; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=armv7a-none-eabi %s -o - | FileCheck %s

declare i32 @many_args_callee(i32 %0, i32 %1, i32 %2, i32 %3, i32 %4, i32 %5)

define i32 @many_args_tail(i32 %0, i32 %1, i32 %2, i32 %3, i32  %4, i32  %5) {
; CHECK-LABEL: many_args_tail:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    mov r0, #5
; CHECK-NEXT:    mov r1, #2
; CHECK-NEXT:    str r0, [sp]
; CHECK-NEXT:    mov r0, #6
; CHECK-NEXT:    str r0, [sp, #4]
; CHECK-NEXT:    mov r0, #1
; CHECK-NEXT:    mov r2, #3
; CHECK-NEXT:    mov r3, #4
; CHECK-NEXT:    b many_args_callee
  %ret = tail call i32 @many_args_callee(i32 1, i32 2, i32 3, i32 4, i32 5, i32 6)
  ret i32 %ret
}

define i32 @many_args_musttail(i32 %0, i32 %1, i32 %2, i32 %3, i32  %4, i32  %5) {
; CHECK-LABEL: many_args_musttail:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    mov r0, #5
; CHECK-NEXT:    mov r1, #2
; CHECK-NEXT:    str r0, [sp]
; CHECK-NEXT:    mov r0, #6
; CHECK-NEXT:    str r0, [sp, #4]
; CHECK-NEXT:    mov r0, #1
; CHECK-NEXT:    mov r2, #3
; CHECK-NEXT:    mov r3, #4
; CHECK-NEXT:    b many_args_callee
  %ret = musttail call i32 @many_args_callee(i32 1, i32 2, i32 3, i32 4, i32 5, i32 6)
  ret i32 %ret
}

; This function has more arguments than it's tail-callee. This isn't valid for
; the musttail attribute, but can still be tail-called as a non-guaranteed
; optimisation, because the outgoing arguments to @many_args_callee fit in the
; stack space allocated by the caller of @more_args_tail.
define i32 @more_args_tail(i32 %0, i32 %1, i32 %2, i32 %3, i32  %4, i32 %5, i32 %6) {
; CHECK-LABEL: more_args_tail:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    mov r0, #5
; CHECK-NEXT:    mov r1, #2
; CHECK-NEXT:    str r0, [sp]
; CHECK-NEXT:    mov r0, #6
; CHECK-NEXT:    str r0, [sp, #4]
; CHECK-NEXT:    mov r0, #1
; CHECK-NEXT:    mov r2, #3
; CHECK-NEXT:    mov r3, #4
; CHECK-NEXT:    b many_args_callee
  %ret = tail call i32 @many_args_callee(i32 1, i32 2, i32 3, i32 4, i32 5, i32 6)
  ret i32 %ret
}

; Again, this isn't valid for musttail, but can be tail-called in practice
; because the stack size if the same.
define i32 @different_args_tail(i64 %0, i64 %1, i64 %2) {
; CHECK-LABEL: different_args_tail:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    mov r0, #5
; CHECK-NEXT:    mov r1, #2
; CHECK-NEXT:    str r0, [sp]
; CHECK-NEXT:    mov r0, #6
; CHECK-NEXT:    str r0, [sp, #4]
; CHECK-NEXT:    mov r0, #1
; CHECK-NEXT:    mov r2, #3
; CHECK-NEXT:    mov r3, #4
; CHECK-NEXT:    b many_args_callee
  %ret = tail call i32 @many_args_callee(i32 1, i32 2, i32 3, i32 4, i32 5, i32 6)
  ret i32 %ret
}

