>- What is the status of the EBSA-285 port? Have most of the bugs been worked
>out?
It seems stable enough. Most of the code is shared with the NetWinder and
CATS ports. There have been reports recently of lockups under heavy network
activity but that's the only problem I'm aware of.
>- Although there are Linux drivers for the 53C895 and 82559 (82558, which is
>pin/software compatible), what problems will we have in porting the drivers
>over to the platform?
They'll need appropriate cache synchronisation primitives adding, since the
StrongARM doesn't do this in hardware. Other than that there shouldn't be any
problems. PCI drivers are usually pretty well behaved.
>- It was mentioned that the Tulip driver (Intel's ethernet chipset) has bugs
>in it. Will those bugs be relevant to our SCSI/Ethernet port?
The Tulip is the 21040 family that Intel bought from Digital. Regardless of
whether it's buggy or not (I've never had ay serious problems with it) the
82558 won't be affected.
>- What is the general feeling on the performance of the StrongArm?
Depends too much on your application. If you want to try for yourself you can
beg, steal, borrow or buy one of the various StrongARM machines and have a play
with it. Typically you get 200+MIPS of integer performance. I don't know if
anybody's measured floating point speed with the currently available emulators.
>What is the story with Intel? My supplier says Intel is not quoting lead
I can't answer that. MIPS would probably be my next choice as well, or maybe
PowerPC. Neither of those, as far as I know, will give you anything like the
integer performance of the StrongARM at the same price, though in terms of
floating point performance and integrated peripherals they do rather better.
Again it all depends on your application.
p.
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
