Yes the wonderfull paper we all know about from Eric says You nicely how
to get features INTO open source projects but users seldomly request
things, which should be REMOVED.
So here it goes:
1. The old getdents stuff. It's ages old and akin to *very* old libc-5.0.xx
If somebody needs compatibility which gets thus far, he/she should
just simple stick with the older kernel.
(4k)
2. The whole console selection related ioctl stuff.
It was broken from the beginning on. There is currently a new
replacement mechanism: /dev/vcs, which should provide the same
funcionality. Removing it will only break the in my oppinion
anyway toyish mouse selection on text based consoles. This is
something that will not break the system since this
functionality isn't essential for operation.
(8k)
3. Due to 2. we could even remove the wholy broken und useless unicde
stuff in the current console code.
It's only used in code related to 2. With the new mechanism for
console selection this should get handled in user space. And
please remember I'm not talking about unicode or whatever here.
It's only about the selection stuff in the current kernel.
(8k)
4. There is plenty of debugging code in the serial driver, which remains
enabled, even after *years* of testing. (16k)
5. The slab allocator contains may many glorious features not used by
anybody in the current kernel.
Maybe that's the darvinistic proof that theyr are not as usefull
as one would think in first place?
Linus please put the trash listed above, where it belongs -- in the trash
can and help preventing Linux becoming in the OS area, what emacs is in
the area of editors :-).
Many thank's in advance...
Marcin
PS. And BTW: Theodore -- where is the /dev/random in Solaris ?????!!!!
=========================================================================
In real life: System Programmer at AIS AXON GmbH
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu