https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=205163
--- Comment #7 from [email protected] --- (In reply to Andriy Gapon from comment #6) The predicate that was turned into a dynamic variable was runtime tunable without any sort of mutex locking; a reasonable perusal of the code disclosed no reason to not do it. However, the potential for a sign inversion (which incidentally MIGHT occur in the base code as well, or to a corner case that could result in a divide-by-zero) is easily removed (and I have done so and sent the patch to do so to the reporter, asking if it stops the panics.) Given the stack trace involved here I don't believe that's implicated as the cause, however (since the panic is a divide-by-zero trap yet there is no division by zero evidenced by examination) however, the potential for the stack to get smashed by an integer division that sets the sign bit and leads to misbehavior somewhere else in the code is plausible and defending against both that and the potential for an actual divide-by-zero obviously should be done. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ [email protected] mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "[email protected]"
