On Sun, Nov 18, 2012 at 08:50:07PM -0500, Walter Dnes wrote: > On Sat, Nov 17, 2012 at 11:52:22PM -0800, Greg KH wrote > > > > Yes, I know all about the firmware issue with media drivers. It's now > > resolved and fixed, in two different ways (the kernel now loads firmware > > directly, and on older kernels, udev has fixed the issue.) So that's no > > longer an issue for anyone. > > The fact that they went ahead with changes, knowing full well it would > break stuff, is reason enough to distrust them in future. It should not > require a rant from Linus, or a workaround in the kernel, to get them to > fix their bugs.
That's the "fun" of working with people you don't have direct control over. Bugs get fixed on different schedules than what you sometimes like. This specific issue, as it was hit by only a very small number of people, and two distros had work-around patches in their udev packages, was missed by a lot of people, myself included. I honestly thought that it had been fixed months ago. Sometimes a rant, or just reminding people, is all that is needed to get issues fixed. And it worked here quite well, don't you think? Actually, I would argue that it worked even better than if the issue had been worked-around in udev in the very beginning when it first came up. Now the kernel has changed to allow udev to remove the whole firmware loading logic, which arguably, should have been done in the very beginning. So you might say that because of people forgetting about this, and people ranting, everyone is much better off in the end. It's a bizarre development model, I know. :) > > It's also a pretty simple set of patches that Gentoo can keep around > > if it's really a serious issue for people. > > That may be true today. But as udev gets more tightly integrated into > systemd, those patches will become a "dead end", to use Lennart's words. What patches? udevd builds for me just fine without building the systemd binary. The developers even have a whole web page set up for how to do this properly if you need to do so. > > Note, a separate /usr has been broken for a while now, udev is just > > pointing the issue out. And again, if you want a separate /usr, just > > use an initrd, the solution is simple. > > ???? I have 4 "broken" Gentoo systems running just fine, without an > initrd, thank you. There have always been a few edge-case setups that > won't work with a separate /usr, without an initrd. What annoys me is > this dog-in-the-manger attitude that if a separate /usr is broken for a > few people, then by golly, it should be broken for everybody. Again, udev isn't the problem here. It hasn't broken the standalone /usr issue at all. There isn't anything in udev to change for this. I don't understand why you are thinking that udev has anything to do with this issue at all. It's other packages that are the problem here. Are people forking and changing them to resolve the problem? If not, why not? thanks, greg k-h
