[llvm] c68d35d - [ExpandMemCmp] Add more tests to show missing constant folding.
Clement Courbet via llvm-commits
llvm-commits at lists.llvm.org
Tue Mar 3 05:58:06 PST 2020
Author: Clement Courbet
Date: 2020-03-03T14:57:11+01:00
New Revision: c68d35d78c8bb53c8f262d676522acebccf70a65
URL: https://github.com/llvm/llvm-project/commit/c68d35d78c8bb53c8f262d676522acebccf70a65
DIFF: https://github.com/llvm/llvm-project/commit/c68d35d78c8bb53c8f262d676522acebccf70a65.diff
LOG: [ExpandMemCmp] Add more tests to show missing constant folding.
Added:
Modified:
llvm/test/CodeGen/X86/memcmp.ll
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/X86/memcmp.ll b/llvm/test/CodeGen/X86/memcmp.ll
index 793b4e39284f..76b5f6ff228f 100644
--- a/llvm/test/CodeGen/X86/memcmp.ll
+++ b/llvm/test/CodeGen/X86/memcmp.ll
@@ -93,6 +93,65 @@ define i32 @length2(i8* %X, i8* %Y) nounwind {
ret i32 %m
}
+define i32 @length2_const(i8* %X, i8* %Y) nounwind {
+; X86-LABEL: length2_const:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: movzwl (%eax), %eax
+; X86-NEXT: movzwl .L.str+1, %ecx
+; X86-NEXT: rolw $8, %ax
+; X86-NEXT: rolw $8, %cx
+; X86-NEXT: movzwl %ax, %eax
+; X86-NEXT: movzwl %cx, %ecx
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: retl
+;
+; X64-LABEL: length2_const:
+; X64: # %bb.0:
+; X64-NEXT: movzwl (%rdi), %eax
+; X64-NEXT: movzwl .L.str+{{.*}}(%rip), %ecx
+; X64-NEXT: rolw $8, %ax
+; X64-NEXT: rolw $8, %cx
+; X64-NEXT: movzwl %ax, %eax
+; X64-NEXT: movzwl %cx, %ecx
+; X64-NEXT: subl %ecx, %eax
+; X64-NEXT: retq
+ %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([513 x i8], [513 x i8]* @.str, i32 0, i32 1), i64 2) nounwind
+ ret i32 %m
+}
+
+define i1 @length2_gt_const(i8* %X, i8* %Y) nounwind {
+; X86-LABEL: length2_gt_const:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: movzwl (%eax), %eax
+; X86-NEXT: movzwl .L.str+1, %ecx
+; X86-NEXT: rolw $8, %ax
+; X86-NEXT: rolw $8, %cx
+; X86-NEXT: movzwl %ax, %eax
+; X86-NEXT: movzwl %cx, %ecx
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: testl %eax, %eax
+; X86-NEXT: setg %al
+; X86-NEXT: retl
+;
+; X64-LABEL: length2_gt_const:
+; X64: # %bb.0:
+; X64-NEXT: movzwl (%rdi), %eax
+; X64-NEXT: movzwl .L.str+{{.*}}(%rip), %ecx
+; X64-NEXT: rolw $8, %ax
+; X64-NEXT: rolw $8, %cx
+; X64-NEXT: movzwl %ax, %eax
+; X64-NEXT: movzwl %cx, %ecx
+; X64-NEXT: subl %ecx, %eax
+; X64-NEXT: testl %eax, %eax
