On Dec 13, 2017, Jakub Jelinek <ja...@redhat.com> wrote: > In particular, this testcase is using selective scheduling, therefore > we turn off -fvar-tracking-assignments, but the debug stmt markers are > emitted anyway.
*nod*, that much was intended (though I could be convinced to change it ;-) > -fvar-tracking is still true, so the var-tracking pass does everything it > normally does (successfully), then the free_cfg pass removes all > BLOCK_FOR_INSN notes, then some targets in their machine reorg recompute > those, but sparc apparently doesn't, and finally in final.c: > /* Turn debug markers into notes. */ > if (!MAY_HAVE_DEBUG_BIND_INSNS && MAY_HAVE_DEBUG_MARKER_INSNS) > variable_tracking_main (); > Eeek, this runs all of the var tracking again, Eeek, indeed ;-) /me takes a mental note that flag_var_tracking != > MAY_HAVE_DEBUG_BIND_INSNS, and underlines it several times ;-D Thanks for spotting the deeper problem and for fixing it! I'm glad the patch I posted to fix the shallower one still serves as a basis for the ongoing attempts to fix one of the remaining ia64 issues. I'm on it. -- Alexandre Oliva, freedom fighter http://FSFLA.org/~lxoliva/ You must be the change you wish to see in the world. -- Gandhi Be Free! -- http://FSFLA.org/ FSF Latin America board member Free Software Evangelist|Red Hat Brasil GNU Toolchain Engineer
