SV: Accessing serial ports from a loadable module (from kernel space)

From: Jörgen Overgaard (overgaard@mailbox.ffc.se)
Date: Tue Feb 22 2000 - 16:48:49 EST


Hi Tommi,

Yes, you're absolutely right! :-)

I've already decided to do it in user space instead by using a
named pipe and a daemon.
Much more simple! :-) Hihi

I have some other project where I can "learning by doing".
For example the antenna which is controlled from an I/O ISA card.

Anyway... Thanks to all who been answering my mail, and I hope I will be
as much welcome when I start digging into the kernel further on. :-)

Best rergards and 73's de SM4WWG

P.S.
    For thoose interested I will post a message on the mailing list when I'm finnished, 'cause I think
    there might be more than me who want's to use SatTrack to controll the radio via CI-V
    (using the CT-17, alike or howbrew levelconvertet)
D.S
 
-----Ursprungligt meddelande-----
Från: Tomi Manninen <tpm@prkele.tky.hut.fi>
Till: linux-hams@vger.rutgers.edu <linux-hams@vger.rutgers.edu>
Datum: den 22 februari 2000 22:30
Ämne: Re: Accessing serial ports from a loadable module (from kernel space)

On Tue, 22 Feb 2000, Ivo Simicevic wrote:

> It is always better to have driver running in kernel space. That is why
> it is called driver ?

No. A "driver" like the one Jörgen talks about only needs to be in kernel
space if it needs to access special hardware (ie. needs to receive
interrupts, needs DMA etc) or has special timing requirements that cannot
be met in user space.

Since from what I gather Jörgen's "driver" would only be a simple shim to
translate from SatTrack commands to CI-V commands and it only needs to
talk to a serial port (for which there already is a very sophisticated
driver in the kernel) I don't see a reason to do that in kernel space.

In user space the implementation would be much simpler to do and much more
flexible (like said it could talk to all kinds of serial ports including
USB serial ports for example) and easier to maintain.

Learning by doing is of course always a good reason to do things that
otherwise aren't that reasonable but I suspect this one would be so simple
to do in user space and so difficult to do in kernel space that it makes
me wonder...

-- 
Tomi Manninen           Internet:  oh2bns@sral.fi
OH2BNS                  AX.25:     oh2bns@oh2rbi.fin.eu
KP20ME04                Amprnet:   oh2bns@oh2rbi.ampr.org

- 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/



This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:31 EST