Strange DMA behavior under 5.3-RELEASE

[Sean Farley]

On Wed, 17 Nov 2004, Sean Welch wrote:

> I originally sent a message about this to the questions list but I've
> not received any responses.  Possibly the people who could answer my
> question do not read that list.  Hopefully someone here can help!
>
> The message in question can be viewed here:
>
> http://docs.FreeBSD.org/cgi/mid.cgi?20041115000519.GA2127
>
> Is anyone else seeing this odd "half way" behavior?  The hardware
> seems to take the atacontrol command just fine but somehow the mode
> isn't being set at boot even though the proper sysctl is set.

That is odd.  I can see it happening on my old Dell PII.  The sysctl
only enable DMA capability I believe.  It is enabled by default.  If you
disable it during boot, I would guess you would not be able to set WDMA2
with atacontrol.

dmesg:
atapci0: <Intel PIIX4 UDMA33 controller> port
0x1420-0x142f,0x376,0x170-0x177,0x 3f6,0x1f0-0x1f7 at device 7.1 on pci0
ata0: channel #0 on atapci0
ata1: channel #1 on atapci0

atacontrol list:
ATA channel 0:
     Master:  ad0 <IBM-DTTA-371440/T71OA70B> ATA/ATAPI revision 4
     Slave:       no device present
ATA channel 1:
     Master: acd0 <HITACHI GD-2000/0056> ATA/ATAPI revision 0
     Slave:  afd0 <IOMEGA ZIP 100 ATAPI/23.D> ATA/ATAPI revision 0

I can later set it to WDMA2 via atacontrol.

Also, I get from the Zip drive:
ata1-slave: FAILURE - SETFEATURES SET TRANSFER MODE status=1<ERROR>
error=4<ABORTED>

Sean
-- 
sean-freebsd at farley.org