Hi and HNY!
I have one question about the "existence of collected ideas".
It is obvious that by using only proven concepts and standards have led
the Linux on the top of the world. But it is hard to believe that you
guys are absolutely agree with every agreement or standard ever created.
Is there a place for linux kernel community where ideas about the needed
advancements in those proven concepts can be presented to the world in
nice, structured and easily readable way? Or is this the professional
property of the kernel programmer to not ask a question about the
"healtiness" of some POSIX thing or not to give a hint for better
hardware implementation? Well, this last question is a little nagging.
For example I found this message yesterday:
Re: ptrace single-stepping change breaks Wine
From: Linus Torvalds
Date: Sat Jan 01 2005 - 17:15:54 EST
...
<snip>
It would have been nice if Intel had added a
"single-step" bit to %db7, and then just or'ed in the values of TF and the
new flag when deciding to single-step.
<snip>
...
Maybe this idea have been said out just for fun (btw, I have one book
which have similar title :) ) but why not collect such things together?
For sure there are ideas about every tool, technology or standard which
is used for kernel development or with which the kernel have to cope
with. Is it possible to build up such a knowledge base or is there
similar already?
thanks,
Indrek
On Sun, Jan 02, 2005 at 02:07:11PM +0200, Indrek Kruusa wrote:
> Is there a place for linux kernel community where ideas about the needed
> advancements in those proven concepts can be presented to the world in
> nice, structured and easily readable way? Or is this the professional
Ottawa Linux Symposium comes closest, together with Linux Conf Australia.
OLD publishes proceedings where people present their ideas, sometimes
already implemented, sometimes planned.
Otherwise, people have been known to present their ideas to this list. An
example is the epoll work which was presented on this page
http://www.xmailserver.org/linux-patches/nio-improve.html
> property of the kernel programmer to not ask a question about the
> "healtiness" of some POSIX thing or not to give a hint for better
> hardware implementation? Well, this last question is a little nagging.
A lot of it is informal. I know some linux kernel developers are pretty
close with hardware manufacturers and especially AMD invested a lot of time
and effort in making sure the Opteron design was fit for linux, which mostly
worked.
Linus mentioned a few years ago that a lot of these exchanges occurred in
hallways of conferences, suddenly some people would jump him and say "we're
from [ big company ] and we'd like linux to do this" or "we're from [ big
company ] and we're wondering how we should design our hardware". Linus then
called for opening up this process, which has happened for a bit. OSDL may
also be a venue.
LT> It would have been nice if Intel had added a
LT> "single-step" bit to %db7, and then just or'ed in the values of TF and the
LT> new flag when deciding to single-step.
>
> Maybe this idea have been said out just for fun (btw, I have one book
> which have similar title :) ) but why not collect such things together?
Rest assured that this list has subscribers @ all relevant processor
manufacturers. But there is nothing formal, it is not the linux way to set
up committees or whatever, except perhaps for OSDL.
Regards,
bert
--
http://www.PowerDNS.com Open source, database driven DNS Software
http://lartc.org Linux Advanced Routing & Traffic Control HOWTO
> Is there a place for linux kernel community where ideas about the
> needed advancements in those proven concepts can be presented to the
> world in nice, structured and easily readable way?
Are you perhaps thinking about a new mailing list?
Greetings,
cam.
bert hubert wrote:
>On Sun, Jan 02, 2005 at 02:07:11PM +0200, Indrek Kruusa wrote:
>
>
>LT> It would have been nice if Intel had added a
>LT> "single-step" bit to %db7, and then just or'ed in the values of TF and the
>LT> new flag when deciding to single-step.
>
>
>>Maybe this idea have been said out just for fun (btw, I have one book
>>which have similar title :) ) but why not collect such things together?
>>
>>
>
>Rest assured that this list has subscribers @ all relevant processor
>manufacturers. But there is nothing formal, it is not the linux way to set
>up committees or whatever, except perhaps for OSDL.
>
>
>
Of course by using this list you can reach all the knowledge which
kernel community has. But maybe some essential conclusions should be
made more visible and more public? This is not needed for kernel
developers itselves but for example for students or IT project leaders
or investors. But okay, I got answer from you.
And thanks for god that this have never happened: "Linux kernel
developers summit 2004", 23.-28. december, Sri Lanka
regards,
Indrek