[lto][patch] Fix how we handle clones in wpa

[Rafael Espindola]

> Unfortunately, I had to revert this patch.  It was causing several new
> failures in our internal builds.  We will phase it back in after we
> figure out what the problem was.

So, where a function is called from is completely irrelevant to
whether or not it should be in the set that is passed to ltrans. The
attached patch fixes my original one.

Bootstrapped and regression tested.

The reduced testcase I have fails with an assert without the patch and
(correctly) fails to link with it. Not sure how to add it to the
testsuite.


2009-02-20  Rafael Avila de Espindola  <espindola@google.com>

	* lto.c (lto_add_all_inlinees): Ignore where a function is called from.

>
> Diego.
>

Cheers,
-- 
Rafael Avila de Espindola

Google | Gordon House | Barrow Street | Dublin 4 | Ireland
Registered in Dublin, Ireland | Registration Number: 368047
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fix.patch
Type: text/x-diff
Size: 1086 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20090220/cad8e5df/attachment.bin>