So what do Linus, Alan, David, Ted, et al have to say about these
ideas and criticisms? It's no use chattering about it unless they
like it.
Well here is what I have to say.
Making more threads of kernel releases is stupid, more complex for
both the developer and the users, and is not going to happen.
Most module authors have some sort of maintainer, if structures change
and the author can't find the time to type friggin' make and put up
new binaries so that it works with the new kernel release, that
maintainer is overworked and needs some more assistance.
Finally, I utterly refuse to hold back on a bug/security fix, or other
change of similar importance just because it requires changing
structure layouts and thus some modules, in fact the issue does not
even enter my mind at all.
Remember, the whole win of Linux is that it is free and you are not
locked into any of the sorts of binary dependancies wrt. kernel
modules and similar, that you would see with some other system. These
binary only modules which break every time a kernel structure changes
are defeating that purpose, you the user are once again locked into
some proprietary technology because the source is unavailable.
---------------------------------------------////
Yow! 11.26 MB/s remote host TCP bandwidth & ////
199 usec remote TCP latency over 100Mb/s ////
ethernet. Beat that! ////
-----------------------------------------////__________ o
David S. Miller, davem@caip.rutgers.edu /_____________/ / // /_/ ><