https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128846
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Component|tree-optimization |target
--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
So the testcase failed before the commit when you pass -fno-vect-cost-model?
It seems this is a target issue where the vectorizer force-aligns the TLS
global to 16 bytes but the target fails to ensure such alignment (or correctly
claim it cannot ensure such alignment).
We currently determine whether we can adjust alignment via
vect_can_force_dr_alignment_p which checks
symtab_node::can_increase_alignment_p
which doesn't seem to have a TLS specific check here.
So - how does TLS allocation on solaris guarantee DECL_ALIGN?