Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 12 Oct 1998 12:36:25 -0700
From:      Mike Smith <mike@smith.net.au>
To:        "Kenneth D. Merry" <ken@plutotech.com>
Cc:        mike@smith.net.au (Mike Smith), mjacob@feral.com, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG
Subject:   Re: cvs commit: src/sys/cam/scsi scsi_da.c 
Message-ID:  <199810121936.MAA07717@dingo.cdrom.com>
In-Reply-To: Your message of "Mon, 12 Oct 1998 13:24:53 MDT." <199810121924.NAA26149@panzer.plutotech.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
> Mike Smith wrote...
> > > We're going to put in a mechanism in the next day or two to allow
> > > specifying SCSI revision in quirk entries.  This will allow us to disable
> > > cache sync, multi-lun probing and serial number inquiries for SCSI-1
> > > devices.
> > 
> > Hmm.  Would you consider adding a facility to parse a text-format quirk 
> > description and add it to the compiled-in tables (if unique?).
> 
> Yes, I'll certainly think about doing it.  It would probably be best to do
> it via camcontrol.

That would be too late.

> It'll require mallocing memory in the kernel to store the additional quirk
> entries, and therefore could be abused by a malicious hacker.  But then
> again, it would be much easier to just send a nice format unit command..:)

Indeed.

> Another difficulty will be handling the fact that we've got a number of
> quirk tables, not one master quirk table like the old code.  This will
> probably require some thought/design to implement.

No problem. 

> This won't happen before 3.0.

I had no delusions about this. 8)

> > If you were to do this, it would be possible to read quirk entries from 
> > a file during the boot phase, decoupling the quirk data from the kernel 
> > code.  (This would eg. make it possible to boot/install on a system 
> > which wouldn't run without a quirk entry.)
> 
> Hmm, how early in the boot phase?  The earlier in the boot phase it goes,
> the more difficult it would be to do.  If you're talking about from
> /etc/rc, then sure, it could be done via camcontrol or something similar.
> Doing it before the CAM probe phase would be much more difficult.

Why?  You'd just have a set of linked lists of quirk tables, initialised
by a SYSINIT very early on.  This would know how to look for sources of 
quirk data, including the compiled-in "standard" tables, and possible 
extra sources.

The real plus here is decoupling the quirk data from the code itself; 
the two don't belong together.  You should, ideally, be able to add a 
line to a configuration file describing the quirk and have things "just 
work" without having to rebuild the kernel.

I realise this is not likely to be a high priority for you folks, and 
in reality it's probably not massively important, but it's one of many 
small steps towards organising the kernel.

-- 
\\  Sometimes you're ahead,       \\  Mike Smith
\\  sometimes you're behind.      \\  mike@smith.net.au
\\  The race is long, and in the  \\  msmith@freebsd.org
\\  end it's only with yourself.  \\  msmith@cdrom.com



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199810121936.MAA07717>