'borrowing' 6.1 drivers for use in 6.2 (sata)

[Steve Franks]

On 2/3/07, Ted Mittelstaedt <tedm at toybox.placo.com> wrote:
> Hi Steve,
>
>   I should have been more clear, sorry!  You cannot just replace a .ko file
> that is a binary file.  You have to install the kernel sources, then go into
> the
> directory that the driver .c and .h files are, back those up, and replace
> them,
> then recompile the kernel.
>
>   Here's probably what you want to do:
>
> Go to here:
>
> http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/ata/
>
> examine the files.  For example, assume it was ata-disk.c  Examine the
> history
> by clicking on the file you will see the different changes.  For example
> according to this revision 1.189.2.5  was used for 6.2-release, revision
>  1.189.2.4  was used for 6.1-release.  If the bug is in this file it was
> introduced
> in one of the intermediate revisions between these 2.
>
> You will note this cooresponds with:
>
> /sys/dev/ata
>
> on your system assuming you installed the kernel sources.  If you wanted to
> backrev this driver you would install sources, recompile the kernel and
> install it and reboot to make sure you know how to build a running
> kernel, then you would go to this directory, replace the file you want
> (ata-disk.c for example) and recompile the kernel and install it and reboot
> and
> cross your fingers.  Rebuilding the kernel is covered in the online handbook
> on the website.
>
> It most likely won't work the first time.  You will need to try this with
> several files.  You also want to try newer revisions of this file.  Or you
> can
> get bold and tar up the entire directory from a 6.1-release system and
> move /sys/dev/ata to /sys/dev/ata.bak and untar the old directory into your
> 6.2 system than try that.  Obviously you need a 6.1 system around for this.
>
> If Soren has a guess to this he may send you patches to one or more
> of these files for testing, you will need to apply them and rebuild and
> reboot to see
> if it works.
>
> This sort of thing can take a lot of time.  But it is really critical that
> you get
> on it right away.  Since right now there wern't many changes between
> the drivers and it will be very easy to narrow it down to precisely what
> revision change causes the problem.  It is much worse when people post
> things like such-and-such piece of hardware worked under FreeBSD 2.2
> 5 years ago and not under 6.2 today.
>
> Ted
>
> ----- Original Message -----
> From: "Steve Franks" <stevefranks at ieee.org>
> To: "Ted Mittelstaedt" <tedm at toybox.placo.com>
> Cc: "FreeBSD Users Questions" <freebsd-questions at freebsd.org>
> Sent: Saturday, February 03, 2007 10:19 AM
> Subject: Re: 'borrowing' 6.1 drivers for use in 6.2 (sata)
>
>
> > Ted,
> >
> > I emailed sos at freebsd.org with that info, since that was who was under
> > "man ata".
> >
> > In the meantime, I replaced /boot/kernel/atapci.ko with the one from
> > my 6.1 install cd, and it seriously #$%#^'ed my system.  Restored now,
> > but, do you suggest I need ata*.ko, not just atapci.ko?  Or am I on
> > the wrong track entirely?  Also, where do I find info on the server
> > and folder structure I should be looking for in cvs when you mention
> > getting the latest thing?
> >
> > Steve
> >
> > On 2/3/07, Ted Mittelstaedt <tedm at toybox.placo.com> wrote:
> > >
> > > You shouldn't have any trouble downgrading to the 6.1 drivers but
> > > I would suggest instead that you e-mail Soren the maintainer of the SATA
> > > driver before doing anything.  He may have patched it already.  At
> least,
> > > try the current driver from CVS first.
> > >
> > > All you really have to do is copy the current driver to a backup file
> > > then download the older driver from the cvs webinterface on freebsd.org
> > > and recompile your kernel.
> > >
> > > If you really want this fixed, find the exact revision in cvs where the
> > > support broke, obviously it will be between 6.1-release and 6.2-release,
> > > and let Soren know.  It will take a few kernel recompiles to do that.
> > > Fortunately your system boot isn't on the controller, not many folks I'm
> > > sure have that setup.  When you do figure it out, please create a PR
> > > using send-pr
> > >
> > > Ted
> > >
> > > ----- Original Message -----
> > > From: "Steve Franks" <stevefranks at ieee.org>
> > > To: "FreeBSD Users Questions" <freebsd-questions at freebsd.org>
> > > Sent: Saturday, February 03, 2007 7:57 AM
> > > Subject: 'borrowing' 6.1 drivers for use in 6.2 (sata)
> > >
> > >
> > > > If anyone has read my earlier posts, they know the acer/uli/asus
> > > > motherboard sata chipset simply won't work in 6.2.  It worked fine in
> > > > 6.1, so I wonder if I can steal the driver from a 6.1 install?  Or do
> > > > I just have to downgrade?  Having my server offline (even though it's
> > > > only my personal files) is simply not an option.  FYI, several others
> > > > have complained about the same chipset with no resolution, so I'm
> > > > fairly certain it's not something I did.
> > > >
> > > > Steve
> > > > _______________________________________________
> > > > freebsd-questions at freebsd.org mailing list
> > > > http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> > > > To unsubscribe, send any mail to
> > > "freebsd-questions-unsubscribe at freebsd.org"
> > > >
> > >
> > >
> >
> >
> > --
> > Steve Franks, KE7BTE
> > Staff Engineer
> > La Palma Devices, LLC
> > http://www.lapalmadevices.com
> > (520) 312-0089
> >
>
>

Ted,

Being the naive sort, after my earlier failure, I decided that maybe
atapci.ko from the 6.1 just wasn't compatible with the 6.2 ata.ko, so
I restored everything, and replaced ata.ko, atapci.ko, and atadisk.ko
with the ones off of the 6.1 disk.  Reboot, and !!! now, it boots
instead of a big crash like with just 6.1-atapci.ko, but **it exhibits
exactly the same problem as with the 6.2 ata drivers** !!!  Now, maybe
I'm still being naive, and that's just a bogus thing to do, but I
thought it could be a useful piece of information.  Between that and
the "reset failure - error 6", I'm thinking someone familiar with the
drivers might just be able to narrow this down.  I'm beginning to
suspect it is some other aspect of the system fouling the ata drivers,
possibly not the drivers themselves....?

Steve