William Kucharski wrote: > For those looking for a more explanative text on adding a system call to > OpenSolaris, Sun Labs' Eric Schrock covered this topic in his blog: > > <http://blogs.sun.com/eschrock/date/20050614> > > he also contributed the same information to Appendix B of the second > edition of > "Solaris Internals" by Richard McDougall and Jim Mauro: > > <http://safari.oreilly.com/0131482092> > <http://www.amazon.com/Solaris-Internals-TM-OpenSolaris-Architecture/dp/0131482092> > > William Kucharski
Cool. The truth is out there. You just have to know where to look. And, Eric Schrock has a lot of coverage of the why of things, like the limit of 256 system calls. Oh, and he has a bunch of steps I left out, like updateing /etc/name_to_sysnum, updating truss, and procnames.c. I didn't do those. But then, I didn't notice any difference in behavior of truss or dtrace on my ODW board. :-) But, I ought to finish the job, just for the sake of correctness. Anyway, I perpared my own roadmap, including some kprintf() specific information. It was easy enough that I didn't think to go looking for a tutorial online or in one the books. But, I need to do it every time, as part of the preparation to actually do the code. I need to go through the motions. I did not set out to write a tutorial. It was timely, because of the need for kprintf(). I had a real need to do kprintf(). It has been about the only way I could get any information out the early run-time loader. It's funny how you could see an appendix in a 1000 page book, with a "toy" example -- nothing that would be needed in the real world. Then, it turns out that I do need something like it, desperately. Then, the information that was there suddenly has new meaning. Also, I still like "live" slices of source code. -- Guy Shaw
