Hi,
The attached patch cleans up the 'autotune' concept used in the current 2.4
IDE driver. It also adds support for using pure BIOS IDE timings with DMA/PIO.
On some systems the BIOS has a far better overview on how things are connected
(some chipsets don't support >ata66 speed detection, etc).
The patch introduces 2 new boot parameters in addtion to the current two
autotune options (autotune vs. noautotune):
ide0=autotune -> kernel IDE timing setup
ide0=noautotune -> BIOS IDE timing setup (PIO only)
ide0=noautotune_dma -> BIOS IDE timing setup (DMA/PIO)
ide0=noautotune_force -> BIOS IDE (DMA/PIO), chipset *cannot* override.
Default is autotune (as always) and chipsets can override the other options
unless *_force is used.
All chipset drivers updated. Tested with PIIX and Serverworks.
Regards,
Torben Mathiasen
On Sat, 2002-11-02 at 00:02, Torben Mathiasen wrote:
> Hi,
>
> The attached patch cleans up the 'autotune' concept used in the current 2.4
> IDE driver. It also adds support for using pure BIOS IDE timings with DMA/PIO.
> On some systems the BIOS has a far better overview on how things are connected
> (some chipsets don't support >ata66 speed detection, etc).
>
Interesting idea. However you are working on whats effectively a dead
codebase.
On Sat, Nov 02 2002, Alan Cox wrote:
> On Sat, 2002-11-02 at 00:02, Torben Mathiasen wrote:
> > Hi,
> >
> > The attached patch cleans up the 'autotune' concept used in the current 2.4
> > IDE driver. It also adds support for using pure BIOS IDE timings with DMA/PIO.
> > On some systems the BIOS has a far better overview on how things are connected
> > (some chipsets don't support >ata66 speed detection, etc).
> >
>
> Interesting idea. However you are working on whats effectively a dead
> codebase.
Oh :). Let me just port it over to AC and 2.5 for the new IDE layer.
Torben