On Sat, Jun 19, 1999 at 09:17:46PM -0500, Brian Stark wrote:
> Hello,
> 
> I'm getting some problems on a 1.4/i386 system (Pentium 166MHz) and maybe
> some here can tell me what is going on.
> 
> 
> PROBLEM:
> 
> While copying a large file (about 17.8MB in size) from a file system 
> located on a Seagate disk (disk wd1) to a file system on a Western 
> Digital disk (disk wd2) the following error messages appear in the 
> window for the xconsole program (X Window System is running):
> 
> pciide0:0:1: lost interrupt
>         type: ata
>         c_bcount: 32768
>         c_skip: 0
> pciide0:0:1: Bus-Master DMA error: missing interrupt, status=0x61
> wd1e: DMA error reading fsbn 275929 of 275929-275992 (wd1 bn 275992; cn 273 tn 12 sn 52), retrying
> wd1: soft error (corrected)
> pciide0:0:1: lost interrupt
>         type: ata
>         c_bcount: 32768
>         c_skip: 0
> pciide0:0:1: Bus-Master DMA error: missing interrupt, status=0x61
> wd1e: DMA error reading fsbn 520601 of 520601-520664 (wd1 bn 520664; cn 516 tn 8 sn 32), retrying
> wd1: soft error (corrected)
> 
> 
> This occurs on a single user system running NetBSD 1.4, but the problem has
> also been observed in the past when the machine was running NetBSD 1.3.3.

You had the IDE DMA patches on the 1.3.3 system then ? Or did you
just get "lost interrupt" without the "DMA error" messages ?

This is an hardware problem, caused by signal distortions on the IDE bus.
I assume your primary channel is properly cabled, with the controller at
one end of the cable and not in the middle.
What you can try is to remplace your IDE cable (if it's an old one maybe its
electric properties are not good enouth), and swap your seagate drive with
the WDC of the second IDE channel.

> Anyone know why this is happening?? Sometimes the system allows me to
> continmue after these messages come out, other times the system is
> "frozen" and I can't do anything. When that happens, I have to reboot and
> then deal with problems like fsck'ing the file systems.

The freeses are not normal, the driver should be able to recover from this
properly. Can you try to reproduce this in text mode ? Maybe there is a
panic hidden by the Xwindow system ?

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--