Hi, On 4/18/07, Shakthi Kannan <shakthimaan@xxxxxxxxx> wrote:
Hi, On 4/18/07, black hole <neyrith@xxxxxxxxx> wrote: > I did not manage to google out anything interesting > on this topic. Something like this? http://www.gelato.unsw.edu.au/~dsw/public-files/kernel-docs/kernel-api/index.html
Actually no.
> The documentation > should also point to how the kernel implements the API You could look for implementation here: http://lxr.linux.no
Well, I was thinking about documenting the relevant undocumented kernel souce code. An example: To be able to use kernel RPC, one has to learn it first. The only way to programatically learn it is to study the source code of the RPC implementation and also the NFS implementation and try it out (to program an example myself). Reading only a manual page of a particular function (if it has any comment at all) is still not enough for me to be able to use it properly. I also need to understand the context (like preconditions, post conditions). Also, many kernel data structures have flags that influence the behaviour of the API functions. Finally, I also sometimes miss the documentation of the allowed function-call sequence diagrams. And by pointing out how a kernel accomplishes some feature I meant "look, this source line inside this function does the job!". Maybe understanding the source code is a precondition to be met prior to do any kernel programming :) I still disagree. Taken to an extreme, by starting reading the source code one could ultimately recurse into reading the manuals for a given architecture to understand, for example, how the kernel module is copied into the kernel memory and has its relocations done. (When all that was needed was to write a simple kernel module). That's the point. BlackHole -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
