; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=loongarch64 -O2 < %s | FileCheck %s -check-prefix=LA64
; RUN: llc -mtriple=loongarch32 -O2 < %s | FileCheck %s -check-prefix=LA32

;; Tests copied from PowerPC.

;; Free probe
define i8 @f0() #0 {
;
; LA64-LABEL: f0:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    addi.d $sp, $sp, -80
; LA64-NEXT:    .cfi_def_cfa_offset 80
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    addi.d $sp, $sp, 80
; LA64-NEXT:    ret
;
; LA32-LABEL: f0:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    addi.w $sp, $sp, -64
; LA32-NEXT:    .cfi_def_cfa_offset 64
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 0
; LA32-NEXT:    ld.b $a0, $sp, 0
; LA32-NEXT:    addi.w $sp, $sp, 64
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 64
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

define i8 @f1() #0 {
;
; LA64-LABEL: f1:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 1
; LA64-NEXT:    sub.d $sp, $sp, $a0
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    .cfi_def_cfa_offset 4096
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    .cfi_def_cfa_offset 4112
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    lu12i.w $a1, 1
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f1:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 1
; LA32-NEXT:    sub.w $sp, $sp, $a0
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    .cfi_def_cfa_offset 4096
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    .cfi_def_cfa_offset 4112
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 16
; LA32-NEXT:    ld.b $a0, $sp, 16
; LA32-NEXT:    lu12i.w $a1, 1
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 4096
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

define i8 @f2() #0 {
;
; LA64-LABEL: f2:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 16
; LA64-NEXT:    sub.d $t1, $sp, $a0
; LA64-NEXT:    .cfi_def_cfa 13, 65536
; LA64-NEXT:    lu12i.w $t2, 1
; LA64-NEXT:  .LBB2_1: # %entry
; LA64-NEXT:    # =>This Inner Loop Header: Depth=1
; LA64-NEXT:    sub.d $sp, $sp, $t2
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bne $sp, $t1, .LBB2_1
; LA64-NEXT:  # %bb.2: # %entry
; LA64-NEXT:    .cfi_def_cfa_register 3
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    .cfi_def_cfa_offset 65552
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    lu12i.w $a1, 16
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f2:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 16
; LA32-NEXT:    sub.w $t1, $sp, $a0
; LA32-NEXT:    .cfi_def_cfa 13, 65536
; LA32-NEXT:    lu12i.w $t2, 1
; LA32-NEXT:  .LBB2_1: # %entry
; LA32-NEXT:    # =>This Inner Loop Header: Depth=1
; LA32-NEXT:    sub.w $sp, $sp, $t2
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bne $sp, $t1, .LBB2_1
; LA32-NEXT:  # %bb.2: # %entry
; LA32-NEXT:    .cfi_def_cfa_register 3
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    .cfi_def_cfa_offset 65552
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 16
; LA32-NEXT:    ld.b $a0, $sp, 16
; LA32-NEXT:    lu12i.w $a1, 16
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 65536
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

define i8 @f3() #0 "stack-probe-size"="32768" {
;
; LA64-LABEL: f3:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 8
; LA64-NEXT:    sub.d $sp, $sp, $a0
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    .cfi_def_cfa_offset 32768
; LA64-NEXT:    lu12i.w $a0, 8
; LA64-NEXT:    sub.d $sp, $sp, $a0
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    .cfi_def_cfa_offset 65536
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    .cfi_def_cfa_offset 65552
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    lu12i.w $a1, 16
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f3:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 8
; LA32-NEXT:    sub.w $sp, $sp, $a0
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    .cfi_def_cfa_offset 32768
; LA32-NEXT:    lu12i.w $a0, 8
; LA32-NEXT:    sub.w $sp, $sp, $a0
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    .cfi_def_cfa_offset 65536
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    .cfi_def_cfa_offset 65552
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 16
; LA32-NEXT:    ld.b $a0, $sp, 16
; LA32-NEXT:    lu12i.w $a1, 16
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 65536
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

