On Thu, Jul 01, 2004 at 12:03:36PM -0700, Florin Andrei wrote: > On Thu, 2004-07-01 at 10:56, akonstam@xxxxxxxxxxx wrote: > > On Wed, Jun 30, 2004 at 03:15:52PM -0700, Mike Fedyk wrote: > > > Greg Swallow wrote: > > > > > > >During installation of the new 4kstacks compatible nvidia driver, am > > > >informed it will not operate correctly with the kernel module rivafb > > > >loaded/active. Tried modprobe -r rivafb to remove, but still have > > > >garbage cursor. > > > I made this work by what may be the kludge of the year. I placed the: > > modprobe -r rivafb as the next to last line in the rc.sysinit file. > > Well nvidia module seems to work so unless someone has a better idea I > > leave it the way it is, > > I have a better idea: > > cat "install rivafb /bin/true" >> /etc/modprobe.conf > > This will stop it from loading at all. cat or echo? # echo "install rivafb /bin/true" >> /etc/modprobe.conf The comment made by the nVidia installer is that the module has been marked to be loaded dynamicaly. i.e. if any of it's functions are called. It smells as if there is a symbol collision between the two driver objects and a reference to one of the riva symbols missing in the nVidia driver could cause the second driver to be linked causing a tangle. If and only if you are having lockups it might pay to move the 'rivafb.ko' object out of the way so that it cannot be loaded no matter what. An error should signal that there was an attempt to load the unwanted object and that this was the problem. Do this only to one kernel and keep the other as a recovery path. I expect that a stub symbol or some other code change will be added in a future package to make this problem a non-issue. -- T o m M i t c h e l l /dev/null the ultimate in secure storage.
