ata regression

[Andrew Pantyukhin]

On 4/9/07, Andrew Pantyukhin <infofarmer at freebsd.org> wrote:
> After today's update ata(4) is unusable on both my
> desktop (i386, via chipset) and my laptop (amd64,
> ati chipset).
>
> Everything works fine with kernel from 1 April.
>
> VIA box says:
> ata0: reset tp1 mask=03 ostat0=58 ostat1=00
> ata0: stat0=0x80 err=0x80 lsb=0x80 msb=0x80
> ...
> ad0: setting PIO4 on 8237 chip
> ad0: setting UDMA100 on 8237 chip
> ata0: reinit done ..
> ad0: FAILURE - WRITE_DMA time out LBA=8511
> g_vfs_done():ad0s1a[WRITE(offset=4325376, length=2948)]error = 5
> <delay>
> <once again>
>
> ATI says something similar, but drops to kgdb after the
> first cycle.
>
> Sorry, I have no time to provide more info right now.

Oops, didn't see the cvs-src thread, sorry. Will report
if anything stays broken after another update.

Cheers!