; Same as f2, but without protection.
define i8 @f4() {
;
; LA64-LABEL: f4:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 16
; LA64-NEXT:    ori $a0, $a0, 16
; LA64-NEXT:    sub.d $sp, $sp, $a0
; LA64-NEXT:    .cfi_def_cfa_offset 65552
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    lu12i.w $a1, 16
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f4:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 16
; LA32-NEXT:    ori $a0, $a0, 16
; LA32-NEXT:    sub.w $sp, $sp, $a0
; LA32-NEXT:    .cfi_def_cfa_offset 65552
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 16
; LA32-NEXT:    ld.b $a0, $sp, 16
; LA32-NEXT:    lu12i.w $a1, 16
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 65536
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

define i8 @f5() #0 "stack-probe-size"="65536" {
;
; LA64-LABEL: f5:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 256
; LA64-NEXT:    sub.d $t1, $sp, $a0
; LA64-NEXT:    .cfi_def_cfa 13, 1048576
; LA64-NEXT:    lu12i.w $t2, 16
; LA64-NEXT:  .LBB5_1: # %entry
; LA64-NEXT:    # =>This Inner Loop Header: Depth=1
; LA64-NEXT:    sub.d $sp, $sp, $t2
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bne $sp, $t1, .LBB5_1
; LA64-NEXT:  # %bb.2: # %entry
; LA64-NEXT:    .cfi_def_cfa_register 3
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    .cfi_def_cfa_offset 1048592
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    lu12i.w $a1, 256
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f5:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 256
; LA32-NEXT:    sub.w $t1, $sp, $a0
; LA32-NEXT:    .cfi_def_cfa 13, 1048576
; LA32-NEXT:    lu12i.w $t2, 16
; LA32-NEXT:  .LBB5_1: # %entry
; LA32-NEXT:    # =>This Inner Loop Header: Depth=1
; LA32-NEXT:    sub.w $sp, $sp, $t2
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bne $sp, $t1, .LBB5_1
; LA32-NEXT:  # %bb.2: # %entry
; LA32-NEXT:    .cfi_def_cfa_register 3
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    .cfi_def_cfa_offset 1048592
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 16
; LA32-NEXT:    ld.b $a0, $sp, 16
; LA32-NEXT:    lu12i.w $a1, 256
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 1048576
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

define i8 @f6() #0 {
;
; LA64-LABEL: f6:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 262144
; LA64-NEXT:    sub.d $t1, $sp, $a0
; LA64-NEXT:    .cfi_def_cfa 13, 1073741824
; LA64-NEXT:    lu12i.w $t2, 1
; LA64-NEXT:  .LBB6_1: # %entry
; LA64-NEXT:    # =>This Inner Loop Header: Depth=1
; LA64-NEXT:    sub.d $sp, $sp, $t2
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bne $sp, $t1, .LBB6_1
; LA64-NEXT:  # %bb.2: # %entry
; LA64-NEXT:    .cfi_def_cfa_register 3
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    .cfi_def_cfa_offset 1073741840
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 16
; LA64-NEXT:    ld.b $a0, $sp, 16
; LA64-NEXT:    lu12i.w $a1, 262144
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f6:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 262144
; LA32-NEXT:    sub.w $t1, $sp, $a0
; LA32-NEXT:    .cfi_def_cfa 13, 1073741824
; LA32-NEXT:    lu12i.w $t2, 1
; LA32-NEXT:  .LBB6_1: # %entry
; LA32-NEXT:    # =>This Inner Loop Header: Depth=1
; LA32-NEXT:    sub.w $sp, $sp, $t2
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bne $sp, $t1, .LBB6_1
; LA32-NEXT:  # %bb.2: # %entry
; LA32-NEXT:    .cfi_def_cfa_register 3
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    .cfi_def_cfa_offset 1073741840
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 16
; LA32-NEXT:    ld.b $a0, $sp, 16
; LA32-NEXT:    lu12i.w $a1, 262144
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 1073741824
  %b = getelementptr inbounds i8, ptr %a, i64 63
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

