Edit: There were four copies of the same logic to keep in sync;
Originally this patch deduplicated each file, but leaving a copy of
the new function in each, thus reducing the duplication but not
eliminating it. A later patch moved it into a shared function file
following further enhancements to the code in question. This has
since been revised to have the function moved to a shared file here,
which simplifies and gives a cleaner diff.
Gbp-Dch: Short
Closes: #952908
the archive content file downloaded to obtain a list of firmware packages
is always deleted and downloaded afresh currently. it may not be ideal that
we do not make use of the cache here, however while that remains
unaddressed, we might as well delete the file after we've used it in order
to not pointlessly waste disk space.
note that this file is ~613 MB for sid-amd64 currently.
Gbp-Dch: Short
Closes: #952907
where multiple archive areas are used, the code here on each loop is:
1) fetching the archive area contents file (compressed)
2) **appending** the output to that of the previous loop
3) searching the file for firmware references, adding to the list
since it appends rather than replaces, entries found in each loop get
re-added on each subsequent loop, resulting in duplication in the
resulting list
below I evaluate the possible solutions to explain why I chose the one
I chose, however the reader should not waste too much time worrying about
whether one of the other solution would have actually been better because
things are changed significantly in further commits shortly!
possible solutions:
a) switching to output (>) rather than append (>>), but this might fail
against an existing file
b) removing the file on each loop, but this will complicate any future
caching improvements that might be made here (currently the files are
always deleted and thus downloaded fresh)
c) allow the appending, evaluating the complete file after the loop
solution C warrants consideration of disk space consumption; currently the
compressed 'main' archive (for sid on amd64) expands to 592.3 MB (feb-2020),
'contrib' is 3.1 MB, and 'non-free' is 18.5 MB.
solution C was chosen here; the difference of accumulated file size vs.
max-single was minor enough to not be of particular concern (~613 vs.
~592 MB).
Gbp-Dch: Short
Closes: #952906
These functions are specific to handling packages stored in the
cache, not other files. They are also always used with the same
`cache/packages.` prefix to the path.
Gbp-Dch: Short
Closes: #952916
instead of trying all from derived mirror then falling back to parent upon
failure, which as pointed out by a message printed out can result in a
load of spurious 404 errors; actually get each udeb from the mirror it is
supposed to be retrieved from.
Partial fix for #952914, this is the last commit for it so closes it
Gbp-Dch: Short
Closes: #952914
the existing logic just bundled the entire parent and derived udeb lists
together, ignoring the fact that there might thus be two instances of some
packages, and relying upon getting derived ones first and checking file
existence to avoid handling the overridden parent instances.
here we now actually filter the list of parent udebs to exclude packages
that are to be obtained from the derivative.
this enables avoiding the file existence checking
Partial fix for #952914
Gbp-Dch: Short
move the code that checks whether a version of a given package has already
been obtained (to account for parent and derived both listing the same
package) to guard the copy from cache action also, not just the download
action.
in rare but possible scenarios it would have been possible to end up with
both the parent and derived copies of a package included.
Partial fix for #952914
Gbp-Dch: Short
this script outputs a series of progressive dots as progress is made,
finally terminated with ' done.'. however if an error occurs then the
error would end up being printed directly on the end of this instead of
on a new line since newlines are not printed after each dot. this fixes
this oversight.
Gbp-Dch: Short
Closes: #952883
the '.' progress stuff has been left as is. perhaps that should be removed
since its use is inconsistent compared to other scripts?
Gbp-Dch: Short
Closes: #952882
Avoid all of the duplication for each installer case
Note, what is done for the netboot case (which was previously missing)
still needs addressing
Gbp-Dch: Short
Closes: #952865
LB_DEBIAN_INSTALLER_GUI defines whether or not to provide the graphical
installer. the installer_debian-installer script pays attention to it and
does not download it if not wanted. the actual bootloaders however
(both grub2/loopback and syslinux) ignore it, which leaves broken and
unwanted menu entries. this fixes that.
