]>
Commit | Line | Data |
---|---|---|
c3e270f4 FB |
1 | --- |
2 | title: Hacking on systemd | |
4cdca0af | 3 | category: Contributing |
b41a3f66 | 4 | layout: default |
0aff7b75 | 5 | SPDX-License-Identifier: LGPL-2.1-or-later |
c3e270f4 FB |
6 | --- |
7 | ||
5a8a9dee FA |
8 | # Hacking on systemd |
9 | ||
10 | We welcome all contributions to systemd. If you notice a bug or a missing | |
54080482 DDM |
11 | feature, please feel invited to fix it, and submit your work as a |
12 | [GitHub Pull Request (PR)](https://github.com/systemd/systemd/pull/new). | |
5a8a9dee | 13 | |
75e09908 ZJS |
14 | Please make sure to follow our [Coding Style](CODING_STYLE.md) when submitting |
15 | patches. Also have a look at our [Contribution Guidelines](CONTRIBUTING.md). | |
5a8a9dee FA |
16 | |
17 | When adding new functionality, tests should be added. For shared functionality | |
18 | (in `src/basic/` and `src/shared/`) unit tests should be sufficient. The general | |
19 | policy is to keep tests in matching files underneath `src/test/`, | |
20 | e.g. `src/test/test-path-util.c` contains tests for any functions in | |
21 | `src/basic/path-util.c`. If adding a new source file, consider adding a matching | |
22 | test executable. For features at a higher level, tests in `src/test/` are very | |
1e268f42 | 23 | strongly recommended. If that is not possible, integration tests in `test/` are |
5a8a9dee FA |
24 | encouraged. |
25 | ||
75e09908 ZJS |
26 | Please also have a look at our list of [code quality tools](CODE_QUALITY.md) we |
27 | have setup for systemd, to ensure our codebase stays in good shape. | |
5a8a9dee FA |
28 | |
29 | Please always test your work before submitting a PR. For many of the components | |
4bb37359 | 30 | of systemd testing is straightforward as you can simply compile systemd and |
5a8a9dee FA |
31 | run the relevant tool from the build directory. |
32 | ||
8f1a581e | 33 | For some components (most importantly, systemd/PID 1 itself) this is not |
5a8a9dee | 34 | possible, however. In order to simplify testing for cases like this we provide |
54080482 DDM |
35 | a set of `mkosi` build files directly in the source tree. |
36 | [mkosi](https://github.com/systemd/mkosi) is a tool for building clean OS images | |
37 | from an upstream distribution in combination with a fresh build of the project | |
f478b6e9 DDM |
38 | in the local working directory. To make use of this, please install `mkosi` v18 |
39 | or newer using your distribution's package manager or from the | |
40 | [GitHub repository](https://github.com/systemd/mkosi). `mkosi` will build an | |
41 | image for the host distro by default. First, run `mkosi genkey` to generate a key | |
42 | and certificate to be used for secure boot and verity signing. After that is done, | |
43 | it is sufficient to type `mkosi` in the systemd project directory to generate a disk | |
44 | image you can boot either in `systemd-nspawn` or in a UEFI-capable VM: | |
5a8a9dee | 45 | |
b2c9da05 | 46 | ```sh |
2edcf8e7 | 47 | $ sudo mkosi boot # nspawn still needs sudo for now |
5a8a9dee FA |
48 | ``` |
49 | ||
50 | or: | |
51 | ||
b2c9da05 DDM |
52 | ```sh |
53 | $ mkosi qemu | |
5a8a9dee FA |
54 | ``` |
55 | ||
56 | Every time you rerun the `mkosi` command a fresh image is built, incorporating | |
2fe24ccc | 57 | all current changes you made to the project tree. |
c38667f7 | 58 | |
f478b6e9 DDM |
59 | Putting this all together, here's a series of commands for preparing a patch |
60 | for systemd: | |
61 | ||
62 | ```sh | |
