From owner-freebsd-hackers Tue Jan 24 17:31:16 1995 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id RAA19952 for hackers-outgoing; Tue, 24 Jan 1995 17:31:16 -0800 Received: from ns.gte.com (ns.gte.com [132.197.8.9]) by freefall.cdrom.com (8.6.9/8.6.6) with SMTP id RAA19941 for ; Tue, 24 Jan 1995 17:31:10 -0800 Received: from bunny.gte.com by ns.gte.com (5.65c/IDA-1.4.4) id AA29039; Tue, 24 Jan 1995 20:29:07 -0500 Received: by bunny.gte.com (8.6.9/GTEL2.19) id UAA07953; Tue, 24 Jan 1995 20:30:31 -0500 Received: from localhost by genesis.nred.ma.us (8.3/genesis0.0) id UAA12517; Tue, 24 Jan 1995 20:19:31 -0800 Date: Tue, 24 Jan 1995 20:19:31 -0800 From: steve2@genesis.nred.ma.us (Steve Gerakines) Message-Id: <199501250419.UAA12517@genesis.nred.ma.us> To: freebsd-hackers@FreeBSD.org Subject: Re: Floppy tape probe patch Sender: hackers-owner@FreeBSD.org Precedence: bulk > Hmmmm. Perhaps fd should also poll. On the other hand, polling the > FDC is dangerous (the danger is only being eliminated by in_fdc() > watching thoroughly for the ready-for-input bit), since many FDC's > tend to lock up if someone requests ``unusual'' things (e.g. SENSEI > where no interrupt happened) from them. Unfortunately looking at dio didn't help. Apparently once a command has been issued, dio doesn't change until a sense interrupt is issued and you pull off st0 and the cylinder number. This is why ft was blasting sense interrupt commands until it finally succeeded instead of just testing msr. The timeouts seem to work fine for fd so I wouldn't bother messing with it. Guaranteed we would end up breaking something that wasn't broken in the first place. :-) - Steve steve2@genesis.nred.ma.us