; Here, the caller requires less stack space for it's arguments than the
; callee, so it would not ba valid to do a tail-call.
define i32 @fewer_args_tail(i32 %0, i32 %1, i32 %2, i32 %3, i32  %4) {
; CHECK-LABEL: fewer_args_tail:
; CHECK:       @ %bb.0:
; CHECK-NEXT:    .save {r11, lr}
; CHECK-NEXT:    push {r11, lr}
; CHECK-NEXT:    .pad #8
; CHECK-NEXT:    sub sp, sp, #8
; CHECK-NEXT:    mov r1, #6
; CHECK-NEXT:    mov r0, #5
; CHECK-NEXT:    strd r0, r1, [sp]
; CHECK-NEXT:    mov r0, #1
; CHECK-NEXT:    mov r1, #2
; CHECK-NEXT:    mov r2, #3
; CHECK-NEXT:    mov r3, #4
; CHECK-NEXT:    bl many_args_callee
; CHECK-NEXT:    add sp, sp, #8
; CHECK-NEXT:    pop {r11, pc}
  %ret = tail call i32 @many_args_callee(i32 1, i32 2, i32 3, i32 4, i32 5, i32 6)
  ret i32 %ret
}

declare void @sret_callee(ptr sret({ double, double }) align 8)

; Functions which return by sret can be tail-called because the incoming sret
; pointer gets passed through to the callee.
define void @sret_caller_tail(ptr sret({ double, double }) align 8 %result) {
; CHECK-LABEL: sret_caller_tail:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    b sret_callee
entry:
  tail call void @sret_callee(ptr sret({ double, double }) align 8 %result)
  ret void
}

define void @sret_caller_musttail(ptr sret({ double, double }) align 8 %result) {
; CHECK-LABEL: sret_caller_musttail:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    b sret_callee
entry:
  musttail call void @sret_callee(ptr sret({ double, double }) align 8 %result)
  ret void
}

; Clang only uses byval for arguments of 65 bytes or larger, but we test with a
; 20 byte struct to keep the tests more readable. This size was chosen to still
; make sure that it will be split between registers and the stack, to test all
; of the interesting code paths in the backend.
%twenty_bytes = type { [5 x i32] }
declare void @large_callee(%twenty_bytes* byval(%twenty_bytes) align 4)

; Functions with byval parameters can be tail-called, because the value is
; actually passed in registers and the stack in the same way for the caller and
; callee. Within @large_caller the first 16 bytes of the argument are spilled
; to the local stack frame, but for the tail-call they are passed in r0-r3, so
; it's safe to de-allocate that memory before the call.
; TODO: The SUB and STM instructions are unnecessary and could be optimised
; out, but the behaviour of this is still correct.
define void @large_caller(%twenty_bytes* byval(%twenty_bytes) align 4 %a) {
; CHECK-LABEL: large_caller:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #16
; CHECK-NEXT:    sub sp, sp, #16
; CHECK-NEXT:    stm sp!, {r0, r1, r2, r3}
; CHECK-NEXT:    b large_callee
entry:
  musttail call void @large_callee(%twenty_bytes* byval(%twenty_bytes) align 4 %a)
  ret void
}

; As above, but with some inline asm to test that the arguments in r0-r3 are
; re-loaded before the call.
define void @large_caller_check_regs(%twenty_bytes* byval(%twenty_bytes) align 4 %a) {
; CHECK-LABEL: large_caller_check_regs:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #16
; CHECK-NEXT:    sub sp, sp, #16
; CHECK-NEXT:    stm sp, {r0, r1, r2, r3}
; CHECK-NEXT:    @APP
; CHECK-NEXT:    @NO_APP
; CHECK-NEXT:    pop {r0, r1, r2, r3}
; CHECK-NEXT:    b large_callee
entry:
  tail call void asm sideeffect "", "~{r0},~{r1},~{r2},~{r3}"()
  musttail call void @large_callee(%twenty_bytes* byval(%twenty_bytes) align 4 %a)
  ret void
}