Gbp-Dch: Short
Closes: #952890
available values currently are memtest86+|memtest86|none; "false" is
presumably handled for backwards compatibility
there is no need to handle this in individual scripts. the right place
to handle it is in Set_Defaults as now done
Gbp-Dch: Short
Closes: #952866
`false` and `none` make no sense as choices for this option. Here we
replace `false` with `none`, and remove `true`.
Note that `true` was treated as an alias for netinst (see the changes to
source_disk and and binary_disk).
For backwards compatibility we still allow `true` and `false` by converting
them to `netinst` and `none` respectively, whilst printing a warning to
encourage users to move to `netinst`/`none`.
Gbp-Dch: Short
Closes: #952864
including:
- spaces replaced with tabs for consistency
- alignment of `;;` in some case statements changed for consistency
Gbp-Dch: Short
Closes: #952857
The specifics of what whould be done for it, beyond the disk info
label, needs addressing. (fixme comment left). (#952854).
Partial fix for #952846
Gbp-Dch: Short
Closes: #952846
update feb-2020: the bug report is marked as fixed in version 5.0-a7-1,
however there is no such patch merged for that tagged release, nor at all
in master, so this properly closes it.
Gbp-Dch: Short
Closes: #776532
Update feb-2020:
Both grub2 and syslinux were affected by this issue. Someone else's
patch fixing syslinux only just recently got merged in
2735f3bd38 but grub2 was still
affected. I had made patches for both in 2015 which never got
merged. This is the grub2 one.
Gbp-Dch: Short
Closes: #952843
all uses of this script pass in a suitable 'pass' param, and a previous
commit added a validation check. Having a fallback to 'binary' should a
param not be supplied is completely unnecessary.
fyi, this script is only ever called from the bootstrap stage and with
'binary' as the param, thus supporting being called with 'source' and
indeed having a 'pass' param at all is surely utterly pointless in itself.
Gbp-Dch: Ignore
18e0a2f325 "rebranded live sources" which
included renaming `source_debian-live` to `source_live` but accidentally
left the old file in place.
(Closes#952834)
apt-ftparchive is not able to differentiate between .deb and .udeb so
we have to install them in different pool directories so that we can
regenerate the Packages files without having the .udeb show up
unexpectedly.
Since binary_package-lists can overwrite the Packages files generated
in installer_debian-installer we have to ensure that it also updates
the Release file created formerly.
Ideally we should find a way to avoid the duplication of this logic.
Gbp-Dch: Full
In the binary stage, chroot is restored from cache/bootstrap, which contains
neither custom repo lists, nor keys. If local packages are present,
chroot_archives will call 'Apt chroot update' after adding custom repo lists
without adding keys. apt-get will then fail instead of warn as of apt version
1.5.
Closes: #941691
This makes it possible to build an image against a first distribution
(--distribution-chroot) and have the resulting image point to another
distribution (--distribution-binary). We can use this to build against a
snapshot and have the result use the original distribution that was
snapshotted.
Closes: #888507
Before Stretch there was an special amd64 kernel in the i386 arch repo.
So if you wanted to install an amd64 kernel alongside an i386 system
you did not need an additional arch repo.
Debian added multiarch support. That way you can install library packages
from multiple architectures on the same machine.
So there is no longer a need for having an amd64 kernel in i386 arch repo.
You can add an amd64 arch repo to an i386 arch system and fetch the amd64
kernel from the am64 arch repo.
live-build can be setup to use several linux kernel flavours in a single
image.
So in the days previous to this patch you could issue:
lb config --linux-flavours "486 amd64"
to use both 486 and amd64 kernel flavours.
Adding additional arch support to linux flavours poses two problems:
* Packages need to have its arch suffix (e.g. amd64:amd64).
If the suffix is not there apt-get insists on search amd64 kernel
package on i386 arch repo and, of course, fails to find it.
* The rest of the code which handles labels (bootloader config files)
or installed filenames (kernel images themselves) do not use the arch suffix.
This patch adds foreign architecture package support to
linux kernel flavours having taken those problems into account.
Practical example usage: i386 system and extra amd64 kernel.
First add amd64 foreign architecture in your i386 system
thanks to:
dpkg --add-architecture amd64
apt-get update
.
