On 20011117 Juan Quintela wrote: >>>>>> "magallon" == J A Magallon <[EMAIL PROTECTED]> writes: > >Hi > >magallon> What should be done is to split the public part of the kernel headers from >magallon> the private one that changes so much. Do you know if there are any >magallon> plans about that for 2.5 ? > >Linus stated in a lot of ocassions that glibc _should_ not include >_any_ kernel headers, if they need any definition from the kernel, >glibc needs to copy it. > >Notice that this is really a cool idea if you think that glibc needs >to work with several kernel versions, that can have some of the >definitions changed, that just mean that they have copies of then for >that, Linus just asked for them to copy _everything_ that they need. >
A plain copy will lead to inconsistencies for sure. The problem is that glibc maintainers will not scan through kernel headers to extract public IOCTLs and so on. They just will copy the full kernel headers. So you can end up compiling an external module with the wrong headers (those of glibc instead of those of running kernel) if you miss the -I/usr/src/linux-???/include. The ideal (do not know if my children will see it) would be to join kernel and glibc developers for a week in a tropical beach, and define a common set of public kernel headers, to put in a place like /usr/include/linux or /usr/include/kernel. IE, define the interface between glibc and kernel. And let kernel and glibc use it instead of duplicate it. By. -- J.A. Magallon # Let the source be with you... mailto:[EMAIL PROTECTED] Mandrake Linux release 8.2 (Cooker) for i586 Linux werewolf 2.4.15-pre5-beo #1 SMP Fri Nov 16 16:38:03 CET 2001 i686
