Collective resource limits status report #7
Gabor Kovesdan
gabor at FreeBSD.org
Sun Jul 18 17:12:43 UTC 2010
Since last report, I temporarily changed to 32-bit jid_t to solve that
problem I described. It worked but then I had similar issues with a
single int variable... I still don't know the actual reason but I could
solve it by adding syscalls to compat32 and declaring jid_t in the
proper place. So the basic syscalls work now, the process number limit
also seems to work, so now I can actually test the existing stuff. I've
added some test cases to the tester app. I've found a bug in killjob,
somehow it results in a page fault. Plan for next week is to track this
down and fix, and then start adding support for more limits.
Latest sources in //depot/projects/soc2010/gabor_jobs/irix_jobs/, full
patch here: http://kovesdan.org/patches/jobs-20100718.diff
Gabor
More information about the soc-status
mailing list