Finally enable amd64 kernel from amd64 arch alongside the
i386 system's 686 kernel thanks to:
lb config --architectures i386 --linux-flavours "686 amd64:amd64"
Turns out gcd works fine after adding /boot/grub/grub.cfg in the img,
as that's the path that gets hardcoded, and adding the EFI/debian/
grub.cfg was not necessary, so remove it.
For secured boot in binary_grub-efi, the gcdx64.efi.signed is
the boot loader for removable device, like CD or USB flash drive,
while grubx64.efi.signed is for hard drive. Therefore for live system,
use gcdx64.efi.signed for amd64 and gcdaa64.efi.signed for arm64.
With the merged /usr feature now enabled in debootstrap, we have
/bin and /lib which are symlinks to /usr/bin and /usr/lib. If
the live-build configuration provides files in
config/includes.chroot/lib/ or config/includes.chroot/bin/ then
lb chroot_includes will replace the /lib or /bin symlink with real
directories having only the content provided in chroot.includes and
problems will follow... the build will usually fail later with a
cryptic error message (for example a hook failing to execute
/usr/bin/env).
Work around the issue by creating a tarball of the files to install
and by unpacking that tarball from within the chroot with appropriate
options.
When using Secure Boot, grub2 as built by Debian will now load a config
file from EFI/$VENDOR instead of having EFI/debian hardcoded.
$VENDOR comes from dpkg-vendor or from the user building grub2.
The vendor string is stored in the control metadata as Efi-Vendor, so
retrieve it when building the EFI image.
Only gnupgv is part of the deboostrap set these days, but apt-key needs
the full gpg (with gpg-agent) which is just a recommends.
Instead just drop the key with an .asc suffix in /etc/apt/trusted.gpg.d
which is supported since apt version 1.4
live-build supports preseeding configuration, but the configuration has
to be installed after bootstrapping since it needs the debconf tools to
be applied. But packages that have already been installed and configured
in the bootstrap step will then ignore those preseeded configs.
After applying each preseed file, parse the package list and manually
reconfigure the relevant package(s) so that the configuration will be
applied.
The binary_onie script works on the host, not in the chroot (if used),
so don't check that the required packages for the script are installed
in the chroot (if present) as they won't be useful.
Check instead on the host.
Instead of hard-coding the decompression and compression formats,
detect them at runtime.
Install the required dependencies as well - they were mistakenly left
out.
Gbp-Dch: Ignore
The apt-secure option does not work anymore when building a sid image,
as with apt 1.6 the existing options are no longer enough to get apt
to accept an unsigned repository, which is necessary when using a
local cached repository (offline build).
Pass Acquire::AllowInsecureRepositories "true"; together with the
other options when --apt-secure false is used to fix the issue.
Open Network Install Environment is an open image format used by
networking vendor to ship a standardised image for networking white
box switches.
ONIE hardware takes this image at boot and a script to chain load
into the final environment via kexec. We can support Debian and
derivatives on such systems by packing an ISO which then gets
unpacked, kexec'ed and live-booted.
A base ONIE system can be tested in QEMU by building a VM following
these instrunctions:
https://github.com/opencomputeproject/onie/blob/master/machine/kvm_x86_64/INSTALL
Once built, boot onie-recovery-x86_64-kvm_x86_64-r0.iso in QEMU/libvirt
and on the console there will be the terminal prompt. Check the IP
assigned by libvirt and then scp the live image (ssh access is enabled
as root without password...). Then the .bin can be booted with:
ONIE-RECOVERY:/ # onie-nos-install /tmp/live.hybrid.iso-ONIE.bin
The implementation is inspired by ONIE's own scripts that can be found
at:
https://github.com/opencomputeproject/onie/blob/master/contrib/debian-iso/cook-bits.sh
A new option, --onie (false by default) can be set to true to enable
building this new format in addition to an ISO.
An additional option, --onie-kernel-cmdline can be used to specify
additional options that the ONIE system should use when kexec'ing the
final image.
Note that only iso or hybrid-iso formats are supported.
For more information about the ONIE ecosystem see:
http://onie.org
Signed-off-by: Erik Ziegenbalg <eziegenb@Brocade.com>
Signed-off-by: Luca Boccassi <bluca@debian.org>
The Tianocore reference UEFI implementation, used for example by Qemu,
wants the EFI directory name to be uppercase in the fat32 partition
when Secure Boot is enabled, and will fail to load otherwise.
