tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Status of revivesa



On Mon, Sep 29, 2008 at 02:23:19PM -0500, James Chacon wrote:
>
> On Sep 29, 2008, at 2:16 PM, Quentin Garnier wrote:
>
>> On Mon, Sep 29, 2008 at 11:43:17AM -0700, Bill Stouder-Studenmund  
>> wrote:
>>> On Mon, Sep 29, 2008 at 10:39:18AM -0400, Christos Zoulas wrote:
>> [...]
>>> What still isn't clear to me is what exactly the negatives are of
>>> revivesa.
>>>
>>> The biggest one I'm hearing is that a number of people HATE it. Flat 
>>> out
>>> HATE it. I'm not really sure what to do with this one, since it's  
>>> hard to
>>> understand. It's an emotion, and we usually try to stick to technical
>>> points.
>>>
>>> The SA that's on revivesa is a kernel option. If you don't want it,  
>>> don't
>>> enable it. If we find a catastrophic flaw in it (or security issue)  
>>> late
>>> in the 5.0 release proces, we turn it off in the default kernels and 
>>> tell
>>> people to only re-enable it with caution.
>>>
>>> SA is not becoming, nor do I ever envision it becoming, the default
>>> threading out-of-the-box for NetBSD. Some sites may eventually  
>>> prefer it,
>>> but that's a specialized situation. And something an admin would  
>>> have to
>>> explicitly select.
>>
>> The main issue with SA is maintainability.  We failed to maintain it
>> once, how likely are we to succeed now?  I have a lot of respect for
>> your work on that, but I don't really see anyone maintaining it if for
>> some reason you can't really do it anymore.
>>
>> Supposing it will be enabled by default, it won't be at risk of simply
>> rotting;  but merely compiling doesn't equal working, especially for a
>> piece of code that is affected by many different areas of the kernel,
>> including MD ones.  And once 5.0 is out, it will hardly be tested  
>> until
>> 6.0 is ready for release because you don't upgrade a system from a
>> release to current the same way you upgrade a system from one release 
>> to
>> another.  The GDT issue is a perfect example of that.
>>
>
> The same issue about maintenance can be made of a lot of code in the  
> system. Beyond Andrew just how many people are diving in and maintaining 
> the replacement thread code? It's always a small number...

Yes, but NFS, Andrew's threading stuff, UVM, all of those are used
constantly in -current, that won't be the case of SA...  And please note
I am just voicing what I think is a legitimate concern;  that doesn't
mean it is something impossible to live with.

-- 
Quentin Garnier - cube%cubidou.net@localhost - cube%NetBSD.org@localhost
"See the look on my face from staying too long in one place
[...] every time the morning breaks I know I'm closer to falling"
KT Tunstall, Saving My Face, Drastic Fantastic, 2007.

Attachment: pgpTRr0AYNI8P.pgp
Description: PGP signature



Home | Main Index | Thread Index | Old Index