maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/test/CodeGenCXX/dynamic-cast-hint.cpp

56 lines
3.1 KiB
C++
Raw Normal View History

CodeGen: Implement hint values for dynamic_cast as described in the Itanium C++ ABI. This can yield dramatic speedups of dynamic_cast for simple inheritance trees, at least with libsupc++. Neither libcxxabi nor libcxxrt make use of this hint currently, it was never implemented because clang didn't support it. There was some concern about the number of class hierarchy walks this change introduces. If it turns out to be an issue we can add caching either at the cast pair level or even deeper, but we also do a lot of walks in Sema so this codepath is probably fairly optimized already. llvm-svn: 174293
2013-02-04 03:59:25 +08:00
// RUN: %clang_cc1 -triple x86_64-apple-darwin12 -emit-llvm -o - %s | FileCheck %s
[Test] Make Lit tests C++11 compatible #9 [Test] Make Lit tests C++11 compatible #9 Differential Revision: https://reviews.llvm.org/D20710 llvm-svn: 296184
2017-02-25 06:22:05 +08:00
// RUN: %clang_cc1 -triple x86_64-apple-darwin12 -emit-llvm -std=c++98 -o - %s | FileCheck %s
// RUN: %clang_cc1 -triple x86_64-apple-darwin12 -emit-llvm -std=c++11 -o - %s | FileCheck %s
CodeGen: Implement hint values for dynamic_cast as described in the Itanium C++ ABI. This can yield dramatic speedups of dynamic_cast for simple inheritance trees, at least with libsupc++. Neither libcxxabi nor libcxxrt make use of this hint currently, it was never implemented because clang didn't support it. There was some concern about the number of class hierarchy walks this change introduces. If it turns out to be an issue we can add caching either at the cast pair level or even deeper, but we also do a lot of walks in Sema so this codepath is probably fairly optimized already. llvm-svn: 174293
2013-02-04 03:59:25 +08:00
[Test] Make Lit tests C++11 compatible #9 [Test] Make Lit tests C++11 compatible #9 Differential Revision: https://reviews.llvm.org/D20710 llvm-svn: 296184
2017-02-25 06:22:05 +08:00
class A { protected: virtual ~A() {} };
class B { protected: virtual ~B() {} };
CodeGen: Implement hint values for dynamic_cast as described in the Itanium C++ ABI. This can yield dramatic speedups of dynamic_cast for simple inheritance trees, at least with libsupc++. Neither libcxxabi nor libcxxrt make use of this hint currently, it was never implemented because clang didn't support it. There was some concern about the number of class hierarchy walks this change introduces. If it turns out to be an issue we can add caching either at the cast pair level or even deeper, but we also do a lot of walks in Sema so this codepath is probably fairly optimized already. llvm-svn: 174293
2013-02-04 03:59:25 +08:00
class C : A { char x; };
class D : public A { short y; };
class E : public A, public B { int z; };
class F : public virtual A { long long w; };
class G : virtual A { long long w; };
class H : public E { int a; };
class I : public F { char b; };
class J : public H { char q; };
class K : public C, public B { char q; };
class XA : public A { };
class XB : public A { };
class XC : public virtual A { };
class X : public XA, public XB, public XC { };
void test(A *a, B *b) {
volatile C *ac = dynamic_cast<C *>(a);
Drop value names from test to get -Asserts builds back to green. llvm-svn: 174294
2013-02-04 04:22:35 +08:00
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64 }* @_ZTI1C to i8*), i64 -2)
CodeGen: Implement hint values for dynamic_cast as described in the Itanium C++ ABI. This can yield dramatic speedups of dynamic_cast for simple inheritance trees, at least with libsupc++. Neither libcxxabi nor libcxxrt make use of this hint currently, it was never implemented because clang didn't support it. There was some concern about the number of class hierarchy walks this change introduces. If it turns out to be an issue we can add caching either at the cast pair level or even deeper, but we also do a lot of walks in Sema so this codepath is probably fairly optimized already. llvm-svn: 174293
2013-02-04 03:59:25 +08:00
volatile D *ad = dynamic_cast<D *>(a);
Drop value names from test to get -Asserts builds back to green. llvm-svn: 174294
2013-02-04 04:22:35 +08:00
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i8* }* @_ZTI1D to i8*), i64 0)
CodeGen: Implement hint values for dynamic_cast as described in the Itanium C++ ABI. This can yield dramatic speedups of dynamic_cast for simple inheritance trees, at least with libsupc++. Neither libcxxabi nor libcxxrt make use of this hint currently, it was never implemented because clang didn't support it. There was some concern about the number of class hierarchy walks this change introduces. If it turns out to be an issue we can add caching either at the cast pair level or even deeper, but we also do a lot of walks in Sema so this codepath is probably fairly optimized already. llvm-svn: 174293
2013-02-04 03:59:25 +08:00
volatile E *ae = dynamic_cast<E *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64, i8*, i64 }* @_ZTI1E to i8*), i64 0)
volatile F *af = dynamic_cast<F *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64 }* @_ZTI1F to i8*), i64 -1)
volatile G *ag = dynamic_cast<G *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64 }* @_ZTI1G to i8*), i64 -2)
volatile H *ah = dynamic_cast<H *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i8* }* @_ZTI1H to i8*), i64 0)
volatile I *ai = dynamic_cast<I *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i8* }* @_ZTI1I to i8*), i64 -1)
volatile J *aj = dynamic_cast<J *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i8* }* @_ZTI1J to i8*), i64 0)
volatile K *ak = dynamic_cast<K *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64, i8*, i64 }* @_ZTI1K to i8*), i64 -2)
volatile X *ax = dynamic_cast<X *>(a);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1A to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64, i8*, i64, i8*, i64 }* @_ZTI1X to i8*), i64 -1)
volatile E *be = dynamic_cast<E *>(b);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1B to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64, i8*, i64 }* @_ZTI1E to i8*), i64 8)
volatile G *bg = dynamic_cast<G *>(b);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1B to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64 }* @_ZTI1G to i8*), i64 -2)
volatile J *bj = dynamic_cast<J *>(b);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1B to i8*), i8* bitcast ({ i8*, i8*, i8* }* @_ZTI1J to i8*), i64 8)
volatile K *bk = dynamic_cast<K *>(b);
// CHECK: i8* bitcast ({ i8*, i8* }* @_ZTI1B to i8*), i8* bitcast ({ i8*, i8*, i32, i32, i8*, i64, i8*, i64 }* @_ZTI1K to i8*), i64 16)
}