Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 26 Apr 1995 14:50:01 -0700
From:      Paul Traina <pst@Shockwave.COM>
To:        freebsd-bugs
Subject:   kern/367: ep0 driver in generic kernel hangs during attach
Message-ID:  <199504262150.OAA27712@freefall.cdrom.com>
In-Reply-To: Your message of Wed, 26 Apr 1995 14:42:21 -0700 <199504262142.OAA05212@precipice.shockwave.com>

next in thread | previous in thread | raw e-mail | index | archive | help

>Number:         367
>Category:       kern
>Synopsis:       kernel hangs after ep0 attach message appears
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs (FreeBSD bugs mailing list)
>State:          open
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Apr 26 14:50:00 1995
>Originator:     Bruce Cole
>Organization:
cisco Systems, Inc.
>Release:        FreeBSD 2.1.0-Development i386
>Environment:

April 24th FreeBSD snapshot generic kernel running on a Pentium based
unix box with a 3COM 3C509 ethernet card at 0x300 / irq 10.

>Description:

The system starts to boot, and hangs after the 3c509 attach message shows
up.

From:       Bruce Cole <bcole@cisco.com>
Subject:    Re: 386bsd hint? 
Date:       Wed, 26 Apr 1995 14:26:55 -0700
To:         Paul Traina <pst@cisco.com>
cc:         Bruce Cole <bcole@cisco.com>

> Isn't boot -c wonderful? :-)
> What has likely occured is that some other card has wedged the controller
> first.  I would use boot -c and disable anything that's probing at 0x300 or
> on irq 10, leaving just the 3c509 enabled.

I had tried disabling all the ones that came after it.  Just tried it more
completely like you suggested, and still no luck.  Then I tried a binary
search on getting rid of drivers...  Found that it doesn't hang if I disable
*both* eg0 and lnc0, but not just one or the other.  Either I've got a time
dependency or something weird is going on...


>How-To-Repeat:

Boot a generic kernel on Bruce's box.

>Fix:
	
Well, it's pretty clear that the probe routines for the eg0 and lnc0 drivers
are hosing over the 3c509.  I would suggest changing the order of the probes
in the ethernet driver section, since the ep driver is in much greater common
use than the eg and lnc drivers,  and if the ep driver gets attached first,
the probes for the eg and lnc drivers will never be executed.
>Audit-Trail:
>Unformatted:





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504262150.OAA27712>