annotate fs/jffs2/TODO @ 912:dd42cdb0ab89

[IA64] Build blktap2 driver by default in x86 builds.

add CONFIG_XEN_BLKDEV_TAP2=y to buildconfigs/linux-defconfig_xen_ia64.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
author Isaku Yamahata <yamahata@valinux.co.jp>
date Mon Jun 29 12:09:16 2009 +0900 (2009-06-29)
parents 831230e53067
rev   line source
ian@0 1 $Id: TODO,v 1.18 2005/09/22 11:24:56 dedekind Exp $
ian@0 2
ian@0 3 - support asynchronous operation -- add a per-fs 'reserved_space' count,
ian@0 4 let each outstanding write reserve the _maximum_ amount of physical
ian@0 5 space it could take. Let GC flush the outstanding writes because the
ian@0 6 reservations will necessarily be pessimistic. With this we could even
ian@0 7 do shared writable mmap, if we can have a fs hook for do_wp_page() to
ian@0 8 make the reservation.
ian@0 9 - disable compression in commit_write()?
ian@0 10 - fine-tune the allocation / GC thresholds
ian@0 11 - chattr support - turning on/off and tuning compression per-inode
ian@0 12 - checkpointing (do we need this? scan is quite fast)
ian@0 13 - make the scan code populate real inodes so read_inode just after
ian@0 14 mount doesn't have to read the flash twice for large files.
ian@0 15 Make this a per-inode option, changable with chattr, so you can
ian@0 16 decide which inodes should be in-core immediately after mount.
ian@0 17 - test, test, test
ian@0 18
ian@0 19 - NAND flash support:
ian@0 20 - almost done :)
ian@0 21 - use bad block check instead of the hardwired byte check
ian@0 22
ian@0 23 - Optimisations:
ian@0 24 - Split writes so they go to two separate blocks rather than just c->nextblock.
ian@0 25 By writing _new_ nodes to one block, and garbage-collected REF_PRISTINE
ian@0 26 nodes to a different one, we can separate clean nodes from those which
ian@0 27 are likely to become dirty, and end up with blocks which are each far
ian@0 28 closer to 100% or 0% clean, hence speeding up later GC progress dramatically.
ian@0 29 - Stop keeping name in-core with struct jffs2_full_dirent. If we keep the hash in
ian@0 30 the full dirent, we only need to go to the flash in lookup() when we think we've
ian@0 31 got a match, and in readdir().
ian@0 32 - Doubly-linked next_in_ino list to allow us to free obsoleted raw_node_refs immediately?
ian@0 33 - Remove totlen from jffs2_raw_node_ref? Need to have totlen passed into
ian@0 34 jffs2_mark_node_obsolete(). Can all callers work it out?
ian@0 35 - Remove size from jffs2_raw_node_frag.
ian@0 36
ian@0 37 dedekind:
ian@0 38 1. __jffs2_flush_wbuf() has a strange 'pad' parameter. Eliminate.
ian@0 39 2. get_sb()->build_fs()->scan() path... Why get_sb() removes scan()'s crap in
ian@0 40 case of failure? scan() does not clean everything. Fix.