Thanks, Anders. That was great! It builds now! I've still got to hack the mospost not accepting non-GPL modules problem (paravirtualisation being stock in the debian kernels, apparently this propagates a GPL violation through the use of the usleep function or something... http://www.nvnews.net/vbulletin/showthread.php?t=90214 -- but we are promised it'll be fixed in 2.6.22... http://www.nvnews.net/vbulletin/showthread.php?t=89844 (at the very end))
Actually, all this started because I'm trying to write a java applet to play some midi notes, http://markov.muisc.gla.ac.uk/Click-Chopin/click-chopin.html (the CGI it invokes isn't finished yet) and there is a huge latency on my very old kernel. So I did a *big* upgrade (including a new sun jdk!) to see if things would be better. This is the last remaining problem. I don't like java much, but the applet has to "run anywhere" (actually, it won't run on Windows until the user downloads a soundfont, but that's out of my control). Annoyingly, it works fine on Mac OS-X, even though I've never found much latency difference between their kernels and a properly set-up linux one. I suppose it is a java run-time problem, but there's no point trying to fix anything on a less-than up-to-date platform. The last Java I wrote was over 5 years ago. Perhaps I'm getting too old for this whole unstable thing :) Once this is sorted out there's the SVG rendering and the lilypond hacking, neither of which really work yet, but I can blame my team for that :) Anyway, thank you so much for your refresher course, and getting me a nice new module built. I'm sure that will decide things one way or the other. Nick/. http://cmt.gla.ac.uk http://www.n-ism.org On Saturday 16 June 2007 20:06:53 Anders Boström wrote: > >>>>> "NB" == Nick Bailey <[EMAIL PROTECTED]> writes: > > Hi Nick! > > NB> Dear Anders, > NB> I saw your bug report at > NB> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=409204 and the fix, > which I NB> am trying to apply so that I can evaluate a recent kernel. > > NB> I have built the binary package using m-a i-a etc many times before, > but I NB> don't understand how I can stop the process at the right stage to > apply this NB> patch. > > NB> Since the bug has been extant for 135 days, I wonder if you would > consider NB> amplifying this for those of us less conversant with the > module building side NB> of things, maybe by adding an additional post to > this bug with the build NB> instructions? I'm guessing there's a dpkg > "only-unpack" -x/-b variant that NB> could be used? This would be very > helpful for those prepared to read through NB> the bug pages and act on > them, but whose developer skills do not lie in the NB> kernel modules area. > > NB> Thank you very much! > > OK, first install nvidia-kernel-legacy-source: > > As root: > # apt-get install nvidia-kernel-legacy-source > > Then save my patch in a file. Use cut & paste from the bug report or > save the included attachment. > Lets call the file /tmp/nvidia-kernel-legacy-source.patch . > > Then read the instructions in > /usr/share/doc/nvidia-kernel-legacy-source/README.Debian . > > After step 1 in README.Debian, patch the source: > > # cd /usr/src/modules/nvidia-kernel-legacy/nv > # patch -p1 </tmp/nvidia-kernel-legacy-source.patch > > Then proceed with the remaining steps described in README.Debian. > > / Anders
signature.asc
Description: This is a digitally signed message part.