mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-26 04:32:44 +01:00
9779454568
This reverts, "r213024 - Revert r212572 "improve BasicAA CS-CS queries", it causes PR20303." with a fix for the bug in pr20303. As it turned out, the relevant code was both wrong and over-conservative (because, as with the code it replaced, it would return the overall ModRef mask even if just Ref had been implied by the argument aliasing results). Hopefully, this correctly fixes both problems. Thanks to Nick Lewycky for reducing the test case for pr20303 (which I've cleaned up a little and added in DSE's test directory). The BasicAA test has also been updated to check for this error. Original commit message: BasicAA contains knowledge of certain intrinsics, such as memcpy and memset, and uses that information to form more-accurate answers to CallSite vs. Loc ModRef queries. Unfortunately, it did not use this information when answering CallSite vs. CallSite queries. Generically, when an intrinsic takes one or more pointers and the intrinsic is marked only to read/write from its arguments, the offset/size is unknown. As a result, the generic code that answers CallSite vs. CallSite (and CallSite vs. Loc) queries in AA uses UnknownSize when forming Locs from an intrinsic's arguments. While BasicAA's CallSite vs. Loc override could use more-accurate size information for some intrinsics, it did not do the same for CallSite vs. CallSite queries. This change refactors the intrinsic-specific logic in BasicAA into a generic AA query function: getArgLocation, which is overridden by BasicAA to supply the intrinsic-specific knowledge, and used by AA's generic implementation. This allows the intrinsic-specific knowledge to be used by both CallSite vs. Loc and CallSite vs. CallSite queries, and simplifies the BasicAA implementation. Currently, only one function, Mac's memset_pattern16, is handled by BasicAA (all the rest are intrinsics). As a side-effect of this refactoring, BasicAA's getModRefBehavior override now also returns OnlyAccessesArgumentPointees for this function (which is an improvement). llvm-svn: 213219
75 lines
3.5 KiB
LLVM
75 lines
3.5 KiB
LLVM
; RUN: opt -basicaa -dse -S < %s | FileCheck %s
|
|
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
|
|
target triple = "x86_64-unknown-linux-gnu"
|
|
|
|
%class.basic_string = type { %"class.__gnu_cxx::__versa_string" }
|
|
%"class.__gnu_cxx::__versa_string" = type { %"class.__gnu_cxx::__sso_string_base" }
|
|
%"class.__gnu_cxx::__sso_string_base" = type { %"struct.__gnu_cxx::__vstring_utility<char, std::char_traits<char>, std::allocator<char> >::_Alloc_hider", i64, %union.anon }
|
|
%"struct.__gnu_cxx::__vstring_utility<char, std::char_traits<char>, std::allocator<char> >::_Alloc_hider" = type { i8* }
|
|
%union.anon = type { i64, [8 x i8] }
|
|
|
|
; Function Attrs: nounwind
|
|
declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) #0
|
|
|
|
; Function Attrs: noinline nounwind readonly uwtable
|
|
declare zeroext i1 @callee_takes_string(%class.basic_string* nonnull) #1 align 2
|
|
|
|
; Function Attrs: nounwind uwtable
|
|
define weak_odr zeroext i1 @test() #2 align 2 {
|
|
|
|
; CHECK-LABEL: @test
|
|
|
|
bb:
|
|
%tmp = alloca %class.basic_string, align 8
|
|
%tmp1 = alloca %class.basic_string, align 8
|
|
%tmp3 = getelementptr inbounds %class.basic_string* %tmp, i64 0, i32 0, i32 0, i32 2
|
|
%tmp4 = bitcast %union.anon* %tmp3 to i8*
|
|
%tmp5 = getelementptr inbounds %class.basic_string* %tmp, i64 0, i32 0, i32 0, i32 0, i32 0
|
|
%tmp6 = getelementptr inbounds %class.basic_string* %tmp, i64 0, i32 0, i32 0, i32 1
|
|
%tmp7 = getelementptr inbounds i8* %tmp4, i64 1
|
|
%tmp8 = bitcast %class.basic_string* %tmp to i8*
|
|
%tmp9 = bitcast i64 0 to i64
|
|
%tmp10 = getelementptr inbounds %class.basic_string* %tmp1, i64 0, i32 0, i32 0, i32 2
|
|
%tmp11 = bitcast %union.anon* %tmp10 to i8*
|
|
%tmp12 = getelementptr inbounds %class.basic_string* %tmp1, i64 0, i32 0, i32 0, i32 0, i32 0
|
|
%tmp13 = getelementptr inbounds %class.basic_string* %tmp1, i64 0, i32 0, i32 0, i32 1
|
|
%tmp14 = getelementptr inbounds i8* %tmp11, i64 1
|
|
%tmp15 = bitcast %class.basic_string* %tmp1 to i8*
|
|
br label %_ZN12basic_stringIcSt11char_traitsIcESaIcEEC2EPKcRKS2_.exit
|
|
|
|
_ZN12basic_stringIcSt11char_traitsIcESaIcEEC2EPKcRKS2_.exit: ; preds = %bb
|
|
store i8* %tmp4, i8** %tmp5, align 8
|
|
store i8 62, i8* %tmp4, align 8
|
|
store i64 1, i64* %tmp6, align 8
|
|
store i8 0, i8* %tmp7, align 1
|
|
%tmp16 = call zeroext i1 @callee_takes_string(%class.basic_string* nonnull %tmp)
|
|
br label %_ZN9__gnu_cxx17__sso_string_baseIcSt11char_traitsIcESaIcEED2Ev.exit3
|
|
|
|
_ZN9__gnu_cxx17__sso_string_baseIcSt11char_traitsIcESaIcEED2Ev.exit3: ; preds = %_ZN12basic_stringIcSt11char_traitsIcESaIcEEC2EPKcRKS2_.exit
|
|
|
|
; CHECK: _ZN9__gnu_cxx17__sso_string_baseIcSt11char_traitsIcESaIcEED2Ev.exit3:
|
|
|
|
; The following can be read through the call %tmp17:
|
|
store i8* %tmp11, i8** %tmp12, align 8
|
|
store i8 125, i8* %tmp11, align 8
|
|
store i64 1, i64* %tmp13, align 8
|
|
store i8 0, i8* %tmp14, align 1
|
|
|
|
; CHECK: store i8* %tmp11, i8** %tmp12, align 8
|
|
; CHECK: store i8 125, i8* %tmp11, align 8
|
|
; CHECK: store i64 1, i64* %tmp13, align 8
|
|
; CHECK: store i8 0, i8* %tmp14, align 1
|
|
|
|
%tmp17 = call zeroext i1 @callee_takes_string(%class.basic_string* nonnull %tmp1)
|
|
call void @llvm.memset.p0i8.i64(i8* %tmp11, i8 -51, i64 16, i32 8, i1 false) #0
|
|
call void @llvm.memset.p0i8.i64(i8* %tmp15, i8 -51, i64 32, i32 8, i1 false) #0
|
|
call void @llvm.memset.p0i8.i64(i8* %tmp4, i8 -51, i64 16, i32 8, i1 false) #0
|
|
call void @llvm.memset.p0i8.i64(i8* %tmp8, i8 -51, i64 32, i32 8, i1 false) #0
|
|
ret i1 %tmp17
|
|
}
|
|
|
|
attributes #0 = { nounwind }
|
|
attributes #1 = { noinline nounwind readonly uwtable }
|
|
attributes #2 = { nounwind uwtable }
|
|
|