In the last days I spend my free time with mentoring Boris and the SoC students. Boris is maybe ready to fly solo now (and he already got implicit aproval for several things from me), but I will keep an eye on him from time to time for some more weeks.
The linuxolator SoC project progresses faster now. We may have support for TLS, TID and futexes in the comming week (this should allow us to run FC4/FC5 with a linux kernel version of 2.6.x, time will tell).
In the sound SoC project, Ryan created an user repository. He is more comfortable with committing “micro-changes” to the user branch first to not spam the perforce mailinglist. I don’t think this matters, but when he feels more comfortable this way, I don’t object. This makes it a little bit harder for me to follow, but I will nag him more often for a status report then.
I also MFCed the fixes for Oracle 10, it should work on 6‑stable too now.
well… the linuxolator is in the state where I should be just fixing some last remaining things and it should be done pretty soon.. currently I am in a process of implementing CLONE_SETTLS semantics in clone() syscall. this is a bit tricky
status;
we have working fc4 on programs that doesnt use threads (my theory) because of the missing CLONE_SETTLS. it remains to implement tid/pid handling (that should be very trivial) and the CLONE_SETTLS and we should have full NPTL support. there’s still a problem with memory leak (which seems to lead to a panic) and I dont know how to address this ;(
the weather is against me as it reaches almost 40C here ;(