|  | ; Test 128-bit addition in which the second operand is a zero-extended i32. | 
|  | ; | 
|  | ; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s | 
|  |  | 
|  | ; Check register additions.  The XOR ensures that we don't instead zero-extend | 
|  | ; %b into a register and use memory addition. | 
|  | define void @f1(ptr %aptr, i32 %b) { | 
|  | ; CHECK-LABEL: f1: | 
|  | ; CHECK: slgfr {{%r[0-5]}}, %r3 | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Like f1, but using an "in-register" extension. | 
|  | define void @f2(ptr %aptr, i64 %b) { | 
|  | ; CHECK-LABEL: f2: | 
|  | ; CHECK: slgfr {{%r[0-5]}}, %r3 | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %trunc = trunc i64 %b to i32 | 
|  | %bext = zext i32 %trunc to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Test register addition in cases where the second operand is zero extended | 
|  | ; from i64 rather than i32, but is later masked to i32 range. | 
|  | define void @f3(ptr %aptr, i64 %b) { | 
|  | ; CHECK-LABEL: f3: | 
|  | ; CHECK: slgfr {{%r[0-5]}}, %r3 | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %bext = zext i64 %b to i128 | 
|  | %and = and i128 %bext, 4294967295 | 
|  | %sub = sub i128 %xor, %and | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Test SLGF with no offset. | 
|  | define void @f4(ptr %aptr, ptr %bsrc) { | 
|  | ; CHECK-LABEL: f4: | 
|  | ; CHECK: slgf {{%r[0-5]}}, 0(%r3) | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %b = load i32, ptr %bsrc | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Check the high end of the SLGF range. | 
|  | define void @f5(ptr %aptr, ptr %bsrc) { | 
|  | ; CHECK-LABEL: f5: | 
|  | ; CHECK: slgf {{%r[0-5]}}, 524284(%r3) | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %ptr = getelementptr i32, ptr %bsrc, i64 131071 | 
|  | %b = load i32, ptr %ptr | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Check the next word up, which must use separate address logic. | 
|  | ; Other sequences besides this one would be OK. | 
|  | define void @f6(ptr %aptr, ptr %bsrc) { | 
|  | ; CHECK-LABEL: f6: | 
|  | ; CHECK: agfi %r3, 524288 | 
|  | ; CHECK: slgf {{%r[0-5]}}, 0(%r3) | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %ptr = getelementptr i32, ptr %bsrc, i64 131072 | 
|  | %b = load i32, ptr %ptr | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Check the high end of the negative aligned SLGF range. | 
|  | define void @f7(ptr %aptr, ptr %bsrc) { | 
|  | ; CHECK-LABEL: f7: | 
|  | ; CHECK: slgf {{%r[0-5]}}, -4(%r3) | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %ptr = getelementptr i32, ptr %bsrc, i128 -1 | 
|  | %b = load i32, ptr %ptr | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Check the low end of the SLGF range. | 
|  | define void @f8(ptr %aptr, ptr %bsrc) { | 
|  | ; CHECK-LABEL: f8: | 
|  | ; CHECK: slgf {{%r[0-5]}}, -524288(%r3) | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %ptr = getelementptr i32, ptr %bsrc, i128 -131072 | 
|  | %b = load i32, ptr %ptr | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Check the next word down, which needs separate address logic. | 
|  | ; Other sequences besides this one would be OK. | 
|  | define void @f9(ptr %aptr, ptr %bsrc) { | 
|  | ; CHECK-LABEL: f9: | 
|  | ; CHECK: agfi %r3, -524292 | 
|  | ; CHECK: slgf {{%r[0-5]}}, 0(%r3) | 
|  | ; CHECK: slbgr | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %ptr = getelementptr i32, ptr %bsrc, i128 -131073 | 
|  | %b = load i32, ptr %ptr | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } | 
|  |  | 
|  | ; Check that SLGF allows an index. | 
|  | define void @f10(ptr %aptr, i64 %src, i64 %index) { | 
|  | ; CHECK-LABEL: f10: | 
|  | ; CHECK: slgf {{%r[0-5]}}, 524284({{%r4,%r3|%r3,%r4}}) | 
|  | ; CHECK: br %r14 | 
|  | %a = load i128, ptr %aptr | 
|  | %xor = xor i128 %a, 127 | 
|  | %add1 = add i64 %src, %index | 
|  | %add2 = add i64 %add1, 524284 | 
|  | %ptr = inttoptr i64 %add2 to ptr | 
|  | %b = load i32, ptr %ptr | 
|  | %bext = zext i32 %b to i128 | 
|  | %sub = sub i128 %xor, %bext | 
|  | store i128 %sub, ptr %aptr | 
|  | ret void | 
|  | } |