Hmm,

I'm not seeing the first line of messages, but I'm seeing the rest of it,
often with varying amounts of "xfer":

Jan  8 16:21:34 alchemy /netbsd: dmaintr: DMA xfer of zero xferred 65509
Jan  8 16:21:34 alchemy /netbsd: esp0: !TC on DATA XFER [intr 10, stat 83, step
4
] prevphase 0, resid 1f0
Jan  8 16:21:34 alchemy /netbsd: dmaintr: DMA xfer of zero xferred 65493
Jan  8 16:21:34 alchemy /netbsd: esp0: !TC on DATA XFER [intr 10, stat 83, step
4
] prevphase 0, resid 7e0

This is on a Quadra 650 with kernel from 1.5Q sources as of about Jan 3rd.
These particular messages were seen while fsck'ing the filesystem during
boot.  I usually see these now when I try to make a kernel or do any compiling.

Not sure if that helps any or muddles the picture.

Steve

Allen Briggs wrote:

> > While using anoncvs I got the following messages on my quadra 650:
> >
> > esp_quick_dma_setup called with 141e000, 0, 0, 0
> > dmaintr: DMA xfer of zero xferred 65504
> > esp0: !TC on DATA XFER [intr 10, stat 83, step 4] prevphase 0, resid 15f0
> >
> > sd0(esp0:0:0): esp0: timed out [ecb 0x1043068 (flags 0x1, dleft 15f0, stat
> > 0)], <state 5, nexus 0x1043068, phase(l 3, c 3, p 3), resid 0, msg(q 0,o 0)
> > >
> >
> > I remember seeing something recently on the list about stuff like that.
> > I've reproduced this several times with different numbers here and there.
>
> There does seem to be a small epidemic of this going around.  I'm
> interested to know if the 65504 number is different for anyone, or
> if the second 0 in:
>
> > esp_quick_dma_setup called with 141e000, 0, 0, 0
>
> is ever a 1.  That would mean that this is happening on reads as well
> as on writes.
>
> Thanks,
> -allen
>
> --
>  Allen Briggs                     briggs@wasabisystems.com
>  http://www.wasabisystems.com/    Quality NetBSD CDs, Sales, Support, Service