[llvm] r278532 - ADT: Remove the ilist_nextprev_traits customization point

[Duncan P. N. Exon Smith via llvm-commits]

FYI, I'm running tests on the attached patch.  It's fairly naïve but might work.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: x86-frame-lowering-dereference-end.patch
Type: application/octet-stream
Size: 2114 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160812/1e3e576d/attachment.obj>
-------------- next part --------------


> On 2016-Aug-12, at 15:27, Duncan P. N. Exon Smith via llvm-commits <llvm-commits at lists.llvm.org> wrote:
> 
> Thanks!
> 
> In the meantime, I've started looking as well and this doesn't look as trivial as I'd hoped :(.  Let me know if you want me to revert while we figure this out.
> 
>> On 2016-Aug-12, at 15:24, Hans Wennborg <hans at chromium.org> wrote:
>> 
>> I'll give fixing it a shot.
>> 
>> On Fri, Aug 12, 2016 at 3:15 PM, Duncan P. N. Exon Smith
>> <dexonsmith at apple.com> wrote:
>>> 
>>>> On 2016-Aug-12, at 15:11, Hans Wennborg <hans at chromium.org> wrote:
>>>> 
>>>> On Fri, Aug 12, 2016 at 3:01 PM, Duncan P. N. Exon Smith
>>>> <dexonsmith at apple.com> wrote:
>>>>> 
>>>>>> On 2016-Aug-12, at 14:48, Hans Wennborg <hans at chromium.org> wrote:
>>>>>> 
>>>>>> Just a heads up I'm getting compiler crashes building Chromium on
>>>>>> Windows with this (well, r278539 really). Still not sure what's going
>>>>>> on, though.. http://crbug.com/637413
>>>>>> 
>>>>>> It's crashing here:
>>>>>> 
>>>>>> void X86FrameLowering::emitEpilogue(MachineFunction &MF,
>>>>>>                                 MachineBasicBlock &MBB) const {
>>>>>> const MachineFrameInfo &MFI = MF.getFrameInfo();
>>>>>> X86MachineFunctionInfo *X86FI = MF.getInfo<X86MachineFunctionInfo>();
>>>>>> MachineBasicBlock::iterator MBBI = MBB.getFirstTerminator();
>>>>>> unsigned RetOpcode = MBBI->getOpcode();    <--- CRASH
>>>>>> 
>>>>>> 
>>>>>> I assume MBBI is an ilist_iterator<> of some kind, but I still haven't
>>>>>> been able to figure out what's up, or reproduce the crash on Linux :-/
>>>>> 
>>>>> Weird, I don't know why that would be caused by this commit.  Must affect the optimizer?
>>>>> 
>>>>> Should be reproducible on Linux by adding:
>>>>> 
>>>>>  assert(MBBI != MBB.end());
>>>>> 
>>>>> before using it.
>>>> 
>>>> Thanks! The assert fires on Linux, so now I'm in business :-)
>>> 
>>> Great!  Are you going to fix it too?  (Great if so (thanks!), but otherwise I should dig in myself.)
> 
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits