aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2019-10-08Remove maximum queue backlog guestimation.David Oberhollenzer
Partially revers 2b7df394057c013fd042b85a4d5fd0104ba4a9be. Making the queue fill the entire RAM had some unintended side effects that need further investigation. For now, revert back to the old behaviour. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-08Fix: propperly clear index entryDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-08Fix storing of the directory index in the inode, size accountingDavid Oberhollenzer
After the last modification, the directory index size counter no longer counts the size, but the number of index entries. To avoid future confusion, remove the thing entirely and compute the size on the fly. Furthermore, eliminate unaligned memory writes and fix the darn counter once again. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-08Fix recovery of directory index in inode readerDavid Oberhollenzer
Apparently mksquashfs writes an actual usage count (1 = 1 entry, 2 = 2 entries; i.e. NOT off by one). Also, if it does happen to be garbage, guard against an overflow. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-08Fix unaligned reads in write_dir_indexDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-08Make sure the statistics counters are initialized to 0David Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Cleanup: consistent, fixed size data types for generic flag fieldsDavid Oberhollenzer
Instead of using int or unsigned int for generic function flag arguments, consistently use an unsigned, fixed size type. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Do an explicit "is filename sane" checkDavid Oberhollenzer
Until now, filenames containing '/' or being equal to '..' or '.' where not handled explicitly, because they are canonicalized later, which will then fail. This commit adds an explicit check to make those fail immediately with a clear, specific error message. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Cleanup: move libutil related headers to "util" sub directoryDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Cleanup: move write_data to libtarDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Replace write_data in libcommon with less generic, opencoded oneDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Move print_version declaration from util.h to common.hDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Cleanup: move directory handling code to libcommonDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Cleanup: move read_data function to libtarDavid Oberhollenzer
Its the only user. The other code doesn't touch raw file descriptors anymore. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Cleanup: Move padd_file function to libtarDavid Oberhollenzer
It's only ever used for padding tarballs. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Add an implementation for an sqfs_file_t writing to stdoutDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Rename libsqfshelper to libcommonDavid Oberhollenzer
That is IMO less confusing and express what it is (i.e. what it has become) more clearly, i.e. common code shared by the utilities. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Rename "hihglevel.h" to the more appropriate "common.h"David Oberhollenzer
It only contains helpers for _common_ stuff for all the utilities. The actual high level stuff has been moved to libsquashfs a while ago. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Fix spelling of "align"David Oberhollenzer
Before the misspelled version has a chance to become stable API. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-07Improve ABI backward & forward compatibillityDavid Oberhollenzer
- Padd the compressor config union - 128 bytes aught to be enough for everyone, i.e. future compressors. - Insist that the padding space is initialized to 0. If a field gets added to an existing compressor, it can test for 0 as a sentinel value. - Add a size field to the hook structure, aka "the Microsoft way". - The explanation is in the comment. - Don't make the Microsoft mistake of checking for >=, insist on *exact* size match. Future users will need a fallback if their hooks are rejected. But at least they will be rejected instead of silently not being used. - Add an unsupported flag check to the dir tree reader. - Add a basic abi unit test that, for now, checks the size of the compressor config struct fields. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-06Fix missing stdlib.h includeDavid Oberhollenzer
Problem only showed up on a different machine and went unnoticed earlier. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-06Cleanup: move padd_sqfs to helper libraryDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-06Fix resource leaks reported by coverityDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-06Improve error reporting for sqfs2tar, rdsquashfs, sqfsdiffDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-05Set sane defaults for backlog and job countDavid Oberhollenzer
Instead of defaulting to -j 1 and -Q 10 times the job count if no option is provided, let tar2sqfs and gensquashfs use all available CPUs and about roughly half the available RAM, if this information can be gathered. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-05Add more diagnostic messagesDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-05Improve error reporting for tar2sqfs and gensquashfsDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-05Add a perror like function for libsquashfs error codesDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-05Merge all the common code for generating imagesDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-03Cleanup: simplify tree_node_to_inodeDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-03Remove accidental internal use of stdint typesDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-03Add a counter-part to sqfs_inode_get_xattr_indexDavid Oberhollenzer
Combines all the type depenend attribute setting and inode type promotion into a single function. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-03Cleanup: merge common code paths in tree serializationDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-03Move tree_node_to_inode function to fstree serialization codeDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-10-03Store directory index in generic inode structureDavid Oberhollenzer
This commit modifies the dir writer to store the directory index directly in the payload area of the inode it creates. The size of the index in bytes is stored in a seperate field. Storing the index is moved to the write inode function. The read inode function is in turn also modified to read the directory index from disk and actually store it in the inode structure in memory for later use. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-30Add flag field to directory writer begin for future expansionDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Cleanup: rename "compress.h" to "compressor.h"David Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Fix str_table_t error behaviour, commentsDavid Oberhollenzer
- str_table_t is used by libsquashfs; Don't write to stderr, report an error code instead. - Fix the comments about that and fix the SPDX license identifier while we're at it. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Fix inode numbering: always start with 1, use 0 as parent for the rootDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Fix inode writer: do not store the file mode, only permissionsDavid Oberhollenzer
SquashFS restores the file mode from the type, so those bits are completely redundant. On a side note: this also reduces entropy, allowing the compressor to *maybe* squeeze a few more bytes out. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Cleanup: fstree no longer has any use for the block sizeDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Cleanup: fstree_from_file does not need to change working directoryDavid Oberhollenzer
The file_info_t no longer stores the size or other such information, so there is no need to do a stat on the input file. This also means that gensquashfs no longer needs to change the working directory when using the function. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Bring back the don't fragment flagDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Make the data writer padding hookableDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-29Fix data writer: remember where we stored the paddingDavid Oberhollenzer
If we don't store the padding location in the deduplication list, the block deduplication might try to deduplicate across two files with padding in between, which can't be done. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-28Cleanup: remove string allocation helper functionDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-28Replace fstree/sqfshelper xattr code with sqfs_xattr_writer_tDavid Oberhollenzer
Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-28Move fstree selinux code to gensquashfsDavid Oberhollenzer
Same rational as for the dir-scanner code: It's actually the only user and it is going to get a lot closer integerated with libsquashfs. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-28Move fstree_from_dir to gensquashfs codeDavid Oberhollenzer
It's actually the only user and the dir-scanner xattr code is going to get a lot closer integerated with libsquashfs. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>
2019-09-28Fix xattr writer: don't generate an xattr table if there is noneDavid Oberhollenzer
If there were no key-value pairs recoreded during the lifetime of the xattr writer, take a short cut out of the flush function and set the super block values to indicate that. Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>