Adrian is right. Building a kernel for Debian isn‘t really difficult. I work for a first level Linux support and I can do it too.
— Christian Sent from my iPhone > On 12. Oct 2018, at 16:39, Dennis Clarke <dcla...@blastwave.org> wrote: > >> On 10/12/2018 10:28 AM, John Paul Adrian Glaubitz wrote: >>> On 10/12/18 4:10 PM, Dennis Clarke wrote: >>> That page speaks of many things but clear instructions are not among them. >>> >>> That whole page speaks in riddles and strange incantations that don't >>> really mean anything. The next full moon is 24th of October and I >>> may be willing to give it a try. However the process will fail as it >>> has before over and over and as a user I lose interest in poorly written >>> instructions that speak in half truths and worthless jargon. >> Well, you can't really expect that someone will give you a full course ... > > Sure I can. > >> on the basics. If it was written like that, the whole manual would probably >> have several thousand pages. > > Yep. > >>> fakeroot ? >>> >>> That alone is something from the distant past that bothers me. >>> >>> Nope ... not interested. >> What's wrong with looking things up that are not familiar to you? > > I was using fakeroot back in 2001 or maybe it was 2000. Can't recall. It > isn't unfamiliar. I just don't see the need to get my own kernel in > place. Debian needs it ... I don't. > > ... > >> I think it's fair to expect that if a user is willing to install an >> unsupported >> version of Debian on an unsupported machine, then the developers can expect >> some manual work from the user. > > Yep. No issue there at all. > >> It's not that people are not willing to help. It's simply that time and >> resources >> are limited and in Debian Ports, we don't have the manpower to provide a >> polished >> product where we have each and every corner-case covered. > > Right. You are busy elsewhere and so leave the long verbose crud to > schmucks like me. I have done it before over and over. It isn't fun > but it helps the next person and isn't that the whole point? BTW I > wrote the original Solaris Zone docs and OpenSolaris kernel build docs > also and they were entirely step by step. Better language. However > very hand holding. Don't expect anyone to look at your technology and > play with it if you make it secret and special and impossible to play > with. Such is life. > >> Building your own kernel isn't really difficult. It's mostly a matter of >> installing >> the build dependencies for the kernel with "apt-get build-dep linux", then >> downloading >> the source tarball of the kernel you want to use, unpacking it, copying the >> configuration >> from /boot/ which you are currently using to $KERNEL_SRC_ROOT/.config, >> running "make oldconfig" >> and applying any patches you want to test. Then just "make", "make modules", >> "make install" >> and "make modules_install". There isn't anything more to it, really. > > Well let's see if that is really true. I have yet to see it work. > So there must be secret magic in there somewhere. > >>> Attached are detailed step by step and clear instuctions on how ... >>> >>> Here is a blunt force trauma set of steps : >>> >>> https://node000.genunix.com/deb_ppc64/debian_ppc64_kernel_build.txt >> Writing a document like this takes really a long time ... > > days. Yep. Coffee. Curse. re-coffee. re-curse. > >> and as already said, this is >> just something we can't do - at least I can't. > > Don't worry about it. If there is interest .. then people will make the > effort. I need to clean that up and re-write it and get the Debian way > of things in there but it can be done. By someone else. Not you. > Relax. > > My real interest is in RISC-V anyways. Just wait until I climb on top > of that. > > >> There is certainly enough documentation >> on the internet on how to build your own kernel, either in Debian or any >> other >> Linux distribution. However, for a generic kernel taken from upstream, you >> can just >> take any howto. > > yep > >>> Would be nice if the "Debian way" were written up in a step by step fashion. >> I think there are better ways to spend so much time than writing >> documentation >> that already exists. > > Maybe. I have a really nice coffee machine. > > Dennis >
