Re: Liberating structs from memory

[Greg McGary <greg at mcgary dot org>]

Jeffrey A Law <law@cygnus.com> writes:

> I think this is most easily done at the tree level -- except that we do not
> have functions as trees for most of the front-ends yet, so it's hard to know
> if an auto has had its address taken until its too late.

I assume you refer to `x_whole_function_mode_p'.  What, if anything,
prevents us from doing this in the C front end today?  If doable, that
sounds like a useful enabling step.  I plan to put BPs only in C and
C++ front ends.  The next natural choice is ObjC, but I'll leave that
to someone who cares about it.  I don't think BPs make sense for other
languages.

> Seems generally reasonable, though you've got to fight the problem of
> dumping stuff to memory *after* you've already creates its DECL_RTL.  In
> a number of ways it's similar to the ADDRESSOF problem where we have these
> magic tidbits lying around for a while then after various optimizations we
> have to convert them to sensible RTL.

Wouldn't whole-function mode would conveniently solve this problem?
We could determine if the struct needs to be in memory before we
make its DECL_RTL, no?

Greg