Lin­ux­o­la­tor bug hunting

Sev­er­al peo­ple help debugging/testing the new lin­ux­o­la­tor. We already got patch­es from kib@ and from the con­trib­u­tor which uses “Intron” as his alias in com­mu­ni­ca­tion. On the amd64 test­ing side bsam@ helps a lot. So far we fixed a pan­ic with linux_vfork (I got hit by it while run­ning the glibc regres­sion tests) and synced the MI parts of the amd64 stuff. We’re also in the process to get some more of the uncom­mit­ted stuff of the SoC into the tree. This stuff are changes which are done based upon reviews by jhb@ and ssouhlal@. Unfor­tu­nate­ly jhb@ seems to be busy ATM, so we have to wait a lit­tle bit for his review of the changes.

We still have the futex prob­lem, but there are patch­es float­ing around which at least fix some prob­lems with realplay.

Roman is going on hol­i­day at the end of the com­ming week, I hope we are able to fix some prob­lems until then.

lin­ux­o­la­tor SoC work committed

I com­mit­ted most of Romans work in the lin­ux­o­la­tor to cur­rent. The new syscalls aren’t used until you run

sysctl compat.linux.osrelease=2.6.16

to switch back (after exit­ing all lin­ux pro­grams) you just have to run

sysctl compat.linux.osrelease=2.4.2

But you have to do this on i386. Amd64 sup­port is not com­plete (and besides this, amd64 is still bro­ken and nobody pro­vid­ed the nec­ces­sary debug­ging info to jhb@).

There are some known prob­lems with osrelease=2.6.16, e.g., prob­lems with futex­es (vis­i­ble in acrore­ad, realplay and skype), but some pro­grams already run with­out obvi­ous prob­lems (linux-firefox, linux-opera).

Any reports about new prob­lems to netchild@ and rdivacky@ please. Reviews, debug­ging info and patch­es are wel­come too.

I got some­thing done in the last days

Wow, some of the TODO items in the last post are done:

  • emu10k1 low­er pri­or­i­ty on attach (allows emu10kx to attach when loaded too)
  • sade com­mit­ted (and code improved)

And I did some more things:

  • emu10kx fix­es (sub­mit­ted by Yuriy)
  • enhance­ment of the bsd to lin­ux errno map­ping (sub­mit­ted by “Intron”)
  • added some more files to ObsoleteFiles.inc (sub­mit­ted by kris@)
  • had a look at linux_kdump and deter­mined that we should update it with what we have in ‑cur­rent

There’s also a sub­mis­sion of sup­port for Envy24HT chips from Kon­stan­tin (he also has a page with a lot of docs for envy* chips).

Sum­ma­ry of the last days

Not much hap­pened in the last days. I talked with my SoC 2006 stu­dents and helped them a lit­tle bit with knowl­edge and test­ing. I also did com­mit some fix­es for the envy24 dri­ver to current.

ATM I have some hol­i­days. So I think I should get some more time do to some stuff in FreeB­SD… at least when it stays rainy and I’m able to find some new fur­ni­ture for the liv­ing room soon (either it’s com­fort­able, or it does look good; in case it is com­fort­able and does look good, it’s either not avail­able in the con­fig­u­ra­tion you want, or it is at least 3 times more expen­sive than you are will­ing to spend… sigh).

On my TODO list is to com­mit is “sade”, the break out of the fdisk and diskla­bel edi­tor in sysin­stall. I have patch­es for this, but need to fix-up a lit­tle part before I can com­mit it. After that sysin­stall can be replaced with some­thing else, and we still have the nice curs­es fron­tend for disk par­ti­tion­ing avail­able. Anoth­er TODO item is to give a low­er pri­or­i­ty to the emu10k1 dri­ver, so that the emu10kx dri­ver has prece­dence in all cas­es. And there are some easy patch­es in some PRs I want to deal with.