Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 14 Nov 1997 17:49:28 +1100
From:      Bruce Evans <bde@zeta.org.au>
To:        bde@zeta.org.au, mouth@ibm.net
Cc:        hackers@FreeBSD.ORG
Subject:   Re: Status of 650 UART support
Message-ID:  <199711140649.RAA20946@godzilla.zeta.org.au>

next in thread | raw e-mail | index | archive | help
>>>Why not start from scratch and develop siov2.c which uses elastic
>>>buffers, 650 polled vs. interrupt mode switching,  yada, yada, yada.
>>
>>High costs/benefits.  It can't be made more than about 10% faster in
>>that way on a reasonably fast CPU, since most of the overheads are
>>for waiting for the ISA bus.
>
>I recall reading that 16-bit ISA I/O is faster, per byte, than 8-bit
>ISA I/O.  I  wonder how much better you could do with a card which
>transferred two bytes at once on the 16-bit bus.  You would just need

It's about twice as fast.  Under Linux, Rocketport devices are reported
to be about twice as efficient as the 16550 devices.  This is possible
because they are 16-bit devices.  I haven't seen any reports of the
efficiencies of Rocketports under FreeBSD but expect they are less
efficient than 16550s for termios because they use the very inefficient
ttyinput() routine.  They should be better for slip and ppp.

>some kind of buffer between the UARTs and the bus.  Maybe even shared
>memory ... oh well I guess I could buy a Cyclades card or such rather
>than go to all that trouble.

Be sure to buy one with a fast bus interface.  My old ISA one is about
as efficient as a 16550.

Bruce



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199711140649.RAA20946>