| E2fsprogs 1.45.5 (January 7, 2020) |
| ================================== |
| |
| Updates/Fixes since v1.45.4: |
| |
| Fixes |
| ----- |
| |
| E2fsck will no longer force a full file system check if time-based |
| forced checks are disabled and the last mount time or last write time in |
| the superblock are in the future. |
| |
| Fix a potential out of bounds write when checking a maliciously |
| corrupted file system. This is probably not exploitable on 64-bit |
| platforms, but may be exploitable on 32-bit binaries depending on how |
| the compiler lays out the stack variables. (Addresses CVE-2019-5188) |
| |
| Fixed spurious weekly e-mails when e2scrub_all is run via a cron job |
| on non-systemd systems. (Addresses Debian Bug: #944033) |
| |
| Remove an unnecessary sleep in e2scrub which could add up to an |
| additional two second delay during the boot up. Also, avoid trying |
| to reap aborted snapshots if it has been disabled via e2scrub.conf. |
| (Addresses Debian Bug: #948193) |
| |
| If a mischievous system administrator mounts a pseudo-file system such |
| as tmpfs with a device name that duplicates another mounted file system, |
| this could potentially confuse resize2fs when it needs to find the mount |
| point of a mounted file system. (Who would have guessed?) Add some |
| sanity checking so that we can make libext2fs more robust against such |
| insanity, at least on Linux. (GNU HURD doesn't support st_rdev.) |
| |
| Tune2fs now prohibits enabling or disabling uninit_bg if the file system |
| is mounted, since this could result in the file system getting |
| corrupted, and there is an unfortunate AskUbuntu article suggesting this |
| as a way to modify a file system's UUID on a live file system. (Ext4 |
| now has a way to do this safely, using the metadata_csum_seed feature, |
| which was added in the 4.4 Linux kernel.) |
| |
| Fix potential crash in e2fsck when rebuilding very large directories on |
| file systems which have the new large_dir feature enable. |
| |
| Fix support of 32-bit uid's and gid's in fuse2fs and in mke2fs -d. |
| |
| Fix mke2fs's setting bad blocks to bigalloc file systems. |
| |
| Fix a bug where fuse2fs would incorrectly report the i_blocks fields for |
| bigalloc file systems. |
| |
| Resize2fs's minimum size estimates (via resize2fs -M) estimates are now |
| more accurate when run on mounted file systems. |
| |
| Fixed potential memory leak in read_bitmap() in libext2fs. |
| |
| Fixed various UBsan failures found when fuzzing file system images. |
| (Addresses Google Bug: #128130353) |
| |
| Updated and clarified various man pages. |
| |
| |
| Performance, Internal Implementation, Development Support etc. |
| -------------------------------------------------------------- |
| |
| Speed up e2fsck on file systems with a very large number of inodes |
| caused by repeated calls to gettext(). |
| |
| The inode_io io_manager can now support files which are greater than |
| 2GB. |
| |
| The ext2_off_t and ext2_off64_t are now signed types so that |
| ext2fs_file_lseek() and ext2fs_file_llseek() can work correctly. |
| |
| Reserve codepoint for the fast_commit feature. |
| |
| Fixed various Debian packaging issues. |
| |
| Fix portability problems for Illumous and on hurd/i386 (Addresses Debian |
| Bug: #944649) |
| |
| Always compile the ext2fs_swap_* functions even on little-endian |
| architectures, so that debian/libext2fs.symbols can be consistent across |
| architectures. |
| |
| Synchronized changes from Android's AOSP e2fsprogs tree. |
| |
| Updated config.guess and config.sub with newer versions from the FSF. |
| |
| Update the Chinese and Malay translations from the translation project. |