r222433 - Prepare for the DIBuilder populating DISubprogram's Variables with null.

[David Blaikie]

(& now I see your follow-up commit... :))

On Thu, Nov 20, 2014 at 9:05 AM, David Blaikie <dblaikie at gmail.com> wrote:

>
>
> On Thu, Nov 20, 2014 at 7:46 AM, Frederic Riss <friss at apple.com> wrote:
>
>> Author: friss
>> Date: Thu Nov 20 09:46:59 2014
>> New Revision: 222433
>>
>> URL: http://llvm.org/viewvc/llvm-project?rev=222433&view=rev
>> Log:
>> Prepare for the DIBuilder populating DISubprogram's Variables with null.
>>
>> This is a followup to r222373. A better solution to the problem solved
>> there is to not create the leaked nodes at all (we know that they will
>> never be used for forward declared functions anyway). To avoid bot
>> breakage in the interval between the cfe and llvm commits, add a check
>> that the nMDNode is not null before deleting it. This code can completely
>> go away after the LLVM part is in.
>>
>> Modified:
>>     cfe/trunk/lib/CodeGen/CGDebugInfo.cpp
>>
>> Modified: cfe/trunk/lib/CodeGen/CGDebugInfo.cpp
>> URL:
>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CGDebugInfo.cpp?rev=222433&r1=222432&r2=222433&view=diff
>>
>> ==============================================================================
>> --- cfe/trunk/lib/CodeGen/CGDebugInfo.cpp (original)
>> +++ cfe/trunk/lib/CodeGen/CGDebugInfo.cpp Thu Nov 20 09:46:59 2014
>> @@ -3417,7 +3417,8 @@ void CGDebugInfo::finalize() {
>>      // to get RAUWed upon DIBuilder finalization. Do not leak these
>>      // nodes for the temporary functions we are about to delete.
>>      if (FwdDecl.isSubprogram())
>> -
>> llvm::MDNode::deleteTemporary(llvm::DISubprogram(FwdDecl).getVariablesNodes());
>> +      if (llvm::MDNode *Vars =
>> llvm::DISubprogram(FwdDecl).getVariablesNodes())
>> +        llvm::MDNode::deleteTemporary(Vars);
>>
>
> Why do you need this deleteTemporary at all? We must always RAUW the
> variables node somewhere (where?) and can delete this temporary node there
> (without the need for extra/duplicate conditionals) maybe? Yeah, this seems
> like it could be handled down in DIBuilder::finalize when it RAUW's the
> variable node (& it'd be more symmetric that way - DIBuilder creates them,
> DIBuilder destroys them, Clang (& any other DIBuilder client) doesn't
> know/care about them)
>
>
>>
>>      FwdDecl.replaceAllUsesWith(CGM.getLLVMContext(),
>>
>> llvm::DIDescriptor(cast<llvm::MDNode>(VH)));
>>
>>
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20141120/d9a330c1/attachment.html>