> here is updated patch that assumes by default that RTL SYMBOL_REFs may overlap
> in arbitrary way and only disambiguate refs with declarations (same way as
> tree
> oracle does) and the anchors.
>
> Bootstrapped/regtested ppc64-linux.
> Does it look resonable?
The logic looks ok to me FWIW. I think it might be better (and easier
to follow) if we do the XSTR test first though. I.e.:
if (XSTR (x_base, 0) == XSTR (y_base, 0))
return 1;
if (x_decl && y_decl)
return compare_base_decls (x_decl, y_decl);
...
Also:
Jan Hubicka <hubi...@ucw.cz> writes:
> + /* If we have variable in section anchor, we can compare by offset. */
> + if (SYMBOL_REF_HAS_BLOCK_INFO_P (x_base)
> + && SYMBOL_REF_HAS_BLOCK_INFO_P (y_base))
> + {
> + if (SYMBOL_REF_BLOCK (x_base) != SYMBOL_REF_BLOCK (y_base))
> + return 0;
> + if (SYMBOL_REF_BLOCK_OFFSET (x_base) == SYMBOL_REF_BLOCK_OFFSET
> (y_base))
> + return binds_def ? 1 : -1;
> + /* Mixing anchors and non-anchors may result to false negative.
> + We probably never do that. */
> + gcc_assert (SYMBOL_REF_ANCHOR_P (x_base) == SYMBOL_REF_ANCHOR_P
> (y_base));
...although that's the usual case, I don't think there's any practical
way of enforcing an absolute guarantee. I think instead we should just
return -1 if either symbol is not an anchor.
Nit, but: there were a couple of 'return false's that I think
should be 'return 0's.
Thanks,
Richard