+; X64-NEXT: setg %al
+; X64-NEXT: retq
+ %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([513 x i8], [513 x i8]* @.str, i32 0, i32 1), i64 2) nounwind
+ %c = icmp sgt i32 %m, 0
+ ret i1 %c
+}
+
define i1 @length2_eq(i8* %X, i8* %Y) nounwind {
; X86-LABEL: length2_eq:
; X86: # %bb.0:
@@ -238,14 +297,14 @@ define i32 @length3(i8* %X, i8* %Y) nounwind {
; X86-NEXT: rolw $8, %dx
; X86-NEXT: rolw $8, %si
; X86-NEXT: cmpw %si, %dx
-; X86-NEXT: jne .LBB9_1
+; X86-NEXT: jne .LBB11_1
; X86-NEXT: # %bb.2: # %loadbb1
; X86-NEXT: movzbl 2(%eax), %eax
; X86-NEXT: movzbl 2(%ecx), %ecx
; X86-NEXT: subl %ecx, %eax
; X86-NEXT: popl %esi
; X86-NEXT: retl
-; X86-NEXT: .LBB9_1: # %res_block
+; X86-NEXT: .LBB11_1: # %res_block
; X86-NEXT: setae %al
; X86-NEXT: movzbl %al, %eax
; X86-NEXT: leal -1(%eax,%eax), %eax
@@ -259,13 +318,13 @@ define i32 @length3(i8* %X, i8* %Y) nounwind {
; X64-NEXT: rolw $8, %ax
; X64-NEXT: rolw $8, %cx
; X64-NEXT: cmpw %cx, %ax
-; X64-NEXT: jne .LBB9_1
+; X64-NEXT: jne .LBB11_1
; X64-NEXT: # %bb.2: # %loadbb1
; X64-NEXT: movzbl 2(%rdi), %eax
; X64-NEXT: movzbl 2(%rsi), %ecx
; X64-NEXT: subl %ecx, %eax
; X64-NEXT: retq
-; X64-NEXT: .LBB9_1: # %res_block
+; X64-NEXT: .LBB11_1: # %res_block
; X64-NEXT: setae %al
; X64-NEXT: movzbl %al, %eax
; X64-NEXT: leal -1(%rax,%rax), %eax
@@ -453,14 +512,14 @@ define i32 @length5(i8* %X, i8* %Y) nounwind {
; X86-NEXT: bswapl %edx
; X86-NEXT: bswapl %esi
; X86-NEXT: cmpl %esi, %edx
-; X86-NEXT: jne .LBB16_1
+; X86-NEXT: jne .LBB18_1
; X86-NEXT: # %bb.2: # %loadbb1
; X86-NEXT: movzbl 4(%eax), %eax
; X86-NEXT: movzbl 4(%ecx), %ecx
; X86-NEXT: subl %ecx, %eax
; X86-NEXT: popl %esi
; X86-NEXT: retl
-; X86-NEXT: .LBB16_1: # %res_block
+; X86-NEXT: .LBB18_1: # %res_block
; X86-NEXT: setae %al
; X86-NEXT: movzbl %al, %eax
; X86-NEXT: leal -1(%eax,%eax), %eax
@@ -474,13 +533,13 @@ define i32 @length5(i8* %X, i8* %Y) nounwind {
; X64-NEXT: bswapl %eax
; X64-NEXT: bswapl %ecx
; X64-NEXT: cmpl %ecx, %eax
-; X64-NEXT: jne .LBB16_1
+; X64-NEXT: jne .LBB18_1
; X64-NEXT: # %bb.2: # %loadbb1
; X64-NEXT: movzbl 4(%rdi), %eax
; X64-NEXT: movzbl 4(%rsi), %ecx
; X64-NEXT: subl %ecx, %eax
; X64-NEXT: retq
-; X64-NEXT: .LBB16_1: # %res_block
+; X64-NEXT: .LBB18_1: # %res_block
; X64-NEXT: setae %al
; X64-NEXT: movzbl %al, %eax
; X64-NEXT: leal -1(%rax,%rax), %eax
@@ -529,17 +588,17 @@ define i1 @length5_lt(i8* %X, i8* %Y) nounwind {
