[lld] r194860 - [PECOFF] Use INT3 instead of NOP.

[Reid Kleckner]

On Tue, Nov 19, 2013 at 12:52 PM, Sean Silva <silvas at purdue.edu> wrote:

>
>
>
> On Tue, Nov 19, 2013 at 3:48 PM, Reid Kleckner <rnk at google.com> wrote:
>
>> MSVC uses int3 fill for alignment everywhere in .text.  I agree we should
>> benchmark eventually, but there's nothing wrong with being consistent here
>> for now.
>>
>
> Wow, even for aligning loops? Seems like a pointless waste of decoding
> bandwidth vs. an appropriately sized nop.
>

Sorry, no, I mean for unreachable regions between functions and things like
that.

-- Sean Silva
>
>
>>
>>
>> On Sat, Nov 16, 2013 at 11:31 PM, Sean Silva <silvas at purdue.edu> wrote:
>>
>>> The CPU is probably still decoding these and it might hiccup. Could you
>>> try running some microbenchmarks to verify that this isn't a performance
>>> issue? (2 otherwise identical tight asm loops under `perf stat` is probably
>>> enough)
>>> It may not be a problem, but in case it is, you're going to have a
>>> *really* hard time tracking down this performance bug later; all you would
>>> basically have to go on from a bug report is "when I link my program with
>>> lld, it runs slower".
>>>
>>> -- Sean Silva
>>>
>>>
>>> On Fri, Nov 15, 2013 at 5:11 PM, Rui Ueyama <ruiu at google.com> wrote:
>>>
>>>> Author: ruiu
>>>> Date: Fri Nov 15 16:11:43 2013
>>>> New Revision: 194860
>>>>
>>>> URL: http://llvm.org/viewvc/llvm-project?rev=194860&view=rev
>>>> Log:
>>>> [PECOFF] Use INT3 instead of NOP.
>>>>
>>>> This patch does not change the meaning of the program, but if
>>>> something's wrong
>>>> in the linker or the compiler and the control reaches to the gap of
>>>> imported
>>>> function table, it will stop immediately because of the presence of
>>>> INT3. If
>>>> NOP, it'd fall through to the next call instruction, which is usually a
>>>> completely foreign function call.
>>>>
>>>> Modified:
>>>>     lld/trunk/lib/ReaderWriter/PECOFF/ReaderImportHeader.cpp
>>>>
>>>> Modified: lld/trunk/lib/ReaderWriter/PECOFF/ReaderImportHeader.cpp
>>>> URL:
>>>> http://llvm.org/viewvc/llvm-project/lld/trunk/lib/ReaderWriter/PECOFF/ReaderImportHeader.cpp?rev=194860&r1=194859&r2=194860&view=diff
>>>>
>>>> ==============================================================================
>>>> --- lld/trunk/lib/ReaderWriter/PECOFF/ReaderImportHeader.cpp (original)
>>>> +++ lld/trunk/lib/ReaderWriter/PECOFF/ReaderImportHeader.cpp Fri Nov 15
>>>> 16:11:43 2013
>>>> @@ -152,7 +152,7 @@ namespace {
>>>>
>>>>  uint8_t FuncAtomContent[] = {
>>>>    0xff, 0x25, 0x00, 0x00, 0x00, 0x00,  // jmp *0x0
>>>> -  0x90, 0x90                           // nop; nop
>>>> +  0xcc, 0xcc                           // int 3; int 3
>>>>  };
>>>>
>>>>  /// The defined atom for jump table.
>>>>
>>>>
>>>> _______________________________________________
>>>> llvm-commits mailing list
>>>> llvm-commits at cs.uiuc.edu
>>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>>>>
>>>
>>>
>>> _______________________________________________
>>> llvm-commits mailing list
>>> llvm-commits at cs.uiuc.edu
>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20131119/0703a852/attachment.html>