kernel/linux.git/usr/gen_init_cpio.c, branch v6.19.11

gen_init_cpio: Ignore fsync() returning EINVAL on pipes

2025-10-07T16:53:05+00:00

The reproducer: echo | ./usr/gen_init_cpio /dev/stdin > /dev/null fsync() on a pipe fd returns -EINVAL, which makes gen_init_cpio fail. Ignore -EINVAL from fsync(). Fixes: ae18b94099b0 ("gen_init_cpio: support -o parameter") Cc: David Disseldorp Cc: Nicolas Schier Signed-off-by: Dmitry Safonov Reviewed-by: David Disseldorp Link: https://patch.msgid.link/20251007-gen_init_cpio-pipe-v2-1-b098ab94b58a@arista.com Signed-off-by: Nathan Chancellor

gen_init_cpio: add -a as reflink optimization

2025-08-20T23:02:56+00:00

As described in buffer-format.rst, the existing initramfs.c extraction logic works fine if the cpio filename field is padded out with trailing zeros, with a caveat that the padded namesize can't exceed PATH_MAX. Add filename zero-padding logic to gen_init_cpio, which can be triggered via the new -a parameter. Performance and storage utilization is improved for Btrfs and XFS workloads, as copy_file_range can reflink the entire source file into a filesystem block-size aligned destination offset within the cpio archive. Btrfs benchmarks run on 6.15.8-1-default (Tumbleweed) x86_64 host: > truncate --size=2G /tmp/backing.img > /sbin/mkfs.btrfs /tmp/backing.img ... Sector size: 4096 (CPU page size: 4096) ... > sudo mount /tmp/backing.img mnt > sudo chown $USER mnt > cd mnt mnt> dd if=/dev/urandom of=foo bs=1M count=20 && cat foo >/dev/null ... mnt> echo "file /foo foo 0755 0 0" > list mnt> perf stat -r 10 gen_init_cpio -o unaligned_btrfs list ... 0.023496 +- 0.000472 seconds time elapsed ( +- 2.01% ) mnt> perf stat -r 10 gen_init_cpio -o aligned_btrfs -a 4096 list ... 0.0010010 +- 0.0000565 seconds time elapsed ( +- 5.65% ) mnt> /sbin/xfs_io -c "fiemap -v" unaligned_btrfs unaligned_btrfs: EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS 0: [0..40967]: 695040..736007 40968 0x1 mnt> /sbin/xfs_io -c "fiemap -v" aligned_btrfs aligned_btrfs: EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS 0: [0..7]: 26768..26775 8 0x0 1: [8..40967]: 269056..310015 40960 0x2000 2: [40968..40975]: 26776..26783 8 0x1 mnt> /sbin/btrfs fi du unaligned_btrfs aligned_btrfs Total Exclusive Set shared Filename 20.00MiB 20.00MiB 0.00B unaligned_btrfs 20.01MiB 8.00KiB 20.00MiB aligned_btrfs XFS benchmarks run on same host: > sudo umount mnt && rm /tmp/backing.img > truncate --size=2G /tmp/backing.img > /sbin/mkfs.xfs /tmp/backing.img ... = reflink=1 ... data = bsize=4096 blocks=524288, imaxpct=25 ... > sudo mount /tmp/backing.img mnt > sudo chown $USER mnt > cd mnt mnt> dd if=/dev/urandom of=foo bs=1M count=20 && cat foo >/dev/null ... mnt> echo "file /foo foo 0755 0 0" > list mnt> perf stat -r 10 gen_init_cpio -o unaligned_xfs list ... 0.011069 +- 0.000469 seconds time elapsed ( +- 4.24% ) mnt> perf stat -r 10 gen_init_cpio -o aligned_xfs -a 4096 list ... 0.001273 +- 0.000288 seconds time elapsed ( +- 22.60% ) mnt> /sbin/xfs_io -c "fiemap -v" unaligned_xfs unaligned_xfs: EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS 0: [0..40967]: 106176..147143 40968 0x0 1: [40968..65023]: 147144..171199 24056 0x801 mnt> /sbin/xfs_io -c "fiemap -v" aligned_xfs aligned_xfs: EXT: FILE-OFFSET BLOCK-RANGE TOTAL FLAGS 0: [0..7]: 120..127 8 0x0 1: [8..40967]: 192..41151 40960 0x2000 2: [40968..40975]: 236728..236735 8 0x0 3: [40976..106495]: 236736..302255 65520 0x801 The alignment is best-effort; a stderr message is printed if alignment can't be achieved due to PATH_MAX overrun, with fallback to non-padded filename. This allows it to still be useful for opportunistic alignment, e.g. on aarch64 Btrfs with 64K block-size. Alignment failure messages provide an indicator that reordering of the cpio-manifest may be beneficial. Archive read performance for reflinked initramfs images may suffer due to the effects of fragmentation, particularly on spinning disks. To mitigate excessive fragmentation, files with lengths less than data_align aren't padded. Signed-off-by: David Disseldorp Reviewed-by: Nicolas Schier Link: https://lore.kernel.org/r/20250819032607.28727-8-ddiss@suse.de Signed-off-by: Nathan Chancellor

