From owner-freebsd-mobile Mon Jan 25 16:20:23 1999 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA10095 for freebsd-mobile-outgoing; Mon, 25 Jan 1999 16:20:23 -0800 (PST) (envelope-from owner-freebsd-mobile@FreeBSD.ORG) Received: from jli.com (jli.com [199.2.111.1]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id QAA10089 for ; Mon, 25 Jan 1999 16:20:20 -0800 (PST) (envelope-from trost@cloud.rain.com) Received: (qmail 19140 invoked by uid 4); 26 Jan 1999 00:20:14 -0000 Received: (qmail 11816 invoked from network); 26 Jan 1999 00:16:28 -0000 Received: from localhost.cloud.rain.com (HELO grey.cloud.rain.com) (127.0.0.1) by localhost.cloud.rain.com with SMTP; 26 Jan 1999 00:16:28 -0000 To: mobile@FreeBSD.ORG Subject: Re: Reclaiming irqs for unsupported PCI hardware? References: <199901220158.RAA12743@dingo.cdrom.com> In-reply-to: Your message of Thu, 21 Jan 1999 17:58:34 PST. <199901220158.RAA12743@dingo.cdrom.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <11812.917309787.1@grey.cloud.rain.com> Date: Mon, 25 Jan 1999 16:16:27 -0800 Message-ID: <11813.917309787@grey.cloud.rain.com> From: Bill Trost Sender: owner-freebsd-mobile@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Mike Smith writes: Perhaps we're just looking at this the wrong way? We don't try to detect when floppies are stupidly removed, perhaps we shouldn't try to do it with pccard/cardbus cards either? Commentary? I recently pulled a MS-DOS floppy while copying a file to it. The net result of doing so was that the cp command hung and the console got spewed with error messages like fd0c: hard error writing fsbn 897 (No status) fd0c: hard error writing fsbn 1 of 1-3 (No status) I would expect the pccard stuff to behave similarly when its cards get pulled (remember, PCMCIA was originally just a replacement for floppies (-: ). The process accessing a removed card can reasonably be expected to hang, but there's no reason the whole machine should turn into a brick just because a card was popped -- at least in the case of network and serial cards. SCSI cards should probably behave that way, too, but I realize that is an awful lot to ask. Incidentally, is there any reason why "gone" is not checked in the inner loop of the sio interrupt routine? My guess is that the driver would be less apt to hang if this check were moved. Then again, I don't even know where "gone" gets set, so maybe this is moot. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message