Skip to content

Conversation

koenkooi
Copy link
Contributor

Upstream OE changed the defaults, so the original stanza prevented all types of tarballs. Enable tar.gz to be able to populate external media like microSD cards a bit easier.

Gzip has been picked over xz because of its speed, both in compression as in decompression. With pigz installed it scales linearly with the number of CPU cores, removing it as the bottleneck.

@koenkooi koenkooi self-assigned this Mar 18, 2025
# Avoid to duplicate the rootfs tarball by generating both tar.gz/tar.xz
IMAGE_FSTYPES:remove = "tar.gz"
IMAGE_FSTYPES:remove = "tar.xz"
IMAGE_FSTYPES:append = " tar.gz"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If it is a current default, do we need to append it manually?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

See previous comment

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Append is kind of evil as it will affect every image related recipe, including initramfs images (since they just replace the value with a normal set '=').

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually... What is the primary reason for enabling/changing tarball generation at all? It looks like a user decision instead of a setting that we should be enforcing in the distro configuration.

Enable `tar.gz` to be able to populate external media like microSD cards
a bit easier. OE-core has `tar.gz` as its default, but BSP layers tend
to overwrite that, so DISTROs need to use `:append` to ensure the
desired types get generated.

Gzip has been picked over xz because of its speed, both in compression
as in decompression. With `pigz` installed it scales linearly with the
number of CPU cores, removing it as the bottleneck.

Signed-off-by: Koen Kooi <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants