[PATCH] D18479: [CodeGenCXX] Fix ItaniumCXXABI::getAlignmentOfExnObject to return 8-byte alignment on Darwin

[Akira Hatanaka via cfe-commits]

> On Mar 25, 2016, at 2:23 PM, Akira Hatanaka via cfe-commits <cfe-commits at lists.llvm.org> wrote:
> 
>> 
>> On Mar 25, 2016, at 2:06 PM, David Majnemer via cfe-commits <cfe-commits at lists.llvm.org <mailto:cfe-commits at lists.llvm.org>> wrote:
>> 
>> 
>> 
>> On Fri, Mar 25, 2016 at 12:57 PM, Akira Hatanaka via cfe-commits <cfe-commits at lists.llvm.org <mailto:cfe-commits at lists.llvm.org>> wrote:
>> ahatanak created this revision.
>> ahatanak added a reviewer: rjmccall.
>> ahatanak added a subscriber: cfe-commits.
>> 
>> r246985 made changes to give a higher alignment for exception objects on the grounds that Itanium says _Unwind_Exception should be "double-word" aligned and the structure is normally declared with __attribute__((aligned)) guaranteeing 16-byte alignment. It turns out that libc++abi doesn't declare the structure with __attribute__((aligned)) and therefore only guarantees 8-byte alignment on 32-bit and 64-bit platforms. This caused a crash in some cases when the backend emitted SIMD store instructions that requires 16-byte alignment (such as movaps).
>> 
>> This patch makes ItaniumCXXABI::getAlignmentOfExnObject return an 8-byte alignment on Darwin to fix the crash.
>> 
>> http://reviews.llvm.org/D18479 <http://reviews.llvm.org/D18479>
>> 
>> Files:
>>   lib/CodeGen/ItaniumCXXABI.cpp
>>   test/CodeGenCXX/eh.cpp
>> 
>> Index: test/CodeGenCXX/eh.cpp
>> ===================================================================
>> --- test/CodeGenCXX/eh.cpp
>> +++ test/CodeGenCXX/eh.cpp
>> @@ -448,5 +448,27 @@
>>    }
>>  }
>> 
>> +namespace test17 {
>> +class BaseException {
>> +private:
>> +  int a[4];
>> +public:
>> +  BaseException() {};
>> +};
>> +
>> +class DerivedException: public BaseException {
>> +};
>> +
>> +int foo() {
>> +  throw DerivedException();
>> +  // The alignment passed to memset is 8, not 16, on Darwin.
>> +
>> +  // CHECK: [[T0:%.*]] = call i8* @__cxa_allocate_exception(i64 16)
>> +  // CHECK-NEXT: [[T1:%.*]] = bitcast i8* [[T0]] to %"class.test17::DerivedException"*
>> +  // CHECK-NEXT: [[T2:%.*]] = bitcast %"class.test17::DerivedException"* [[T1]] to i8*
>> +  // CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* [[T2]], i8 0, i64 16, i32 8, i1 false)
>> +}
>> +}
>> +
>>  // CHECK: attributes [[NUW]] = { nounwind }
>>  // CHECK: attributes [[NR]] = { noreturn }
>> Index: lib/CodeGen/ItaniumCXXABI.cpp
>> ===================================================================
>> --- lib/CodeGen/ItaniumCXXABI.cpp
>> +++ lib/CodeGen/ItaniumCXXABI.cpp
>> @@ -163,8 +163,17 @@
>>    /// we assume that alignment here.  (It's generally 16 bytes, but
>>    /// some targets overwrite it.)
>>    CharUnits getAlignmentOfExnObject() {
>> -    auto align = CGM.getContext().getTargetDefaultAlignForAttributeAligned();
>> -    return CGM.getContext().toCharUnitsFromBits(align);
>> +    unsigned Align;
>> +
>> +    // Alignment is 8 for darwin since libc++abi doesn't declare
>> +    // _Unwind_Exception with __attribute__((aligned)) and therefore doesn't
>> +    // guarantee 16-byte alignment.
>> +    if (CGM.getContext().getTargetInfo().getTriple().isOSDarwin())
>> 
>> What about Linux/FreeBSD targets which use libc++abi?
>>  
> 
> Is there a way to detect whether libc++abi is used?
> 

I wasn’t planning to fix this for Linux or FreeBSD because I don’t know whether that is what people want. If the library being used returns an object that is 16-byte aligned (does libsupc++ return an aligned object?), my change would unnecessarily pessimize the code.

>> +      Align = 64;
>> +    else
>> +      Align = CGM.getContext().getTargetDefaultAlignForAttributeAligned();
>> +
>> +    return CGM.getContext().toCharUnitsFromBits(Align);
>>    }
>> 
>>    void emitRethrow(CodeGenFunction &CGF, bool isNoReturn) override;
>> 
>> 
>> 
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at lists.llvm.org <mailto:cfe-commits at lists.llvm.org>
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits <http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits>
>> 
>> 
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at lists.llvm.org <mailto:cfe-commits at lists.llvm.org>
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits <http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits>
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at lists.llvm.org <mailto:cfe-commits at lists.llvm.org>
> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits <http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20160328/22fe3c7e/attachment-0001.html>