boot-utils: Initial bring up #1
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
There are a few places where we report errors. Instead of open coding it, create a die() function that prints a message in bold red and exits. This change allows us to simplify and consolidate some statements and add more informative messages to future changes. Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
… no arguments to the script This script is not intended to be run by many people thus no usage text but I frequently forget that I need to pass in 'all' to have it do anything. Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
We are going to rebuild all of the images so we should make sure that we have the latest LTS Buildroot release. Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
My benchmarks show that zstd has the second best decompression ratio/speed, whereas the best decompression ratio algorithm is not the same as the best decompression speed algorithm. Compression speed is not awful either but that is not as much of a concern because compression only happens during image build time, where it is a very small fraction of the overall build process. Check that the zstd tool is installed and bail out if it is not. Link: ClangBuiltLinux/continuous-integration#184 (comment) Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
$ ./buildroot/rebuild.sh all Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
This was referenced Mar 27, 2020
Merged
We have had issues with Debian's skiboot.lid so we have to build our own. To make sure we can always build it, add a script that handles this for us. $ ./build-skiboot.sh 6.5.4 Requires CROSS=<path_to_cross_compiler> or for powerpc64le-linux-gnu-gcc to be available in PATH. Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Built with GCC 9.3.0 and tested with current CI checkout. Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
To quote Nick:
"Orthogonally, I've been finding it frustrating to:
1. jump on a new machine without shell history, and try to recreate
our QEMU commands from driver.sh. Even sending commands to newbies
is a PITA.
2. boot more than just our scripts that immediately power off a
machine."
driver.sh is a very basic kernel build script, it is not good for
extensive testing because it only builds a defconfig kernel with no
modules. However, the boot testing that it provides is invaluable
because we have figured out all of the necessary QEMU flags.
Enter boot-qemu.sh, a split out version of driver.sh's booting
framework. There are two required arguments: the architecture being
booted (which expects the same value as driver.sh does) and a path to
the kernel's object folder (either the kernel source itself if an
in-tree build or the path to O= if an out-of-tree build).
By default, the rootfs will launch /init, which will execute our init
script that just prints the kernel version and shuts the machine down.
The script has an option to dump into an interactive shell (busybox's
shell) so that more testing can be done.
Run:
$ ./boot-qemu.sh -h
for more information.
Suggested-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
nickdesaulniers
approved these changes
Mar 27, 2020
| -k | --kbuild-folder: | ||
| The kernel build folder, as an absolute path or relative path | ||
| from wherever the script is being run. This is wherever the | ||
| compiled vmlinux image lives, not the architecture's boot folder. |
There was a problem hiding this comment.
Why the folder and not path to kernel image?
There was a problem hiding this comment.
For a couple of the ARM devices, we need the dtb to boot:
boot-qemu.sh
Outdated
| function sanity_check() { | ||
| # Kernel build folder and architecture are required paramters | ||
| # We are not interactive by default and the default timeout is 3m | ||
| : "${ARCH:?}" "${KBUILD_DIR:?}" "${INTERACTIVE:=false}" "${TIMEOUT:=3m}" |
There was a problem hiding this comment.
It is basically a shell builtin that acts like true. The point here is so that I can avoid a bunch of:
[[ -z ${ARCH} ]] && die "ARCH is undefined"
and such.
There was a problem hiding this comment.
But you only need it for ARCH and KBUILD_DIR because of the :??
When we start using sigils, I get nervous that the code starts to look like Perl, and there's good jokes about Perl being unreadable for a reason.
There was a problem hiding this comment.
Yes, I prefer to be explicit about the variables that I expect to be defined rather than using set -u.
Otherwise, I can do the more verbose thing I mentioned if it is more readable.
There was a problem hiding this comment.
Specially if the symbols are not that common in other languages.
There was a problem hiding this comment.
Alright, I pushed 936f954, which should be less cryptic.
There was a problem hiding this comment.
That looks way better for both user and dev, thanks!
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Suggested-by: Nick Desaulniers <ndesaulniers@google.com> Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
tpimh
approved these changes
Mar 30, 2020
tpimh
reviewed
Mar 30, 2020
tpimh
reviewed
Mar 30, 2020
|
I will leave this up until Monday afternoon PDT/MST to allow for any last comments. Once we get this merged, we can merge the CI one so that this starts getting used there too. |
This is a bit more clear for the user and anyone reading the script. Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
nathanchance
added a commit
to nathanchance/boot-utils
that referenced
this pull request
Aug 4, 2022
CI reported a crash when attempting to boot arm64 kernels:
Traceback (most recent call last):
File ".../boot-qemu.py", line 727, in <module>
cfg = get_qemu_args(cfg)
File ".../boot-qemu.py", line 451, in get_qemu_args
linux_ver_code = get_linux_ver_code(gzip_kernel_cmd)
File ".../boot-qemu.py", line 338, in get_linux_ver_code
return create_version_code(linux_version)
File ".../boot-qemu.py", line 262, in create_version_code
major, minor, patch = [int(version[i]) for i in (0, 1, 2)]
File ".../boot-qemu.py", line 262, in <listcomp>
major, minor, patch = [int(version[i]) for i in (0, 1, 2)]
ValueError: invalid literal for int() with base 10: '%s'
Looking at the strings output reveals:
$ gzip -d -c Image.gz &| strings &| grep "Linux version "
Linux version %s (%s)
Linux version 4.19.254 (tuxmake@tuxmake) (Debian clang version 14.0.6-++20220622053050+f28c006a5895-1~exp1~20220622173135.152, Debian LLD 14.0.6) ClangBuiltLinux#1 SMP PREEMPT @1659593940
Use re.search() plus a regular expression for a more precise match. I
had originally tried to do this but I could not get it to work for some
reason. I tested this against all the kernels that had issues in CI with
no further issues.
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The commit messages should break everything down well. Please let me know if you have any questions.
If this all looks good, I'll write a README after.