define i8 @f7() #0 "stack-probe-size"="65536" {
;
; LA64-LABEL: f7:
; LA64:       # %bb.0: # %entry
; LA64-NEXT:    lu12i.w $a0, 244128
; LA64-NEXT:    sub.d $t1, $sp, $a0
; LA64-NEXT:    .cfi_def_cfa 13, 999948288
; LA64-NEXT:    lu12i.w $t2, 16
; LA64-NEXT:  .LBB7_1: # %entry
; LA64-NEXT:    # =>This Inner Loop Header: Depth=1
; LA64-NEXT:    sub.d $sp, $sp, $t2
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bne $sp, $t1, .LBB7_1
; LA64-NEXT:  # %bb.2: # %entry
; LA64-NEXT:    .cfi_def_cfa_register 3
; LA64-NEXT:    lu12i.w $a0, 12
; LA64-NEXT:    ori $a0, $a0, 2576
; LA64-NEXT:    sub.d $sp, $sp, $a0
; LA64-NEXT:    .cfi_def_cfa_offset 1000000016
; LA64-NEXT:    ori $a0, $zero, 3
; LA64-NEXT:    st.b $a0, $sp, 9
; LA64-NEXT:    ld.b $a0, $sp, 9
; LA64-NEXT:    lu12i.w $a1, 244140
; LA64-NEXT:    ori $a1, $a1, 2576
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ret
;
; LA32-LABEL: f7:
; LA32:       # %bb.0: # %entry
; LA32-NEXT:    lu12i.w $a0, 244128
; LA32-NEXT:    sub.w $t1, $sp, $a0
; LA32-NEXT:    .cfi_def_cfa 13, 999948288
; LA32-NEXT:    lu12i.w $t2, 16
; LA32-NEXT:  .LBB7_1: # %entry
; LA32-NEXT:    # =>This Inner Loop Header: Depth=1
; LA32-NEXT:    sub.w $sp, $sp, $t2
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bne $sp, $t1, .LBB7_1
; LA32-NEXT:  # %bb.2: # %entry
; LA32-NEXT:    .cfi_def_cfa_register 3
; LA32-NEXT:    lu12i.w $a0, 12
; LA32-NEXT:    ori $a0, $a0, 2576
; LA32-NEXT:    sub.w $sp, $sp, $a0
; LA32-NEXT:    .cfi_def_cfa_offset 1000000016
; LA32-NEXT:    ori $a0, $zero, 3
; LA32-NEXT:    st.b $a0, $sp, 9
; LA32-NEXT:    ld.b $a0, $sp, 9
; LA32-NEXT:    lu12i.w $a1, 244140
; LA32-NEXT:    ori $a1, $a1, 2576
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ret
entry:
  %a = alloca i8, i64 1000000007
  %b = getelementptr inbounds i8, ptr %a, i64 101
  store volatile i8 3, ptr %a
  %c = load volatile i8, ptr %a
  ret i8 %c
}

; alloca + align < probe_size
define i32 @f8(i64 %i) local_unnamed_addr #0 {
;
; LA64-LABEL: f8:
; LA64:       # %bb.0:
; LA64-NEXT:    addi.d $sp, $sp, -896
; LA64-NEXT:    .cfi_def_cfa_offset 896
; LA64-NEXT:    st.d $ra, $sp, 888 # 8-byte Folded Spill
; LA64-NEXT:    st.d $fp, $sp, 880 # 8-byte Folded Spill
; LA64-NEXT:    .cfi_offset 1, -8
; LA64-NEXT:    .cfi_offset 22, -16
; LA64-NEXT:    addi.d $fp, $sp, 896
; LA64-NEXT:    .cfi_def_cfa 22, 0
; LA64-NEXT:    bstrins.d $sp, $zero, 5, 0
; LA64-NEXT:    slli.d $a0, $a0, 2
; LA64-NEXT:    addi.d $a1, $sp, 64
; LA64-NEXT:    ori $a2, $zero, 1
; LA64-NEXT:    stx.w $a2, $a0, $a1
; LA64-NEXT:    ld.w $a0, $sp, 64
; LA64-NEXT:    addi.d $sp, $fp, -896
; LA64-NEXT:    ld.d $fp, $sp, 880 # 8-byte Folded Reload
; LA64-NEXT:    ld.d $ra, $sp, 888 # 8-byte Folded Reload
; LA64-NEXT:    addi.d $sp, $sp, 896
; LA64-NEXT:    ret
;
; LA32-LABEL: f8:
; LA32:       # %bb.0:
; LA32-NEXT:    addi.w $sp, $sp, -832
; LA32-NEXT:    .cfi_def_cfa_offset 832
; LA32-NEXT:    st.w $ra, $sp, 828 # 4-byte Folded Spill
; LA32-NEXT:    st.w $fp, $sp, 824 # 4-byte Folded Spill
; LA32-NEXT:    .cfi_offset 1, -4
; LA32-NEXT:    .cfi_offset 22, -8
; LA32-NEXT:    addi.w $fp, $sp, 832
; LA32-NEXT:    .cfi_def_cfa 22, 0
; LA32-NEXT:    bstrins.w $sp, $zero, 5, 0
; LA32-NEXT:    slli.w $a0, $a0, 2
; LA32-NEXT:    addi.w $a1, $sp, 0
; LA32-NEXT:    add.w $a0, $a1, $a0
; LA32-NEXT:    ori $a1, $zero, 1
; LA32-NEXT:    st.w $a1, $a0, 0
; LA32-NEXT:    ld.w $a0, $sp, 0
; LA32-NEXT:    addi.w $sp, $fp, -832
; LA32-NEXT:    ld.w $fp, $sp, 824 # 4-byte Folded Reload
; LA32-NEXT:    ld.w $ra, $sp, 828 # 4-byte Folded Reload
; LA32-NEXT:    addi.w $sp, $sp, 832
; LA32-NEXT:    ret
  %a = alloca i32, i32 200, align 64
  %b = getelementptr inbounds i32, ptr %a, i64 %i
  store volatile i32 1, ptr %b
  %c = load volatile i32, ptr %a
  ret i32 %c
}

