; Test spilling using MVC. The tests here assume z10 register pressure,
; without the high words being available.
;
; RUN: llc < %s -mtriple=s390x-linux-gnu -mcpu=z10 -verify-machineinstrs | FileCheck %s
declare void @foo()
@g0 = dso_local global i32 0
@g1 = dso_local global i32 1
@g2 = dso_local global i32 2
@g3 = dso_local global i32 3
@g4 = dso_local global i32 4
@g5 = dso_local global i32 5
@g6 = dso_local global i32 6
@g7 = dso_local global i32 7
@g8 = dso_local global i32 8
@g9 = dso_local global i32 9
@h0 = dso_local global i64 0
@h1 = dso_local global i64 1
@h2 = dso_local global i64 2
@h3 = dso_local global i64 3
@h4 = dso_local global i64 4
@h5 = dso_local global i64 5
@h6 = dso_local global i64 6
@h7 = dso_local global i64 7
@h8 = dso_local global i64 8
@h9 = dso_local global i64 9
; This function shouldn't spill anything
define dso_local void @f1(ptr %ptr0) {
; CHECK-LABEL: f1:
; CHECK: stmg
; CHECK: aghi %r15, -160
; CHECK-NOT: %r15
; CHECK: brasl %r14, foo@PLT
; CHECK-NOT: %r15
; CHECK: lmg
; CHECK: br %r14
%ptr1 = getelementptr i32, ptr %ptr0, i32 2
%ptr2 = getelementptr i32, ptr %ptr0, i32 4
%ptr3 = getelementptr i32, ptr %ptr0, i32 6
%ptr4 = getelementptr i32, ptr %ptr0, i32 8
%ptr5 = getelementptr i32, ptr %ptr0, i32 10
%ptr6 = getelementptr i32, ptr %ptr0, i32 12
%val0 = load i32, ptr %ptr0
%val1 = load i32, ptr %ptr1
%val2 = load i32, ptr %ptr2
%val3 = load i32, ptr %ptr3
%val4 = load i32, ptr %ptr4
%val5 = load i32, ptr %ptr5
%val6 = load i32, ptr %ptr6
call void @foo()
store i32 %val0, ptr %ptr0
store i32 %val1, ptr %ptr1
store i32 %val2, ptr %ptr2
store i32 %val3, ptr %ptr3
store i32 %val4, ptr %ptr4
store i32 %val5, ptr %ptr5
store i32 %val6, ptr %ptr6
ret void
}
; Test a case where at least one i32 load and at least one i32 store
; need spills.
define dso_local void @f2(ptr %ptr0) {
; CHECK-LABEL: f2:
; CHECK: mvc [[OFFSET1:16[04]]](4,%r15), [[OFFSET2:[0-9]+]]({{%r[0-9]+}})
; CHECK: brasl %r14, foo@PLT
; CHECK: mvc [[OFFSET2]](4,{{%r[0-9]+}}), [[OFFSET1]](%r15)
; CHECK: br %r14
%ptr1 = getelementptr i32, ptr %ptr0, i64 2
%ptr2 = getelementptr i32, ptr %ptr0, i64 4
%ptr3 = getelementptr i32, ptr %ptr0, i64 6
%ptr4 = getelementptr i32, ptr %ptr0, i64 8
%ptr5 = getelementptr i32, ptr %ptr0, i64 10
%ptr6 = getelementptr i32, ptr %ptr0, i64 12
%ptr7 = getelementptr i32, ptr %ptr0, i64 14
%ptr8 = getelementptr i32, ptr %ptr0, i64 16
%val0 = load i32, ptr %ptr0
%val1 = load i32, ptr %ptr1
%val2 = load i32, ptr %ptr2
%val3 = load i32, ptr %ptr3
%val4 = load i32, ptr %ptr4
%val5 = load i32, ptr %ptr5
%val6 = load i32, ptr %ptr6
%val7 = load i32, ptr %ptr7
%val8 = load i32, ptr %ptr8
call void @foo()
store i32 %val0, ptr %ptr0
store i32 %val1, ptr %ptr1
store i32 %val2, ptr %ptr2
store i32 %val3, ptr %ptr3
store i32 %val4, ptr %ptr4
store i32 %val5, ptr %ptr5
store i32 %val6, ptr %ptr6
store i32 %val7, ptr %ptr7
store i32 %val8, ptr %ptr8
ret void
}
; Test a case where at least one i64 load and at least one i64 store
; need spills.
