Just a heads-up that on at least one box (mine), gcc-4.7.1 might be problematic for the kernel in certain circumstances.
I've been using 3.5.0 on the LFS-7.1 system on this box, and it is fine. Didn't have time to upgrade the kernel, but 3.5.2 came out during my latest build - so, I dropped that in. Thought I had a kernel problem - when I use my KVM switch to change to my other desktop box, _and_ that box is running X, I lose a tall chunk in the bottom left of the screen. Now that I've got X in use on this machine, I can see that the bottom right is in fact repeated at the bottom left - on a 133x54 text screen, that part at the right is usually empty, so it looks as if someone put a card over the corner of the screen. Tonight, I've started building other kernels to try to see if it has already been fixed. Both 3.5.3-rc1 and 3.6-rc3 also show the problem. And then I tried switching the KVM while using 3.5.0 - the problem is there too. :-( So, not a kernel regression, if I'm reading my results correctly it's a gcc problem (4.6.2 on the LFS-7.1 installation, 4.7.1 now - both with binutils-2.22). Serves me right for breaking my rule to *always* try the same version of the kernel on the (LFS) host system, I suppose. With luck, it's not going to affect many people - we'll find out during the -rc testing. ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
