by Ciro Santilli (@cirosantilli, 34)
Ciro's tutorial: Linux Kernel Module Cheat.
Ciro's word of caution for 2019 aspiring system programmers: Should you waste your life with systems programming?
This is basically a direct consequence of backward design.
The higher the level you can operate at, the better.
C is better than assembly, userland better than kernelland.
The ideal level to operate at, and one of humankind's greatest ambitions is "AGI, make me money", the highest possible level.
Only go down a level when it seems necessary.
Magic software that allows you to write a single program that runs on a wide range of hardware.
Bare metal programming is to run a program without an operating system below it.
Or in other words, it is basically implementing an operating system/firmware yourself ad hoc, together with your actual program.
Zephyr is cool. Its installation setup is annoying. But the project is cool.
Real hardware is for newbs. Real hardware is for newbs.
Tested on Ubuntu 23.10 we approximately follow instructions from: docs.zephyrproject.org/3.4.0/develop/getting_started/index.html stopping before the "Flash the sample" section, as we don't flash QEMU. We just run it.
sudo apt install --no-install-recommends git cmake ninja-build gperf \
  ccache dfu-util device-tree-compiler wget \
  python3-dev python3-pip python3-setuptools python3-tk python3-wheel xz-utils file \
  make gcc gcc-multilib g++-multilib libsdl2-dev libmagic1
python3 -m venv ~/zephyrproject/.venv
source ~/zephyrproject/.venv/bin/activate
pip install west
west init ~/zephyrproject
cd ~/zephyrproject
west update
west zephyr-export
cd ~
wget https://github.com/zephyrproject-rtos/sdk-ng/releases/download/v0.16.1/zephyr-sdk-0.16.1_linux-x86_64.tar.xz
tar xvf zephyr-sdk-0.16.1_linux-x86_64.tar.xz
cd zephyr-sdk-0.16.1
./setup.sh
The installation procedure install all compiler toolchains for us, so we can then basically compile for any target. It also fetches the latest Git source code of Zephyr under:
~/zephyrproject/zephyr
The "most default" blinky hello world example which blinks an LED is a bit useless for us because QEMU doesn't have LEDs, so instead we are going to use one of the UART examples which will print characters we can see on QEMU stdout.
Let's start with the hello world example on an x86 target:
cd ~/zephyrproject/zephyr
west build -b qemu_x86 samples/hello_world -t run
and it outputs:
Hello World! qemu_x86
The qemu_x64 on the output comes from the CONFIG_BOARD macro github.com/zephyrproject-rtos/zephyr/blob/c15ff103001899ba0321b2c38013d1008584edc0/samples/hello_world/src/main.c#L11
#include <zephyr/kernel.h>

