> I do so hope you mean 2.6.24.2 or greater. OK, not everyone has > untrusted users, but why build something with a known vulnerability. > OK, I know that the latest stable was .3 last time I looked, but > that fails headers_check because of a bogus change. > > Apart from regressions between kernel versions, particularly on > less-common equipment, there is also the need to update a working > config. Somewhere after 2.6.16, the IDE config details changed > (people using libata can refer to /dev/sda instead of /dev/hda) > which can make it interesting when you want to be able to boot both > old and new kernels (typically, mount by label - still need to pass > the correct root= in the bootargs). > > In general, I totally agree that people should update to a > newer stable kernel, but until they have a config which works, it > probably isn't the most productive thing to attempt. Even then, it > can sometimes go wrong (new options get wrongly used/ignored, or there > are new regressions). > > ?en > No, I did mean 2.6.24. I realise now that there is a local exploit with this kernel version but I am the only one using it. Clearly, I'll be updating the kernel as and when I get round to it.
Regarding the .config file, surely it is a worthwhile learning experience to build your own from scratch. You're right Ken, it is incredibly tedious, but it allows one to select the correct parameters to suit one's own hardware. When I built my first one it seemed to take forever, but as I was building a small, fanless, solid-state computer (Via ML6000 m/board) it seemed to make sense, particularly as I wanted a static kernel, and I was not using any swap. To be honest I've found the process to be fairly forgiving; even my first attempt at producing a .config file was a success. The first draft, if successful, can always be tweaked later by scrutinising dmesg. Richard Melville -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
