From owner-freebsd-hackers  Wed Jul  5 07:46:47 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id HAA25554
          for hackers-outgoing; Wed, 5 Jul 1995 07:46:47 -0700
Received: from mail.htp.com (mail.htp.com [199.171.4.2])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id HAA25546
          for <freebsd-hackers@freebsd.org>; Wed, 5 Jul 1995 07:46:44 -0700
Received: from et.htp.com (et.htp.com [199.171.4.228]) by mail.htp.com (8.6.5/8.6.5) with SMTP id KAA12288 for <freebsd-hackers@freebsd.org>; Wed, 5 Jul 1995 10:45:16 -0400
Date: Wed, 5 Jul 1995 10:45:16 -0400
Message-Id: <199507051445.KAA12288@mail.htp.com>
X-Sender: dennis@mail.htp.com
X-Mailer: Windows Eudora Version 2.0.3
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: freebsd-hackers@freebsd.org
From: dennis@et.htp.com (dennis)
Subject: Re: Booting with no kbd
Sender: hackers-owner@freebsd.org
Precedence: bulk

>>On 1 machine the system won't even try to load the kernel. With boot manager
>>installed it just sits at the prompt and never times out. As soon as the
>>keyboard is plugged in it continues ok. Without boot manager it just hangs
>>without ever displaying the initial boot: prompt. The machine boots MS-DOS
>>OK in the same configuration.
>
>Does DOS + boot manager boot OK?
>
I don't have a dos partition on this machine, but I can boot from a floppy
so the system will boot with no keyboard installed.

Any suggestions for a hack to make this work? Which particular module is
involved?

>>The other machine loads the kernel ok, but then hangs at the sc0 probe. As
>>soon as the kbd is plugged in, it continues.
>
scprobe() has 2 possibly endless loops:
>
>E1) At the start, to "flush any noise in the buffer".  Perhaps endless noise
>is picked up by the unattached-to keyboard socket.  Unlikely.
>
>E2) If `gotack' is reached (for the first character of the ack or for
>the following bug) then there is an endless loop if the second character
>of the ack never arrives.
>
>and at least 2 logic bugs:
>
>B1) `retries' ends up as -1 if the `while (retries--)' loop is executed the
>maximum number of times, but the test for this is `if (!retries)'.
>
>B2) The value read from the keyboard in the loop in B2) is not waited for
>so it may be garbage.  Some garbage values would result in E2) being
>reached even if B1) is fixed.
>
>On my test system, the keyboard controller apparently responds to the
>KB_RESET command with 2 KB_RESEND's when there is no keyboard attached,
>so there is no problem.
>

I'll test this and see what I find.

thanks,

dennis