Follow-up to 09794ef832 -
initfs file does not have flavor now, too.
Without this 'pmbootstrap initfs extract' fails with:
(rootfs_samsung-klte) % cp /boot/initramfs-postmarketos-qcom-msm8974 /tmp/initfs-extracted/_initfs.gz
cp: cannot stat '/boot/initramfs-postmarketos-qcom-msm8974': No such file or directory
With this option you can run
$ pmbootstrap kconfig migrate --arch <arch> linux-postmarketos-xxx-xxx
to perform safe kconfig upgrades between kernel releases.
"make oldconfig" will ask question for every new/renamed kconfig option,
so you have no chance to miss anything.
Newly generated binutils package has file conflicsts with
non-cross binutils.
ERROR: binutils-2.37-r3: trying to overwrite usr/lib/bfd-plugins/libdep.so
owned by binutils-aarch64-2.37-r3.
Fix the generator to produce package that does not include these plugins.
Verify that using a Mediatek recovery boot.img file works with
bootimg_analyze. The mtk_mkimage-boot.img file is the twrp boot.img file
posted in pmbootstrap issue 2066.
Some Mediatek recovery images have headers named
"RECOVERY". Accept those too.
Signed-off-by: Boris Lysov <arzamas-16@mail.ee>
Co-Authored-By: Alexey Min <alexeymin@postmarketos.org>
Co-Authored-By: Oliver Smith <ollieparanoid@postmarketos.org>
Move all code that verifies the labels of the kernel and ramdisk inside
the boot.img file into a separate function, so it is easier to extend it
to allow recovery images too.
Replace the previous mtkimage-boot.img file, which were only the first 4
KiB of a mediatek boot.img file, with a full one. I made this full image
with pmbootstrap install for fairphone-fp1, it contains the kernel
packaged in pmaports.git master.
Reason for the replacement is, that the previous image did not contain
the label for the ramdisk. The current code ignored this, but upcoming
refactoring will test this more strictly and the dummy image will not
pass the test anymore.
Fix the issue of having the postmarketOS binary repository key deleted
from the chroots after "abuild undeps" removes the postmarketos-keys
package. This happens if e.g. building two device packages in a row (as
they depend on postmarketos-base, which depends on postmarketos-keys in
current pmaports.git master).
I've also considered installing the postmarketos-keys next to
alpine-base in new chroots. But this would introduce a bootstrap
problem, since you can't install the postmarketos-keys package unless
it already exists in the repository. We'd run into that when building
the next release.
Move code to install Alpine and postmarketOS keys into its own function,
so it can be called outside of pmb.chroot.init too (next patch).
Describe why this is needed in the first place and, while at it, tweak
the function to only copy the key if it does not exist in the target
directory.
The -p arg is suppose to point to the folder where the dtc program
resides (if it is not found in PATH). If we use an _outdir, which the
template currently does, then the arg needs to be -p
$_outdir/scripts/dtc/, and not -p scripts/dtc/, as scripts/dtc/ only
contains the dtc source code.
The python variant of dtbTool, that we currently use, does not use
this arg as it does not run dtc in a shell. I suppose the argument
was added only for compatibility with other dtbtools.
Before this commit, package folders were copied into the chroot one by
one in order to run apkbuild-lint on them. This logic is replaced by
mounting pmaports.git into the chroot and using a single apkbuild-lint
invocation to lint the supplied packages.
Both of these changes result in a performance improvement, especially
when linting multiple packages at once.
Before this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m5,261s
user 0m7,046s
sys 0m1,842s
Using the pmaports.git mount but calling apkbuild-lint in a loop:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m4,089s
user 0m6,418s
sys 0m1,219s
After this change:
$ time ./pmbootstrap.py -q lint $(cd ../pmaports/cross; echo *) \
> /dev/null
real 0m3,518s
user 0m5,968s
sys 0m0,959s
Additionally, running apkbuild-lint from the pmaports.git mount point
has the benefit that every printed violation contains a nice source
identifier à la "./cross/grub-x86/APKBUILD". This makes it possible to
differentiate between different packages even though only a single
apkbuild-lint invocation is used.
Relates: postmarketOS/pmaports#564
kernel is named /boot/vmlinuz now, looking at the filename will no
longer tell us what flavor it is. This now will look at
/usr/share/kernel, which has always contained the kernel 'flavor', and
since we currently only install 1 kernel these days, guarding this with
pmaports.cfg should be unnecessary. In the worst case (if there are
multiple kernel 'flavors' installed), it'll just grab the first one and
return it.
