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/microsoft-abi-byval-sret.cpp

71 lines
2.3 KiB
C++
Raw Normal View History

[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
// RUN: %clang_cc1 -emit-llvm %s -o - -triple=i686-pc-win32 -mconstructor-aliases -fno-rtti | FileCheck %s
struct A {
A() : a(42) {}
A(const A &o) : a(o.a) {}
~A() {}
int a;
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
};
struct B {
[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
A foo(A o);
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
A __cdecl bar(A o);
A __stdcall baz(A o);
A __fastcall qux(A o);
[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
};
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
A B::foo(A x) {
return x;
[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
}
Bring r325915 back. The tests that failed on a windows host have been fixed. Original message: Start setting dso_local for COFF. With this there are still some GVs where we don't set dso_local because setGVProperties is never called. I intend to fix that in followup commits. This is just the bare minimum to teach shouldAssumeDSOLocal what it should do for COFF. llvm-svn: 325940
2018-02-24 03:30:48 +08:00
// CHECK-LABEL: define dso_local x86_thiscallcc %struct.A* @"\01?foo@B@@QAE?AUA@@U2@@Z"
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: (%struct.B* %this, <{ %struct.A*, %struct.A }>* inalloca)
Update Clang tests to handle explicitly typed gep changes in LLVM. llvm-svn: 230783
2015-02-28 03:18:17 +08:00
// CHECK: getelementptr inbounds <{ %struct.A*, %struct.A }>, <{ %struct.A*, %struct.A }>* %{{.*}}, i32 0, i32 0
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: load %struct.A*, %struct.A**
MS ABI: Return sret parameters when using inalloca Previously the X86 backend would look for the sret attribute and handle this for us. inalloca takes that all away, so we have to do the return ourselves now. llvm-svn: 202097
2014-02-25 08:59:14 +08:00
// CHECK: ret %struct.A*
[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
A B::bar(A x) {
return x;
}
Bring r325915 back. The tests that failed on a windows host have been fixed. Original message: Start setting dso_local for COFF. With this there are still some GVs where we don't set dso_local because setGVProperties is never called. I intend to fix that in followup commits. This is just the bare minimum to teach shouldAssumeDSOLocal what it should do for COFF. llvm-svn: 325940
2018-02-24 03:30:48 +08:00
// CHECK-LABEL: define dso_local %struct.A* @"\01?bar@B@@QAA?AUA@@U2@@Z"
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: (<{ %struct.B*, %struct.A*, %struct.A }>* inalloca)
Update Clang tests to handle explicitly typed gep changes in LLVM. llvm-svn: 230783
2015-02-28 03:18:17 +08:00
// CHECK: getelementptr inbounds <{ %struct.B*, %struct.A*, %struct.A }>, <{ %struct.B*, %struct.A*, %struct.A }>* %{{.*}}, i32 0, i32 1
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: load %struct.A*, %struct.A**
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: ret %struct.A*
A B::baz(A x) {
return x;
}
Bring r325915 back. The tests that failed on a windows host have been fixed. Original message: Start setting dso_local for COFF. With this there are still some GVs where we don't set dso_local because setGVProperties is never called. I intend to fix that in followup commits. This is just the bare minimum to teach shouldAssumeDSOLocal what it should do for COFF. llvm-svn: 325940
2018-02-24 03:30:48 +08:00
// CHECK-LABEL: define dso_local x86_stdcallcc %struct.A* @"\01?baz@B@@QAG?AUA@@U2@@Z"
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: (<{ %struct.B*, %struct.A*, %struct.A }>* inalloca)
Update Clang tests to handle explicitly typed gep changes in LLVM. llvm-svn: 230783
2015-02-28 03:18:17 +08:00
// CHECK: getelementptr inbounds <{ %struct.B*, %struct.A*, %struct.A }>, <{ %struct.B*, %struct.A*, %struct.A }>* %{{.*}}, i32 0, i32 1
Update Clang tests to handle explicitly typed load changes in LLVM. llvm-svn: 230795
2015-02-28 05:19:58 +08:00
// CHECK: load %struct.A*, %struct.A**
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: ret %struct.A*
A B::qux(A x) {
return x;
}
Bring r325915 back. The tests that failed on a windows host have been fixed. Original message: Start setting dso_local for COFF. With this there are still some GVs where we don't set dso_local because setGVProperties is never called. I intend to fix that in followup commits. This is just the bare minimum to teach shouldAssumeDSOLocal what it should do for COFF. llvm-svn: 325940
2018-02-24 03:30:48 +08:00
// CHECK-LABEL: define dso_local x86_fastcallcc void @"\01?qux@B@@QAI?AUA@@U2@@Z"
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: (%struct.B* inreg %this, %struct.A* inreg noalias sret %agg.result, <{ %struct.A }>* inalloca)
// CHECK: ret void
[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
int main() {
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
B b;
A a = b.foo(A());
a = b.bar(a);
a = b.baz(a);
a = b.qux(a);
[ms-cxxabi] Use inalloca on win32 when passing non-trivial C++ objects When a non-trivial parameter is present, clang now gathers up all the parameters that lack inreg and puts them into a packed struct. MSVC always aligns each parameter to 4 bytes and no more, so this is a pretty simple struct to lay out. On win64, non-trivial records are passed indirectly. Prior to this change, clang was incorrectly using byval on win64. I'm able to self-host a working clang with this change and additional LLVM patches. Reviewers: rsmith Differential Revision: http://llvm-reviews.chandlerc.com/D2636 llvm-svn: 200597
2014-02-01 08:04:45 +08:00
}
MS ABI: Ensure 'this' is first for byval+sret methods Previously we were building up the inalloca struct in the usual pattern of return type followed by arguments. However, on Windows, 'this' always precedes the 'sret' parameter, so we need to insert it into the struct first as a special case. llvm-svn: 213990
2014-07-26 08:12:26 +08:00
// CHECK: call x86_thiscallcc %struct.A* @"\01?foo@B@@QAE?AUA@@U2@@Z"
// CHECK: (%struct.B* %{{[^,]*}}, <{ %struct.A*, %struct.A }>* inalloca %{{[^,]*}})
// CHECK: call %struct.A* @"\01?bar@B@@QAA?AUA@@U2@@Z"
// CHECK: (<{ %struct.B*, %struct.A*, %struct.A }>* inalloca %{{[^,]*}})
// CHECK: call x86_stdcallcc %struct.A* @"\01?baz@B@@QAG?AUA@@U2@@Z"
// CHECK: (<{ %struct.B*, %struct.A*, %struct.A }>* inalloca %{{[^,]*}})
// CHECK: call x86_fastcallcc void @"\01?qux@B@@QAI?AUA@@U2@@Z"
// CHECK: (%struct.B* inreg %{{[^,]*}}, %struct.A* inreg sret %{{.*}}, <{ %struct.A }>* inalloca %{{[^,]*}})