On Sun, 2025-05-11 at 10:45 +0000, 李扬韬 wrote: > Hi Slava, > > > Frankly speaking, if we would like to rework the debugging > > framework in HFS/HFS+, then I prefer to switch on pr_debug() and to > > use dynamic debug framework of Linux kernel [1]. It will provide > > the more flexible solution. > > I'll try it. Sounds good! As far as I can see, it should be pretty simple fix. And I think we we need to unify the code for HFS and HFS+. I am considering to introduce HFS folder in include/linux where we can gather small duplicated code patterns of HFS/HFS+. However, for more complex duplicated code patterns, maybe, we need to consider of introduction an fs/hfs_common module or shared HFS subsystem. > > By the way, I plan to export disk, mem and some statistics related > information to debugfs, just like f2fs does. Any suggestions? > I think there is no troubles with it. But could you please share which particular stats would you like to keep in debugfs? Which in-core and on-disk objects fields will be stored into debugfs? Thanks, Slava.