int main(void)
{
	printk("Hello World! %s\n", CONFIG_BOARD);
	return 0;
}
You can also first cd into the directory that you want to build in to avoid typing samples/hello_world all the time:
cd ~/zephyrproject/zephyr/samples/hello_world
zephyr west build -b qemu_x86 -t run
You can also build and run separately with:
west build -b qemu_x86
west build -t run
Another important option is:
west build -t menuconfig
But note that it does not modify your prj.conf automatically for you.
Let's try on another target:
rm -rf build
zephyr west build -b qemu_cortex_a53 -t run
and same output, but on a completely different board! The qemu_cortex_a53 board is documented at: docs.zephyrproject.org/3.4.0/boards/arm64/qemu_cortex_a53/doc/index.html
The list of all examples can be seen under:
ls ~/zephyrproject/zephyr/samples
which for example contains:
zephyrproject/zephyr/samples/hello_world
So run another sample simply select it, e.g. to run zephyrproject/zephyr/samples/synchronization:
west build -b qemu_cortex_a53 samples/synchronization -t run
It ain't perfect, but it's decent enough.
From a technical point of view, it can do anything that Microsoft Windows can. Except being forcefully installed on every non-MacOS 2019 computer you can buy.
Ciro Santilli's conversion to Linux happened around 2012, and was a central part of Ciro Santilli's Open Source Enlightenment, since it fundamentally enables the discovery and contribution to open source software. Because what awesome open source person would waste time porting their amazing projects to closed source OSes?
Ciro's modest nature can be seen as he likes to compare this event Buddha's Great Renunciation.
Particularly interesting in the history of Linux is how it won out over the open competitors that were coming up in the time: MINIX (see the chat) and BSD Operating System that got legally bogged down at the critical growth moment.
Figure 1.
xkcd 619: Supported Features
. Source. This perfectly illustrates Linux development. First features that matter. Then useless features.
Video 1.
Bill Gates vs Steve Jobs by Epic Rap Battles of History (2012)
Source. Just stop whatever you are doing, and watch this right now. "I'm on Linux, bitch, I thought you GNU". Fandom explanations. It is just a shame that the Bill Gates actor looks absolutely nothing like the real gates. Actually, the entire Gates/Jobs parts are good, but not genial. But the Linux one is.
The fact that this foundation has a bunch of paid, closed, certification courses makes Ciro Santilli not respect them at all. They should be making open access content instead!
Documents the Linux kernel. Somewhat of a competitor to Linux Kernel Module Cheat, but more wordy and less automated.
As of 2020, no one knows how to build the major desktop distros fully from source into the ISO, and especially so in a reproducible build way. Everything is done in build servers somewhere with complicated layers of prebuilds. It's crap.
merlijn.sebrechts.be/blog/2020-08-02-why-one-snap-store/ has some very good comments on how snap is more closed than Flatpak.
Snap's permission system is extremely annoying, notably restricting access to home files. They need a popup that says "permission required by app, accept?" urgently!!!
However, many, many, many terrible horrors come with it:
The video folder is DCIM/Camera.
Respect. Big respect. Those people are hardcore from scratch hackers, and their wiki is amazing: wiki.archlinux.org/
It's just that Buildroot is more hardcore ;-)
But can you build the ISO full from source: Linux distribution buildable from source
Buildroot is good.
This thing is sexy.
Ciro Santilli's Linux distro of choice as of 2019.
It ain't perfect, but it's decent enough.
The greatest advantage of it being that it has the likely largest desktop user base, and therefore the highest likelihood that your problems are solved on Ask Ubuntu, and goes together with Ciro's philosophy that "people should do everything in the same way to factor stuff out", especially the open source losers.
Ciro considers that the killer flaw of Ubuntu, and most desktop distros of 2020, is that no one under the Sun knows how to build them fully from source: Linux distribution buildable from source. This is why Ciro based the Linux Kernel Module Cheat on Buildroot, see also: Linux distribution buildable from source.
On Ubuntu 23.10, a crash led to the creation of:
/var/crash/_usr_bin_gnome-shell.1000.crash
After that simply running apport-cli as:
apport-cli gnome-shell
led to the creation of: bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/2049368 a bug on the gnome-shell package.
Their crash system does not have an amazing user interface.
Tested on Ubuntu 21.10.
After something crashes, look under /var/crash for a crash file, which helps to determine which package to report under on Launchpad.
E.g. a file /var/crash/_usr_sbin_gdm3.0.crash makes you want to file the bug under gdm at: bugs.launchpad.net/ubuntu/+source/gdm/+filebug
