[llvm] r336178 - [ADT] Disable the single callback optimization on Windows.

[Chandler Carruth via llvm-commits]

On Tue, Jul 3, 2018 at 2:12 PM Friedman, Eli <efriedma at codeaurora.org>
wrote:

> On 7/3/2018 1:25 PM, Chandler Carruth wrote:
>
> On Tue, Jul 3, 2018 at 1:05 PM Friedman, Eli <efriedma at codeaurora.org>
> wrote:
>
>> On 7/3/2018 1:19 AM, Chandler Carruth via llvm-commits wrote:
>> > Author: chandlerc
>> > Date: Tue Jul  3 01:19:10 2018
>> > New Revision: 336178
>> >
>> > URL: http://llvm.org/viewvc/llvm-project?rev=336178&view=rev
>> > Log:
>> > [ADT] Disable the single callback optimization on Windows.
>> >
>> > It appears that the function pointer we use there isn't reliably 4-byte
>> > aligned. I have no idea why or how we could correct this, so for now we
>> > just regress the Windows performance some.
>>
>> I don't know why you'd expect functions to be aligned; x86 doesn't
>> require any alignment. clang usually prefers to align functions to 16
>> bytes because the instruction decoder fetches 16 bytes at a time, but it
>> won't align the entry point at all at -Oz.
>
>
> I thought there was at least some ABI requirements, but I'm happy to add
> an explicit alignment attribute which should be sufficient outside of
> Windows?
>
>
> The alignment attribute should work on most targets, yes...
>

One interesting problem with the alignment attribute, and the reason I
didn't use it initially, is that it may *reduce* the alignment.

For example, I don't really want to use `__attribute__((aligned(4))))`
because while that's all the alignment we *require*, I think that will
actually override any default alignment that is likely higher, and result
in worse performance.

I guess I can use `__attribute__((aligned(16)))` for now and that won't
likely regress performance on any targets, but it will result in a small
amount of binary bloat at -Oz and such.

but it won't work on ARM or certain MIPS variants.
>
>
>
>
>> And even on platforms where
>> functions are aligned, function pointers might not be aligned; for
>> example, ARM uses the bottom bit for interworking.
>>
>
> Yes, but I think we can just insist "don't use such function pointers
> here"?
>
>
> I would guess most users of LLVM on ARM use Thumb2, since it's smaller
> (and therefore faster). So functions pointers will generally have the least
> significant bit set.  But we can work around that with some target-specific
> hacks if we control the callee.
>
> Not sure about the details for MIPS.
>
>
> -Eli
>
> --
> Employee of Qualcomm Innovation Center, Inc.
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180703/be6799dc/attachment.html>