Re: Compiling an embedded driver

[Martin Fuzzey <mfuzzey@xxxxxxxxx>]

Roger wrote:
> My studies of the driver and your info has convinced me
> that I don't have to do anything with the a91_serial.c driver.
> Instead I should modify the module that lies above and
> where the character replacement takes place. Perhaps add
> a special ioctl to disable this modifying of the input
> data. I will you use the serial port to
> transfer binary data so any tampering with data is not
> acceptable.
>   
Just set it to IGNPAR.

If you really _need_ to do stuff in the kernel (and I agree with Russell
and others who say that for your use case you do _not_) you can probably
do it without _modifying_ anything by creating a new line discipline
(again this is an overkill for the current requirement).

Indeed the standard parity error marking code is implemented in the
default line discipline (n_tty.c  n_tty_receive_parity_error()).  Line
disciplines are "kernel plugins" for the tty layer. See
http://lwn.net/Kernel/LDD3/   [chapter 18]

> When I write for Windows I find all info I need in
> Microsofts MSDL.
>   
Well I guess that varies. I have had the misfortune to do some windows
work too and curse it every time I can't find the info in the MSDN (or
it's wrong) and I can't just grab the source and see how it _really_ works.

> The success of a Linux project seems to depend on
> who you know, or to hunt down the one man that has
> that piece of information you need.
>   
There are plenty of books on linux / device drivers / embedded linux
(some even available online like the link I posted above).
http://www.linuxdevices.com/articles/AT2969812114.html  lists a few.

Martin


-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php