; alloca > probe_size, align > probe_size
define i32 @f9(i64 %i) local_unnamed_addr #0 {
;
; LA64-LABEL: f9:
; LA64:       # %bb.0:
; LA64-NEXT:    addi.d $sp, $sp, -2032
; LA64-NEXT:    .cfi_def_cfa_offset 2032
; LA64-NEXT:    st.d $ra, $sp, 2024 # 8-byte Folded Spill
; LA64-NEXT:    st.d $fp, $sp, 2016 # 8-byte Folded Spill
; LA64-NEXT:    .cfi_offset 1, -8
; LA64-NEXT:    .cfi_offset 22, -16
; LA64-NEXT:    addi.d $fp, $sp, 2032
; LA64-NEXT:    .cfi_def_cfa 22, 0
; LA64-NEXT:    lu12i.w $a1, 1
; LA64-NEXT:    sub.d $sp, $sp, $a1
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    sub.d $sp, $sp, $a1
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    bstrins.d $sp, $zero, 10, 0
; LA64-NEXT:    slli.d $a0, $a0, 2
; LA64-NEXT:    ori $a1, $zero, 2048
; LA64-NEXT:    add.d $a1, $sp, $a1
; LA64-NEXT:    ori $a2, $zero, 1
; LA64-NEXT:    stx.w $a2, $a0, $a1
; LA64-NEXT:    ori $a0, $zero, 2048
; LA64-NEXT:    add.d $a0, $sp, $a0
; LA64-NEXT:    ld.w $a0, $a0, 0
; LA64-NEXT:    lu12i.w $a1, 2
; LA64-NEXT:    ori $a1, $a1, 2048
; LA64-NEXT:    sub.d $sp, $fp, $a1
; LA64-NEXT:    lu12i.w $a1, 2
; LA64-NEXT:    ori $a1, $a1, 16
; LA64-NEXT:    add.d $sp, $sp, $a1
; LA64-NEXT:    ld.d $fp, $sp, 2016 # 8-byte Folded Reload
; LA64-NEXT:    ld.d $ra, $sp, 2024 # 8-byte Folded Reload
; LA64-NEXT:    addi.d $sp, $sp, 2032
; LA64-NEXT:    ret
;
; LA32-LABEL: f9:
; LA32:       # %bb.0:
; LA32-NEXT:    addi.w $sp, $sp, -2032
; LA32-NEXT:    .cfi_def_cfa_offset 2032
; LA32-NEXT:    st.w $ra, $sp, 2028 # 4-byte Folded Spill
; LA32-NEXT:    st.w $fp, $sp, 2024 # 4-byte Folded Spill
; LA32-NEXT:    .cfi_offset 1, -4
; LA32-NEXT:    .cfi_offset 22, -8
; LA32-NEXT:    addi.w $fp, $sp, 2032
; LA32-NEXT:    .cfi_def_cfa 22, 0
; LA32-NEXT:    lu12i.w $a1, 1
; LA32-NEXT:    sub.w $sp, $sp, $a1
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    sub.w $sp, $sp, $a1
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    bstrins.w $sp, $zero, 10, 0
; LA32-NEXT:    slli.w $a0, $a0, 2
; LA32-NEXT:    ori $a1, $zero, 2048
; LA32-NEXT:    add.w $a1, $sp, $a1
; LA32-NEXT:    add.w $a0, $a1, $a0
; LA32-NEXT:    ori $a1, $zero, 1
; LA32-NEXT:    st.w $a1, $a0, 0
; LA32-NEXT:    ori $a0, $zero, 2048
; LA32-NEXT:    add.w $a0, $sp, $a0
; LA32-NEXT:    ld.w $a0, $a0, 0
; LA32-NEXT:    lu12i.w $a1, 2
; LA32-NEXT:    ori $a1, $a1, 2048
; LA32-NEXT:    sub.w $sp, $fp, $a1
; LA32-NEXT:    lu12i.w $a1, 2
; LA32-NEXT:    ori $a1, $a1, 16
; LA32-NEXT:    add.w $sp, $sp, $a1
; LA32-NEXT:    ld.w $fp, $sp, 2024 # 4-byte Folded Reload
; LA32-NEXT:    ld.w $ra, $sp, 2028 # 4-byte Folded Reload
; LA32-NEXT:    addi.w $sp, $sp, 2032
; LA32-NEXT:    ret
  %a = alloca i32, i32 2000, align 2048
  %b = getelementptr inbounds i32, ptr %a, i64 %i
  store volatile i32 1, ptr %b
  %c = load volatile i32, ptr %a
  ret i32 %c
}

