xref: /OK3568_Linux_fs/kernel/fs/jffs2/TODO (revision 4882a59341e53eb6f0b4789bf948001014eff981)
1*4882a593Smuzhiyun
2*4882a593Smuzhiyun - support asynchronous operation -- add a per-fs 'reserved_space' count,
3*4882a593Smuzhiyun   let each outstanding write reserve the _maximum_ amount of physical
4*4882a593Smuzhiyun   space it could take. Let GC flush the outstanding writes because the
5*4882a593Smuzhiyun   reservations will necessarily be pessimistic. With this we could even
6*4882a593Smuzhiyun   do shared writable mmap, if we can have a fs hook for do_wp_page() to
7*4882a593Smuzhiyun   make the reservation.
8*4882a593Smuzhiyun - disable compression in commit_write()?
9*4882a593Smuzhiyun - fine-tune the allocation / GC thresholds
10*4882a593Smuzhiyun - chattr support - turning on/off and tuning compression per-inode
11*4882a593Smuzhiyun - checkpointing (do we need this? scan is quite fast)
12*4882a593Smuzhiyun - make the scan code populate real inodes so read_inode just after
13*4882a593Smuzhiyun	mount doesn't have to read the flash twice for large files.
14*4882a593Smuzhiyun	Make this a per-inode option, changeable with chattr, so you can
15*4882a593Smuzhiyun	decide which inodes should be in-core immediately after mount.
16*4882a593Smuzhiyun - test, test, test
17*4882a593Smuzhiyun
18*4882a593Smuzhiyun - NAND flash support:
19*4882a593Smuzhiyun	- almost done :)
20*4882a593Smuzhiyun	- use bad block check instead of the hardwired byte check
21*4882a593Smuzhiyun
22*4882a593Smuzhiyun - Optimisations:
23*4882a593Smuzhiyun   - Split writes so they go to two separate blocks rather than just c->nextblock.
24*4882a593Smuzhiyun	By writing _new_ nodes to one block, and garbage-collected REF_PRISTINE
25*4882a593Smuzhiyun	nodes to a different one, we can separate clean nodes from those which
26*4882a593Smuzhiyun	are likely to become dirty, and end up with blocks which are each far
27*4882a593Smuzhiyun	closer to 100% or 0% clean, hence speeding up later GC progress dramatically.
28*4882a593Smuzhiyun   - Stop keeping name in-core with struct jffs2_full_dirent. If we keep the hash in
29*4882a593Smuzhiyun     the full dirent, we only need to go to the flash in lookup() when we think we've
30*4882a593Smuzhiyun     got a match, and in readdir().
31*4882a593Smuzhiyun   - Doubly-linked next_in_ino list to allow us to free obsoleted raw_node_refs immediately?
32*4882a593Smuzhiyun   - Remove size from jffs2_raw_node_frag.
33*4882a593Smuzhiyun
34*4882a593Smuzhiyundedekind:
35*4882a593Smuzhiyun1. __jffs2_flush_wbuf() has a strange 'pad' parameter. Eliminate.
36*4882a593Smuzhiyun2. get_sb()->build_fs()->scan() path... Why get_sb() removes scan()'s crap in
37*4882a593Smuzhiyun   case of failure? scan() does not clean everything. Fix.
38