* $Id: TODO,v 3.8 2005/01/05 16:16:52 dedekind Exp $

 - support asynchronous operation -- add a per-fs 'reserved_space' count,
   let each outstanding write reserve the _maximum_ amount of physical
   space it could take. Let GC flush the outstanding writes because the
   reservations will necessarily be pessimistic. With this we could even
   do shared writable mmap, if we can have a fs hook for do_wp_page() to
   make the reservation.
 - disable compression in commit_write()?
 - fine-tune the allocation / GC thresholds
 - chattr support - turning on/off and tuning compression per-inode
 - checkpointing (do we need this? scan is quite fast)
 - make the scan code populate real inodes so read_inode just after 
	mount doesn't have to read the flash twice for large files.
	Make this a per-inode option, changable with chattr, so you can
	decide which inodes should be in-core immediately after mount.
 - test, test, test

 - NAND flash support:
	- almost done :)
	- use bad block check instead of the hardwired byte check

 - Optimisations:
   - Split writes so they go to two separate blocks rather than just c->nextblock.
	By writing _new_ nodes to one block, and garbage-collected REF_PRISTINE
	nodes to a different one, we can separate clean nodes from those which
	are likely to become dirty, and end up with blocks which are each far
	closer to 100% or 0% clean, hence speeding up later GC progress dramatically.
   - Stop keeping name in-core with struct jffs3_full_dirent. If we keep the hash in 
     the full dirent, we only need to go to the flash in lookup() when we think we've
     got a match, and in readdir(). 
   - Doubly-linked next_in_ino list to allow us to free obsoleted raw_node_refs immediately?
   - Remove totlen from jffs3_raw_node_ref? Need to have totlen passed into
	jffs3_mark_node_obsolete(). Can all callers work it out?
   - Remove size from jffs3_raw_node_frag. 

   Joakim Tjernlund <joakim.tjernlund@lumentis.se>  Thu, 16 Dec 2004 14:20:43:
   - Consider another checksum algorithm. Crc32 is very expensive
     and JFFS2 suffered severely in the early days. Now that crc32 is
     very optimized that problem is less visible, but crc32 is still
     expensive. Maybe an Adler32 checksum is good enough or a crc16?
   - Don't calculate a Adler32 checksum when compressing with zlib.
     JFFS2 already has its own checksum.
   - Consider changing the start seed to crc from 0 to -1. Zero
     is not a good start seed for crc

   - Currently when JFFS3 finds any CRC error it prints warning message. This is right from one
     hand. CRC errors may appear because of unclean reboots (1) and flash media corruptions (2).
     Unclean reboots happen rather often and, as result, one node may have CRC error (JFFS2 was in
     the middle of writing it when was rebooted uncleanly). Bu JFFS2 still prints CRC error warning.
     That makes users worry. So, it is good to detect such situations and not to print the CRC error
     warning message. Possibly, we may not print that if the bogus node is the last node in the current
     block (may there be one more bogus node in the same block, if there were several consequent
     unclean reboots? - not sure ...)

   - Would be nice to reformat sources and keep lines length < 80
     characters. Currently it is hard to read JFFS3 sources 80xX
     terminals.

   - __totlen elimination: when adding obsolete nodes, we just allocate
     new node_ref and call add_physical_node_ref. That is sux. It
     doesn't merge obsolete nodes. Fix it.
     Wasted space accounting is broken. Pads are treated as dirty with
     new changes. Fix it. 

   - Summaries:
     1. In get_inode_nodes - do not check nodes CRC if their block
	contains summary node (when checking).
     2. For NOR do not check nodes CRC at all (never) when there is summary in
	the block.
     3. For NOR summaries needed either (not only for NAND). But for not
	- no need to keep lots of information in summaries. May be just
	  keep summary only header (no data), may be keep only
	  referenced to nodes within the block. Anyway, most important
	  is just to know that there is summary at the end of block. In
	  this case we may be sure that there was no unclean reboot when
	  this block was written, and we may trust this block - that
	  means we may not check CRCs there.

   - Wasted space: treat pads as wasted only if their size > ISDIRTY
     (__flush_wbuf).