Re: Errors on 2th ide channel of promise ultra100 tx2

[Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>]

On Friday 27 of February 2004 23:44, you wrote:
> On Fri Feb 27, 2004 at 09:14:23PM +0100, Bartlomiej Zolnierkiewicz wrote:
> > ide-disk.c sends WIN_READ_NATIVE_MAX_{EXT} without checking
> > if HPA feature set is supported, this is fixed in 2.6.x for a long time.
> >
> > We need 2.4<->2.6 IDE sync monkey... a really smart one...
>
> Dunno if I qualify as sufficiently 'really smart' enough but the
> last time I put in the considerable effort needed to re-sync the
> 2.4 and 2.6 IDE layers, and merge in the useful -ac bits that
> never made it into mainstream, nothing whatsoever came of my
> efforts...

Did you actually split and send out your patches? :)

> My 2.4.x patches are in daily use by a large group of people
> and they work fine, for what it is worth.  My IDE merging
> patches are the following:
>
>     http://codepoet.org/kernel/
>
>     020_ide_layer_2.4.22-ac4.bz2

Needs splitting and most of this stuff needs new re-sync with 2.6. :-(

>     021_ide_geom_hpa_capacity64.bz2

Now I remember why this wasn't applied.
It breaks braindamaged HDIO_GETGEO_BIG_RAW ioctl
(because changes way drive->cyls is calculated).
We workaround-ed it in 2.6 by removing this ioctl. :)
I think we really should do the same for 2.4.

>     022-extra-ide-drives.bz2

This hack to allow > 10 interfaces
is useless without additional major numbers.

>     025-cenatek.patch.bz2

IDE controller for solid state disks?  Cool.

Bartlomiej