Support for UEFI Secure Boot is modelled after how it currently works
in Ubuntu and on how it is going to work on Debian.
A minimal bootloader, shim, is used as the first-stage and it then
loads grub. Both have to be signed.
shim-signed is already available in Debian so the filenames are
already established, and the grub2 repository and packaging is common
between the 2 distros so we can already be reasonably sure of what it
is going to be.
So if both are available, copy /usr/lib/shim/shim[x64|aa64].efi.signed
as boot[x64|aa64].efi so that UEFI loads it first, and copy
/usr/lib/grub/[x86_64|arm64]-efi-signed/grub[x64|aa64].efi.signed as
grub[x64|aa64].efi.
This grub2 EFI monolithic image is currently hard-coded in grub2's
repository to look for a config file in efi/debian, so make a copy
of the previously added minimal grub.cfg that loads the real one in
that directory in both the fat32 and ISO 9660 partitions.
The new option --uefi-secure-boot can be set to auto (default,
enable or disable.
In auto, the lack of the signed EFI binaries is intentionally left as a
soft failure - live-build will simply fallback to using the locally
generated non-signed grub2 monolithic EFI binary as the only
bootloader. Given the difficulties surrounding the Secure Boot
signing infrastructure this approach gives the most flexibility and
makes sure things will "just work" once the packages are available,
without the need to change anything in the configuration.
This will also greatly help downstream distributions and users who
want to do self-signing.
The enable or disable options work as expected.
Closes: #821084
On some UEFI implementations, like the AMI found in the Supermicro
X10SDV-TP8F development board, the fat32 partition will be loaded
first and so Grub will set it the root, and then drop to the console
as it cannot find any config on it.
Add a minimal grub.cfg that allows Grub to find the main config on
the ISO 9660 partition and load it.
Closes: #892406
Machines tend to become unresponsive during the mksquashfs step.
Avoid this by lowering the priority of the process.
Thanks: Ronny Standtke for the patch.
Closes: #867539
Commit a15b579652 (#775989) dropped an early exit from the
chroot_archives remove step in case the parent mirror chroot and binary
parameters are the same and introduced a regression, as with the
following live-build now fails when the parent mirror is using a file:/
local apt repository (for example when the build worker is offline and
uses a pre-built cache of packages).
Example config:
lb config --mirror-bootstrap "file:/pkgs" \
--mirror-chroot "file:/pkgs/" \
--mirror-binary "file:/pkgs" \
--parent-mirror-bootstrap "file:/pkgs" \
--parent-mirror-chroot "file:/pkgs/" \
--parent-mirror-binary "file:/pkgs" \
...
with /pkgs being a directory with the packages for the installation and
the apt metadata (Packages/Sources/Release).
The problem is that, with such a setup, the /pkgs directory is bind
mounted inside the chroot as an optimisation in the install step,
and umounted as one of the first actions in the remove step for
chroot_archives.
Before that fix, the script terminated immediately. But now it
progresses and at the end it tries to run apt update inside the chroot
which will fail since the repository directory has been umounted, and
thus the packages and the apt metadata are no longer available, while
still being listed in /etc/apt/sources.list.
The proposed solution is to umount the local directory at the end of
the remove step, rather than at the beginning.
Closes: #891206
Now grub.cfg shows all the kernel options. Before this patch when you
had more than two kernels it only showed the auto option.
Signed-off-by: Raphaël Hertzog <hertzog@debian.org>
To generate an hdd image, binary_hdd first estimates the needed size of
the image using du. By default, when du finds multiple hardlinked copies
of a file, it counts them only once. However, when the target filesystem
is FAT, which does not support hardlinks, these files will take up more
space when finally copying the contents, breaking the build:
P: Copying binary contents into image...
cp: error writing 'chroot/binary.tmp/live/initrd.img-4.9.0-3-amd64': No space left on device
cp: error writing 'chroot/binary.tmp/efi/boot/bootx64.efi': No space left on device
cp: error writing 'chroot/binary.tmp/efi/boot/bootia32.efi': No space left on device
cp: cannot create directory 'chroot/binary.tmp/boot/grub': No space left on device
cp: cannot create directory 'chroot/binary.tmp/isolinux': No space left on device
To fix this, pass --count-links to du when the target is FAT, to make
the space estimation correct.