define dso_local void @f3(ptr %ptr0) {
; CHECK-LABEL: f3:
; CHECK: mvc 160(8,%r15), [[OFFSET:[0-9]+]]({{%r[0-9]+}})
; CHECK: brasl %r14, foo@PLT
; CHECK: mvc [[OFFSET]](8,{{%r[0-9]+}}), 160(%r15)
; CHECK: br %r14
%ptr1 = getelementptr i64, ptr %ptr0, i64 2
%ptr2 = getelementptr i64, ptr %ptr0, i64 4
%ptr3 = getelementptr i64, ptr %ptr0, i64 6
%ptr4 = getelementptr i64, ptr %ptr0, i64 8
%ptr5 = getelementptr i64, ptr %ptr0, i64 10
%ptr6 = getelementptr i64, ptr %ptr0, i64 12
%ptr7 = getelementptr i64, ptr %ptr0, i64 14
%ptr8 = getelementptr i64, ptr %ptr0, i64 16
%val0 = load i64, ptr %ptr0
%val1 = load i64, ptr %ptr1
%val2 = load i64, ptr %ptr2
%val3 = load i64, ptr %ptr3
%val4 = load i64, ptr %ptr4
%val5 = load i64, ptr %ptr5
%val6 = load i64, ptr %ptr6
%val7 = load i64, ptr %ptr7
%val8 = load i64, ptr %ptr8
call void @foo()
store i64 %val0, ptr %ptr0
store i64 %val1, ptr %ptr1
store i64 %val2, ptr %ptr2
store i64 %val3, ptr %ptr3
store i64 %val4, ptr %ptr4
store i64 %val5, ptr %ptr5
store i64 %val6, ptr %ptr6
store i64 %val7, ptr %ptr7
store i64 %val8, ptr %ptr8
ret void
}
; Test a case where at least at least one f32 load and at least one f32 store
; need spills. The 8 call-saved FPRs could be used for 8 of the %vals
; (and are at the time of writing), but it would really be better to use
; MVC for all 10.
define dso_local void @f4(ptr %ptr0) {
; CHECK-LABEL: f4:
; CHECK: mvc [[OFFSET1:16[04]]](4,%r15), [[OFFSET2:[0-9]+]]({{%r[0-9]+}})
; CHECK: brasl %r14, foo@PLT
; CHECK: mvc [[OFFSET2]](4,{{%r[0-9]+}}), [[OFFSET1]](%r15)
; CHECK: br %r14
%ptr1 = getelementptr float, ptr %ptr0, i64 2
%ptr2 = getelementptr float, ptr %ptr0, i64 4
%ptr3 = getelementptr float, ptr %ptr0, i64 6
%ptr4 = getelementptr float, ptr %ptr0, i64 8
%ptr5 = getelementptr float, ptr %ptr0, i64 10
%ptr6 = getelementptr float, ptr %ptr0, i64 12
%ptr7 = getelementptr float, ptr %ptr0, i64 14
%ptr8 = getelementptr float, ptr %ptr0, i64 16
%ptr9 = getelementptr float, ptr %ptr0, i64 18
%val0 = load float, ptr %ptr0
%val1 = load float, ptr %ptr1
%val2 = load float, ptr %ptr2
%val3 = load float, ptr %ptr3
%val4 = load float, ptr %ptr4
%val5 = load float, ptr %ptr5
%val6 = load float, ptr %ptr6
%val7 = load float, ptr %ptr7
%val8 = load float, ptr %ptr8
%val9 = load float, ptr %ptr9
call void @foo()
store float %val0, ptr %ptr0
store float %val1, ptr %ptr1
store float %val2, ptr %ptr2
store float %val3, ptr %ptr3
store float %val4, ptr %ptr4
store float %val5, ptr %ptr5
store float %val6, ptr %ptr6
store float %val7, ptr %ptr7
store float %val8, ptr %ptr8
store float %val9, ptr %ptr9
ret void
}
; Similarly for f64.
