On Mar 25, 2008, Dave Jones <davej@xxxxxxxxxx> wrote: > The kernel team is already buried alive in backlogs of bugs. This > serves to do nothing but introduce new failure modes unless it's > done properly. Hence my concern. Honest question: what failure mode would you find it easier to recognize, "module not present" or "module loads but doesn't work"? I'm asking because I'm working on software that may simplify my life of maintaining patches for kernel-libre for Fedora. Removing entire files has been the practice of people who started these efforts (not for Fedora) a long time ago, but this creates some pain in dealing with patch files that later attempt to modify these files in harmless points, especially while tracking the rc/git stuff in rawhide. OTOH, removing only the blobs from the source code would be relatively easy to do now, with deblob-check (URL below), but then the modules could still be enabled, still compile, but not quite work inasmuch as the blobs are necessary for proper functioning. http://www.fsfla.org/svn/fsfla/software/linux-libre/ I'd probably keep on disabling these modules as useless in my own builds, but people might re-enable them, as they might appear to be available and even compile correctly. -- Alexandre Oliva http://www.lsd.ic.unicamp.br/~oliva/ FSF Latin America Board Member http://www.fsfla.org/ Red Hat Compiler Engineer aoliva@{redhat.com, gcc.gnu.org} Free Software Evangelist oliva@{lsd.ic.unicamp.br, gnu.org} -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
