This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Kill pointers keeping gimple function body around after RTL is generated
> While I would encourage identifying and fixing any leaks of SSA_NAMEs
> which are easily resolved. You might even reach a point where we
> do not leak SSA_NAMEs, which would be good. However, I don't think
> we should be looking at deallocating statement memory on the fly.
I was bit looking into this but hit an interesting problem. I extended
release_ssa_name to kill pointers within the SSA_NAME structure so we
don't keep pointing to them and this broke repeated SSA renaming. The
code first releases the phi nodes via remove_all_phi_nodes_for and later
travels the virtual operands that still possibly reffer to the old
SSA_NAMEs. Is this an expected behavioiur?
It looks like quite dubious practice at least.
Honza