This problem is exposed by commit 9c974b26b (Instead of renaming kernel
for syslinux, create hardlinks), which might need to be separately fixed
(to not waste space on FAT targets), but binary_hdd should at least
handle hardlinks more gracefully.
Since commit fdc9250bc (Changing package dependency checks within chroot
to work outside as well), Check_package automatically checks for
LB_BUILD_WITH_CHROOT and works inside as well as outside of the chroot,
so no need to check LB_BUILD_WITH_CHROOT before calling them.
Install_package and Remove_package are just a no-op when building
without chroot, so they can also be called unconditionally.
Restore_cache and Save_cache do not check LB_BUILD_WITH_CHROOT but it
it should not hurt to call them when not needed (which already happened
in some cases).
This commit makes all Check_package calls unconditional on
LB_BUILD_WITH_CHROOT.
For binary_syslinux, this fixes the check (which used outdated paths
outside the chroot since 7b6dfd9d1), for binary_grub-efi,
binary_package-lists and chroot_package-lists this simplifies the code
(but also causes the check to become package-based instead of file-based
on apt-based systems), and for binary_loadlin and binary_win32-loader
this adds the check outside the chroot which was previously missing.
* Use only long kernel names.
* Put advanced options in a submenu.
* Use distro-agnostic labels.
* Don't generate entries with kernel version when we have a single
version.
This option lets you use an alternate bootstrap script when running
debootstrap. Thanks to Sjoerd Simons <sjoerd@debian.org> for the initial
patch.
Closes: #790033
Commit e24e4b in debootstrap fixed setup_available to work in the
--foreign case (iotw at the second stage). Unfortunately this breaks
things if components aren't passed to the second stage _and_ your main
component isn't called main.
To fix this, pass --components to both the first and second stage
debootstrap when needed.
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Up to now we created a temporary GPG key that we registered with apt-key
but with the switch to GnuPG 2 by default, this code broke. Now we stop
doing that but we add the “trusted=yes“ attribute in sources.list so
that APT knows that the repository can be trusted even if it's unsigned.
Signed-off-by: Raphaël Hertzog <hertzog@debian.org>
This work is based on debian-cd team work and uses,
as much as possible, the same mkisofs options
than the Debian Installation CD disk does.
It assumes that /boot/grub/grub.cfg (and other design items)
is generated by: binary_loopback_cfg .
It relies on efi-image and grub-cpmodules being setup
as build scripts on live-build package.
In the future event of these two files being moved
to a binary package (they are originally from:
src: live-installer) the binary_grub-efi script would have
to be rewritten to take the new paths into account.
These two scripts simplify the creation of efi images based on grub-efi.
I have decided to simply steal them. If I had to include them thanks to a source package that would have mean that an src repo would have to be defined by default.
TODO: Ask in a bug a RFE so that these two scripts are put into a binary that could be consumed by both live-installer and live-build packages.
The binary parts of grub-pc are left for the original binary_grub-pc.
As a consequence both /boot/grub/grub.cfg and /boot/grub/loopback.cfg files will be present in any Debian Live CD.
This might be useful to be reused from binary_grub-* bootloaders.
* Added: functions/bootloaders.sh . This file adds bootloader functions that are heavily used in efi scenarios where a bootloader can act as a first or an extra bootloader.
Since the introduction of the new switch:
--bootloaders
you can setup it like this:
--bootloaders=syslinux,grub-efi
.
This means that syslinux is the first bootloader and grub-efi is the extra bootloader.
* Added new bootloader functions: Check_Non_First_Bootloader and Check_Non_Extra_Bootloader.
These functions let each one of the bootloaders abort the build because
they cannot perform a role either as a first bootloader or as an extra bootloader.
* Added bootloader functions: Check_First_Bootloader_Role, Check_Extra_Bootloader_Role and Check_Any_Bootloader_Role
These functions let bootloaders to force their default role in a single line.
