28dc57 Compare October 16, 2024 15:05

Support s390x z13 vector ABI

5fc1014

taiki-e force-pushed the s390x-vector-abi branch from c28dc57 to 5fc1014 Compare October 25, 2024 19:07

taiki-e commented Oct 25, 2024

View reviewed changes

taiki-e marked this pull request as ready for review October 25, 2024 19:32

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-blocked Status: Marked as blocked ❌ on something else such as an RFC or other implementation work. labels Oct 25, 2024

taiki-e requested a review from uweigand October 25, 2024 19:32

rustbot added the S-blocked Status: Marked as blocked ❌ on something else such as an RFC or other implementation work. label Oct 25, 2024

compiler-errors removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 4, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support s390x z13 vector ABI #131586

Support s390x z13 vector ABI #131586

taiki-e commented Oct 12, 2024 •

edited

Loading

rustbot commented Oct 12, 2024

This comment has been minimized.

jieyouxu commented Oct 12, 2024

This comment has been minimized.

This comment was marked as outdated.

This comment was marked as outdated.

uweigand left a comment

uweigand Oct 14, 2024

taiki-e Oct 14, 2024

uweigand Oct 14, 2024

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

bors commented Oct 16, 2024

bors commented Oct 25, 2024

taiki-e Oct 25, 2024

taiki-e Oct 25, 2024

rustbot commented Oct 25, 2024

taiki-e commented Oct 25, 2024

taiki-e commented Oct 25, 2024

compiler-errors commented Oct 25, 2024

bors commented Oct 26, 2024

	// FIXME: The assumes we're using the non-vector ABI, i.e., compiling
	// for a pre-z13 machine or using -mno-vx.

	// FIXME: The ABI implementation in abi/call/s390x.rs is for now hard-coded to assume the no-vector
	// ABI. Pass the -vector feature string to LLVM to respect this assumption.
	base.features = "-vector".into();

		// FIXME: +soft-float itself doesn't set -vector
		//@[z13_soft_float] compile-flags: --target s390x-unknown-linux-gnu -C target-cpu=z13 -C target-feature=-vector,+soft-float

Support s390x z13 vector ABI #131586

Are you sure you want to change the base?

Support s390x z13 vector ABI #131586

Conversation

taiki-e commented Oct 12, 2024 • edited Loading

rustbot commented Oct 12, 2024

This comment has been minimized.

jieyouxu commented Oct 12, 2024

This comment has been minimized.

This comment was marked as outdated.

This comment was marked as outdated.

uweigand left a comment

Choose a reason for hiding this comment

uweigand Oct 14, 2024

Choose a reason for hiding this comment

taiki-e Oct 14, 2024

Choose a reason for hiding this comment

uweigand Oct 14, 2024

Choose a reason for hiding this comment

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

bors commented Oct 16, 2024

bors commented Oct 25, 2024

taiki-e Oct 25, 2024

Choose a reason for hiding this comment

taiki-e Oct 25, 2024

Choose a reason for hiding this comment

rustbot commented Oct 25, 2024

taiki-e commented Oct 25, 2024

taiki-e commented Oct 25, 2024

compiler-errors commented Oct 25, 2024

bors commented Oct 26, 2024

taiki-e commented Oct 12, 2024 •

edited

Loading