[PATCH] - Fix PR 21956 and PR 22003

[Caroline Tice]

Sorry I didn't answer this sooner; I've gotten a little behind on
reading the gcc mailing list.  I don't really understand how
what exactly a thunk is in gcc, so it's hard for me to answer your
question.  Hot/cold partitioning only changes from one section to
another if it encounters a note telling it to, and the note can only
be inserted if there are basic blocks, so it seems to me by default
that the thunk should land in the same section as the function.
But I'm not making any guarantees at this point.  if you could
give me a  thunk to test things with, I could give you a better
answer...

-- Caroline
ctice@apple.com


On Jul 27, 2005, at 2:00 PM, Steven Bosscher wrote:

> On Wednesday 27 July 2005 22:55, Richard Henderson wrote:
>
>> On Wed, Jul 27, 2005 at 02:41:08AM +0200, Steven Bosscher wrote:
>>
>>>     PR c++/22003
>>>     * varasm.c (assemble_start_function): Don't do anything that may
>>>     require a CFG if the current function is a thunk.
>>>
>>
>> Not entirely ok.  Several of the target backend thunks require that
>> the thunk and the function reside in the same section.  Though I
>> suspect that this may already be broken by -ffunction-sections,
>>
>
> Yeah, that is why I asked Caroline whether she knows where the thunks
> end up.  I'm not sure how to create a test case for this, so I hope
> Caroline has some ideas.
>
>
>> so
>> I guess this can probably go in as-is.
>>
>
> I'll commit it, then.
>
> Gr.
> Steven
>
>