The new mkinitfs does not have options for kernel flavor or version
A new pmaports.cfg variable, "supported_mkinitfs_without_flavors" is used to
determine how to invoke mkinitfs
This improves handling input in pmbootstrap init.
Various config options in "Additional options" are int, thus whole input
must be checked:
This fixes obvious error:
Jobs [9]: 5e
and less obvious errors:
Extra space size: 600.5
Extra space size: 600m
...
$ pmbootstrap install
[17:28:23] *** (3/4) PREPARE INSTALL BLOCKDEVICE ***
[17:28:23] ERROR: invalid literal for int() with base 10: '600.5'
NOTE: I suppose moving ^ $ to ReadlineTabCompleter.ask() would break
some of already defined regexes.
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
gitlab runners seem to be registering binfmt now (?), so this fixes an
issue where register() might return immediately because the OS (or ??)
took care of registration, but the qemu-<arch> binary wasn't added to
the chroot.
It seems like the gitlab runners will now automatically register archs
when binfmt is mounted (I'm not completely sure, it's really
hard/annoying to confirm using gitlab CI jobs and no direct access to
the runner)
In any case, checking if it's already registered fixes a problem where
CI fails with "File exists" when it tries to register the arch.
While zram has uses outside of swap-on-zram, I would believe that the
main use of zram in postmarketOS is swap-on-zram. In addition to that,
I imagine many probably expect swap-on-zram to work if the zram kconfig
succeeds.
In order to support FAT32 as boot partition, the label is shortened from
pmOS_inst_boot to pmOS_i_boot. Read the value from pmaports.cfg and fall
back to the old value, so both are supported (-> building v21.06 and
v21.03 will use the old label).
Remove the old codepath that would set "pmOS_boot" as label for the
install partition, if the postmarketos-ondev package was older than
0.4.0. This is only the case on the long unsupported v20.05 branch.
The Git version suffixes usually generated automatically by the Linux
build system break packaging in Alpine because the flavor (e.g.
"-postmarketos-qcom-msm8916") will no longer match the module and
zImage path. So far setting LOCALVERSION= and disabling
CONFIG_LOCALVERSION_AUTO was sufficient to avoid the Git version
suffixes but this is no longer enough in Linux 5.14.
Instead, add a better fix by creating an empty .scmversion file
that will be used by the kernel instead of checking the Git status.
The advantage is that this works on Linux 5.14 and it should even
work when CONFIG_LOCALVERSION_AUTO is set.
The .scmversion code can be found here:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/scripts/setlocalversion?h=v5.14-rc1#n38
Install osk-sdl in the installer OS's boot partition for now. I forgot
about a code path earlier, which could render an encrypted target OS
without osk-sdl in the initramfs (and being unable to boot).
The target OS gets embeded in the installer OS as image file. This can
happen in two formats:
a) a full image with partition header and the boot and root partition
This is what bpo is doing when building the official, pre-built images,
as this method allows having the exact same image available separately
without the installer. Basically:
pmbootstrap install \
--ondev \
--no-rootfs \
--cp path/to/rootfs:/var/lib/rootfs.img
b) an image with just the root partition, no partition header and no
boot partition. This is what you get when running regular
"pmbootstrap install --ondev". It's slightly smaller, as there is no
duplicate boot partition.
If b) was done, the installer will copy the contents of the installer's
boot partition to the target OS. And that means: if osk-sdl is missing
from the installer's boot partition (the initramfs generated there), it
will also be missing in the boot partition of the target OS!
I think we should get rid of the b) code path to avoid confusion in the
future/make maintenance. But until that is done, always install osk-sdl
into the installer OS.
Remove the "(y/n) " in the question:
Choose default locale for installation (y/n) [C.UTF-8]:
So it becomes:
Choose default locale for installation [C.UTF-8]:
Fix this weird help output:
-mp URL, --mirror-pmOS URL
postmarketOS mirror, disable with: -mp='', specify
multiple with: -mp='one' -mp='two', default: h, t, t,
p, :, /, /, m, i, r, r, o, r, ., p, o, s, t, m, a, r,
k, e, t, o, s, ., o, r, g, /, p, o, s, t, m, a, r, k,
e, t, o, s, /
Remove /in-pmbootstrap inside chroots in "pmbootstrap shutdown" instead
of having it at a specific part of "pmbootstrap install".
Reasoning:
* With current approach, it didn't get removed in the on-device
installer chroot.
* This is less error prone than calling it multiple times in
"pmbootstrap install"