From 980daea857719b3eeb7dc88fe013049d361d955b Mon Sep 17 00:00:00 2001 From: Jakob Stoklund Olesen Date: Mon, 3 Aug 2009 20:08:18 +0000 Subject: Fix Bug 4657: register scavenger asserts with subreg lowering When LowerSubregsInstructionPass::LowerInsert eliminates an INSERT_SUBREG instriction because it is an identity copy, make sure that the same registers are alive before and after the elimination. When the super-register is marked this requires inserting an IMPLICIT_DEF instruction to make sure the super register is live. Fix a related bug where a kill flag on the inserted sub-register was not transferred properly. Finally, clear the undef flag in MachineInstr::addRegisterKilled. Undef implies dead and kill implies live, so they cant both be valid. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77989 91177308-0d34-0410-b5e6-96231b3b80d8 --- .../ARM/2009-08-02-RegScavengerAssert-Neon.ll | 29 ++++++++++++++++++++++ 1 file changed, 29 insertions(+) create mode 100644 test/CodeGen/ARM/2009-08-02-RegScavengerAssert-Neon.ll (limited to 'test/CodeGen/ARM/2009-08-02-RegScavengerAssert-Neon.ll') diff --git a/test/CodeGen/ARM/2009-08-02-RegScavengerAssert-Neon.ll b/test/CodeGen/ARM/2009-08-02-RegScavengerAssert-Neon.ll new file mode 100644 index 0000000..bdf5131 --- /dev/null +++ b/test/CodeGen/ARM/2009-08-02-RegScavengerAssert-Neon.ll @@ -0,0 +1,29 @@ +; RUN: llvm-as < %s | llc -march=arm -mattr=+neon +; PR4657 + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:64" +target triple = "armv7-apple-darwin9" + +define arm_apcscc <4 x i32> @scale(<4 x i32> %v, i32 %f) nounwind { +entry: + %v_addr = alloca <4 x i32> ; <<4 x i32>*> [#uses=2] + %f_addr = alloca i32 ; [#uses=2] + %retval = alloca <4 x i32> ; <<4 x i32>*> [#uses=2] + %0 = alloca <4 x i32> ; <<4 x i32>*> [#uses=2] + %"alloca point" = bitcast i32 0 to i32 ; [#uses=0] + store <4 x i32> %v, <4 x i32>* %v_addr + store i32 %f, i32* %f_addr + %1 = load <4 x i32>* %v_addr, align 16 ; <<4 x i32>> [#uses=1] + %2 = load i32* %f_addr, align 4 ; [#uses=1] + %3 = insertelement <4 x i32> undef, i32 %2, i32 0 ; <<4 x i32>> [#uses=1] + %4 = shufflevector <4 x i32> %3, <4 x i32> undef, <4 x i32> zeroinitializer ; <<4 x i32>> [#uses=1] + %5 = mul <4 x i32> %1, %4 ; <<4 x i32>> [#uses=1] + store <4 x i32> %5, <4 x i32>* %0, align 16 + %6 = load <4 x i32>* %0, align 16 ; <<4 x i32>> [#uses=1] + store <4 x i32> %6, <4 x i32>* %retval, align 16 + br label %return + +return: ; preds = %entry + %retval1 = load <4 x i32>* %retval ; <<4 x i32>> [#uses=1] + ret <4 x i32> %retval1 +} -- cgit v1.1