llvm dev - Apr 2016 - RFC: New function attribute "patchable-prologue"="<kind>"

I think most function redirection patching schemes are going to be mutually
incompatible, so I'm not sure it makes sense to make this attribute a
comma-separated list.

I think Eric's and Dean's use case may be better addressed by a separate
attribute. My recollection is that they want to add nop slides to the
prologue and epilogue that can be hotpatched to enable and disable
instrumentation at runtime.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160414/a061e542/attachment.html>

On Thu, Apr 14, 2016 at 1:56 PM Reid Kleckner <rnk at google.com> wrote:
> I think most function redirection patching schemes are going to be
> mutually incompatible, so I'm not sure it makes sense to make this
> attribute a comma-separated list.
>
>
Isn't this what I said? :)

> I think Eric's and Dean's use case may be better addressed by a
separate
> attribute. My recollection is that they want to add nop slides to the
> prologue and epilogue that can be hotpatched to enable and disable
> instrumentation at runtime.
>
I'm not sure about this, it'd be nice not to have a bajillion more
attributes.

-eric
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160414/d9d27715/attachment.html>

On Fri, Apr 15, 2016 at 6:58 AM Eric Christopher <echristo at gmail.com>
wrote:
> On Thu, Apr 14, 2016 at 1:56 PM Reid Kleckner <rnk at google.com>
wrote:
>
>> I think most function redirection patching schemes are going to be
>> mutually incompatible, so I'm not sure it makes sense to make this
>> attribute a comma-separated list.
>>
>>
> Isn't this what I said? :)
>
While some schemes may be incompatible, I suspect composing some of them
together makes some sense. For example:

patchable-function=hotpatch-short-prologue,hotpatch-short-epilogue

We can catch some incompatibilities in the implementation if there are
serious problems with mixing them.

>
>
>> I think Eric's and Dean's use case may be better addressed by a
separate
>> attribute. My recollection is that they want to add nop slides to the
>> prologue and epilogue that can be hotpatched to enable and disable
>> instrumentation at runtime.
>>
>
> I'm not sure about this, it'd be nice not to have a bajillion more
> attributes.
>
>
I for one would like less attributes. It seems this is potentially
something we can use for a broader purpose so I'd like to explore that
possibility early (rather than doing it later).

Cheers
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160415/c26bbcc6/attachment.html>