[NCLUG] Looking for Linux device driver book

Tkil tkil at scrye.com
Mon Mar 6 02:29:31 MST 2006


>>>>> "Sean" == Sean Reifschneider <jafo at tummy.com> writes:

Sean> The O'Reily book is ok, though by definition it's dated.  Don't
Sean> know HOW dated currently, I read it years ago.  I know that Jon
Sean> Corbett did at least one update to it, maybe it's more up to
Sean> date now.

It's worth pointing out that:

1. Jon still lurks on blug at least.

2. The home page for LDD is handy, and claims it's current as of
   2.6.10, and has PDFs available:

      http://lwn.net/Kernel/LDD3/

   (Found through the magic of, yes, googling for "linux device
   drivers".  Second hit.  First was the O'Reilly page for the 2nd
   ed.)

3. Through LWN, Jon publishes great articles keeping current on the
   2.6 device development efforts / kernel API.  For cash flow
   consolidation purposes, I've had to let my subscription lapse, but
   LWN content is freely available after a week, so.  A good starting
   point is:

      http://lwn.net/Articles/2.6-kernel-api/

4. If something in his book or LWN articles is out of date and not
   marked as such, I suspect he'd appreciate hearing about it.
   Contact details on the above web site.

Sean> I must say that if I'm going to be doing C development, I'd
Sean> rather do it in the kernel than on the command-line.  In the
Sean> kernel, you get tracebacks!  Yay!  Kernel work isn't really that
Sean> tough.  It tends to be pretty easy to read code.  It's not
Sean> really any worse than hacking any other piece of code.  As I
Sean> said, in some respects, it's better.

It's been interesting following git and watching Linus stretch his C
legs in the "hey, i can just quit here, i don't have to dot every 'i'
and cross every 't'" sense.  Of course people went and are trying to
make 'git' as bullet-proof as the kernel, but still.

t.



More information about the NCLUG mailing list