[llvm] 7b9a3b9 - [RISCV] Precommit testcase to show wrong result of make-compressible optimization
Kito Cheng via llvm-commits
llvm-commits at lists.llvm.org
Fri Jul 8 02:01:29 PDT 2022
Author: Kito Cheng
Date: 2022-07-08T17:01:22+08:00
New Revision: 7b9a3b9d6d98184d3e20c03fceefaac6dc0c3580
URL: https://github.com/llvm/llvm-project/commit/7b9a3b9d6d98184d3e20c03fceefaac6dc0c3580
DIFF: https://github.com/llvm/llvm-project/commit/7b9a3b9d6d98184d3e20c03fceefaac6dc0c3580.diff
LOG: [RISCV] Precommit testcase to show wrong result of make-compressible optimization
Use following example to demo what happened now:
li a1, 1
sd a1, 800(a0)
sd a0, 808(a0) # Store base address into base + offset
li a1, 2
sd a1, 816(a0)
Current will optimizate into:
li a1, 1
addi a2, a0, 768
sd a1, 32(a2)
sd a2, 40(a2) # Wrong replacement for the source register.
li a1, 2
sd a1, 48(a2)
Reviewed By: asb
Differential Revision: https://reviews.llvm.org/D128875
Added:
llvm/test/CodeGen/RISCV/make-compressible-for-store-address.mir
Modified:
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/RISCV/make-compressible-for-store-address.mir b/llvm/test/CodeGen/RISCV/make-compressible-for-store-address.mir
new file mode 100644
index 000000000000..a913fcb6f609
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/make-compressible-for-store-address.mir
@@ -0,0 +1,47 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -o - %s -mtriple=riscv64 -mattr=+c -simplify-mir \
+# RUN: -run-pass=riscv-make-compressible | FileCheck %s
+--- |
+ define void @foo(ptr noundef %ptr) #0 {
+ entry:
+ %0 = ptrtoint ptr %ptr to i64
+ %1 = getelementptr inbounds i64, ptr %ptr, i64 100
+ store i64 1, ptr %1, align 8
+ %2 = getelementptr inbounds i64, ptr %ptr, i64 101
+ store i64 %0, ptr %2, align 8
+ %3 = getelementptr inbounds i64, ptr %ptr, i64 102
+ store i64 2, ptr %3, align 8
+ ret void
+ }
+
+ attributes #0 = { minsize }
+
+...
+---
+name: foo
+alignment: 2
+tracksRegLiveness: true
+liveins:
+ - { reg: '$x10' }
+body: |
+ bb.0.entry:
+ liveins: $x10
+
+ ; CHECK-LABEL: name: foo
+ ; CHECK: liveins: $x10
+ ; CHECK-NEXT: {{ $}}
+ ; CHECK-NEXT: renamable $x11 = ADDI $x0, 1
+ ; CHECK-NEXT: $x12 = ADDI $x10, 768
+ ; CHECK-NEXT: SD killed renamable $x11, $x12, 32 :: (store (s64) into %ir.1)
+ ; CHECK-NEXT: SD $x12, $x12, 40 :: (store (s64) into %ir.2)
+ ; CHECK-NEXT: renamable $x11 = ADDI $x0, 2
+ ; CHECK-NEXT: SD killed renamable $x11, killed $x12, 48 :: (store (s64) into %ir.3)
+ ; CHECK-NEXT: PseudoRET
+ renamable $x11 = ADDI $x0, 1
+ SD killed renamable $x11, renamable $x10, 800 :: (store (s64) into %ir.1)
+ SD renamable $x10, renamable $x10, 808 :: (store (s64) into %ir.2)
+ renamable $x11 = ADDI $x0, 2
+ SD killed renamable $x11, killed renamable $x10, 816 :: (store (s64) into %ir.3)
+ PseudoRET
+
+...
More information about the llvm-commits
mailing list