]> git.ipfire.org Git - thirdparty/e2fsprogs.git/blob - doc/RelNotes/v1.45.5.txt
e2fsck: clean up unwind_pass1() as it's no longer really needed
[thirdparty/e2fsprogs.git] / doc / RelNotes / v1.45.5.txt
1 E2fsprogs 1.45.5 (January 7, 2020)
2 ==================================
3
4 Updates/Fixes since v1.45.4:
5
6 Fixes
7 -----
8
9 E2fsck will no longer force a full file system check if time-based
10 forced checks are disabled and the last mount time or last write time in
11 the superblock are in the future.
12
13 Fix a potential out of bounds write when checking a maliciously
14 corrupted file system. This is probably not exploitable on 64-bit
15 platforms, but may be exploitable on 32-bit binaries depending on how
16 the compiler lays out the stack variables. (Addresses CVE-2019-5188)
17
18 Fixed spurious weekly e-mails when e2scrub_all is run via a cron job
19 on non-systemd systems. (Addresses Debian Bug: #944033)
20
21 Remove an unnecessary sleep in e2scrub which could add up to an
22 additional two second delay during the boot up. Also, avoid trying
23 to reap aborted snapshots if it has been disabled via e2scrub.conf.
24 (Addresses Debian Bug: #948193)
25
26 If a mischievous system administrator mounts a pseudo-file system such
27 as tmpfs with a device name that duplicates another mounted file system,
28 this could potentially confuse resize2fs when it needs to find the mount
29 point of a mounted file system. (Who would have guessed?) Add some
30 sanity checking so that we can make libext2fs more robust against such
31 insanity, at least on Linux. (GNU HURD doesn't support st_rdev.)
32
33 Tune2fs now prohibits enabling or disabling uninit_bg if the file system
34 is mounted, since this could result in the file system getting
35 corrupted, and there is an unfortunate AskUbuntu article suggesting this
36 as a way to modify a file system's UUID on a live file system. (Ext4
37 now has a way to do this safely, using the metadata_csum_seed feature,
38 which was added in the 4.4 Linux kernel.)
39
40 Fix potential crash in e2fsck when rebuilding very large directories on
41 file systems which have the new large_dir feature enable.
42
43 Fix support of 32-bit uid's and gid's in fuse2fs and in mke2fs -d.
44
45 Fix mke2fs's setting bad blocks to bigalloc file systems.
46
47 Fix a bug where fuse2fs would incorrectly report the i_blocks fields for
48 bigalloc file systems.
49
50 Resize2fs's minimum size estimates (via resize2fs -M) estimates are now
51 more accurate when run on mounted file systems.
52
53 Fixed potential memory leak in read_bitmap() in libext2fs.
54
55 Fixed various UBsan failures found when fuzzing file system images.
56 (Addresses Google Bug: #128130353)
57
58 Updated and clarified various man pages.
59
60
61 Performance, Internal Implementation, Development Support etc.
62 --------------------------------------------------------------
63
64 Speed up e2fsck on file systems with a very large number of inodes
65 caused by repeated calls to gettext().
66
67 The inode_io io_manager can now support files which are greater than
68 2GB.
69
70 The ext2_off_t and ext2_off64_t are now signed types so that
71 ext2fs_file_lseek() and ext2fs_file_llseek() can work correctly.
72
73 Reserve codepoint for the fast_commit feature.
74
75 Fixed various Debian packaging issues.
76
77 Fix portability problems for Illumous and on hurd/i386 (Addresses Debian
78 Bug: #944649)
79
80 Always compile the ext2fs_swap_* functions even on little-endian
81 architectures, so that debian/libext2fs.symbols can be consistent across
82 architectures.
83
84 Synchronized changes from Android's AOSP e2fsprogs tree.
85
86 Updated config.guess and config.sub with newer versions from the FSF.
87
88 Update the Chinese and Malay translations from the translation project.