Date: Mon, 21 Aug 1995 10:42:48 -0700 From: "Justin T. Gibbs" <gibbs@freefall.FreeBSD.org> To: Bruce Evans <bde@zeta.org.au> Cc: hackers@freefall.FreeBSD.org Subject: Re: Clock interrupts during probes? Message-ID: <199508211742.KAA02371@freefall.FreeBSD.org> In-Reply-To: Your message of "Tue, 22 Aug 95 03:30:41 %2B1000." <199508211730.DAA02704@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
>>I need access to timeout(), not tsleep() for the work in the upperlevel >>SCSI code, so I don't need a process context. Most of the DELAY() calls >>in the aic7xxx driver stem from that fact that I can't use timeouts during >>boot and must poll in order to do a timeout. > >You need tsleep() for the most time-consuming operation, which is >DELAY(1000000 * SCSI_DELAY) in scsiconf.c (default 15 seconds). This >is very annoying when there are multiple scsi controllers or other >slow to start devices that could be started in parallel. > >Bruce Timeout could be used there with a little manipulation of the code. But I guess I should be more clear on what I need timeout for. Right now, we don't wait between retries of commands that return "target busy". We should really sleep for a second or so. It would be really easy simply to setup a timeout pointing the retry function, but that solution won't work at boot. It seems just so silly to me to have to special case code for boot like this if, as you said before, we could have clock interrupts with minimal effort. -- Justin T. Gibbs =========================================== Software Developer - Walnut Creek CDROM FreeBSD: Turning PCs into workstations ===========================================
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199508211742.KAA02371>