; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=i386-linux-gnu -verify-machineinstrs %s -o - | FileCheck %s --check-prefixes=X86-ALL,X86-GOOD-RA
; RUN: llc -mtriple=i386-linux-gnu -verify-machineinstrs -pre-RA-sched=fast %s -o - | FileCheck %s --check-prefixes=X86-ALL,X86-FAST-RA
; RUN: llc -mtriple=x86_64-linux-gnu -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=X64-ALL
; RUN: llc -mtriple=x86_64-linux-gnu -verify-machineinstrs -pre-RA-sched=fast %s -o - | FileCheck %s --check-prefix=X64-ALL
; RUN: llc -mtriple=x86_64-linux-gnu -verify-machineinstrs -mattr=+sahf %s -o - | FileCheck %s --check-prefix=X64-ALL
; RUN: llc -mtriple=x86_64-linux-gnu -verify-machineinstrs -mattr=+sahf -pre-RA-sched=fast %s -o - | FileCheck %s --check-prefix=X64-ALL
; RUN: llc -mtriple=x86_64-linux-gnu -verify-machineinstrs -mcpu=corei7 %s -o - | FileCheck %s --check-prefix=X64-ALL
declare i32 @foo()
declare i32 @bar(i64)
; In the following case when using fast scheduling we get a long chain of
; EFLAGS save/restore due to a sequence of:
; cmpxchg8b (implicit-def eflags)
; eax = copy eflags
; adjcallstackdown32
; ...
; use of eax
; During PEI the adjcallstackdown32 is replaced with the subl which
; clobbers eflags, effectively interfering in the liveness interval. However,
; we then promote these copies into independent conditions in GPRs that avoids
; repeated saving and restoring logic and can be trivially managed by the
; register allocator.
define i64 @test_intervening_call(ptr %foo, i64 %bar, i64 %baz) nounwind {
; X86-GOOD-RA-LABEL: test_intervening_call:
; X86-GOOD-RA: # %bb.0: # %entry
; X86-GOOD-RA-NEXT: pushl %ebx
; X86-GOOD-RA-NEXT: pushl %esi
; X86-GOOD-RA-NEXT: pushl %eax
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %edx
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %ebx
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %esi
; X86-GOOD-RA-NEXT: lock cmpxchg8b (%esi)
; X86-GOOD-RA-NEXT: setne %bl
; X86-GOOD-RA-NEXT: subl $8, %esp
; X86-GOOD-RA-NEXT: pushl %edx
; X86-GOOD-RA-NEXT: pushl %eax
; X86-GOOD-RA-NEXT: calll bar@PLT
; X86-GOOD-RA-NEXT: addl $16, %esp
; X86-GOOD-RA-NEXT: testb %bl, %bl
; X86-GOOD-RA-NEXT: jne .LBB0_3
; X86-GOOD-RA-NEXT: # %bb.1: # %t
; X86-GOOD-RA-NEXT: movl $42, %eax
; X86-GOOD-RA-NEXT: jmp .LBB0_2
; X86-GOOD-RA-NEXT: .LBB0_3: # %f
; X86-GOOD-RA-NEXT: xorl %eax, %eax
; X86-GOOD-RA-NEXT: .LBB0_2: # %t
; X86-GOOD-RA-NEXT: xorl %edx, %edx
; X86-GOOD-RA-NEXT: addl $4, %esp
; X86-GOOD-RA-NEXT: popl %esi
; X86-GOOD-RA-NEXT: popl %ebx
; X86-GOOD-RA-NEXT: retl
;
; X86-FAST-RA-LABEL: test_intervening_call:
; X86-FAST-RA: # %bb.0: # %entry
; X86-FAST-RA-NEXT: pushl %ebx
; X86-FAST-RA-NEXT: pushl %esi
; X86-FAST-RA-NEXT: pushl %eax
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %esi
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %ebx
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %edx
; X86-FAST-RA-NEXT: lock cmpxchg8b (%esi)
; X86-FAST-RA-NEXT: setne %bl
; X86-FAST-RA-NEXT: subl $8, %esp
; X86-FAST-RA-NEXT: pushl %edx
; X86-FAST-RA-NEXT: pushl %eax
; X86-FAST-RA-NEXT: calll bar@PLT
; X86-FAST-RA-NEXT: addl $16, %esp
; X86-FAST-RA-NEXT: testb %bl, %bl
; X86-FAST-RA-NEXT: jne .LBB0_3
; X86-FAST-RA-NEXT: # %bb.1: # %t
; X86-FAST-RA-NEXT: movl $42, %eax
; X86-FAST-RA-NEXT: jmp .LBB0_2
; X86-FAST-RA-NEXT: .LBB0_3: # %f
; X86-FAST-RA-NEXT: xorl %eax, %eax
; X86-FAST-RA-NEXT: .LBB0_2: # %t
; X86-FAST-RA-NEXT: xorl %edx, %edx
; X86-FAST-RA-NEXT: addl $4, %esp
; X86-FAST-RA-NEXT: popl %esi
; X86-FAST-RA-NEXT: popl %ebx
; X86-FAST-RA-NEXT: retl
;
; X64-ALL-LABEL: test_intervening_call:
; X64-ALL: # %bb.0: # %entry
; X64-ALL-NEXT: pushq %rbx
