On Fri, 23 Feb 2001, Peter Pentchev wrote:
> On Thu, Feb 22, 2001 at 01:19:33PM -0600, Michael C . Wu wrote:
> > On Thu, Feb 22, 2001 at 07:40:16PM +0100, Torbjorn Kristoffersen scribbled:
> > | Hi I'm using 4.2-RELEASE, with a parallel port ZIP drive (100M).
> > | Whenever I copy a large file from the zip drive (for example /dev/da0s1),
> > | the "cp" process eats 98% of the system resources. What's behind all this?
> > | Is there a way to fix it?
> > |
> > | 711 root 54 0 280K 168K RUN 0:45 93.87% 93.21% cp
> > |
> > | A 'renice' won't help.
> >
> >
> > That's natural with "parallel". No way around it.
>
> To clarify a bit, parallel port hardware depends on the system processor
> doing all the data transfers, every single byte, and spending even more
> time checking if it's time for the next byte to go. There's no DMA, there's
> not even a controller you can tell 'here's a 512-byte block, let it fly'.
>
> There's no way around it indeed.
>
> G'luck,
> Peter
So there doesn't exists any controllers (ISA/PCI) that can do the
serialization of parallel data, and pass it to a serial interface (or
UART), so we can use DMA and move Serial Data Units instead of
single bytes?
(Probably a feather-brained question..)
Cheers,
Torbjorn Kristoffersen
[EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
