Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 6 Nov 2009 11:29:51 -0500
From:      Ed Maste <emaste@freebsd.org>
To:        John Baldwin <jhb@freebsd.org>
Cc:        Attilio Rao <attilio@freebsd.org>, freebsd-current@freebsd.org
Subject:   Re: [PATCH] Boot-time entering prompt by typing "123" sequence rather than any button
Message-ID:  <20091106162951.GA77255@sandvine.com>
In-Reply-To: <200911061050.50873.jhb@freebsd.org>
References:  <3bbf2fe10911060712y46684f76sefe195a62083c15@mail.gmail.com> <200911061050.50873.jhb@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Nov 06, 2009 at 10:50:50AM -0500, John Baldwin wrote:

> On Friday 06 November 2009 10:12:59 am Attilio Rao wrote:
> > This patch adds the possibility to enter the prompt at boot time by
> > typing the sequence of buttons "123" rather than a single button:
> > http://www.freebsd.org/~attilio/Sandvine/STABLE_8/boot123/boot123.diff
> > 
> > This is useful in the cases where a serial console is likely going to
> > be used which can carry on spourious character, leding to the prompt
> > erroneously.
> > This option is wrappered into the BOOT_PROMPT_123 option, in order to
> > maintain the current POLA.
> > This patch has been contributed back by Sandvine Incorporated.
> > Please review.
> 
> This seems a bit hackish, but the patch is fine on technical grounds.  On 
> machines where I have a serial console I tend to put "-Dh" in /boot.config 
> (so I can choose an alternate loader if need be) and in that case boot2 would 
> eat the extra input and pause during the boot process (I've had this happen 
> occasionally).  This patch wouldn't help with that case.  Another suggestion 
> made on IRC was to simply drain input at the start of boot2 or loader to 
> avoid accepting early spurious input.  This won't help if you are dealing 
> with noisy serial lines that are just spewing random garbage during the boot 
> process however.

In a specific case we encountered, an unterminated cable plugged into a
serial port caused enough signal to be reflected to produce a lossy
loopback.  This is a similar case to the random garbage case you
suggest.

My first stage loader isn't using the serial console, it's only when
loader(8) starts that it switches.

-Ed



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20091106162951.GA77255>