gen_init_cpio: avoid duplicate strlen calls

2025-08-20T23:02:56+00:00

We determine the filename length for the cpio header, so shouldn't recalculate it when writing out the filename. Signed-off-by: David Disseldorp Reviewed-by: Nicolas Schier Link: https://lore.kernel.org/r/20250819032607.28727-5-ddiss@suse.de Signed-off-by: Nathan Chancellor

gen_init_cpio: attempt copy_file_range for file data

2025-08-20T23:02:55+00:00

The copy_file_range syscall can improve copy performance by cloning extents between cpio archive source and destination files. Existing read / write based copy logic is retained for fallback in case the copy_file_range syscall is unsupported or unavailable due to cross-filesystem EXDEV, etc. Clone or reflink, as opposed to copy, of source file extents into the output cpio archive may (e.g. on Btrfs and XFS) require alignment of the output to the filesystem block size. This could be achieved by inserting padding entries into the cpio archive manifest. Signed-off-by: David Disseldorp Reviewed-by: Nicolas Schier Link: https://lore.kernel.org/r/20250819032607.28727-4-ddiss@suse.de Signed-off-by: Nathan Chancellor

gen_init_cpio: support -o parameter

2025-08-20T23:02:55+00:00

This is another preparatory change to allow for reflink-optimized cpio archives with file data written / cloned via copy_file_range(). The output file is truncated prior to write, so that it maps to usr/gen_initramfs.sh usage. It may make sense to offer an append option in future, for easier archive concatenation. Signed-off-by: David Disseldorp Reviewed-by: Nicolas Schier Link: https://lore.kernel.org/r/20250819032607.28727-3-ddiss@suse.de Signed-off-by: Nathan Chancellor

gen_init_cpio: write to fd instead of stdout stream

2025-08-20T23:02:55+00:00

In preparation for more efficient archiving using copy_file_range(), switch from writing archive data to stdout to using STDOUT_FILENO and I/O via write(), dprintf(), etc. Basic I/O error handling is added to cover cases such as ENOSPC. Partial writes are treated as errors. Signed-off-by: David Disseldorp Reviewed-by: Nicolas Schier Link: https://lore.kernel.org/r/20250819032607.28727-2-ddiss@suse.de Signed-off-by: Nathan Chancellor

gen_init_cpio: Apply mtime supplied by user to all file types

2023-12-29T13:25:21+00:00

Currently gen_init_cpio -d is applied to symlinks, directories and special files. These files are created by gen_init_cpio from their description. Without option current time(NULL) is used. And regular files that go in initramfs are created before cpio generation, so their mtime(s) are preserved. This is usually not an issue as reproducible builds should rebuild everything in the distribution, including binaries, configs and whatever other regular files may find their way into kernel's initramfs. On the other hand, gen_initramfs.sh usage claims: > -d Use date for all file mtime values Ar Arista initramfs files are managed with version control system that preserves mtime. Those are configs, boot parameters, init scripts, version files, platform-specific files, probably some others, too. While it's certainly possible to work this around by copying the file into temp directory and adjusting mtime prior to gen_init_cpio call, I don't see why it needs workarounds. The intended user of -d option is the one that needs to create a reproducible build, see commit a8b8017c34fe ("initramfs: Use KBUILD_BUILD_TIMESTAMP for generated entries"). If a user wants the build reproduction, they use -d , which can be set on all types of files, without surprising exceptions and workarounds. Let's KISS here and just apply the time that user specified with -d option. Based-on-a-patch-by: Baptiste Covolato Link: https://lore.kernel.org/lkml/20181025215133.20138-1-baptiste@arista.com/ Signed-off-by: Dmitry Safonov Signed-off-by: Masahiro Yamada

initramfs: Check negative timestamp to prevent broken cpio archive

2023-04-16T08:37:01+00:00

Similar to commit 4c9d410f32b3 ("initramfs: Check timestamp to prevent broken cpio archive"), except asserts that the timestamp is non-negative. This can happen when the KBUILD_BUILD_TIMESTAMP is a value before UNIX epoch, which may be set when making reproducible builds that don't want to look like they use a valid date. While support for dates before 1970 might not be supported, this is more about preventing undetected CPIO corruption. The printf's use a minimum length format specifier, and will happily make the field longer than 8 characters if they need to. Signed-off-by: Benjamin Gray Reviewed-by: Andrew Donnellan Tested-by: Andrew Donnellan Signed-off-by: Masahiro Yamada

usr/gen_init_cpio.c: remove unnecessary -1 values from int file

2022-10-03T21:21:44+00:00

The file variable is assigned first, it does not need to be initialized. Link: https://lkml.kernel.org/r/20220919014406.3242-1-zeming@nfschina.com Signed-off-by: Li zeming Cc: Li zeming Cc: Masahiro Yamada Cc: Nicolas Schier Signed-off-by: Andrew Morton

gen_init_cpio: support file checksum archiving

2022-05-10T01:29:20+00:00

Documentation/driver-api/early-userspace/buffer-format.rst includes the specification for checksum-enabled cpio archives. Implement support for this format in gen_init_cpio via a new '-c' parameter. Link: https://lkml.kernel.org/r/20220404093429.27570-6-ddiss@suse.de Signed-off-by: David Disseldorp Suggested-by: Matthew Wilcox (Oracle) Cc: Al Viro Cc: Christian Brauner Cc: Martin Wilck Signed-off-by: Andrew Morton