Messages in this thread | | | Date | Thu, 11 Apr 2013 12:07:28 -0700 | From | Greg KH <> | Subject | Re: [PATCH 00/26] Eliminate create_proc_read_entry() [RFC] |
| |
On Thu, Apr 11, 2013 at 02:27:40PM +0100, David Howells wrote: > > Here is a series of patches to eliminate create_proc_read_entry(), replacing it > with proc_create() and proc_create_data() and the use of seq_file facilities. > > This requires file_operations structs and open methods to be provided in most > cases, though single_open() and single_open_size() suffice for the actual > translation into the seq_file environment. I did try and reduce the number of > file_operations structs and open methods added by placing the show function > pointer into seq_file::private in some places. > > Certain places (hostap most notably) required conversion of some files into > full seq_file iterators to handle the possibility that large lists might > overrun a single buffer full of data. > > This reduces the size of the proc_dir_entry by the removal of the read_proc > pointer, plus the support mechanisms for these have been removed from the > procfs code. > > These patches have been compile tested as best I can (apart from SH for which I > don't have a working compiler). No runtime testing has been done outside of > the core procfs code as I don't have the requisite hardware.
The ones I've been cc:ed on, with the exception of the hostap patch, I will be glad to take though my trees now, unless you want to take them yourself?
thanks,
greg k-h
| |