Then, while reporting the bug, you want to give the developpers access to that .crash file. But you can't publicly upload it because it contains memory dumps and could contain secret information. The way to do it is to look at the ID under:
sudo cat /var/crash/_usr_sbin_gdm3.0.uploaded
Ubuntu's crash report system has already uploaded the .crash for you, so you just have to confirm it and give the ID on the ticket.
You can view a list of all your uploaded errors at:
xdg-open https://errors.ubuntu.com/user/$(sudo cat /var/lib/whoopsie/whoopsie-id)
and each of those contain a link to:
https://errors.ubuntu.com/oops/<.uloaded error id>
which you yourself cannot see.
Running:
sudo apport-unpack /var/crash/_usr_sbin_gdm3.0.crash /tmp/app
splits it up into a few files, but does not make any major improvements.
apport-retrace
sudo apt install apport-retrace
sudo chmod 666 /var/crash/_usr_sbin_gdm3.0.crash
apport-retrace -g /var/crash/_usr_sbin_gdm3.0.crash
opens GDB with the core dump. Debug symbols are supplied as separate packages, which is a really cool idea: so you should be able to download them after the crash to see symbols. askubuntu.com/questions/487222/how-to-install-debug-symbols-for-installed-packages mentions how to install them. Official docs at: wiki.ubuntu.com/DebuggingProgramCrash#Debug_Symbol_Packages
Tried:
echo "deb http://ddebs.ubuntu.com $(lsb_release -cs) main restricted universe multiverse" | sudo tee -a /etc/apt/sources.list.d/ddebs.list
echo -e "deb http://ddebs.ubuntu.com $(lsb_release -cs)-updates main restricted universe multiverse\ndeb http://ddebs.ubuntu.com $(lsb_release -cs)-proposed main restricted universe multiverse" | sudo tee -a /etc/apt/sources.list.d/ddebs.list
sudo apt install ubuntu-dbgsym-keyring
but then sudo apt update fails with:
E: The repository 'http://ddebs.ubuntu.com impish-security Release' does not have a Release file.
Ciro Santilli reproduced on Dell Inspiron 15 3520 with factory defaults. This bug prevents "wipe windows" installation without workarounds, one of which is to disable encryption in Windows.
Happens at startup without doing anything, and then keeps happening randomly infinitely many times... on a almost clean 24.04 ISO install on Dell Inspiron 15 3520... God how can it be so bad.
Viewing the apport issue a bit further shows title:
nautilus crashed with sigabrt in g_assertion_message_expr
https://web.archive.org/web/20240521054355if_/https://files.mastodon.social/media_attachments/files/112/477/476/049/270/962/original/d6346176be376ff5.jpg
Had this happen on P14s on Ubuntu 23.10 while causally using Chromium. The screen went blank for a few seconds, but it apparently managed to reboot itself, and things started working again, except that and most windows were killed:
[drm:gfx_v11_0_priv_reg_irq [amdgpu]] *ERROR* Illegal register access in command stream
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=5774109, emitted seq=5774111
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process chrome pid 14023 thread chrome:cs0 pid 14087
amdgpu 0000:64:00.0: amdgpu: GPU reset begin!
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=3
[drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[drm:gfx_v11_0_cp_gfx_enable.isra.0 [amdgpu]] *ERROR* failed to halt cp gfx
Dec 27 15:03:38 ciro-p14s kernel: amdgpu 0000:64:00.0: amdgpu: MODE2 reset
Dec 27 15:03:38 ciro-p14s kernel: amdgpu 0000:64:00.0: amdgpu: GPU reset succeeded, trying to resume
Dec 27 15:03:38 ciro-p14s kernel: [drm] PCIE GART of 512M enabled (table at 0x0000008000900
It appears to be a bug in the AMDGPU open source driver.
I think this was on Wayland. Possibly relatd but on X Window System, crashed the UI, showed message "oh no! Something has gone wrong."
2024-01-13_21-55-07@ciro@ciro-p14s$ cat /var/log/apport.log
ERROR: apport (pid 975172) 2024-01-13 21:41:02,087: host pid 3528 crashed in a separate mount namespace, ignoring
INFO: apport (pid 975227) 2024-01-13 21:41:02,398: called for pid 2728, signal 5, core limit 0, dump mode 1
INFO: apport (pid 975227) 2024-01-13 21:41:02,401: executable: /usr/bin/gnome-shell (command line "/usr/bin/gnome-shell")
INFO: apport (pid 975227) 2024-01-13 21:41:12,667: wrote report /var/crash/_usr_bin_gnome-shell.1000.crash
Happened on P14s on Ubuntu 23.10, which started with fresh Ubuntu 23.10 install.
However it did not happen on Lenovo ThinkPad P51 (2017) also on Ubuntu 23.10 which had been upgraded several times from God knows what starting point... At first one had X11 (forced by Nvidia drivers) and the other Wayland, but moving to p14s X11 changed nothing.
Both were running GNOME Display Manager.
Same happens with Super + L, but also CLI commands: askubuntu.com/questions/7776/how-do-i-lock-the-desktop-screen-via-command-line
Switching to the other installed kernel, 5.9 made boot work.
The solution on kernel 6.2 was:
sudo apt instal nvidia-driver-515
as per comments under: bugs.launchpad.net/ubuntu/+source/linux/+bug/2012559. This also made the nvidia driver work: Find GPU information in Ubuntu.
