In article <Pine.LNX.3.96.980720085558.20344A-100000@mercator.cs.kuleuven.ac.be> Geert.Uytterhoeven@cs.kuleuven.ac.be wrote:
>Wrong :-) He was telling Windows can't justify its bloat...
Yeah, I know... I told you I was babbling :-)
BTW, why does the 2.1 kernel behave worse than the 2.0 kernel?
There has been a lot of talk, but has there been any conclusive tests
yet? Is it the dcache that's fundamentally more memory hungry?
Someone mentioned that the fragmentation might be due to the dcace
allocating different sized qstr objects, in that case it might be a
big win to convert the kmallocs of dentries to use the slab allocator
instead and maybe even extend the slab allocator to give out smaller
objects of 16 or maybe even 8 bytes.
Another problem as far as I can tell is that unswappable/unmovable
memory which stays in one place for a long time (such as a processes
kernel stack) is allocated all over the place so that it is eventually
impossible to allocate temporary buffers such as two page skbs. This
problem might be reduced by adding an GPF_UNSWAPPABLE flag to kmalloc
and hack kmalloc to try really hard to allocate unswappable memory
close to other unswappable memory, keeping the swappable regions
larger. On the oter hand this might just be a case of treating the
symptoms, not the cause.
/Christer (still babbling. I'd better go back to sleep :-)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html