; X64-ALL-NEXT: movq %rsi, %rax
; X64-ALL-NEXT: lock cmpxchgq %rdx, (%rdi)
; X64-ALL-NEXT: setne %bl
; X64-ALL-NEXT: movq %rax, %rdi
; X64-ALL-NEXT: callq bar@PLT
; X64-ALL-NEXT: testb %bl, %bl
; X64-ALL-NEXT: jne .LBB0_2
; X64-ALL-NEXT: # %bb.1: # %t
; X64-ALL-NEXT: movl $42, %eax
; X64-ALL-NEXT: popq %rbx
; X64-ALL-NEXT: retq
; X64-ALL-NEXT: .LBB0_2: # %f
; X64-ALL-NEXT: xorl %eax, %eax
; X64-ALL-NEXT: popq %rbx
; X64-ALL-NEXT: retq
entry:
%cx = cmpxchg ptr %foo, i64 %bar, i64 %baz seq_cst seq_cst
%v = extractvalue { i64, i1 } %cx, 0
%p = extractvalue { i64, i1 } %cx, 1
call i32 @bar(i64 %v)
br i1 %p, label %t, label %f
t:
ret i64 42
f:
ret i64 0
}
; Interesting in producing a clobber without any function calls.
define i32 @test_control_flow(ptr %p, i32 %i, i32 %j) nounwind {
; X86-ALL-LABEL: test_control_flow:
; X86-ALL: # %bb.0: # %entry
; X86-ALL-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-ALL-NEXT: cmpl {{[0-9]+}}(%esp), %eax
; X86-ALL-NEXT: jle .LBB1_6
; X86-ALL-NEXT: # %bb.1: # %loop_start
; X86-ALL-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X86-ALL-NEXT: .p2align 4
; X86-ALL-NEXT: .LBB1_2: # %while.condthread-pre-split.i
; X86-ALL-NEXT: # =>This Loop Header: Depth=1
; X86-ALL-NEXT: # Child Loop BB1_3 Depth 2
; X86-ALL-NEXT: movl (%ecx), %edx
; X86-ALL-NEXT: .p2align 4
; X86-ALL-NEXT: .LBB1_3: # %while.cond.i
; X86-ALL-NEXT: # Parent Loop BB1_2 Depth=1
; X86-ALL-NEXT: # => This Inner Loop Header: Depth=2
; X86-ALL-NEXT: movl %edx, %eax
; X86-ALL-NEXT: xorl %edx, %edx
; X86-ALL-NEXT: testl %eax, %eax
; X86-ALL-NEXT: je .LBB1_3
; X86-ALL-NEXT: # %bb.4: # %while.body.i
; X86-ALL-NEXT: # in Loop: Header=BB1_2 Depth=1
; X86-ALL-NEXT: lock cmpxchgl %eax, (%ecx)
; X86-ALL-NEXT: jne .LBB1_2
; X86-ALL-NEXT: # %bb.5:
; X86-ALL-NEXT: xorl %eax, %eax
; X86-ALL-NEXT: .LBB1_6: # %cond.end
; X86-ALL-NEXT: retl
;
; X64-ALL-LABEL: test_control_flow:
; X64-ALL: # %bb.0: # %entry
; X64-ALL-NEXT: movl %esi, %eax
; X64-ALL-NEXT: cmpl %edx, %esi
; X64-ALL-NEXT: jle .LBB1_5
; X64-ALL-NEXT: .p2align 4
; X64-ALL-NEXT: .LBB1_1: # %while.condthread-pre-split.i
; X64-ALL-NEXT: # =>This Loop Header: Depth=1
; X64-ALL-NEXT: # Child Loop BB1_2 Depth 2
; X64-ALL-NEXT: movl (%rdi), %ecx
; X64-ALL-NEXT: .p2align 4
; X64-ALL-NEXT: .LBB1_2: # %while.cond.i
; X64-ALL-NEXT: # Parent Loop BB1_1 Depth=1
; X64-ALL-NEXT: # => This Inner Loop Header: Depth=2
; X64-ALL-NEXT: movl %ecx, %eax
; X64-ALL-NEXT: xorl %ecx, %ecx
; X64-ALL-NEXT: testl %eax, %eax
; X64-ALL-NEXT: je .LBB1_2
; X64-ALL-NEXT: # %bb.3: # %while.body.i
; X64-ALL-NEXT: # in Loop: Header=BB1_1 Depth=1
; X64-ALL-NEXT: lock cmpxchgl %eax, (%rdi)
; X64-ALL-NEXT: jne .LBB1_1
; X64-ALL-NEXT: # %bb.4:
; X64-ALL-NEXT: xorl %eax, %eax
; X64-ALL-NEXT: .LBB1_5: # %cond.end
; X64-ALL-NEXT: retq
entry:
%cmp = icmp sgt i32 %i, %j
br i1 %cmp, label %loop_start, label %cond.end
loop_start:
br label %while.condthread-pre-split.i
while.condthread-pre-split.i:
%.pr.i = load i32, ptr %p, align 4
br label %while.cond.i
while.cond.i:
%0 = phi i32 [ %.pr.i, %while.condthread-pre-split.i ], [ 0, %while.cond.i ]
%tobool.i = icmp eq i32 %0, 0
br i1 %tobool.i, label %while.cond.i, label %while.body.i
while.body.i:
%.lcssa = phi i32 [ %0, %while.cond.i ]
%1 = cmpxchg ptr %p, i32 %.lcssa, i32 %.lcssa seq_cst seq_cst
%2 = extractvalue { i32, i1 } %1, 1
br i1 %2, label %cond.end.loopexit, label %while.condthread-pre-split.i
cond.end.loopexit:
br label %cond.end
cond.end:
%cond = phi i32 [ %i, %entry ], [ 0, %cond.end.loopexit ]
ret i32 %cond
}
; This one is an interesting case because CMOV doesn't have a chain
; operand. Naive attempts to limit cmpxchg EFLAGS use are likely to fail here.
define i32 @test_feed_cmov(ptr %addr, i32 %desired, i32 %new) nounwind {
; X86-GOOD-RA-LABEL: test_feed_cmov:
; X86-GOOD-RA: # %bb.0: # %entry
; X86-GOOD-RA-NEXT: pushl %ebx
; X86-GOOD-RA-NEXT: pushl %esi
; X86-GOOD-RA-NEXT: pushl %eax
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %esi
; X86-GOOD-RA-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X86-GOOD-RA-NEXT: lock cmpxchgl %esi, (%ecx)
; X86-GOOD-RA-NEXT: sete %bl
; X86-GOOD-RA-NEXT: calll foo@PLT
; X86-GOOD-RA-NEXT: testb %bl, %bl
; X86-GOOD-RA-NEXT: jne .LBB2_2
; X86-GOOD-RA-NEXT: # %bb.1: # %entry
; X86-GOOD-RA-NEXT: movl %eax, %esi
; X86-GOOD-RA-NEXT: .LBB2_2: # %entry
; X86-GOOD-RA-NEXT: movl %esi, %eax
; X86-GOOD-RA-NEXT: addl $4, %esp
; X86-GOOD-RA-NEXT: popl %esi
; X86-GOOD-RA-NEXT: popl %ebx
; X86-GOOD-RA-NEXT: retl
;
; X86-FAST-RA-LABEL: test_feed_cmov:
; X86-FAST-RA: # %bb.0: # %entry
; X86-FAST-RA-NEXT: pushl %ebx
; X86-FAST-RA-NEXT: pushl %esi
; X86-FAST-RA-NEXT: pushl %eax
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %esi
; X86-FAST-RA-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-FAST-RA-NEXT: lock cmpxchgl %esi, (%ecx)
; X86-FAST-RA-NEXT: sete %bl
; X86-FAST-RA-NEXT: calll foo@PLT
; X86-FAST-RA-NEXT: testb %bl, %bl
; X86-FAST-RA-NEXT: jne .LBB2_2
; X86-FAST-RA-NEXT: # %bb.1: # %entry
; X86-FAST-RA-NEXT: movl %eax, %esi
; X86-FAST-RA-NEXT: .LBB2_2: # %entry
; X86-FAST-RA-NEXT: movl %esi, %eax
; X86-FAST-RA-NEXT: addl $4, %esp
; X86-FAST-RA-NEXT: popl %esi
; X86-FAST-RA-NEXT: popl %ebx
; X86-FAST-RA-NEXT: retl
;
; X64-ALL-LABEL: test_feed_cmov:
; X64-ALL: # %bb.0: # %entry
; X64-ALL-NEXT: pushq %rbp
; X64-ALL-NEXT: pushq %rbx
; X64-ALL-NEXT: pushq %rax
; X64-ALL-NEXT: movl %edx, %ebx
; X64-ALL-NEXT: movl %esi, %eax
; X64-ALL-NEXT: lock cmpxchgl %edx, (%rdi)
; X64-ALL-NEXT: sete %bpl
; X64-ALL-NEXT: callq foo@PLT
; X64-ALL-NEXT: testb %bpl, %bpl
; X64-ALL-NEXT: cmovnel %ebx, %eax
; X64-ALL-NEXT: addq $8, %rsp
; X64-ALL-NEXT: popq %rbx
; X64-ALL-NEXT: popq %rbp
; X64-ALL-NEXT: retq
entry:
%res = cmpxchg ptr %addr, i32 %desired, i32 %new seq_cst seq_cst
%success = extractvalue { i32, i1 } %res, 1
%rhs = call i32 @foo()
%ret = select i1 %success, i32 %new, i32 %rhs
ret i32 %ret
}
Codebase Browser
llvm