On Tue, 14 Dec 1999, Philip Blundell wrote:
> >That is a cakewalk to fix.........
> >I did not know that UDMA, ALI M5229 was on any other platform than x86.
> >This is only a ./drivers/block/Config.in issue.
>
> The ARM issue is not quite that simple. Even with Config.in fixed, at
> some point between 2.3.18 and 2.3.26 UDMA stopped working on these machines.
> With DMA enabled it seems that blocks read from the disk always contain
> zeros.
>
> I haven't had the time yet to review the changes that were made in that period
> to try to find the problem.
Can you tell me exactly when?
There was a point there ALI (OEM vendor support) started writing code for
me. It will ve a lighter task to bring you back to speed if I know if
that was the near the break point.
Andre Hedrick
The Linux IDE guy
