From owner-freebsd-bugs  Thu Mar 21 06:43:33 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id GAA29773
          for bugs-outgoing; Thu, 21 Mar 1996 06:43:33 -0800 (PST)
Received: from frig.mt.cs.keio.ac.jp (frig.mt.cs.keio.ac.jp [131.113.32.7])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id GAA29743
          Thu, 21 Mar 1996 06:43:25 -0800 (PST)
Received: (from hosokawa@localhost) by frig.mt.cs.keio.ac.jp (8.6.12+2.4W/3.4Wbeta3) id XAA08476; Thu, 21 Mar 1996 23:43:22 +0900
Date: Thu, 21 Mar 1996 23:43:22 +0900
Message-Id: <199603211443.XAA08476@frig.mt.cs.keio.ac.jp>
To: pst@shockwave.com, bde@freebsd.org, hackers@freebsd.org, bugs@freebsd.org,
        jmz@freebsd.org
Cc: hosokawa@mt.cs.keio.ac.jp
Subject: Re: dev_attach after boot (Re: modloaded block/char device drivers)
In-Reply-To: Your message of Thu, 21 Mar 1996 22:44:16 +0900.
	<199603211344.WAA08147@frig.mt.cs.keio.ac.jp>
From: hosokawa@mt.cs.keio.ac.jp (HOSOKAWA Tatsumi)
X-Mailer: mnews [version 1.18PL3] 1994-08/01(Mon)
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Sorry! I forgot to write about important experiences.

In article <199603211344.WAA08147@frig.mt.cs.keio.ac.jp>
hosokawa@mt.cs.keio.ac.jp writes:

>> When I try to enable (hotplug) Flash ATA (wdc and wd) driver or SCSI
>> devices (cd, sd, etc.), the kernel panic()'s by kernel page fault.
>> 
>> I found it never panics when I commented out the calls of dev_attach()
>> in those drivers.

Not all dynamic dev_attach() causes panic().

For example,

	if_ed	NE2000 compatible PCMCIA Ethernet	OK
	if_nep	3C589x PCMCIA Ethernet			OK
	if_fe	Fujitsu PCMCIA Ethernet			OK
	wdc/wd	Flash ATA / Type 3 ATA disk		NG
	sio	PCMCIA Modem				OK
	aic/cd	SCSI CDROM w/ SlimSCSI PCMCIA		NG
	aic/sd	SCSI HDD w/ SlimSCSI PCMCIA		NG

So, I think adding the controllers dynamically causes this panic(),
but I was confused by reading that qcam has the same problem.

Apparently, qcam is not controller/disk type device.  Of course I read
dev_attach() in kern_devconf.c and it seems to have no problem which
causes such panic()'s.

Currently, I'm planning to write a quick hack patch that PCMCIA
dynamic device enabler never calls dev_attach() if this problem can't
solved for long.

Does any can guess the reason of this fatal problem?

Thanks.

--
	HOSOKAWA, Tatsumi              E-mail: hosokawa@mt.cs.keio.ac.jp
	  WWW homepage: http://www.mt.cs.keio.ac.jp/person/hosokawa.html
	Department of Computer Science, Keio University, Yokohama, Japan