; X86-NEXT: bswapl %edx
; X86-NEXT: bswapl %esi
; X86-NEXT: cmpl %esi, %edx
-; X86-NEXT: jne .LBB18_1
+; X86-NEXT: jne .LBB20_1
; X86-NEXT: # %bb.2: # %loadbb1
; X86-NEXT: movzbl 4(%eax), %eax
; X86-NEXT: movzbl 4(%ecx), %ecx
; X86-NEXT: subl %ecx, %eax
-; X86-NEXT: jmp .LBB18_3
-; X86-NEXT: .LBB18_1: # %res_block
+; X86-NEXT: jmp .LBB20_3
+; X86-NEXT: .LBB20_1: # %res_block
; X86-NEXT: setae %al
; X86-NEXT: movzbl %al, %eax
; X86-NEXT: leal -1(%eax,%eax), %eax
-; X86-NEXT: .LBB18_3: # %endblock
+; X86-NEXT: .LBB20_3: # %endblock
; X86-NEXT: shrl $31, %eax
; X86-NEXT: # kill: def $al killed $al killed $eax
; X86-NEXT: popl %esi
@@ -552,7 +611,7 @@ define i1 @length5_lt(i8* %X, i8* %Y) nounwind {
; X64-NEXT: bswapl %eax
; X64-NEXT: bswapl %ecx
; X64-NEXT: cmpl %ecx, %eax
-; X64-NEXT: jne .LBB18_1
+; X64-NEXT: jne .LBB20_1
; X64-NEXT: # %bb.2: # %loadbb1
; X64-NEXT: movzbl 4(%rdi), %eax
; X64-NEXT: movzbl 4(%rsi), %ecx
@@ -560,7 +619,7 @@ define i1 @length5_lt(i8* %X, i8* %Y) nounwind {
; X64-NEXT: shrl $31, %eax
; X64-NEXT: # kill: def $al killed $al killed $eax
; X64-NEXT: retq
-; X64-NEXT: .LBB18_1: # %res_block
+; X64-NEXT: .LBB20_1: # %res_block
; X64-NEXT: setae %al
; X64-NEXT: movzbl %al, %eax
; X64-NEXT: leal -1(%rax,%rax), %eax
@@ -610,7 +669,7 @@ define i32 @length8(i8* %X, i8* %Y) nounwind {
; X86-NEXT: bswapl %ecx
; X86-NEXT: bswapl %edx
; X86-NEXT: cmpl %edx, %ecx
-; X86-NEXT: jne .LBB20_2
+; X86-NEXT: jne .LBB22_2
; X86-NEXT: # %bb.1: # %loadbb1
; X86-NEXT: movl 4(%esi), %ecx
; X86-NEXT: movl 4(%eax), %edx
@@ -618,13 +677,13 @@ define i32 @length8(i8* %X, i8* %Y) nounwind {
; X86-NEXT: bswapl %edx
; X86-NEXT: xorl %eax, %eax
; X86-NEXT: cmpl %edx, %ecx
-; X86-NEXT: je .LBB20_3
-; X86-NEXT: .LBB20_2: # %res_block
+; X86-NEXT: je .LBB22_3
+; X86-NEXT: .LBB22_2: # %res_block
; X86-NEXT: xorl %eax, %eax
; X86-NEXT: cmpl %edx, %ecx
; X86-NEXT: setae %al
; X86-NEXT: leal -1(%eax,%eax), %eax
-; X86-NEXT: .LBB20_3: # %endblock
+; X86-NEXT: .LBB22_3: # %endblock
; X86-NEXT: popl %esi
; X86-NEXT: retl
;
@@ -818,7 +877,7 @@ define i32 @length12(i8* %X, i8* %Y) nounwind {
; X64-NEXT: bswapq %rcx
; X64-NEXT: bswapq %rdx
; X64-NEXT: cmpq %rdx, %rcx
-; X64-NEXT: jne .LBB27_2
+; X64-NEXT: jne .LBB29_2
; X64-NEXT: # %bb.1: # %loadbb1
; X64-NEXT: movl 8(%rdi), %ecx
; X64-NEXT: movl 8(%rsi), %edx