Previously I had:
nvidia-driver-510
and it blew up before reaching disk decryption.
I also tried:
nvidia-driver-525
but that broke in a different way:
Finished apport-autoreport.service - Process error reports when automatic reporting is enabled.
nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000947d:0:0:407
(1 of 2) Job systemd-backlight@backlight: nvidia_e.service/start running (32s no limit)
GDM crashes sometimes when switching windows right after opening a new window: bugs.launchpad.net/ubuntu/+source/gdm/+bug/1956299
Does not happen every time, only some times. Can't figure out why. Usually happens when has suspended for a longer time.
bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-470/+bug/1946303 sounds like a likely report, Nvidia driver version 470, but can't find those error messages anywhere. The last line of:
journalctl -o short-precise -k -b -1
once was:
PM: suspend entry (deep)
which is when sleep starts.
This suggests that it is not a video bug then, seems that it is not waking up at all? Gotta try to SSH into it. OK. I did SSH into it, and that was fine, so it is just the video that won't start.
PM: suspend exit
bugs.launchpad.net/ubuntu/+source/linux/+bug/1949977 is another possible bug, based on kernel version. I'm running 5.13, which is one of the failing versions on the report. Can't find any interesting dmesg though.
In another crash:
journalctl -o short-precise -k -b -1
had the following interesting lines:
nvidia-modeset: WARNING: GPU:0: Lost display notification (0:0x00000000); continuing.
[24307.640014] NVRM: GPU at PCI:0000:01:00: GPU-18af74bb-7c72-ff70-e447-87d48378ea20
[24307.640018] NVRM: Xid (PCI:0000:01:00): 79, pid=8828, GPU has fallen off the bus.
[24307.640021] NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
[24328.054022] nvidia-modeset: ERROR: GPU:0: The requested configuration of display devices (LGD (DP-4)) is not supported on this GPU.
[repeats several more times]
[24328.056767] nvidia-modeset: ERROR: GPU:0: The requested configuration of display devices (LGD (DP-4)) is not supported on this GPU.
[24328.056951] nvidia-modeset: ERROR: GPU:0: Failed to query display engine channel state: 0x0000927c:0:0:0x0000000f
[24328.056955] nvidia-modeset: ERROR: GPU:0: Failed to query display engine channel state: 0x0000927c:1:0:0x0000000f
[24328.056959] nvidia-modeset: ERROR: GPU:0: Failed to query display engine channel state: 0x0000927c:2:0:0x0000000f
[24328.056962] nvidia-modeset: ERROR: GPU:0: Failed to query display engine channel state: 0x0000927c:3:0:0x0000000f
[24328.056983] nvidia-modeset: ERROR: GPU:0: DP-4: Failed to disable DisplayPort audio stream-0
[24328.056992] nvidia-modeset: ERROR: GPU:0: Failed to query display engine channel state: 0x0000947d:0:0:0x0000000f
and there was a corresponding /var/crash/_usr_sbin_gdm3.0.crash.
Sometimes it just feels like Ubuntu devs don't actually use Ubuntu as a desktop.
Some extremelly anoying problems are introduced and just never get fixed, even though they feel so obvious!
Would never happen on Mac...
This BS started after the move to ZFS. The temporary solution appears to be: askubuntu.com/questions/1293685/out-of-space-on-boot-zpool-and-cant-run-updates-anymore/1374204#1374204
And then this to disable automatic snapshots in the future: askubuntu.com/questions/1233049/disable-automatic-zsys-snapshots-zfs-on-root/1279593#1279593
sudo mv /etc/apt/apt.conf.d/90_zsys_system_autosnapshot /etc/apt/apt.conf.d/90_zsys_system_autosnapshot.disabled
This has annoyed Ciro Santilli for many years, it is just too wasteful of screen space on laptops!
Or likely more generally, on GNOME desktop, which is the default desktop environment as of Ubuntu 22.10.
Legal issues stalled them at the turning point of the Internet, and Linux won. Can't change history.
Did Apple just fork it and made Mac OS X without giving anything back?
This section is about POSIX environment variable that have special effects.
Non-POSIX only here.
If you are a programmer, grep becomes a verb: "to grep" means "to search text files", much like "to Google" means "to search random stuff online".
The OS that the Gods ordered be made.
One is reminded of Ulillillia, see also: www.youtube.com/watch?v=9-79yOZ13qg The Story of Ulillillia by Atrocity Guide (2019)
Video 1.
I like elephants and God likes elephants
. Source.
For a quick and dirty introduction to the format, see: ELF Hello World Tutorial.
Introductory analysis of a simple example of the executable and Linkable Format.
ELF is the dominating file format for Linux. It competes with Mach-O for OS X and PE for Windows.
ELF supersedes .coff, which supersedes a.out.
The LSB basically links to other standards with minor extensions, in particular:
A handy summary can be found at:
man elf
Spin like mad between:
The ELF standard specifies multiple file formats:
  • Object files (.o).
    Intermediate step to generating executables and other formats:
    Source code

