[Bug rtl-optimization/65980] [6 Regression] -fcompare-debug failure building Linux kernel
jakub at gcc dot gnu.org
gcc-bugzilla@gcc.gnu.org
Mon Dec 14 16:26:00 GMT 2015
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65980
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
Assignee|unassigned at gcc dot gnu.org |jakub at gcc dot gnu.org
--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 37029
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=37029&action=edit
gcc6-pr65980.patch
Untested fix. next_real_insn stops on DEBUG_INSNs, so breaks -fcompare-debug
when used for cross-jump possibilities.
Also, it doesn't stop on JUMP_TABLE_DATA nor BARRIER, bet with say
__builtin_unreachable () we might have some fallthru into a BARRIER, at least
hypothetically. We certainly want to stop on INSN/JUMP_INSN/CALL_INSN (both
functions stop on those), and don't want to stop on NOTEs (neither function
stops on those), CODE_LABELs (only next_real_insn skips over those).
More information about the Gcc-bugs
mailing list