[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: pipes & ptys



Warwick Allison
>
> > number of bytes in a pipe. So at least my reading program will not be the
> > bottleneck, if the writing one choses to write in 1 byte chunks, I loose...
> 
> Why?  It seems unlikely that your reader will run after every writer
> writes once.  Normally, you'd expect the writer to be writing into the
> buffer (possibly buffered on their end, but maybe not), and when your
> reader process runs, there will be a few chars in the buffer.  I think
> the timeslice is 20ms or so normally.

Since task switch is done in the VBL interrupt, it occurs app. 60 times a
second on my TT. That makes 16ms per process, if everything is running
"normally". Suppose only my (multibly-byte) reader and my (one-byte) writer
are running, say, ignore the rest, the writer doesn't really have a big chance
to write very much during one time slice, I would estimate app. 30 bytes.

That's what I was trying to say. If the writer outputs in larger chunks, ok,
if the reader waits for more than one timeslice, also ok, but if not, I
loose... :-(

> > But what we _really_ want is a extra device driver for each modem device that
> 
> I haven't really looked at it, but some MiNT modem stuff just appeared at 
> atari.archive.

Wouldn't be the first time I'm coming too late to invent the wheel! ;-)

TeSche