MiataGL doesn't boot in 5.4-RC2

[John Baldwin]

On Apr 15, 2005, at 6:39 AM, Bernd Walter wrote:

> On Fri, Apr 15, 2005 at 02:23:28AM +0200, Victor Balada Diaz wrote:
>> On Thu, Apr 14, 2005 at 08:22:17PM +0200, Wilko Bulte wrote:
>>> But I noticed something else in the system of the original poster:
>>> a TGA2 display card.  Any chance you can give it a try without that 
>>> one
>>> please?  There has been some fun with the various TGA display 
>>> adapters in
>>> the past.
>>
>> I tried without the TGA card and it hangs in the same place. I can 
>> provide
>> shell if needed.
>
> It has the card already done acording to boot log:
>   pci0: <display, VGA> at device 11.0 (no driver attached)
> The symptoms are the same as described by José M. Fandiño in
> <42444922.794080FB at fadesa.es> a few days ago.
> Next in boot order should be ata access.
> In José's case removing ata from kernel helped to boot, but he has
> devices attached that he needs.
>
> Victors machine has the IDE part of the Cypress chip probed the
> following:
> found-> vendor=0x1080, dev=0xc693, revid=0x00
>         bus=0, slot=7, func=1
>         class=01-01-80, hdrtype=0x00, mfdev=1
>         cmdreg=0x0045, statreg=0x0280, cachelnsz=0 (dwords)
>         lattimer=0x00 (0 ns), mingnt=0x00 (0 ns), maxlat=0x00 (0 ns)
>         intpin=a, irq=238
>         map[10]: type 4, range 32, base 00000170, size  3, enabled
>         map[14]: type 4, range 32, base 00000374, size  2, enabled
>         map[20]: type 1, range 32, base 80140000, size 16, memory 
> disabled
> found-> vendor=0x1080, dev=0xc693, revid=0x00
>         bus=0, slot=7, func=2
>         class=01-01-00, hdrtype=0x00, mfdev=1
>         cmdreg=0x0045, statreg=0x0280, cachelnsz=0 (dwords)
>         lattimer=0x00 (0 ns), mingnt=0x00 (0 ns), maxlat=0x00 (0 ns)
>         intpin=b, irq=239
>         map[10]: type 1, range 32, base 80158000, size 12, enabled
>
> You can see a mapping at 0x170 of 8 Bytes, but the second channel
> has no mapping at 0x1f0.
> Wilko, can you compare this with your PC164SX?
> Would be interesting if removing the ata hints will help, as they are
> the one refering to the ISA compatibility adresses.

Our PCI code in 5.x and newer makes certain assumptions about resources 
for ATA controllers as well, so you may want to forward this on to 
Soren and Warner to see if they have any ideas.

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org