Melvyn Sopacua wrote: > On Tuesday 24 February 2004 19:47, David Gilbert wrote: > > >>>>>>>"Daniel" == Daniel Eischen <eischen_at_vigrid.com> writes: >> >>Daniel> Read src/UPDATING and then if you still have questions seach >>Daniel> the -threads and -current archives for anything nvidia >>Daniel> related. >> >>You're so sweet. But you're out of order. Obviously, I read UPDATING >>as I got the nvidia driver kernel module to load. I also follow >>current and neither are particularly covering the issue. >> >>Geez. > > > That's a pretty loaded answer, for somebody who follows and reads stuff, with > an obvious short-term memory problem: > http://lists.freebsd.org/pipermail/freebsd-current/2004-February/020791.html > > If you follow the entire thread, you'll notice, that it contains all the > information you we're asking about, right down to the precise processor > instruction and whether or not nvidia is needed/wanted to fix it. Ok, enough with the snide passive-aggressive comments. 1) This issue at hand is a FAQ. However, just because it is a FAQ doesn't mean that we need to be abusive towards the people that ask it. 2) The issue at hand is that libpthread/libkse and libthr are incompatible with the libGL.so that is used by the nvidia driver. Remember that the nVidia 3D driver solution includes an open-source kernel driver and a custom, closed source libGL library. These two pieces are a matched set. However, the kernel driver itself does not pose a problem, only the library. 3) The nvidia libGL library works fine with libc_r. 4) Some people are lucky and don't experience any problems in this situation. However, they are the exception to the rule. Most people who want to use the nvidia 3D driver set must switch back to using libc_r. 5) The nvidia libGL library is closed source, and there is nothing that we in the community can do to fix it ourselves. To solve this, the project needs to supply a common Thread Local Storage mechanism that doesn't conflict with libpthread and can be used by the nvidia libGL. Until that happens, there is little that can be done to fix the problem. 6) Deeper discussions on the technical issues can be found by searching the mailing lists, and I will not repeat them here. Please be more respectful to people asking these kinds of questions. It's neither an obvious problem nor an obvious solution, and even if it was, there is no need for so many people to cop such an attitude over answering a comon question. If you have the motivation to sit down and write a flame over an innocent question, maybe you can redirect that energy towards something that is actually positive and useful. ScottReceived on Tue Feb 24 2004 - 14:44:14 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:44 UTC