Re: I have information/driver source for winmodem's (no Joke)
Matthew Vanecek (mev0003@unt.edu)
Tue, 13 Apr 1999 12:24:34 -0500
"Michael H. Warfield" wrote:
>
> Alan Cox enscribed thusly:
> > > And if available kernel interfaces are not enough, a patch in source form
> > > that provides the needed interfaces will do the trick.
>
> > Probably only if accepted into the mainstream code.
>
> > I really doubt you want to run a softmodem in kernel space anyway. Its
> > MMX and float heavy, its large (they tend to be 200K+ code) and potentially
> > at that size and for what it has to do very buggy (yes you _can_ have
> > buffer overruns in soft modems 8)).
>
> > Its probably a real time process job
>
> Besides... That way "top" will show you just how much of a CPU
> hog the modem logic is and why a "winmodem" is really a "losemodem" because
> you "lose" more in CPU than you "win" in a cheap crappy piece of hardware.
>
> Other than attempting to profile the kernel code (which I've
> never done) I don't know of any other way to comparmentalize what
> code is burning your CPU if it's running, say, 80% or more CPU. That
> would just make Linux look bad. Make the winmodem logic run in a
> real time process and then it looks bad because it's burning up the CPU,
> not the kernel. Purely from support and appearances sake (why does my
> Linux box run so slow?) this looks like a better way to place preformance
> blame on the performance culprits.
>
> I like that idea. :-)
Yes, that would probably be the best idea. I think there are more
important issues in kernel development than trying to include support
for losemodems. If there is to be support for it, it should be
userspace, not kernel space. My personal opinion, for what it's worth,
is that we should boycott losemodems, and refuse them even on prebuilts.
I.e., demand quality for the money we shell out to vendors.
I don't think it would be a bad thing to develop a module, separate from
the kernel, though, for those poor souls stuck with one of those idiotic
modems, and too broke to buy a decent one. But it *should* be separate,
like some others that I use (e.g., lm_sensors.o), especially if it's not
going to be distributed with the source code.
--
Matthew Vanecek
Course of Study: http://www.unt.edu/bcis
Visit my Website at http://people.unt.edu/~mev0003
For answers type: perl -e 'print
$i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'
*****************************************************************
For 93 million miles, there is nothing between the sun and my shadow
except me. I'm always getting in the way of something...
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/