Re: [PATCH] Re: pdc202xx_old broke boot [was Re: 2.6.9-mm1]

[Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>]

On Wed, 27 Oct 2004 22:19:52 +0000, J.A. Magallon <jamagallon@able.es> wrote:
> 
> On 2004.10.23, Bartlomiej Zolnierkiewicz wrote:
> > On Fri, 22 Oct 2004 17:21:00 -0700, Andrew Morton <akpm@osdl.org> wrote:
> > > "J.A. Magallon" <jamagallon@able.es> wrote:
> > > >
> > > > Hi all...
> > > >
> > > > On 2004.10.22, Andrew Morton wrote:
> > > > >
> > > > > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.9/2.6.9-mm1/
> > > > >
> > > >
> > > > I upgraded from 2.6.9-rc3-mm3 to 2.6.9-mm1 and the system coould not boot.
> > > > What was before hde now was hda (guess ? root is on hde1...)
> > >
> > > yikes.  Perhaps the PCI scanning order was changed?
> >
> > Fortunately, not. :)
> >
> > What happened is that ide-dev-2.6 tree contains a patch which ignores
> > BIOS settings for Promise controllers but ide-dev-2.6 tree is not in
> > 2.6.9-mm1 (due to syncing with -linus -> temporary breakage).

^

> > > > How can I restore the old behaviour ? Plain 2.6.9 booted. So reconfiguring
> > > > fstab to say / == hda1 will make impossible switch between kernels ...
> >
> > In 2.6.9-mm1 CONFIG_PDC202XX_FORCE option can also be used
> > for pdc202xx_old but pdc202xx_new must be enabled (yes it a bug).
> >
> 
> Is this patch OK ? (against 2.6.9-mm1)

Nice attempt but...

There shoud be either separate config options or no options et all.
ide-dev tree contains the latter fix, I just need to sync/rebuild ide-dev
so akpm can merge it into the next -mm kernel.

> --- linux/drivers/ide/Kconfig.orig      2004-10-26 01:26:09.000000000 +0200
> +++ linux/drivers/ide/Kconfig   2004-10-28 00:15:27.936420036 +0200
> @@ -624,6 +624,8 @@
> 
>           Please read the comments at the top of <file:drivers/ide/pci/ns87415.c>.
> 
> +menu "Promise PDC support"
> +
>  config BLK_DEV_PDC202XX_OLD
>         tristate "PROMISE PDC202{46|62|65|67} support"
>         help
> @@ -648,9 +650,15 @@
> 
>           If unsure, say N.
> 
> +config BLK_DEV_PDC202XX_NEW
> +       tristate "PROMISE PDC202{68|69|70|71|75|76|77} support"
> +
> +config PDC202XX_DUMMY
> +       bool
> +
>  config PDC202XX_BURST
>         bool "Special UDMA Feature"
> -       depends on BLK_DEV_PDC202XX_OLD
> +       depends on BLK_DEV_PDC202XX_OLD || BLK_DEV_PDC202XX_NEW
>         help
>           This option causes the pdc202xx driver to enable UDMA modes on the
>           PDC202xx even when the PDC202xx BIOS has not done so.
> @@ -665,15 +673,20 @@
> 
>           If unsure, say N.
> 
> -config BLK_DEV_PDC202XX_NEW
> -       tristate "PROMISE PDC202{68|69|70|71|75|76|77} support"
> -
> -# FIXME - probably wants to be one for old and for new
>  config PDC202XX_FORCE
>         bool "Enable controller even if disabled by BIOS"
> -       depends on BLK_DEV_PDC202XX_NEW
> +       depends on BLK_DEV_PDC202XX_OLD || BLK_DEV_PDC202XX_NEW
>         help
> -         Enable the PDC202xx controller even if it has been disabled in the BIOS setup.
> +         Enable the PDC202xx controller even if it has been disabled, both
> +         manually in the BIOS setup, or because it has no drive connected.
> +         If you do not force the detection and have other IDE busses
> +         in the box, the device names for your other busses (/dev/hdX)
> +         can change depending on drives being connected to the Promise or not.
> +
> +         If unsure, say Y. This will keep your IDE drive device names
> +         consistent with or without drives connected to the PDC.
> +
> +endmenu
> 
>  config BLK_DEV_SVWKS
>         tristate "ServerWorks OSB4/CSB5/CSB6 chipsets support"
> 
>