lkml.org 
[lkml]   [2008]   [Mar]   [26]   [last100]   RSS-feed
Views: [more markup]  [less markup]  [headers]  [forward] 
 
Messages in this thread
Complete thread (EXPERIMENTAL)
/
DateWed, 26 Mar 2008 20:09:25 +0000
FromRussell King <>
SubjectRe: [PATCH 1/3] Clocklib: add generic framework for managing clocks.
Digg This
On Thu, Mar 27, 2008 at 01:14:56AM +0900, Paul Mundt wrote:
Conversely it also has fields that I don't need. If struct clk could have been done generically without growing to insane sizes, it would have been done so in linux/clk.h a long time ago. The main thing there is API consistency for drivers, leaving the details up to the architecture.
It's true that there is significant overlap between the different users of the clock framework, but it's also not clear that there's any clean way to share a common implementation (especially since struct clk means totally different things on different architectures). I suspect everyone in the CC list has been through this before, also.

That's the exact reason why I never implemented any kind of framework and just left it as an API for drivers to use. What's behind the API is very platform specific, and as can be seen from the comments, trying to define something common results in something that just doesn't fit in different ways.

Trying to make it expand to fit someone elses platform makes it unsuitable for another due to it becoming too heavy weight.

Personally, I don't have much interest in these patches - had I been interested in having a common framework behind it when I created the API, I'd have written some code.

However, if folk think that they can solve the complexity problem while still allowing for simple implementations...

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:


\
ISP Services
Valid XHTML 1.0!\ /
Valid CSS! Last update: 2008-03-26 21:17    [W:1.019 / U:0.000 seconds]
©2003-2005 Jasper Spaans