> Hi,
>
> On Fri, Nov 15 2019, luoxhu wrote:
> >
> > Sorry to break the bootstrap. This was my lack of consideration when
> > splitting this small piece of code from the previous patch, the line
> > should be in the first line of the second loop. Could you please add
> > some comments that edge may be freed in resolve_speculation()?
> > Thanks.
>
> First and foremost, resolve_speculation should not free its this
> pointer. This is the third time it has caused us a headache in the last
> few months. If we cannot think of a better solution in the next week or
> three, I will revert it back to a non-member function in this stage 3
The bug in the walker loop here IMO predated C++ conversion, but to
avoid member function freeing THIS we probably want to have:
cgraph_edge::remove (e),
symtab_node::remove (e),
ipa_ref::remove (ref),
cgraph_node::resolve_speculation (e, target)
I suppose good time for the conversion would be after remaining patches
are merget. If you would have time to do that even better ;)
> ...and then we can (and should and will) document that it can free the
> edge.
It speaks about removing edges and return value,
Speculative call edge turned out to be direct call to CALLEE_DECL.
Remove the speculative call sequence and return edge representing the call.
It is up to caller to redirect the call as appropriate.
We could add
Edge E may be freed if it is resolved to other edge of the
speculative call.
Honza
>
> Martin
