summaryrefslogtreecommitdiff
path: root/feature-removal-schedule.txt
blob: f9910f11d5f2ac8e7798239b9d2070f97715f120 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
The following is a list of files and features that are going to be
removed in the mtd-utils source tree.  Every entry should contain what
exactly is going away, why it is happening, and who is going to be doing
the work.  When the feature is removed from the utils, it should also
be removed from this file.

---------------------------
1. Kill --squash-rino-perm and --nosquash-rino-perm parameters of mkfs.ubifs.

Originally, mkfs.ubifs did not have these parameters, and it always set
permissions for the UBIFS root inode to be {uid=gid=root, u+rwx,go+rx}. This was
a bug which was found too late, when mkfs.ubifs had already been used in
production. To fix this bug, 2 new options were introduced:
--squash-rino-perm which preserves the old behavior and --nosquash-rino-perm
which makes mkfs.ubifs use the right permissions for the root
inode.

For a while (releases 1.4.0-1.4.3) --squash-rino-perm was the default, and if
neither --squash-rino-perm nor --nosquash-rino-perm were used,
mkfs.ubifs printed a warning. This was introduced in mtd-utils-1.4.0 (13 Sep 2010).

Now we have removed a warning and made --nosquash-rino-perm to be the
default. Also, both options are declared deprecated, so users should try
to stop using them.

The further step is to remove both of them.

---------------------------
2. Kill -m parameter of mtdinfo

We cannot assume that mtd device names follow the "/dev/mtd%d" pattern,
because it is up to udev rules to name the devices. So we are removing
the -m option. For now, we just have a warning, but the option will be
removed in release 1.4.6.

---------------------------
3. flash_info utility

This is duplicating behavior with the mtdinfo utility.  Now the util
warns when people use it, but it'll be removed in release 1.4.6.

---------------------------

4. Rename --noskipbad option, kill --omitbad, change default BB method for nanddump

Our bad block handling methods are getting disorganized. They will be replaced
by a series of --bb=METHOD options (see `nanddump --help').

In addition, the previous default bad block behavior was to dump 0xFF in place
of bad blocks. This was not a good "inverse operation" of nandwrite. Thus, the
current default behavior has become the (new) option --bb=padbad, and in 1.4.6,
the default will be --bb=skipbad.

The new `--bb=METHOD' options are included in this release, but all feature
removals and changed defaults will occur in 1.4.6. For now, there are
appropriate deprecation warning messages and an alert telling about the
soon-to-be-default settings.

Transition summary table:
-----------------------------------------------------------------------------------------------------------
 Old option     New option                 Comment
-----------------------------------------------------------------------------------------------------------
 <default>      --bb=padbad                dump flash data, substituting 0xFF for any bad blocks
 --noskipbad    --bb=dumpbad               dump flash data, including any bad blocks
 --skipbad      --bb=skipbad, <default>    dump good data, completely skipping any bad blocks (new default)
 --omitbad      N/A                        very similar to `skipbad', will be removed soon

---------------------------