Closed ashthespy closed 3 years ago
This shall be carefully tested before merging...
I know the diff looks large -- but it's due to just a file rename (mkinitramfs-buster -> mkinitramfs-custom
)
We have been using the same mkinitramfs-buster
script since the start :-)
Here is the changeset when comparing with the old mkinitramfs-buster
https://www.diffchecker.com/AW1nLjbM
The main diffs are -- setting gzip level to -9 -- streamline volumio-init-updater addition
What is the advantage of -- setting gzip level to -9 ?
I believe the most cases the best way to reduce mkinitramfs is with using modules= list instead of modules=most. X86 may be an exception, as we currently need to load many storage device driver modules. Perhaps we should have all storage & graphics modules built-in and then also use modules=list (which drops the rest), never tried that...
We did have modules=list in the past, didn't we? what would be your ideal approach?
What is the advantage of -- setting gzip level to -9 ?
It should reduce the initrd size (in theory), with a penalty w.t.r time to compress
man gzip
-# --fast --best Regulate the speed of compression using the specified digit #, where -1 or --fast indicates the fastest compression method (less compression) and -9 or --best indicates the slowest compression method (best compression). The default compression level is -6 (that is, biased towards high compression at expense of speed).
lol, talk about low returns :-D
Using the default -6
[ o.k. ] Created: /boot/volumio.initrd [ 22827 /boot/volumio.initrd ]
Now:
$ ./build.sh -d pi
<snip>
[ dbg ] Compressing volumio.initrd with gzip -9
[ o.k. ] Created: /boot/volumio.initrd [ 22735 /boot/volumio.initrd ]
I believe the most cases the best way to reduce mkinitramfs is with using modules= list instead of modules=most. X86 may be an exception, as we currently need to load many storage device driver modules. Perhaps we should have all storage & graphics modules built-in and then also use modules=list (which drops the rest), never tried that...
It would great if we can get your modular initrd approach working on all devices! Probably once buster is settled?
This should streamline initrmafs creating a bit.
-- Add volumio-init-updater to all devices -- Hopefully reduce initrd size a little -- Speed up initramfs creation