define dso_local void @f5(ptr %ptr0) {
; CHECK-LABEL: f5:
; CHECK: mvc 160(8,%r15), [[OFFSET:[0-9]+]]({{%r[0-9]+}})
; CHECK: brasl %r14, foo@PLT
; CHECK: mvc [[OFFSET]](8,{{%r[0-9]+}}), 160(%r15)
; CHECK: br %r14
%ptr1 = getelementptr double, ptr %ptr0, i64 2
%ptr2 = getelementptr double, ptr %ptr0, i64 4
%ptr3 = getelementptr double, ptr %ptr0, i64 6
%ptr4 = getelementptr double, ptr %ptr0, i64 8
%ptr5 = getelementptr double, ptr %ptr0, i64 10
%ptr6 = getelementptr double, ptr %ptr0, i64 12
%ptr7 = getelementptr double, ptr %ptr0, i64 14
%ptr8 = getelementptr double, ptr %ptr0, i64 16
%ptr9 = getelementptr double, ptr %ptr0, i64 18
%val0 = load double, ptr %ptr0
%val1 = load double, ptr %ptr1
%val2 = load double, ptr %ptr2
%val3 = load double, ptr %ptr3
%val4 = load double, ptr %ptr4
%val5 = load double, ptr %ptr5
%val6 = load double, ptr %ptr6
%val7 = load double, ptr %ptr7
%val8 = load double, ptr %ptr8
%val9 = load double, ptr %ptr9
call void @foo()
store double %val0, ptr %ptr0
store double %val1, ptr %ptr1
store double %val2, ptr %ptr2
store double %val3, ptr %ptr3
store double %val4, ptr %ptr4
store double %val5, ptr %ptr5
store double %val6, ptr %ptr6
store double %val7, ptr %ptr7
store double %val8, ptr %ptr8
store double %val9, ptr %ptr9
ret void
}
; Repeat f2 with atomic accesses. We shouldn't use MVC here.
define dso_local void @f6(ptr %ptr0) {
; CHECK-LABEL: f6:
; CHECK-NOT: mvc
; CHECK: br %r14
%ptr1 = getelementptr i32, ptr %ptr0, i64 2
%ptr2 = getelementptr i32, ptr %ptr0, i64 4
%ptr3 = getelementptr i32, ptr %ptr0, i64 6
%ptr4 = getelementptr i32, ptr %ptr0, i64 8
%ptr5 = getelementptr i32, ptr %ptr0, i64 10
%ptr6 = getelementptr i32, ptr %ptr0, i64 12
%ptr7 = getelementptr i32, ptr %ptr0, i64 14
%ptr8 = getelementptr i32, ptr %ptr0, i64 16
%val0 = load atomic i32, ptr %ptr0 unordered, align 4
%val1 = load atomic i32, ptr %ptr1 unordered, align 4
%val2 = load atomic i32, ptr %ptr2 unordered, align 4
%val3 = load atomic i32, ptr %ptr3 unordered, align 4
%val4 = load atomic i32, ptr %ptr4 unordered, align 4
%val5 = load atomic i32, ptr %ptr5 unordered, align 4
%val6 = load atomic i32, ptr %ptr6 unordered, align 4
%val7 = load atomic i32, ptr %ptr7 unordered, align 4
%val8 = load atomic i32, ptr %ptr8 unordered, align 4
call void @foo()
store atomic i32 %val0, ptr %ptr0 unordered, align 4
store atomic i32 %val1, ptr %ptr1 unordered, align 4
store atomic i32 %val2, ptr %ptr2 unordered, align 4
store atomic i32 %val3, ptr %ptr3 unordered, align 4
store atomic i32 %val4, ptr %ptr4 unordered, align 4
store atomic i32 %val5, ptr %ptr5 unordered, align 4
store atomic i32 %val6, ptr %ptr6 unordered, align 4
store atomic i32 %val7, ptr %ptr7 unordered, align 4
store atomic i32 %val8, ptr %ptr8 unordered, align 4
ret void
}
; ...likewise volatile accesses.
