rdivacky blog

I have implemented epoll() subsystem in the Linuxulator based on kqueue. This was quite simple and the resulting code is about 200 lines of code. Mark Heily sent me his inotify implementation library which works in userland. I looked at the code and its a little buggy so porting to FreeBSD will take a little time. Now I am working on fixing various Linuxulator bugs in preparation for switching 2.6 emulation on default after 7.0R is released. Today I fixed java and my next target is flash9.

Please read emulation@ mailing list and help me with testing.

I basically finished the *at syscalls patch and it’s waiting for a review. Then it hopefully can be commited. I also started to work on epoll/inotify. I am just branching a new branch for the work on it. So far it looks very easy and I hope to get it working (at least the epoll thing) in a week or so Stay tuned!

So, I just finished university and plan to focus 100% on SoC Linuxulator update. My todo for the next few days includes deciding what to do with *at patch (and doing it ;)), Giant removal from futexes (should be pretty easy but needs coordination with a few people). I am also cooperating with Gabor on a Linuxulator article which will be hopefully included in our official documentation. Last but not least I helped netchild review his linuxulator MFC patch.

After all this is solved I hope to finally bite into the epoll stuff. And when I get into the mood I hope to resolve whats wrong with asyncronous pthread_cancel.

Konstantin commited the correct atomic operations implementation of futexes so we are one step closer to the holy grail. In the meantime I got a mail from a “YYJ” submitting a testing program that exhibits ill behaviour with 2.6 futexes. I am going to investigate this real soon and hopefully fix it. Unfortunatelly I have to focus on school these days as I am having my final exam on June 19th. So wish me luck!

I got a little bored today so I looked at futexes. The progress was quite fast and now we pass futex testing program. The only downside is that with definitely wrong kernel code But I hope to fix this really soon now. I asked Konstantin Belousov to help me with those things that remain. But now when I cracked the problem I guess we will have fully working futexes in like a week or so…

My *at patch (currently implementing 3 syscalls) is pending review from Robert Watson but I hope it was dont right this time and will be commited. I also have experimental patch “fixing” SysV SHM memory usage in Linuxulator but no time to test it please help if you can!

At this pace I think 2.6 emulation in 7.0R will be pretty usable and almost production ready!

So I am here again this year to mess up linux emulation layer in FreeBSD. Now I want to focus on finishing *at set of sycalls which are necessary for 2.6.16 Linux emulation. Then I have to pause for a while because of school things (final exams coming) then the plan is not very clear but I’d like to work on epoll and possibly inotify mapper to kqueue. I’d love to fix futexes in cooperation with Konstantin Belousov (my mentor this year) and maybe with Attilio Rao (if time permits). I am writing a thesis that describes Linuxulator in quite a detail and it is planned to publish this as an FreeBSD documentation article.

You can watch things at http://wiki.freebsd.org/linux-soc2007