In <Pine.SUN.3.96.1000702203330.260F-100000@invisible.eskimo.com> Robert Dinse (nanook@eskimo.com) wrote:
RD> On Sun, 2 Jul 2000, Khimenko Victor wrote:
>>
>> In short: mostly patch is bunch of band-aids over band-aids. Go read what
>> Linus thinks about band-aids and SHUT UP!
RD> Heaven forbid should anybody express an oppossing viewpoint!
Huh ? I REALLY eagerly wait your proof of otherwise. Since Linus said "maybe
people in the end can convince me that every single scheduling point makes
100% sense, and is not a hack at all but a natural thing" it's possible
(in theory) to convince his that Solar Designer and/or low latency patches
are not pile of crap but "a natural thing". So far this was tried only for
non-executable stack of Solar Designer patch and it WAS NOT convincing.
May be other parts are no so stupid - just noone ever tried to explain why
it should be in kernel (from technical viewpoint - there were enough cries
"oh, that's so great patch, let's include it" from peoples who
a) don't understand patches
b) don't understand the kernel internals
c) don't understand the problems with the patches
d) have no time/desire to change (a)--(c)
How the hell it can EVER lead to inclusion of said patch in kernel ? It's
true, some parts of oldier Solar Designer patch was adopted by kernel (and
I sure some parts of low-latency patch will be adopted as well). But if you
need MORE then change a)-c), then come back.
RD> This is starting to feel like the boys at M$.
>> P.S. Yes, there are exist few interesting points. But mostly it's MESS. Worse
>> then low-latency patch. And whe you'll think about it is authors of patch do
>> not want to clean up that mess (and throw out stupid things like non-exec
>> stack altogether) then why Linus should care ?
RD> I'm afraid I don't share your viewpoint here, but you know it's less of
RD> an issue for me because it's well maintained and easy to apply, so it's not a
RD> big deal to apply for each new release.
RD> The low-latency patch is more difficult, it won't apply as is for Sparc,
RD> so I have to hand remove portions for every new release which is a PITA.
RD> But it's become abundantly clear that neither security nor latency are
RD> important issues to the vast majority here.
You are speaking from bitterness so I will forgive you but it's FAR from true.
Latency IS important issue and security IS important issue (recall recent
capabilities history if you need proof). Just kernel clearity and maintability
ARE also important issues (sometimes may be even more import the latency and
for sure more important than any band-aid solutions in "security" and/or
latency).
RD> Nor is stability on any platform other than Intel.
Stability on any platform other than Intel is simple problem: the only ones
who can investigate problems in non-major platform for ANY OS is the ones
who work with said platform. Take look on FreeBSD for Alpha, Windows NT for
Alpha or even Solaris for Intel. All are sucks rocks (in comparision to
main version). It's not conspiracy, it's just lack of testers/developers.
You can screem as load as you can but you will not change this till there are
more developers with access to Sparc then Intel.
RD> I give up. Have a nice life.
-
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.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Jul 07 2000 - 21:00:11 EST