At the same time many binary bootloaders were rewritten to make use of the new bootloader role functions explained above.
These roles were enforced:
binary_grub-legacy : First bootloader
binary_grub-pc : Either first or extra bootloader
binary_syslinux : Either first or extra bootloader
If a bootloader is tried to be used in a role that it's not meant to be used then the build fails because that might lead to a non-bootable system.
The fix in a294a46fb9 was not enough.
This should finally resolve the problem when a package list ends
up empty (most notably due to #if evaluating to false).
Sponsored-By: Offensive Security
Some BIOSes dont't boot from partitions starting at sector 1024.
Some are even more peculiar and only start from sector 63.
This patch adds an option for the binary_hdd target to manually
configure the partition start.
Development details
--------------------
* This patch has been based originally on: binary_grub2 . It has been
improved thanks to some binary_syslinux bits.
* This patch ensures that binary_loopback_cfg needs is run
before binary_syslinux is run.
The reason is that it reuses some code from binary_syslinux to avoid
problems when binary_syslinux renames the kernel filenames.
* This patch already supports CPU detection
* I haven't tested all the possible scenarios for the script (with the
latest version). When amd64 and 486 Gnu/Linux flavours is used and the
bootloader is syslinux it works ok.
* I have not implemented a disable switch for not generating it.
Usually you always want loopback.cfg to be there.
* Compared to binary_grub2 script I have removed the installation
entries because I did not see any of them in binary_syslinux.
How to test
-----------
These are some steps to easily test if Looback cfg support is working ok.
0) We assume you have generated an iso
1) Make sure you have a partition that Grub understands. Plain ext4 or
vfat should do it.
2) Create directory: /boot/boot-isos/
3) Put the iso file into that directory making sure it has an ISO or iso
extension.
4) Setup your computer to boot from cdrom and use: Super Grub2 Disk
2.00s2 (Hybrid version recommended)
5) Choose Boot manually...
6) Choose Bootable ISOs (in /boot- ... )
7) Choose (the detected) GRUB Loopback Config
(hdN,msodsN)/boot/boot-ios/name-of-the.iso
8) You will be presented your loopback.cfg. Choose anyone of the entries
(unless it does not match your cpu architecture of course).
9) You should boot into your Debian Live without problems (thanks to
findiso boot parametre).
If you ever wanted to test from your grub2 installation instead from
Super Grub2 Disk check: http://www.supergrubdisk.org/wiki/Loopback.cfg
for an example.
Note: OLDIFS use makes IFS to be reset to "" instead to it being unset.
Either we need to detect if old IFS was unset to unset it
or we need a proper way of setting it as a local variable.
Even more IFS it's not currently used in
Check_package (which it's called from: binary_hdd).
we should have a clean way of resetting/unsetting IFS when calling Check_package.
The other approach it's to explicitly define IFS with its default value in the
places inside live-build code where we implicitly suppose that it's going to have
its default value.
Future live-build versions will still allow to use casper,
but its configuration will be done differently by a custom
config tree, rather than embedded and maintenance intensive
code in live-build itself.
Future live-build versions will still allow to use casper,
but its configuration will be done differently by hooks
in the config tree, rather than embedded and maintenance intensive
code in live-build itself.
debootstrap is the official tool to bootstrap debian,
cdebootstrap has had the one or other bug making it
broken for times during the release cycles.
The extra effort of supporting both debootstrap
and cdebootstrap is hardly worth it since the bootstrap
stage is cached anyway.
If a local repository path is given as a mirror URL lets bind it into the
chroot. The local repository will be unmounted while processing "remove" or
latest by the exit function.
Lyndon Brown
David Hewitt
Raphaël Hertzog
John Estabrook
Ronny Standtke
Roland Clobus
adrian15
Adrian Gibanel Lopez
Luca Boccassi
Steven Shiau
Marcel Partap
Raphaël Hertzog
Rohan Garg
Balint Reczey
Matthijs Kooijman
Sjoerd Simons
Sophie Brun
Kristian Klausen
Ben Armstrong
Maximilian Mehnert
Carlos Zuferri
Daniel Baumann
Daniel Reichelt
Raphaël Hertzog
Jan Blunck