Messages in this thread | | | Date | Fri, 3 Jul 2009 18:10:21 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH] vt: add an event interface |
| |
* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> [...] Quite frankly the way some people behave with it is a > disgrace and it puts people off contributing to the kernel when > their 500 line driver gets nothing but emails from people saying > "that space is wrong". [...]
I'd like to address your generic argument here because i think it's interesting.
The thing is, IMO if the author or the affected maintainer(s) are not willing to do _basic_ cleanup of new drivers, they simply dont deserve more substantial review.
Why? Because their lack of basic care is a sign that they will likely be unwilling to expend the (far bigger) effort of actually do ongoing maintenance of the driver.
drivers/isdn/ is one example of unclean code, which, once it got upstream, never got cleaned up.
Its also a positive feedback loop: lack of basic cleanups deters contributors (for example i personally try to stay away from 'weird looking' code as 'not worth the effort'), which makes the code even worse ... which then bitrots as kernel facilities slowly change as the months go on.
So such unclean drivers you mention above should probably go to drivers/staging/ - that is an effort that tries to de-crappify drivers before they get mixed into core drivers.
All in one, i dont see at all the harm from people looking at stylistic issues first - it's an obvious first easy level of trust to inject into an unknown piece of code - if the reply to that minimal stylistic review is fixes then it makes sense to inject more effort into reviewing the driver. If the reply is defiance or inaction then the reviewer probably should not bother because nobody cares about his feedback.
Thanks,
Ingo
| |