llvm-6502/test/CodeGen/PowerPC/crsave.ll
Ulrich Weigand 33efedc048 [PowerPC] Use mtocrf when available
Just as with mfocrf, it is also preferable to use mtocrf instead of
mtcrf when only a single CR register is to be written.

Current code however always emits mtcrf.  This probably does not matter
when using an external assembler, since the GNU assembler will in fact
automatically replace mtcrf with mtocrf when possible.  It does create
inefficient code with the integrated assembler, however.

To fix this, this patch adds MTOCRF/MTOCRF8 instruction patterns and
uses those instead of MTCRF/MTCRF8 everything.  Just as done in the
MFOCRF patch committed as 185556, these patterns will be converted
back to MTCRF if MTOCRF is not available on the machine.

As a side effect, this allows to modify the MTCRF pattern to accept
the full range of mask operands for the benefit of the asm parser.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@185561 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-03 17:59:07 +00:00

63 lines
1.7 KiB
LLVM

; RUN: llc -O0 -disable-fp-elim -mtriple=powerpc-unknown-linux-gnu -mcpu=g5 < %s | FileCheck %s -check-prefix=PPC32
; RUN: llc -O0 -mtriple=powerpc64-unknown-linux-gnu -mcpu=g5 < %s | FileCheck %s -check-prefix=PPC64
declare void @foo()
define i32 @test_cr2() nounwind uwtable {
entry:
%ret = alloca i32, align 4
%0 = call i32 asm sideeffect "\0A\09mtcr $4\0A\09cmp 2,$2,$1\0A\09mfcr $0", "=r,r,r,r,r,~{cr2}"(i32 1, i32 2, i32 3, i32 0) nounwind
store i32 %0, i32* %ret, align 4
call void @foo()
%1 = load i32* %ret, align 4
ret i32 %1
}
; PPC32: stw 31, -4(1)
; PPC32: stwu 1, -32(1)
; PPC32: mfcr 12
; PPC32-NEXT: stw 12, 24(31)
; PPC32: lwz 12, 24(31)
; PPC32-NEXT: mtocrf 32, 12
; PPC64: .cfi_startproc
; PPC64: mfcr 12
; PPC64: stw 12, 8(1)
; PPC64: stdu 1, -[[AMT:[0-9]+]](1)
; PPC64: .cfi_def_cfa_offset 128
; PPC64: .cfi_offset lr, 16
; PPC64: .cfi_offset cr2, 8
; PPC64: addi 1, 1, [[AMT]]
; PPC64: lwz 12, 8(1)
; PPC64: mtocrf 32, 12
; PPC64: .cfi_endproc
define i32 @test_cr234() nounwind {
entry:
%ret = alloca i32, align 4
%0 = call i32 asm sideeffect "\0A\09mtcr $4\0A\09cmp 2,$2,$1\0A\09cmp 3,$2,$2\0A\09cmp 4,$2,$3\0A\09mfcr $0", "=r,r,r,r,r,~{cr2},~{cr3},~{cr4}"(i32 1, i32 2, i32 3, i32 0) nounwind
store i32 %0, i32* %ret, align 4
call void @foo()
%1 = load i32* %ret, align 4
ret i32 %1
}
; PPC32: stw 31, -4(1)
; PPC32: stwu 1, -32(1)
; PPC32: mfcr 12
; PPC32-NEXT: stw 12, 24(31)
; PPC32: lwz 12, 24(31)
; PPC32-NEXT: mtocrf 32, 12
; PPC32-NEXT: mtocrf 16, 12
; PPC32-NEXT: mtocrf 8, 12
; PPC64: mfcr 12
; PPC64: stw 12, 8(1)
; PPC64: stdu 1, -[[AMT:[0-9]+]](1)
; PPC64: addi 1, 1, [[AMT]]
; PPC64: lwz 12, 8(1)
; PPC64: mtocrf 32, 12
; PPC64: mtocrf 16, 12
; PPC64: mtocrf 8, 12