    |
    | Compilation
    |
    v

Object file

    |
    | Linking
    |
    v

Executable
    Object files exist to make compilation faster: with make, we only have to recompile the modified source files based on timestamps.
    We have to do the linking step every time, but it is much less expensive.
  • Executable files (no standard Linux extension).
    This is what the Linux kernel can actually run.
  • Archive files (.a).
    Libraries meant to be embedded into executables during the Linking step.
  • Shared object files (.so).
    Libraries meant to be loaded when the executable starts running.
  • Core dumps.
    Such files may be generated by the Linux kernel when the program does naughty things, e.g. segfault.
    They exist to help debugging the program.
In this tutorial, we consider only object and executable files.
  • Compiler toolchains generate and read ELF files.
    Sane compilers should use a separate standalone library to do the dirty work. E.g., Binutils uses BFD (in-tree and canonical source).
  • Operating systems read and run ELF files.
    Kernels cannot link to a library nor use the C stlib, so they are more likely to implement it themselves.
    This is the case of the Linux kernel 4.2 which implements it in th file fs/binfmt_elf.c.
It is non-trivial to determine what is the smallest legal ELF file, or the smaller one that will do something trivial in Linux.
In this example we will consider a saner hello world example that will better capture real life cases.
Let's break down a minimal runnable Linux x86-64 example:
hello_world.asm
section .data
    hello_world db "Hello world!", 10
    hello_world_len  equ $ - hello_world
section .text
    global _start
    _start:
        mov rax, 1
        mov rdi, 1
        mov rsi, hello_world
        mov rdx, hello_world_len
        syscall
        mov rax, 60
        mov rdi, 0
        syscall
Compiled with:
nasm -w+all -f elf64 -o 'hello_world.o' 'hello_world.asm'
ld -o 'hello_world.out' 'hello_world.o'
TODO: use a minimal linker script with -T to be more precise and minimal.
Versions:
  • NASM 2.10.09
  • Binutils version 2.24 (contains ld)
  • Ubuntu 14.04
We don't use a C program as that would complicate the analysis, that will be level 2 :-)
Articles were limited to the first 100 out of 148 total. Click here to view all children of Systems programming.

Ancestors (6)

  1. Software
  2. Computer
  3. Information technology
  4. Area of technology
  5. Technology
  6. Home

Incoming links (5)

View article source

Discussion (0)

+ New discussion

There are no discussions about this article yet.

Articles by others on the same topic (0)

There are currently no matching articles.
See all articles in the same topic + Create my own version