Date: Thu, 5 Dec 1996 13:01:17 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: smp@csn.net (Steve Passe) Cc: terry@lambert.org, ccsanady@friley216.res.iastate.edu, peter@spinner.dialix.com, smp@freebsd.org Subject: Re: make locking more generic? Message-ID: <199612052001.NAA19785@phaeton.artisoft.com> In-Reply-To: <199612051940.MAA14434@clem.systemsix.com> from "Steve Passe" at Dec 5, 96 12:40:55 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> >If CPU 1 is servicing INT 12 from a disk controller, and you get INT 4 > >from a serial port, don't you *want* CPU 2 to service the INT 4 in > >parallel??? > > yes, you do. but the problem is that CPU 1 holds the lock, so if CPU2 > tries to service INT 4, it starts spinning in the un-obtainable lock till CPU1 > finishes servicing INT 12. At this point CPU1 releases the lock, CPU2 > finally gets it and can proceed to service INT 4. Might as well let CPU2 > alone to do whatever it is doing instead. If the INT 4 is of higher priority > than INT 12, CPU1 could start servicing it immediatly. > > The proper form of fine-grainularity could eliminate this problem. Ah. I see. We are still talking about the global kernel entrancy lock in this case. Never mind then (the global entrancy lock is bogus). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199612052001.NAA19785>