; alloca < probe_size, align < probe_size, alloca + align > probe_size
define i32 @f10(i64 %i) local_unnamed_addr #0 {
;
; LA64-LABEL: f10:
; LA64:       # %bb.0:
; LA64-NEXT:    addi.d $sp, $sp, -2032
; LA64-NEXT:    .cfi_def_cfa_offset 2032
; LA64-NEXT:    st.d $ra, $sp, 2024 # 8-byte Folded Spill
; LA64-NEXT:    st.d $fp, $sp, 2016 # 8-byte Folded Spill
; LA64-NEXT:    .cfi_offset 1, -8
; LA64-NEXT:    .cfi_offset 22, -16
; LA64-NEXT:    addi.d $fp, $sp, 2032
; LA64-NEXT:    .cfi_def_cfa 22, 0
; LA64-NEXT:    addi.d $sp, $sp, -2048
; LA64-NEXT:    addi.d $sp, $sp, -1040
; LA64-NEXT:    bstrins.d $sp, $zero, 9, 0
; LA64-NEXT:    slli.d $a0, $a0, 2
; LA64-NEXT:    addi.d $a1, $sp, 1024
; LA64-NEXT:    ori $a2, $zero, 1
; LA64-NEXT:    stx.w $a2, $a0, $a1
; LA64-NEXT:    ld.w $a0, $sp, 1024
; LA64-NEXT:    lu12i.w $a1, 1
; LA64-NEXT:    ori $a1, $a1, 1024
; LA64-NEXT:    sub.d $sp, $fp, $a1
; LA64-NEXT:    addi.d $sp, $sp, 2032
; LA64-NEXT:    addi.d $sp, $sp, 1056
; LA64-NEXT:    ld.d $fp, $sp, 2016 # 8-byte Folded Reload
; LA64-NEXT:    ld.d $ra, $sp, 2024 # 8-byte Folded Reload
; LA64-NEXT:    addi.d $sp, $sp, 2032
; LA64-NEXT:    ret
;
; LA32-LABEL: f10:
; LA32:       # %bb.0:
; LA32-NEXT:    addi.w $sp, $sp, -2032
; LA32-NEXT:    .cfi_def_cfa_offset 2032
; LA32-NEXT:    st.w $ra, $sp, 2028 # 4-byte Folded Spill
; LA32-NEXT:    st.w $fp, $sp, 2024 # 4-byte Folded Spill
; LA32-NEXT:    .cfi_offset 1, -4
; LA32-NEXT:    .cfi_offset 22, -8
; LA32-NEXT:    addi.w $fp, $sp, 2032
; LA32-NEXT:    .cfi_def_cfa 22, 0
; LA32-NEXT:    addi.w $sp, $sp, -2048
; LA32-NEXT:    addi.w $sp, $sp, -1040
; LA32-NEXT:    bstrins.w $sp, $zero, 9, 0
; LA32-NEXT:    slli.w $a0, $a0, 2
; LA32-NEXT:    addi.w $a1, $sp, 1024
; LA32-NEXT:    add.w $a0, $a1, $a0
; LA32-NEXT:    ori $a1, $zero, 1
; LA32-NEXT:    st.w $a1, $a0, 0
; LA32-NEXT:    ld.w $a0, $sp, 1024
; LA32-NEXT:    lu12i.w $a1, 1
; LA32-NEXT:    ori $a1, $a1, 1024
; LA32-NEXT:    sub.w $sp, $fp, $a1
; LA32-NEXT:    addi.w $sp, $sp, 2032
; LA32-NEXT:    addi.w $sp, $sp, 1056
; LA32-NEXT:    ld.w $fp, $sp, 2024 # 4-byte Folded Reload
; LA32-NEXT:    ld.w $ra, $sp, 2028 # 4-byte Folded Reload
; LA32-NEXT:    addi.w $sp, $sp, 2032
; LA32-NEXT:    ret
  %a = alloca i32, i32 1000, align 1024
  %b = getelementptr inbounds i32, ptr %a, i64 %i
  store volatile i32 1, ptr %b
  %c = load volatile i32, ptr %a
  ret i32 %c
}