@@ -826,13 +885,13 @@ define i32 @length12(i8* %X, i8* %Y) nounwind {
; X64-NEXT: bswapl %edx
; X64-NEXT: xorl %eax, %eax
; X64-NEXT: cmpq %rdx, %rcx
-; X64-NEXT: je .LBB27_3
-; X64-NEXT: .LBB27_2: # %res_block
+; X64-NEXT: je .LBB29_3
+; X64-NEXT: .LBB29_2: # %res_block
; X64-NEXT: xorl %eax, %eax
; X64-NEXT: cmpq %rdx, %rcx
; X64-NEXT: setae %al
; X64-NEXT: leal -1(%rax,%rax), %eax
-; X64-NEXT: .LBB27_3: # %endblock
+; X64-NEXT: .LBB29_3: # %endblock
; X64-NEXT: retq
%m = tail call i32 @memcmp(i8* %X, i8* %Y, i64 12) nounwind
ret i32 %m
@@ -892,6 +951,26 @@ define i1 @length14_eq(i8* %X, i8* %Y) nounwind {
ret i1 %c
}
+define i32 @length15_const(i8* %X, i8* %Y) nounwind {
+; X86-LABEL: length15_const:
+; X86: # %bb.0:
+; X86-NEXT: pushl $0
+; X86-NEXT: pushl $15
+; X86-NEXT: pushl $.L.str+1
+; X86-NEXT: pushl {{[0-9]+}}(%esp)
+; X86-NEXT: calll memcmp
+; X86-NEXT: addl $16, %esp
+; X86-NEXT: retl
+;
+; X64-LABEL: length15_const:
+; X64: # %bb.0:
+; X64-NEXT: movl $.L.str+1, %esi
+; X64-NEXT: movl $15, %edx
+; X64-NEXT: jmp memcmp # TAILCALL
+ %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([513 x i8], [513 x i8]* @.str, i32 0, i32 1), i64 15) nounwind
+ ret i32 %m
+}
+
define i1 @length15_eq(i8* %X, i8* %Y) nounwind {
; X86-LABEL: length15_eq:
; X86: # %bb.0:
@@ -919,6 +998,34 @@ define i1 @length15_eq(i8* %X, i8* %Y) nounwind {
ret i1 %c
}
+define i1 @length15_gt_const(i8* %X, i8* %Y) nounwind {
+; X86-LABEL: length15_gt_const:
+; X86: # %bb.0:
+; X86-NEXT: pushl $0
+; X86-NEXT: pushl $15
+; X86-NEXT: pushl $.L.str+1
+; X86-NEXT: pushl {{[0-9]+}}(%esp)
+; X86-NEXT: calll memcmp
+; X86-NEXT: addl $16, %esp
+; X86-NEXT: testl %eax, %eax
+; X86-NEXT: setg %al
+; X86-NEXT: retl
+;
+; X64-LABEL: length15_gt_const:
+; X64: # %bb.0:
+; X64-NEXT: pushq %rax
+; X64-NEXT: movl $.L.str+1, %esi
+; X64-NEXT: movl $15, %edx
+; X64-NEXT: callq memcmp
+; X64-NEXT: testl %eax, %eax
+; X64-NEXT: setg %al
+; X64-NEXT: popq %rcx
+; X64-NEXT: retq
+ %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([513 x i8], [513 x i8]* @.str, i32 0, i32 1), i64 15) nounwind
+ %c = icmp sgt i32 %m, 0
+ ret i1 %c
+}
+
; PR33329 - https://bugs.llvm.org/show_bug.cgi?id=33329
define i32 @length16(i8* %X, i8* %Y) nounwind {
@@ -939,7 +1046,7 @@ define i32 @length16(i8* %X, i8* %Y) nounwind {
; X64-NEXT: bswapq %rcx
; X64-NEXT: bswapq %rdx
; X64-NEXT: cmpq %rdx, %rcx
-; X64-NEXT: jne .LBB31_2
