This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] New IPA-CP with real function cloning - updated version


> Thanks, however, I'm not confident committing this on Friday when I'm
> going to be offline until Monday noon :-) Moreover, I already have a

I will be flying back to Europe, so you even can't push responsibility to me :)

> newer version that should handle aliases to thunks.  The changes since
> the yesterday's version are:
> 
>   - initialize_node_lattices asserting node has body
>   - propagate_constants_accross_call more sensible with regard to
>     thunks (and aliases).
>   - gather_caller_stats and has_undead_caller_from_outside_scc_p
>     (hooks to cgraph_for_node_and_aliases) are careful about aliases
>     to thunks too.
>   - all cgraph_function_or_thunk_node turned into cgraph_function_node
>     because we should be propagating across these calls fine now.
> 
> All these changes are only in ipa-cp.c so only that file is verbatim
> below now (and a traditional context diff in an attachment).  I have
> bootstrapped and profiledbootstrapped and tested a very similar patch
> on x86_64-linux, bootstrap and profiledbootstrap of exactly this patch
> underway.  I hope this will be OK for trunk on Monday too if both
> pass.

These changes seems fine.  I declare the patch OK on Monday then ;))

Honza


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]