mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-12-15 20:29:48 +00:00
11729224bf
This is a preliminary patch for fast instruction selection on PowerPC. Code generation can differ between DAG isel and fast isel. Existing tests that specify -O0 were written to expect DAG isel. Make this explicit by adding -fast-isel=false to the tests. In some cases specifying -fast-isel=false produces different code even when there isn't a fast instruction selector specified. This is because TM.Options.EnableFastISel = 1 at -O0 whether or not a FastISel object exists. Thus disabling fast isel can actually produce less conservative code. Because of this, some of the expected code generation in the -O0 tests needs to be adjusted. In particular, handling of function arguments is less conservative with -fast-isel=false (see isOnlyUsedInEntryBlock() in SelectionDAGBuilder.cpp). This results in fewer stack accesses and, in some cases, reduced stack size as uselessly loaded values are no longer stored back to spill locations in the stack. No functional change with this patch; test case adjustments only. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@183939 91177308-0d34-0410-b5e6-96231b3b80d8
50 lines
1.9 KiB
LLVM
50 lines
1.9 KiB
LLVM
; RUN: llc -mcpu=pwr7 -O0 -fast-isel=false < %s | FileCheck %s
|
|
|
|
; This tests correct handling of empty aggregate parameters and return values.
|
|
; An empty parameter passed by value does not consume a protocol register or
|
|
; a parameter save area doubleword. An empty parameter passed by reference
|
|
; is treated as any other pointer parameter. An empty aggregate return value
|
|
; is treated as any other aggregate return value, passed via address as a
|
|
; hidden parameter in GPR3. In this example, GPR3 contains the return value
|
|
; address, GPR4 contains the address of e2, and e1 and e3 are not passed or
|
|
; received.
|
|
|
|
target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
|
|
target triple = "powerpc64-unknown-linux-gnu"
|
|
|
|
%struct.empty = type {}
|
|
|
|
define void @callee(%struct.empty* noalias sret %agg.result, %struct.empty* byval %a1, %struct.empty* %a2, %struct.empty* byval %a3) nounwind {
|
|
entry:
|
|
%a2.addr = alloca %struct.empty*, align 8
|
|
store %struct.empty* %a2, %struct.empty** %a2.addr, align 8
|
|
%0 = load %struct.empty** %a2.addr, align 8
|
|
%1 = bitcast %struct.empty* %agg.result to i8*
|
|
%2 = bitcast %struct.empty* %0 to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* %2, i64 0, i32 1, i1 false)
|
|
ret void
|
|
}
|
|
|
|
; CHECK: callee:
|
|
; CHECK: std 4,
|
|
; CHECK-NOT: std 5,
|
|
; CHECK-NOT: std 6,
|
|
; CHECK: blr
|
|
|
|
declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind
|
|
|
|
define void @caller(%struct.empty* noalias sret %agg.result) nounwind {
|
|
entry:
|
|
%e1 = alloca %struct.empty, align 1
|
|
%e2 = alloca %struct.empty, align 1
|
|
%e3 = alloca %struct.empty, align 1
|
|
call void @callee(%struct.empty* sret %agg.result, %struct.empty* byval %e1, %struct.empty* %e2, %struct.empty* byval %e3)
|
|
ret void
|
|
}
|
|
|
|
; CHECK: caller:
|
|
; CHECK: addi 4,
|
|
; CHECK-NOT: std 5,
|
|
; CHECK-NOT: std 6,
|
|
; CHECK: bl callee
|