On Wed, Sep 20, 2017 at 10:40:49AM -0600, Hans Malissa wrote: > Hi, > > ‘make defconfig’ did a very good job, thanks for the advice. My idea of using > .config of the host distribution was completely wrong. LFS is up and running > now, and booting is much faster than with my host distribution. > I still wonder where the error message came from - I was assuming that the > kernel should compile in either case after using the standard configuration > tools.
If you started with a distro, they may have turned on defines for drivers which do not exist on real x86 hardware, but where kernel developers want the extra build coverage. What built for a distro in some previous kernel version, perhaps with additional patches they are carrying, is not guaranteed to build with a different kernel. > Regarding the (working) configuration that is a result of ‘make defconfig’ > and customization with ‘make menuconfig’, I have two questions: > - How do I know if one or more modules/configuration settings are missing > (even though the kernel runs)? You will find out if you ever need to do something and the kernel doesn't support it. For things covered in BLFS, look at the Kernel Configuration part of the longindex.html (it's near the bottom) http://www.linuxfromscratch.org/blfs/view/systemd/longindex.html for systemd svn. Also look at About Firmware in BLFS chapter 3. > - How can I tell whether there are extra modules or options turned on that > are not really used in my system? For built-in options, trial and error (turn things off in menuconfig, change the EXTRAVERSION in the Makefile, recompile, see if it boots and still does what you need). After a while, the benefits of doing that are probably not worth the time. My recent vmlinuz files on the intel machine where I'm writing this reply are 6.6 MB for 4.13.1 and 6.5 MB for 4.12.8, on an old AMD phenom - despite building in firmware - 5.8 to 5.9 MB, on a less-old AMD Kaveri 6.7 MB - that one has bigger firmare, and probably still has MD options for testing mdadm when I'm editing. For modules, make a list of which modules exist in /lib/modules for the current kernel version, then compare it to lsmod. Anything not loaded is either something you might use occasionally (e.g. filesystems, support for peripherals not presently connected) or a candidate for removal. If in doubt (e.g. crypto modules), err on the side of caution. You might need extra selections beyond what is enough to boot LFS by the time you get to a fully-built system. > Thanks a lot, > > Hans > > > On Sep 19, 2017, at 9:08 AM, Bruce Dubbs <bruce.du...@gmail.com> wrote: > > > > Hans Malissa wrote: And please don't top post on LFS lists. Snipping here. ĸen -- Truth, in front of her huge walk-in wardrobe, selected black leather boots with stiletto heels for such a barefaced truth. - Unseen Academicals -- http://lists.linuxfromscratch.org/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page Do not top post on this list. A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail? http://en.wikipedia.org/wiki/Posting_style