; The IR for this one looks dodgy, because it has an alloca passed to a
; musttail function, but it is passed as a byval argument, so will be copied
; into the stack space allocated by @large_caller_new_value's caller, so is
; valid.
define void @large_caller_new_value(%twenty_bytes* byval(%twenty_bytes) align 4 %a) {
; CHECK-LABEL: large_caller_new_value:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #36
; CHECK-NEXT:    sub sp, sp, #36
; CHECK-NEXT:    add r12, sp, #20
; CHECK-NEXT:    stm r12, {r0, r1, r2, r3}
; CHECK-NEXT:    mov r0, #4
; CHECK-NEXT:    add r1, sp, #36
; CHECK-NEXT:    str r0, [sp, #16]
; CHECK-NEXT:    mov r0, #3
; CHECK-NEXT:    str r0, [sp, #12]
; CHECK-NEXT:    mov r0, #2
; CHECK-NEXT:    str r0, [sp, #8]
; CHECK-NEXT:    mov r0, #1
; CHECK-NEXT:    str r0, [sp, #4]
; CHECK-NEXT:    mov r0, #0
; CHECK-NEXT:    str r0, [sp]
; CHECK-NEXT:    mov r0, sp
; CHECK-NEXT:    add r0, r0, #16
; CHECK-NEXT:    mov r3, #3
; CHECK-NEXT:    ldr r2, [r0], #4
; CHECK-NEXT:    str r2, [r1], #4
; CHECK-NEXT:    mov r0, #0
; CHECK-NEXT:    mov r1, #1
; CHECK-NEXT:    mov r2, #2
; CHECK-NEXT:    add sp, sp, #36
; CHECK-NEXT:    b large_callee
entry:
  %y = alloca %twenty_bytes, align 4
  store i32 0, ptr %y, align 4
  %0 = getelementptr inbounds i8, ptr %y, i32 4
  store i32 1, ptr %0, align 4
  %1 = getelementptr inbounds i8, ptr %y, i32 8
  store i32 2, ptr %1, align 4
  %2 = getelementptr inbounds i8, ptr %y, i32 12
  store i32 3, ptr %2, align 4
  %3 = getelementptr inbounds i8, ptr %y, i32 16
  store i32 4, ptr %3, align 4
  musttail call void @large_callee(%twenty_bytes* byval(%twenty_bytes) align 4 %y)
  ret void
}

declare void @two_byvals_callee(%twenty_bytes* byval(%twenty_bytes) align 4, %twenty_bytes* byval(%twenty_bytes) align 4)
define void @swap_byvals(%twenty_bytes* byval(%twenty_bytes) align 4 %a, %twenty_bytes* byval(%twenty_bytes) align 4 %b) {
; CHECK-LABEL: swap_byvals:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #16
; CHECK-NEXT:    sub sp, sp, #16
; CHECK-NEXT:    .save {r4, r5, r11, lr}
; CHECK-NEXT:    push {r4, r5, r11, lr}
; CHECK-NEXT:    .pad #40
; CHECK-NEXT:    sub sp, sp, #40
; CHECK-NEXT:    add r12, sp, #56
; CHECK-NEXT:    add lr, sp, #20
; CHECK-NEXT:    stm r12, {r0, r1, r2, r3}
; CHECK-NEXT:    add r0, sp, #56
; CHECK-NEXT:    mov r12, sp
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    mov r2, r12
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    add r3, sp, #20
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    add r4, sp, #76
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    add r0, sp, #76
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    mov r2, lr
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldr r1, [r0], #4
; CHECK-NEXT:    str r1, [r2], #4
; CHECK-NEXT:    ldm r3, {r0, r1, r2, r3}
; CHECK-NEXT:    ldr r5, [r12], #4
; CHECK-NEXT:    str r5, [r4], #4
; CHECK-NEXT:    ldr r5, [r12], #4
; CHECK-NEXT:    str r5, [r4], #4
; CHECK-NEXT:    ldr r5, [r12], #4
; CHECK-NEXT:    str r5, [r4], #4
; CHECK-NEXT:    ldr r5, [r12], #4
; CHECK-NEXT:    str r5, [r4], #4
; CHECK-NEXT:    ldr r5, [r12], #4
; CHECK-NEXT:    str r5, [r4], #4
; CHECK-NEXT:    add r5, lr, #16
; CHECK-NEXT:    add r12, sp, #72
; CHECK-NEXT:    ldr r4, [r5], #4
; CHECK-NEXT:    str r4, [r12], #4
; CHECK-NEXT:    add sp, sp, #40
; CHECK-NEXT:    pop {r4, r5, r11, lr}
; CHECK-NEXT:    add sp, sp, #16
; CHECK-NEXT:    b two_byvals_callee
entry:
  musttail call void @two_byvals_callee(%twenty_bytes* byval(%twenty_bytes) align 4 %b, %twenty_bytes* byval(%twenty_bytes) align 4 %a)
  ret void
}

