Date: Mon, 18 Nov 2002 17:15:18 +1100 (EST) From: Bruce Evans <bde@zeta.org.au> To: Robert Watson <rwatson@FreeBSD.ORG> Cc: Thierry Herbelot <thierry@herbelot.com>, <current@FreeBSD.ORG> Subject: Re: SMP stability ? [was Re: more info from panic from running dnet on SMP kernel] Message-ID: <20021118170528.V25596-100000@gamplex.bde.org> In-Reply-To: <Pine.NEB.3.96L.1021117144423.93303G-100000@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 17 Nov 2002, Robert Watson wrote: > I've seen several reports that using a serial break to get into ddb is now > quite a bit more reliable than a keyboard break. If you're not already This is a fact. In RELENG_4, the keyboard interrupt handler is a normal tty interrupt handler so it can't interrupt things blocked by spltty(), while the sio interrupt handler is a fast interrupt handler so it can interrupt almost anything (anything not blocked by disable_intr(), which should be everything except fast interrupt handlers and the entry code for normal interrupt handlers). Things are much more broken in -current: the keyboard interrupt handler is non-MPSAFE so it can't interrupt things blocked by Giant (which is most syscalls), and the sio interrupt handler is a "fast" interrupt handler so it can't interrupt things blocked by critical_enter() (which is too many things for too long, so fast interrupt handlers aren't actually fast). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021118170528.V25596-100000>