Introducing Myself, Looking to Learn
arlie at worldash.org
Tue Sep 3 21:33:59 EDT 2013
On Sep 04 2013, Varad Gautam wrote:
> I want to start working on the Linux kernel but am an absolute
> beginner. I am currently on my way through Robert Love's Linux
> Kernel Development and need help with finding something I can work
> on to get a hang of what it's like.
How much of a beginner are you? In particular, how much do you know
about operating system kernels in general? And how much
multi-threaded programming have you done, in situations where you have
to manage your own locking?
If you don't have good practical understanding of concurrency, you
might want to do yourself a favour and do some work with a large
multi-threaded program until you've got practical experience with race
conditions, lock contention, and deadlocks.
Linux is a fairly mature monster. This means that the simple outlines
of what all OSes do tend to be obscured by layers of complexity.
That's one some folks use BSD for their OS-kernel classes. You'll be
doing more digging in linux. Will you be comfortable with that?
> I have also subscribed to the LKML, but find it completely
My question would be "incomprehensible how?" There's a lot of shared
context, which you won't have, and diffs aren't the best way to learn
what's in a piece of code. That level of "incomprehensible" makes
sense to me.
Beyond that, there's learning you need to do.
> As a beginner, would it be better to work with the
> kernel of a specific OS (I'm running Ubuntu), or work on the
> upstream kernel?
Pick one that you actually run - that way you're set up to try out the
results of your experiments, and more likely to encounter some problem
that motivates you to attack it. Personally I'd prefer not to be
testing on my development system [hate it when that's unstable], but
if you aren't playing with hardware, VMs make great test
Beyond that, my advice would be to do something trivial. How about a
kernel module that prints "Hello world"? It's useless, but you'll
start internalizing the build system. Then do something else trivial,
that let's you poke the kernel somewhere else. Eventually either it'll
start making sense, and your ambitions will increase - or you'll
decide you aren't having fun and try somethign else.
(Arlie Stephens arlie at worldash.org)
More information about the Kernelnewbies