[Intel-gfx] [PATCH] Proc to debugfs migration

Ben Gamari bgamari at gmail.com
Thu Jan 15 01:33:05 CET 2009


Bah, naturally just as I sent this out I realized that anholt sent me
comments on the last iteration of this patch set. I'll incorporate his
proposed changes and any others I receive in the next few hours and send
it out again. Sorry about the spam,

- Ben


On Wed, 2009-01-14 at 19:27 -0500, Ben Gamari wrote:
> Hey everyone,
>     
> Here is my procfs cleanup patchset. As you can see, the first two 
> patches attempt to clean up the existing proc file creation code and 
> port the files' implementations to the seq_file interface. The remaining
> patches focus on implementing a similar system for debugfs files.
> 
> Feedback has generally indicated that while removing drm information 
> from /proc is a worthwhile goal, we need to provide a transition period
> in which we can keep this information available in both debugfs and 
> /proc. The refactoring with seq_file made this almost trivial, requiring
> only that I move the file implementations to a combined file which both
> the proc and debugfs code can link against. I declared these symbols in
> a new header file (drm_info.h) and while I haven't seen any precedent 
> for this in the drm code, I did want to keep from lengthening drmP.h any
> more than it already is, especially given these symbols are only 
> referred to in two places. It is my thought that eventually the proc 
> file code will be pulled out anyways and we'll be able to merge 
> drm_info.[ch] back into drm_debugfs.c. Let me know if anyone has an 
> issue with this.
> 
> Anyways, I think most of it is pretty straightforward. Please pardon the
> less-than-stellar organization of the patchset, I did my best to santize it
> but rebasing while moving whole files around seems to be quite tedious. I
> think that the patch set is pretty much ready for merge. I've been running it
> for several weeks now without incident. Let me know what you think,
> 
>     - Ben
> 




More information about the Intel-gfx mailing list