This is an archive of the discontinued LLVM Phabricator instance.

[RISCV] Precommit a test case that test accessing a fixed object when has rvv vector object existed
ClosedPublic

Authored by StephenFan on Apr 11 2021, 10:29 PM.

Download Raw Diff

Details

Reviewers

rogfer01
craig.topper
HsiangKai

Commits

rG325b454ed8d8: [RISCV] Precommit a test case that test accessing a fixed object when has rvv…

Diff Detail

Repository: rG LLVM Github Monorepo

Unit TestsFailed

	Time	Test
	90 ms	x64 debian > Clang.Driver::debug-pass-structure.c

Event Timeline

StephenFan created this revision.Apr 11 2021, 10:29 PM

Herald added subscribers: vkmr, frasercrmck, evandro and 23 others. · View Herald TranscriptApr 11 2021, 10:29 PM

StephenFan requested review of this revision.Apr 11 2021, 10:29 PM

Herald added a project: Restricted Project. · View Herald TranscriptApr 11 2021, 10:29 PM

Herald added subscribers: llvm-commits, MaskRay. · View Herald Transcript

StephenFan added a child revision: D100286: [RISCV] Fix StackOffset calculation when using sp to access the fixed stack object in the case of rvv vector objects existed.Apr 11 2021, 10:48 PM

Harbormaster completed remote builds in B98211: Diff 336744.Apr 11 2021, 11:03 PM

LGTM.

This revision is now accepted and ready to land.Apr 12 2021, 2:30 AM

frasercrmck added inline comments.Apr 12 2021, 2:44 AM

llvm/test/CodeGen/RISCV/rvv/localvar.ll
290	Maybe if this is wrong code add a little FIXME comment to explain what's wrong? Just in case your follow-up patch takes a while to land.

StephenFan added inline comments.Apr 29 2021, 7:13 PM

llvm/test/CodeGen/RISCV/rvv/localvar.ll
290	Maybe if this is wrong code add a little FIXME comment to explain what's wrong? Just in case your follow-up patch takes a while to land.
290	I'm sorry that I forgot your comment. Recently, the patch that fix this wrong code is accepted.

rebase.

This revision was landed with ongoing or failed builds.Apr 29 2021, 7:37 PM

Closed by commit rG325b454ed8d8: [RISCV] Precommit a test case that test accessing a fixed object when has rvv… (authored by StephenFan). · Explain Why

This revision was automatically updated to reflect the committed changes.

StephenFan added a commit: rG325b454ed8d8: [RISCV] Precommit a test case that test accessing a fixed object when has rvv….

Harbormaster completed remote builds in B101812: Diff 341745.Apr 29 2021, 9:46 PM

Revision Contents

Path

Size

llvm/

test/

CodeGen/

RISCV/

rvv/

localvar.ll

18 lines

Diff 341745

llvm/test/CodeGen/RISCV/rvv/localvar.ll

Show First 20 Lines • Show All 281 Lines • ▼ Show 20 Lines	; RV64IV-NEXT: ret
call void @notdead2(i8* %1, i32* %2, <vscale x 16 x i8>* %local0)		call void @notdead2(i8* %1, i32* %2, <vscale x 16 x i8>* %local0)
load volatile i32, i32* %local_scalar0		load volatile i32, i32* %local_scalar0
load volatile <vscale x 16 x i8>, <vscale x 16 x i8>* %local0		load volatile <vscale x 16 x i8>, <vscale x 16 x i8>* %local0
load volatile <vscale x 16 x i8>, <vscale x 16 x i8>* %local1		load volatile <vscale x 16 x i8>, <vscale x 16 x i8>* %local1
load volatile i32, i32* %local_scalar1		load volatile i32, i32* %local_scalar1
ret void		ret void
}		}

		define i64 @fixed_object(i64 %0, i64 %1, i64 %2, i64 %3, i64 %4, i64 %5, i64 %6, i64 %7, i64 %8) nounwind {
		frasercrmckUnsubmitted Not Done Reply Inline Actions Maybe if this is wrong code add a little FIXME comment to explain what's wrong? Just in case your follow-up patch takes a while to land. frasercrmck: Maybe if this is wrong code add a little FIXME comment to explain what's wrong? Just in case…
		StephenFanAuthorUnsubmitted Done Reply Inline Actions Maybe if this is wrong code add a little FIXME comment to explain what's wrong? Just in case your follow-up patch takes a while to land. StephenFan: > Maybe if this is wrong code add a little FIXME comment to explain what's wrong? Just in case…
		StephenFanAuthorUnsubmitted Done Reply Inline Actions I'm sorry that I forgot your comment. Recently, the patch that fix this wrong code is accepted. StephenFan: I'm sorry that I forgot your comment. Recently, the patch that fix this wrong code is accepted.
		; RV64IV-LABEL: fixed_object:
		; RV64IV: # %bb.0:
		; RV64IV-NEXT: addi sp, sp, -32
		; RV64IV-NEXT: csrr a0, vlenb
		; RV64IV-NEXT: slli a0, a0, 3
		; RV64IV-NEXT: sub sp, sp, a0
		; RV64IV-NEXT: ld a0, 32(sp)
		; RV64IV-NEXT: csrr a1, vlenb
		; RV64IV-NEXT: slli a1, a1, 3
		; RV64IV-NEXT: add sp, sp, a1
		; RV64IV-NEXT: addi sp, sp, 32
		; RV64IV-NEXT: ret
		%fixed_size = alloca i32
		%rvv_vector = alloca <vscale x 8 x i64>, align 8
		ret i64 %8
		}

declare void @notdead(i8, <vscale x 16 x i8>)		declare void @notdead(i8, <vscale x 16 x i8>)
declare void @notdead2(i8, i32, <vscale x 16 x i8>*)		declare void @notdead2(i8, i32, <vscale x 16 x i8>*)