Hi,
  Ok just let let folks know that there is a new kernel 4338 available.

This kernel is running with re-written IRQ subsystem. Interrupts are no longer
executed in IRQ32 mode but in SVC32 mode instead. This should significantly
improve interrupt latencies as interrupts are now disabled for much shorter
periods.
Soft clock interrupts are now working find and are accounted so the number of
interrupts per second will appear to be higher than before.
Note: You will not get as many soft clock interrupts has hardclock interrupts
because sometimes softclock() will be called directly rather than scheduling a
soft clock interrupt.

The networking problems in some of the recent kernels should be fixed. I am no
longer having any problems with SLIP for example.

There is the possibility that the new interrupt code has shown up a bug in the
etherB driver. In the last couple of days I have had a number of receiver
errors that have locked the driver which has then failed to reset itself
requiring the interface to be taken down and the up again to fix things. This
may just be network problems triggering a bug that was already there. If any
body has problems please let me know.

This driver also has ATAPI support in it.

For the record I have not finished with the changes to the IRQ system. The next
stage is to remove the DMA and podule interrupts from the standard IOMD
interrupt masks and create separate interrupts masks for them.
This will then leave spare bits in the IOMD interrupt mask for the ARM7500
interrupt registers thus less condition compilation will be needed for this.
It will also mean that separate irq attachment functions will be provided for
attaching DMA and podule interrupts, more news when I write it.

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