"John P. Burkett" <burk...@uri.edu> posted 4a2ee4c1.8000...@uri.edu, excerpted below, on Tue, 09 Jun 2009 18:40:01 -0400:
> Doing "revdep-rebuild -p" elicited this response: Calculating > dependencies... done! > [ebuild R ] sci-libs/blas-atlas-3.8.0 > [ebuild R ] app-emulation/emul-linux-x86-soundlibs-20080418 > [ebuild R ] mail-client/evolution-2.24.5 > [ebuild R ] x11-drivers/ati-drivers-8.552-r2 > [ebuild R ] sci-libs/lapack-atlas-3.8.0 > [ebuild R ] dev-lang/R-2.8.1 > [ebuild R ] sci-mathematics/octave-3.0.3 > > That looked o.k., so I did "revdep-rebuild" The first 3 packages seem to > have been handled fairly successfully. Problems arose with the > ati-drivers. The final messages from revdep-rebuild read as follows: >>>> Emerging (4 of 7) x11-drivers/ati-drivers-8.552-r2 > * ati-driver-installer-8-11-x86.x86_64.run RMD160 SHA1 SHA256 size ;-) > ... [ ok ] > * checking ebuild checksums ;-) ... [ > ok ] > * checking auxfile checksums ;-) ... [ > ok ] > * checking miscfile checksums ;-) ... [ > ok ] > * Determining the location of the kernel source code > * Found kernel source directory: > * /usr/src/linux > * Could not find a Makefile in the kernel source directory. > * Please ensure that /usr/src/linux points to a complete set of Linux > sources > * > * ERROR: x11-drivers/ati-drivers-8.552-r2 failed. As Josh says, it probably didn't get beyond this one. Drivers like this are a special case for a couple reasons. One is that parts of them are closed source and simply may not build or work correctly with the latest stuff. You're at the mercy of the ATI folks for that, but particularly with the latest xorg, they may have issues, and stabling xorg gets routinely delayed because neither ATI nor nVidia have caught up yet. FWIW I won't run closed source, nor could I even if I wanted, since I cannot and will not agree to the license they try to enforce. Yes, that does limit my hardware choices a bit, but not as much now that AMD/ATI is cooperating a bit better with the open source world. But to each his own. However, that doesn't seem to be the issue here, which brings us to reason #2. The ati-drivers include a kernel driver, which must be compiled against a kernel that has been built so the proper parts of it are exposed for the drivers to use. The kernel must reside at /usr/src/ linux. If you look at the error above, it found the directory all right, but it didn't find what it needed in it. Have you built your kernel using those sources, with your new gcc, yet? Did you leave the working files exposed for ati-drivers to use if so, or did you cleanup using make clean, or something? My guess is that you need to build your kernel with the new gcc, and then ati-drivers will hopefully compile. However, as I said, I do not and will not run them personally, so you may need to get help from someone else that knows their quirks better than I do, on them. > * Messages for package sci-libs/blas-atlas-3.8.0: > > * Please make sure to disable CPU throttling completely > * during the compile of blas-atlas. Otherwise, all atlas > * generated timings will be completely random and the > * performance of the resulting libraries will be degraded > * considerably. > * blas has been eselected to atlas-threads > * cblas has been eselected to atlas-threads Note that you have an eselect for these, so if anything can't find them, you can try eselect setting them back to what works... > * revdep-rebuild failed to emerge all packages. Like Josh said, it didn't finish. You can try remerging just the specific packages it didn't merge, or rerun it after you fix ati-drivers. > * Found some broken files that weren't associated with known packages * > The broken files are: > * /usr/lib64/R/library/BMA/libs/BMA.so (snip more...) All these seem to be R related. If after finishing the revdep-rebuild as far as you can, it still doesn't associate these with anything, try remerging R, and/or deleting (or perhaps saving them somewhere else, just to be safe) these manually and /then/ remerging R. -- Duncan - List replies preferred. No HTML msgs. "Every nonfree program has a lord, a master -- and if you use the program, he is your master." Richard Stallman