; A forwarded byval arg, but at a different offset on the stack, so it needs to
; be copied to the local stack frame first. This can't be musttail because of
; the different signatures, but is still tail-called as an optimisation.
declare void @shift_byval_callee(%twenty_bytes* byval(%twenty_bytes) align 4)
define void @shift_byval(i32 %a, %twenty_bytes* byval(%twenty_bytes) align 4 %b) {
; CHECK-LABEL: shift_byval:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #12
; CHECK-NEXT:    sub sp, sp, #12
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    .pad #20
; CHECK-NEXT:    sub sp, sp, #20
; CHECK-NEXT:    add r0, sp, #28
; CHECK-NEXT:    add lr, sp, #40
; CHECK-NEXT:    stm r0, {r1, r2, r3}
; CHECK-NEXT:    add r0, sp, #28
; CHECK-NEXT:    mov r1, sp
; CHECK-NEXT:    ldr r2, [r0], #4
; CHECK-NEXT:    add r12, r1, #16
; CHECK-NEXT:    str r2, [r1], #4
; CHECK-NEXT:    ldr r2, [r0], #4
; CHECK-NEXT:    str r2, [r1], #4
; CHECK-NEXT:    ldr r2, [r0], #4
; CHECK-NEXT:    str r2, [r1], #4
; CHECK-NEXT:    ldr r2, [r0], #4
; CHECK-NEXT:    str r2, [r1], #4
; CHECK-NEXT:    ldr r2, [r0], #4
; CHECK-NEXT:    str r2, [r1], #4
; CHECK-NEXT:    ldm sp, {r0, r1, r2, r3}
; CHECK-NEXT:    ldr r4, [r12], #4
; CHECK-NEXT:    str r4, [lr], #4
; CHECK-NEXT:    add sp, sp, #20
; CHECK-NEXT:    pop {r4, lr}
; CHECK-NEXT:    add sp, sp, #12
; CHECK-NEXT:    b shift_byval_callee
entry:
  tail call void @shift_byval_callee(%twenty_bytes* byval(%twenty_bytes) align 4 %b)
  ret void
}

; A global object passed to a byval argument, so it must be copied, but doesn't
; need a stack temporary.
@large_global = external global %twenty_bytes
define void @large_caller_from_global(%twenty_bytes* byval(%twenty_bytes) align 4 %a) {
; CHECK-LABEL: large_caller_from_global:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .pad #16
; CHECK-NEXT:    sub sp, sp, #16
; CHECK-NEXT:    .save {r4, lr}
; CHECK-NEXT:    push {r4, lr}
; CHECK-NEXT:    add r12, sp, #8
; CHECK-NEXT:    add lr, sp, #24
; CHECK-NEXT:    stm r12, {r0, r1, r2, r3}
; CHECK-NEXT:    movw r3, :lower16:large_global
; CHECK-NEXT:    movt r3, :upper16:large_global
; CHECK-NEXT:    add r12, r3, #16
; CHECK-NEXT:    ldm r3, {r0, r1, r2, r3}
; CHECK-NEXT:    ldr r4, [r12], #4
; CHECK-NEXT:    str r4, [lr], #4
; CHECK-NEXT:    pop {r4, lr}
; CHECK-NEXT:    add sp, sp, #16
; CHECK-NEXT:    b large_callee
entry:
  musttail call void @large_callee(%twenty_bytes* byval(%twenty_bytes) align 4 @large_global)
  ret void
}