On Sun, Feb 25, 2001 at 07:44:07PM +0000, Doros Eracledes wrote:
> I have a used HP 7100 cd-writer and i'm trying to make it work .
> After creating a test iso image wirh mkisofs , i used 
> " cdrecord -v dev=/dev/rcd0d /cd.iso   " command to burn the cd and i
> get the following: 
> 
> Feb 25 18:17:57 rhino /netbsd: pciide0:1:0: lost interrupt
> Feb 25 18:17:57 rhino /netbsd:  type: atapi tc_bcount: 60 tc_skip: 0
> Feb 25 18:17:57 rhino /netbsd: pciide0:1:0: bus-master DMA error:
> missing interr    upt, status=0x20
> Feb 25 18:17:57 rhino /netbsd: pciide0:1:0: device timeout, c_bcount=60,
> c_skip=    0
> Feb 25 18:17:57 rhino /netbsd: cd0: transfer error, downgrading to PIO
> mode 3
> Feb 25 18:17:57 rhino /netbsd: cd0(pciide0:1:0): using PIO mode 3
> Feb 25 18:21:57 rhino /netbsd: pciide0:1:0: lost interrupt
> Feb 25 18:21:57 rhino /netbsd:  type: atapi tc_bcount: 28 tc_skip: 0
> Feb 25 18:21:57 rhino /netbsd: pciide0:1:0: reset failed
> Feb 25 18:22:02 rhino /netbsd: pciide0:1:0: lost interrupt
> Feb 25 18:22:02 rhino /netbsd:  type: atapi tc_bcount: 0 tc_skip: 0
> Feb 25 18:22:02 rhino /netbsd: pciide0:1:0: piomode timed out
> Feb 25 18:22:02 rhino /netbsd: pciide0:1:0: reset failed
> 
> 
> I believe that the problem is with DMA transfers . i compiled a new
> Kernel with 
> 
> pciide* at pci? dev ? function ? flags 0xf

this is wrong, the pciide flag is only used to force the generic DMA code
to be used on unsupported chipsets.

> cd*     at atapibus? drive ? flags 0xf 

'flags 0xff0' would be better.

--
Manuel Bouyer <bouyer@antioche.eu.org>
--