define void @f11(i32 %vla_size, i64 %i) #0 {
;
; LA64-LABEL: f11:
; LA64:       # %bb.0:
; LA64-NEXT:    addi.d $sp, $sp, -2032
; LA64-NEXT:    .cfi_def_cfa_offset 2032
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    st.d $ra, $sp, 2024 # 8-byte Folded Spill
; LA64-NEXT:    st.d $fp, $sp, 2016 # 8-byte Folded Spill
; LA64-NEXT:    st.d $s8, $sp, 2008 # 8-byte Folded Spill
; LA64-NEXT:    .cfi_offset 1, -8
; LA64-NEXT:    .cfi_offset 22, -16
; LA64-NEXT:    .cfi_offset 31, -24
; LA64-NEXT:    addi.d $fp, $sp, 2032
; LA64-NEXT:    .cfi_def_cfa 22, 0
; LA64-NEXT:    lu12i.w $a2, 15
; LA64-NEXT:    sub.d $t1, $sp, $a2
; LA64-NEXT:    lu12i.w $t2, 1
; LA64-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
; LA64-NEXT:    sub.d $sp, $sp, $t2
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bne $sp, $t1, .LBB11_1
; LA64-NEXT:  # %bb.2:
; LA64-NEXT:    addi.d $sp, $sp, -2048
; LA64-NEXT:    addi.d $sp, $sp, -16
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bstrins.d $sp, $zero, 14, 0
; LA64-NEXT:    move $s8, $sp
; LA64-NEXT:    slli.d $a1, $a1, 2
; LA64-NEXT:    lu12i.w $a2, 8
; LA64-NEXT:    add.d $a2, $s8, $a2
; LA64-NEXT:    ori $a3, $zero, 1
; LA64-NEXT:    stx.w $a3, $a1, $a2
; LA64-NEXT:    bstrpick.d $a0, $a0, 31, 0
; LA64-NEXT:    addi.d $a0, $a0, 15
; LA64-NEXT:    bstrpick.d $a0, $a0, 32, 4
; LA64-NEXT:    slli.d $a0, $a0, 4
; LA64-NEXT:    sub.d $a0, $sp, $a0
; LA64-NEXT:    bstrins.d $a0, $zero, 10, 0
; LA64-NEXT:    lu12i.w $a1, 1
; LA64-NEXT:  .LBB11_3: # =>This Inner Loop Header: Depth=1
; LA64-NEXT:    sub.d $sp, $sp, $a1
; LA64-NEXT:    st.d $zero, $sp, 0
; LA64-NEXT:    bltu $a0, $sp, .LBB11_3
; LA64-NEXT:  # %bb.4:
; LA64-NEXT:    move $sp, $a0
; LA64-NEXT:    ld.b $zero, $a0, 0
; LA64-NEXT:    lu12i.w $a0, 16
; LA64-NEXT:    sub.d $sp, $fp, $a0
; LA64-NEXT:    lu12i.w $a0, 15
; LA64-NEXT:    ori $a0, $a0, 2064
; LA64-NEXT:    add.d $sp, $sp, $a0
; LA64-NEXT:    ld.d $s8, $sp, 2008 # 8-byte Folded Reload
; LA64-NEXT:    ld.d $fp, $sp, 2016 # 8-byte Folded Reload
; LA64-NEXT:    ld.d $ra, $sp, 2024 # 8-byte Folded Reload
; LA64-NEXT:    addi.d $sp, $sp, 2032
; LA64-NEXT:    ret
;
; LA32-LABEL: f11:
; LA32:       # %bb.0:
; LA32-NEXT:    addi.w $sp, $sp, -2032
; LA32-NEXT:    .cfi_def_cfa_offset 2032
