>There I started Netscape Navigator 2.00, and then the computer hang.
>It really hung. Nothing worked.

Without reading the rest of your message I suspect an etherH card and a newish
kernel ...

>- EtherH16 on podule8 (Network slot) using TP

ah ;-)

>- bsd-4634-64MB kernel

what more can I say ;-)

>I booted singleuser and then went into multiuser with ^D, then logged
>in as me and started X. Logged into local Solaris box, started an
>X-term, started a PDM application, started Navigator.
>
>If someone has a clue about why this happened, I would like to
>hear it so that we could test and see why this happened.
>
>If someone wants me to do tests with new versions of X or kernels,
>then please let me know and I will test.
>
>I do not require an action on this as I only want to report that
>this was a problem that should be fixed some time in the future.

Ok This sounds like the duff etherH driver problem rather than a X problem.

EtherH problems :

What has happened is that as various parts of the kernel have improved
(particular IRQ stuff) problems have arisen with the etherH driver.
I first noticed it when my dev machine hung solid while serving X.
Since then I have noticed it using NFS and ftp.

(Currently I run on EtherB)

The problem is that if something goes wrong during DMA xfer with the card it
hangs the bus and thus the machine. The solution is to make sure the driver
never does anything wrong.
This used to be ok but with various irq changes (higher priority interrupt can
interrupt the DMA xfers (note they are DMA xfers as far as the NIC is concerned
but not DMA as far as the RPC is concerned).

The result is that sometimes a DMA xfer will fail jamming the bus and the
machine. At this point there is nothing left to do but reach for the blue
button.

Solution:
1. Fix the etherH driver
2. Write a new etherH driver

1. Acording to Nut fixing the current etherH driver is not practical (I was
thinking of doing this as a temporary measure - see (2)) as too much work would
be involved.

2. Nut has written a new etherH driver from scratch that is much better.

The problem, Although Nut has written a new driver, Nut was in London but his
RPC with the driver were elsewhere in the country. During Sept Nut was going to
be out of London (with is RPC) but would not have net connection.

So, we are waiting for Nut to return to London with the nice new driver so I
can slot it in and all the problems will go away. I am hoping that this will be
in a week or so.

Until then using the etherH driver with recent kernels (4444 and beyond) there
is a risk of the machine hanging if a xfer to/from the etherH card goes wrong.

Of course there could be something else that is wrong but I would suspect this
before I suspect X and other parts of the kernel...

Cheers,
				Mark

-- 
Mark Brinicombe				amb@physig.ph.kcl.ac.uk
Research Associate			http://www.ph.kcl.ac.uk/~amb/
Department of Physics			tel: 0171 873 2894
King's College London			fax: 0171 873 2716