+; X64-NEXT: jne .LBB35_2
; X64-NEXT: # %bb.1: # %loadbb1
; X64-NEXT: movq 8(%rdi), %rcx
; X64-NEXT: movq 8(%rsi), %rdx
@@ -947,13 +1054,13 @@ define i32 @length16(i8* %X, i8* %Y) nounwind {
; X64-NEXT: bswapq %rdx
; X64-NEXT: xorl %eax, %eax
; X64-NEXT: cmpq %rdx, %rcx
-; X64-NEXT: je .LBB31_3
-; X64-NEXT: .LBB31_2: # %res_block
+; X64-NEXT: je .LBB35_3
+; X64-NEXT: .LBB35_2: # %res_block
; X64-NEXT: xorl %eax, %eax
; X64-NEXT: cmpq %rdx, %rcx
; X64-NEXT: setae %al
; X64-NEXT: leal -1(%rax,%rax), %eax
-; X64-NEXT: .LBB31_3: # %endblock
+; X64-NEXT: .LBB35_3: # %endblock
; X64-NEXT: retq
%m = tail call i32 @memcmp(i8* %X, i8* %Y, i64 16) nounwind
ret i32 %m
@@ -1068,7 +1175,7 @@ define i1 @length16_lt(i8* %x, i8* %y) nounwind {
; X64-NEXT: bswapq %rcx
; X64-NEXT: bswapq %rdx
; X64-NEXT: cmpq %rdx, %rcx
-; X64-NEXT: jne .LBB33_2
+; X64-NEXT: jne .LBB37_2
; X64-NEXT: # %bb.1: # %loadbb1
; X64-NEXT: movq 8(%rdi), %rcx
; X64-NEXT: movq 8(%rsi), %rdx
@@ -1076,13 +1183,13 @@ define i1 @length16_lt(i8* %x, i8* %y) nounwind {
; X64-NEXT: bswapq %rdx
; X64-NEXT: xorl %eax, %eax
; X64-NEXT: cmpq %rdx, %rcx
-; X64-NEXT: je .LBB33_3
-; X64-NEXT: .LBB33_2: # %res_block
+; X64-NEXT: je .LBB37_3
+; X64-NEXT: .LBB37_2: # %res_block
; X64-NEXT: xorl %eax, %eax
; X64-NEXT: cmpq %rdx, %rcx
; X64-NEXT: setae %al
; X64-NEXT: leal -1(%rax,%rax), %eax
-; X64-NEXT: .LBB33_3: # %endblock
+; X64-NEXT: .LBB37_3: # %endblock
; X64-NEXT: shrl $31, %eax
; X64-NEXT: # kill: def $al killed $al killed $eax
; X64-NEXT: retq
@@ -1111,7 +1218,7 @@ define i1 @length16_gt(i8* %x, i8* %y) nounwind {
; X64-NEXT: bswapq %rax
; X64-NEXT: bswapq %rcx
; X64-NEXT: cmpq %rcx, %rax
-; X64-NEXT: jne .LBB34_2
+; X64-NEXT: jne .LBB38_2
; X64-NEXT: # %bb.1: # %loadbb1
; X64-NEXT: movq 8(%rdi), %rax
; X64-NEXT: movq 8(%rsi), %rcx
@@ -1119,13 +1226,13 @@ define i1 @length16_gt(i8* %x, i8* %y) nounwind {
; X64-NEXT: bswapq %rcx
; X64-NEXT: xorl %edx, %edx
; X64-NEXT: cmpq %rcx, %rax
-; X64-NEXT: je .LBB34_3
-; X64-NEXT: .LBB34_2: # %res_block
+; X64-NEXT: je .LBB38_3
+; X64-NEXT: .LBB38_2: # %res_block
; X64-NEXT: xorl %edx, %edx
; X64-NEXT: cmpq %rcx, %rax
; X64-NEXT: setae %dl
; X64-NEXT: leal -1(%rdx,%rdx), %edx
-; X64-NEXT: .LBB34_3: # %endblock
+; X64-NEXT: .LBB38_3: # %endblock
; X64-NEXT: testl %edx, %edx
; X64-NEXT: setg %al
; X64-NEXT: retq
More information about the llvm-commits
mailing list