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    st.w $ra, $sp, 2028 # 4-byte Folded Spill
; LA32-NEXT:    st.w $fp, $sp, 2024 # 4-byte Folded Spill
; LA32-NEXT:    st.w $s8, $sp, 2020 # 4-byte Folded Spill
; LA32-NEXT:    .cfi_offset 1, -4
; LA32-NEXT:    .cfi_offset 22, -8
; LA32-NEXT:    .cfi_offset 31, -12
; LA32-NEXT:    addi.w $fp, $sp, 2032
; LA32-NEXT:    .cfi_def_cfa 22, 0
; LA32-NEXT:    lu12i.w $a2, 15
; LA32-NEXT:    sub.w $t1, $sp, $a2
; LA32-NEXT:    lu12i.w $t2, 1
; LA32-NEXT:  .LBB11_1: # =>This Inner Loop Header: Depth=1
; LA32-NEXT:    sub.w $sp, $sp, $t2
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bne $sp, $t1, .LBB11_1
; LA32-NEXT:  # %bb.2:
; LA32-NEXT:    addi.w $sp, $sp, -2048
; LA32-NEXT:    addi.w $sp, $sp, -16
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bstrins.w $sp, $zero, 14, 0
; LA32-NEXT:    move $s8, $sp
; LA32-NEXT:    slli.w $a1, $a1, 2
; LA32-NEXT:    lu12i.w $a2, 8
; LA32-NEXT:    add.w $a2, $s8, $a2
; LA32-NEXT:    add.w $a1, $a2, $a1
; LA32-NEXT:    ori $a2, $zero, 1
; LA32-NEXT:    st.w $a2, $a1, 0
; LA32-NEXT:    addi.w $a0, $a0, 15
; LA32-NEXT:    addi.w $a1, $zero, -16
; LA32-NEXT:    and $a0, $a0, $a1
; LA32-NEXT:    sub.w $a0, $sp, $a0
; LA32-NEXT:    addi.w $a1, $zero, -2048
; LA32-NEXT:    and $a0, $a0, $a1
; LA32-NEXT:    lu12i.w $a1, 1
; LA32-NEXT:  .LBB11_3: # =>This Inner Loop Header: Depth=1
; LA32-NEXT:    sub.w $sp, $sp, $a1
; LA32-NEXT:    st.w $zero, $sp, 0
; LA32-NEXT:    bltu $a0, $sp, .LBB11_3
; LA32-NEXT:  # %bb.4:
; LA32-NEXT:    move $sp, $a0
; LA32-NEXT:    ld.b $zero, $a0, 0
; LA32-NEXT:    lu12i.w $a0, 16
; LA32-NEXT:    sub.w $sp, $fp, $a0
; LA32-NEXT:    lu12i.w $a0, 15
; LA32-NEXT:    ori $a0, $a0, 2064
; LA32-NEXT:    add.w $sp, $sp, $a0
; LA32-NEXT:    ld.w $s8, $sp, 2020 # 4-byte Folded Reload
; LA32-NEXT:    ld.w $fp, $sp, 2024 # 4-byte Folded Reload
; LA32-NEXT:    ld.w $ra, $sp, 2028 # 4-byte Folded Reload
; LA32-NEXT:    addi.w $sp, $sp, 2032
; LA32-NEXT:    ret
  %a = alloca i32, i32 4096, align 32768
  %b = getelementptr inbounds i32, ptr %a, i64 %i
  store volatile i32 1, ptr %b
  %1 = zext i32 %vla_size to i64
  %vla = alloca i8, i64 %1, align 2048
  %2 = load volatile i8, ptr %vla, align 2048
  ret void
}

attributes #0 = { "probe-stack"="inline-asm" }