define dso_local void @f7(ptr %ptr0) {
; CHECK-LABEL: f7:
; CHECK-NOT: mvc
; CHECK: br %r14
%ptr1 = getelementptr i32, ptr %ptr0, i64 2
%ptr2 = getelementptr i32, ptr %ptr0, i64 4
%ptr3 = getelementptr i32, ptr %ptr0, i64 6
%ptr4 = getelementptr i32, ptr %ptr0, i64 8
%ptr5 = getelementptr i32, ptr %ptr0, i64 10
%ptr6 = getelementptr i32, ptr %ptr0, i64 12
%ptr7 = getelementptr i32, ptr %ptr0, i64 14
%ptr8 = getelementptr i32, ptr %ptr0, i64 16
%val0 = load volatile i32, ptr %ptr0
%val1 = load volatile i32, ptr %ptr1
%val2 = load volatile i32, ptr %ptr2
%val3 = load volatile i32, ptr %ptr3
%val4 = load volatile i32, ptr %ptr4
%val5 = load volatile i32, ptr %ptr5
%val6 = load volatile i32, ptr %ptr6
%val7 = load volatile i32, ptr %ptr7
%val8 = load volatile i32, ptr %ptr8
call void @foo()
store volatile i32 %val0, ptr %ptr0
store volatile i32 %val1, ptr %ptr1
store volatile i32 %val2, ptr %ptr2
store volatile i32 %val3, ptr %ptr3
store volatile i32 %val4, ptr %ptr4
store volatile i32 %val5, ptr %ptr5
store volatile i32 %val6, ptr %ptr6
store volatile i32 %val7, ptr %ptr7
store volatile i32 %val8, ptr %ptr8
ret void
}
; Check that LRL and STRL are not converted.
define dso_local void @f8() {
; CHECK-LABEL: f8:
; CHECK-NOT: mvc
; CHECK: br %r14
%val0 = load i32, ptr@g0
%val1 = load i32, ptr@g1
%val2 = load i32, ptr@g2
%val3 = load i32, ptr@g3
%val4 = load i32, ptr@g4
%val5 = load i32, ptr@g5
%val6 = load i32, ptr@g6
%val7 = load i32, ptr@g7
%val8 = load i32, ptr@g8
%val9 = load i32, ptr@g9
call void @foo()
store i32 %val0, ptr@g0
store i32 %val1, ptr@g1
store i32 %val2, ptr@g2
store i32 %val3, ptr@g3
store i32 %val4, ptr@g4
store i32 %val5, ptr@g5
store i32 %val6, ptr@g6
store i32 %val7, ptr@g7
store i32 %val8, ptr@g8
store i32 %val9, ptr@g9
ret void
}
; Likewise LGRL and STGRL.
define dso_local void @f9() {
; CHECK-LABEL: f9:
; CHECK-NOT: mvc
; CHECK: br %r14
%val0 = load i64, ptr@h0
%val1 = load i64, ptr@h1
%val2 = load i64, ptr@h2
%val3 = load i64, ptr@h3
%val4 = load i64, ptr@h4
%val5 = load i64, ptr@h5
%val6 = load i64, ptr@h6
%val7 = load i64, ptr@h7
%val8 = load i64, ptr@h8
%val9 = load i64, ptr@h9
call void @foo()
store i64 %val0, ptr@h0
store i64 %val1, ptr@h1
store i64 %val2, ptr@h2
store i64 %val3, ptr@h3
store i64 %val4, ptr@h4
store i64 %val5, ptr@h5
store i64 %val6, ptr@h6
store i64 %val7, ptr@h7
store i64 %val8, ptr@h8
store i64 %val9, ptr@h9
ret void
}
; This showed a problem with the way stack coloring updated instructions.
