questions for Linus

[Pei Lin]

2011/5/31 Greg KH <greg at kroah.com>:
> On Tue, May 31, 2011 at 02:12:36PM +0700, Mulyadi Santosa wrote:
>> Hi Greg..
>>
>> I came with these questions for Linus, hope you're kind enough to
>> collect it and forward it to Linus in LinuxCon:
>>
>> 1. When will Linux 2.8.x start? and what are the plans regarding the
>> development model? do we back in dual 2.4.x/2.5x era? personally I
>> think that model is nice.... makes one really know which tree to
>> follow when he needs stable one, or the devel one...
>
> This has been answered already.
>
>> 2. I saw task/process scheduler is one that under heavy concern
>> lately, take a look on "200 line patch that does wonder" for example.
>> I am thinking, we better really incorporate pluggable scheduler
>> framework, that enable us to change scheduling core algorithm on the
>> fly. What do you think?
>
> This has been covered many times by the scheduler developers already.
> With cgroups, you can do lots of neat things already, no need to change
> any kernel code.
>


As the newbie, look at here, suddenly one thought rush in my brain.
After twenty years development, the kernel became bigger and bigger.
The cost time for building the kernel becomes much more.
If someone writes one new feature in kernel core, he maybe wants to
load the component directly into running kernel instead of rebuild the
kernel and reboot it.
Could the kernel support all components reloadable dynamically in core
as scheduler, memory management etc?
Now the driver subsystem is modularity, could kernel make everything
modularity and all things be hotplugged?
Even kernel can breakdown to little pieces which running on different
computers or terminals communicate with network or others.
Kernel can be one and more points connection. A little more crazy, :-)

Best Regards
-Lin

>> 3. I read somewhere that once Coverity helps Linux kernel developers
>> found out unseen bugs using sophisticated static code analysis. Is
>> that still happening now? Probably a good deal, squashing bugs in
>> nowadays kernel's size is pretty hard, I could say..
>
> I do not think Coverity is doing anything with the kernel anymore.  But
> we have lots of other good code coverage tools that do static analysis
> that people are using, and fixing bugs with today (sparse, coccinelle,
> etc.)
>
> But this is a good one, I'll try to work it into the questions, thanks,
>
> greg k-h
>
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies at kernelnewbies.org
> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
>