At 10:07 PM 1/30/97 -0800, Jeremy Cooper wrote:
>On Thu, 30 Jan 1997, Kevin P. Neal wrote:
>
>> What causes a le0: memory error?
>
>I too, think that you might have a bad memory chip.
>
>The LANCE chip executes a DMA read roughly every 10 miliseconds during
>normal system operation.  If this cycle returns in an error, it will
>complain with this message.  There are only two ways that this could
>happen on the Sun4m.  The first is an I/O MMU misconfiguration, and 
>the second is just a plain 'old error detected in the read cycle.  This
>could be a hint that your RAM might have problems.

I borrowed my roommates copy of Solaris 2.4.

I'm now ftp'ing to it from a Win95 box, and I'm getting 7-12k per second.
I'm guessing that Solaris is having to reset the chip so often that it
just kills the speed. Anybody see any flaws in that thinking?

Yuck.

I also had Solaris fail to find 'ls' a second ago. This problem was 100
times worse with a bad SIMM, which is now out, but still indicates some
sort of RAM problem.

Since swapping SIMMs around does nothing to cure the problem, it looks
like I have a bad board.

Thanks for the info, guys. I now have something solid to tell the guy I got
my board from, so he'll replace it.

Thanks again!
--
XCOMM Kevin P. Neal, Junior, Comp. Sci.    -   kpneal@pobox.com
XCOMM  House of Retrocomputing:            -   kpneal@eos.ncsu.edu
XCOMM     http://www.pobox.com/~kpn/       -   kevinneal@bix.com
XCOMM "Rebooting with command" -- from a SPARCstation boot prom