; The copy from %val9 to %newval8 can be done using an MVC, which then
; has two frame index operands. Stack coloring chose a valid renumbering
; [FI0, FI1] -> [FI1, FI2], but applied it in the form FI0 -> FI1 -> FI2,
; so that both operands ended up being the same.
define dso_local void @f10() {
; CHECK-LABEL: f10:
; CHECK: lgrl [[REG:%r[0-9]+]], h9
; CHECK: stg [[REG]], [[VAL9:[0-9]+]](%r15)
; CHECK: brasl %r14, foo@PLT
; CHECK: brasl %r14, foo@PLT
; CHECK: mvc [[NEWVAL8:[0-9]+]](8,%r15), [[VAL9]](%r15)
; CHECK: brasl %r14, foo@PLT
; CHECK: lg [[REG:%r[0-9]+]], [[NEWVAL8]](%r15)
; CHECK: stgrl [[REG]], h8
; CHECK: br %r14
entry:
%val8 = load volatile i64, ptr@h8
%val0 = load volatile i64, ptr@h0
%val1 = load volatile i64, ptr@h1
%val2 = load volatile i64, ptr@h2
%val3 = load volatile i64, ptr@h3
%val4 = load volatile i64, ptr@h4
%val5 = load volatile i64, ptr@h5
%val6 = load volatile i64, ptr@h6
%val7 = load volatile i64, ptr@h7
%val9 = load volatile i64, ptr@h9
call void @foo()
store volatile i64 %val0, ptr@h0
store volatile i64 %val1, ptr@h1
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
%check = load volatile i64, ptr@h0
%cond = icmp eq i64 %check, 0
br i1 %cond, label %skip, label %fallthru
fallthru:
call void @foo()
store volatile i64 %val0, ptr@h0
store volatile i64 %val1, ptr@h1
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
store volatile i64 %val8, ptr@h8
br label %skip
skip:
%newval8 = phi i64 [ %val8, %entry ], [ %val9, %fallthru ]
call void @foo()
store volatile i64 %val0, ptr@h0
store volatile i64 %val1, ptr@h1
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
store volatile i64 %newval8, ptr@h8
store volatile i64 %val9, ptr@h9
ret void
}
; This used to generate a no-op MVC. It is very sensitive to spill heuristics.
define dso_local void @f11() {
; CHECK-LABEL: f11:
; CHECK-NOT: mvc [[OFFSET:[0-9]+]](8,%r15), [[OFFSET]](%r15)
; CHECK: br %r14
entry:
%val0 = load volatile i64, ptr@h0
%val1 = load volatile i64, ptr@h1
%val2 = load volatile i64, ptr@h2
%val3 = load volatile i64, ptr@h3
%val4 = load volatile i64, ptr@h4
%val5 = load volatile i64, ptr@h5
%val6 = load volatile i64, ptr@h6
%val7 = load volatile i64, ptr@h7
%altval0 = load volatile i64, ptr@h0
%altval1 = load volatile i64, ptr@h1
call void @foo()
store volatile i64 %val0, ptr@h0
store volatile i64 %val1, ptr@h1
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
%check = load volatile i64, ptr@h0
%cond = icmp eq i64 %check, 0
br i1 %cond, label %a1, label %b1
a1:
call void @foo()
br label %join1
b1:
call void @foo()
br label %join1
join1:
%newval0 = phi i64 [ %val0, %a1 ], [ %altval0, %b1 ]
call void @foo()
store volatile i64 %val1, ptr@h1
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
br i1 %cond, label %a2, label %b2
a2:
call void @foo()
br label %join2
b2:
call void @foo()
br label %join2
join2:
%newval1 = phi i64 [ %val1, %a2 ], [ %altval1, %b2 ]
call void @foo()
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
call void @foo()
store volatile i64 %newval0, ptr@h0
store volatile i64 %newval1, ptr@h1
store volatile i64 %val2, ptr@h2
store volatile i64 %val3, ptr@h3
store volatile i64 %val4, ptr@h4
store volatile i64 %val5, ptr@h5
store volatile i64 %val6, ptr@h6
store volatile i64 %val7, ptr@h7
ret void
}
Codebase Browser
llvm