From owner-freebsd-bugs Wed Jun 19 16:40:48 2002 Delivered-To: freebsd-bugs@hub.freebsd.org Received: from freefall.freebsd.org (freefall.FreeBSD.org [216.136.204.21]) by hub.freebsd.org (Postfix) with ESMTP id 13E5637B404 for ; Wed, 19 Jun 2002 16:40:07 -0700 (PDT) Received: (from gnats@localhost) by freefall.freebsd.org (8.11.6/8.11.6) id g5JNe7K28300; Wed, 19 Jun 2002 16:40:07 -0700 (PDT) (envelope-from gnats) Date: Wed, 19 Jun 2002 16:40:07 -0700 (PDT) Message-Id: <200206192340.g5JNe7K28300@freefall.freebsd.org> To: freebsd-bugs@FreeBSD.org Cc: From: "Crist J. Clark" Subject: Re: kern/39449: wierd ata status Reply-To: "Crist J. Clark" Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org The following reply was made to PR kern/39449; it has been noted by GNATS. From: "Crist J. Clark" To: "Jin Guojun[DSD]" Cc: bug-followup@freebsd.org Subject: Re: kern/39449: wierd ata status Date: Wed, 19 Jun 2002 16:33:07 -0700 On Wed, Jun 19, 2002 at 04:20:36PM -0700, Jin Guojun[DSD] wrote: > "Crist J. Clark" wrote: > > > > On Wed, Jun 19, 2002 at 01:45:02PM -0700, Jin Guojun[DSD] wrote: > > > "Crist J. Clark" wrote: > > > > > > > > On Wed, Jun 19, 2002 at 01:03:17AM -0700, Jin Guojun[DSD] wrote: > > > > [snip] > > > > > > > > > I am not able to access these machines till Friday. > > > > > Unless you think the regular dmesg will help, I will send you full message > > > > > with "boot -v" on Friday. > > > > > > > > The regular boot messages may help. > > > > > > Below is dmesg from 4.6 machine, and a set of -v message from a 4.5 > > > machine with the identical hardware we used before. > > > The message of 4.6 and the first message of 4.5 are from customer built > > > kernel, which has ata1 disabled in kernel config file. The 4.5 has no ata1 > > > configured, but 4.6 has ata1 configured. > > > > > > The rest of messages of 4.5 is from GENERIC kernel (original from > > > the installation) with following conditions: > > > > > > #2, just -v option. We see both ata0 and ata1 configured. > > > > > > #3, the secondary IDE is disbaled in BIOS, and ata1 is not configured. > > > #4, the disk is removed from the secondary IDE bus, and ata1 is not > > > configured. > > > > > > So, in 4.5, we have muliplte ways to disble ata1, but in 4.6 we have none. > > > > OK, I think the dmesg(8) shows what I think is going on. You never > > "disabled" ata1 in 4.5-RELEASE by doing anything to the kernel > > configuration. In 4.5-RELEASE, if there was no hardware attached to > > ata1, the probe turns up nothing, > > > > > #3 ----------- GENERIC kernel with second IDE disabled in BIOS ------ > > > Copyright (c) 1992-2002 The FreeBSD Project. > > > Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 > > > The Regents of the University of California. All rights reserved. > > > FreeBSD 4.5-RELEASE #0: Mon Jan 28 14:31:56 GMT 2002 > > > murray@builder.freebsdmall.com:/usr/src/sys/compile/GENERIC > > [snip] > > > ata1: iobase=0x0170 altiobase=0x0376 bmaddr=0xd808 > > > ata1: mask=00 status0=ff status1=ff > > > ata1: probe allocation failed > > > > But it is still probed. You are not disabling it in the kernel > > configuration, it is disabling itself when no devices are found. > > > > In 4.6-RELEASE, we have support for removable ATA devices. Even if > > ata1 doesn't have anything when probed at boot, we want to attach the > > device in order to have it available for swapping devices. > > > > So, to summarize, you have never had the ability to disable PCI ATA > > devices in the kernel configuration. However, in 4.5-RELEASE, ata(4) > > devices with nothing attached to them were not allocated resources. In > > 4.6-RELEASE, ata(4) devices are allocated in order to support > > removable devices (which is totally wicked-cool BTW, see > > atacontrol(8)). > > It is good to support removable devices. But remember that in 4.5 we have > two ways to disable ata1: > (1) no device attached > or > (2) disbale the secondary IDE in BIOS. No, you couldn't. Only ISA devices. > If 4.6 or later could distinguish these two cases, it will make this > feature more useful. If users want to use removeable devices, enable > the IDE in BIOS, but leave devices either attached or unattached; > if users do not want to use IDE at all, disable IDE in BIOS. > > I do not see why this cannot be implemented as in 4.5. It wasn't. If you can show me dmesg(8) output where you have a device attached to a PCI ata1 and it is not detected, I would be suprised. None of the dmesg(8) outputs that you showed were this case. -- Crist J. Clark | cjclark@alum.mit.edu | cjclark@jhu.edu http://people.freebsd.org/~cjc/ | cjc@freebsd.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message