Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 17 Jan 1996 08:17:08 -0600
From:      Rob Snow <rsnow@lgc.com>
To:        Terry Lambert <terry@lambert.org>
Cc:        Wes Santee <wsantee@wsantee.oz.net>, questions@freebsd.org
Subject:   Re: Setting for Promise 2300+ controller?
Message-ID:  <30FD04E4.6956@lgc.com>
References:  <199601151815.LAA27876@phaeton.artisoft.com>

next in thread | previous in thread | raw e-mail | index | archive | help
[ I missed the first message, hence the follow up to Terry's message ]

I've been using a 2300+ for close to a year now and have found the
following info which might be of use.

Turn off LBA in BIOS and on the 2300+ BEFORE you FDISK and format
a drive.  The Promise docs say something to the effect of:  "If you
FDISK and format a drive with LBA on then it'll do LBA from then
on no matter what you do with the jumper and vice-versa".  This might
be BS however it seems to work similar to that here.

Turn off the onboard BIOS on the 2300+.  The __CRASH_OF_95__ was
caused by jacking with the BIOS.

Set your flags for your drive in your kernel config to turn on 32bit
access and multi-block.  (Flags documented in LINT, I believe I use
80ff for my 0 drive (32bit and biggest possible multi-block?))


Terry Lambert wrote:
> 
> > I asked awhile back on the -stable mailing list what I needed to
> > know about installing a Promise 2300+ EIDE controller in my FreeBSD
> > 2.1-stable box considering the LINT kernel config file says EIDE
> > drives aren't supported.  The only responses I got were people
> > asking me to pass along the information if I figured it out.
> >
> > I've seen people asking questions here about Promise 2300+
> > controllers and ATAPI CD-ROM drives.  To those in the know, what
> > were the settings you set your card at, and what options did you
> > have to compile your kernel with to get the sytem to boot correctly?
> > I don't have an ATAPI CD-ROM drive, just the said controller.
> >
> > I have standard IDE drives formatted using CHS so I've turned off
> > LBA mode on the card (is LBA mode even supported considering the
> > comments from the LINT config file?).  With those setting, my DOS
> > partition boots fine, but the FreeBSD partition stops after about 2
> > rotations of the spinner.
> >
> > If anybody has any info on this, it sounds like more than just
> > myself is looking for the answers.  Any help appreciated.
> 
> OK.  First off, "LBA mode" is irrelevant.
> 
> All LBA refers to is linear sector addressing.
> 
> LBA mode on the controller means that the controller supplies the
> extended interface.
> 
> No DOS software uses this extended interface.
> 
> Instead, there must be an INT 13 redirector, either in the BIOS or
> in OnTrack DiskManger or similar software that cause:
> 
> 1)      Geometry translation.
> 2)      Translated C/H/S values at the INT 13 interface to be
>         computed into LBA values based on the translation and
>         the LBA interface used to actually do the disk access.
> 
> This means that, like SCSI, there is no way to exceed the bit
> limitation on the disk size after translation (8G @ 512b sectors)
> in the DOS INT 21/INT 13 framework.
> 
> Now the BSD second stage boot (loaded by the MBR or MBR replacement,
> like OS-BS) uses the DOS INT 13 interface (and thus can use LBA) to
> load the kernel.
> 
> The problem occurs when the kernel itself comes up and the geometry
> is different than INT 13->LBA translator says the geometry is.
> 
> You can fix this by hacking the sector offsets in the disklabel to
> agree with the values you get using the translated geometry.
> 
> You have to do this right after install before the reboot on the
> Alt-F2 screen (the setup program *might* let you do this inside
> of it; I don't know).
> 
>                                         Terry Lambert
>                                         terry@lambert.org
> ---
> Any opinions in this posting are my own and not those of my present
> or previous employers.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?30FD04E4.6956>