Re: 2.1.102, IDE DMA and OPTi 82C621

Trevor Johnson (
Mon, 18 May 1998 10:24:55 -0700 (PDT)

> > I have a PC with an Amptron PM-7400 motherboard. The OPTi 82C621 chip is
> > built onto it. According to the author of the Linux driver for this chip,
> > it does not support IDE DMA. I compiled Linux 2.1.102 with
> I am the author of the Linux IDE DMA driver,
> which is *the* driver for all IDE PCI DMA chipsets,
> and I don't remember you asking me that.

I had asked Jaromir Koutek, the author of the OPTi 621 driver for Linux

> In front of me, is the OPTi 82C621 chip technical documentation
> (from OPTi), clearly showing full support for IDE Bus-Master DMA.
> ...
> > 00:14.0 IDE interface: OPTi Inc. 82C621 (rev 11) (prog-if 80)
> > Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop-
> Looks like BusMaster DMA to me.


> Perhaps there is a problem with the drives or BIOS intialization.

How can I identify such problems? Do you see anything suspicious in the
output from "hdparm -i /dev/hda /dev/hdc"? It looks to me as though
hdparm is saying both my drives can do DMA. They are different brands,
are on separate IDE buses, and have apparently been working fine in PIO
mode (apart from data being lost from my PPP sessions). DMA in only
mentioned briefly in Documentation/ide.txt--basically a statement that it
is supported. Are you saying that the BIOS must initialize something that
I am not able to override with hdparm? It's an Award BIOS, version 4.51g,
and prints a string containing "OPTi Viper" and "2A5UNH01C-00" at the
bottom of the screen at boot time. It has an "IDE Second Channel Enable"
option, an option to have the parallel port use DMA or not, settings for
two "fast" DMA channels, and other options that seem even less relevant.
Trevor Johnson

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to