63 | $ git clone https://github.com/systemd/mkosi.git # If mkosi v18 or newer is not packaged by your distribution | |
64 | $ ln -s $PWD/mkosi/bin/mkosi /usr/local/bin/mkosi # If mkosi v18 or newer is not packaged by your distribution | |
65 | $ git clone https://github.com/systemd/systemd.git | |
66 | $ cd systemd | |
67 | $ git checkout -b <BRANCH> # where BRANCH is the name of the branch | |
68 | $ vim src/core/main.c # or wherever you'd like to make your changes | |
69 | $ mkosi -f qemu # (re-)build and boot up the test image in qemu | |
70 | $ git add -p # interactively put together your patch | |
71 | $ git commit # commit it | |
72 | $ git push -u <REMOTE> # where REMOTE is your "fork" on GitHub | |
73 | ``` | |
74 | ||
75 | And after that, head over to your repo on GitHub and click "Compare & pull request" | |
76 | ||
c38667f7 | 77 | If you want to do a local build without mkosi, most distributions also provide |
f478b6e9 | 78 | very simple and convenient ways to install most development packages necessary |
b43ed972 | 79 | to build systemd: |
5a8a9dee | 80 | |
b43ed972 DDM |
81 | ```sh |
82 | # Fedora | |
83 | $ sudo dnf builddep systemd | |
84 | # Debian/Ubuntu | |
4df5799f | 85 | $ sudo apt-get build-dep systemd |
b43ed972 | 86 | # Arch |
ba96ba04 MY |
87 | $ sudo pacman -S devtools |
88 | $ pkgctl repo clone --protocol=https systemd | |
89 | $ cd systemd | |
b43ed972 | 90 | $ makepkg -seoc |
5a8a9dee FA |
91 | ``` |
92 | ||
f478b6e9 | 93 | After installing the development packages, systemd can be built from source as follows: |
5a8a9dee FA |
94 | |
95 | ```sh | |
f478b6e9 DDM |
96 | $ meson setup build <options> |
97 | $ ninja -C build | |
98 | $ meson test -C build | |
5a8a9dee FA |
99 | ``` |
100 | ||
5a8a9dee FA |
101 | Happy hacking! |
102 | ||
89f52a78 ZJS |
103 | ## Templating engines in .in files |
104 | ||
105 | Some source files are generated during build. We use two templating engines: | |
106 | * meson's `configure_file()` directive uses syntax with `@VARIABLE@`. | |
107 | ||
ba777d01 ZJS |
108 | See the |
109 | [Meson docs for `configure_file()`](https://mesonbuild.com/Reference-manual.html#configure_file) | |
110 | for details. | |
89f52a78 | 111 | |
c9d311c7 | 112 | {% raw %} |
89f52a78 ZJS |
113 | * most files are rendered using jinja2, with `{{VARIABLE}}` and `{% if … %}`, |
114 | `{% elif … %}`, `{% else … %}`, `{% endif … %}` blocks. `{# … #}` is a | |
115 | jinja2 comment, i.e. that block will not be visible in the rendered | |
c9d311c7 ZJS |
116 | output. `{% raw %} … `{% endraw %}`{{ '{' }}{{ '% endraw %' }}}` creates a block |
117 | where jinja2 syntax is not interpreted. | |
89f52a78 | 118 | |
ba777d01 ZJS |
119 | See the |
120 | [Jinja Template Designer Documentation](https://jinja2docs.readthedocs.io/en/stable/templates.html#synopsis) | |
89f52a78 ZJS |
121 | for details. |
122 | ||
123 | Please note that files for both template engines use the `.in` extension. | |
5a8a9dee | 124 | |
4c8e5f44 ZJS |
125 | ## Developer and release modes |
126 | ||
127 | In the default meson configuration (`-Dmode=developer`), certain checks are | |
128 | enabled that are suitable when hacking on systemd (such as internal | |
89f52a78 ZJS |
129 | documentation consistency checks). Those are not useful when compiling for |
130 | distribution and can be disabled by setting `-Dmode=release`. | |
4c8e5f44 | 131 | |
69d638e6 DDM |
132 | ## Sanitizers in mkosi |
133 | ||
134 | See [Testing systemd using sanitizers](TESTING_WITH_SANITIZERS.md) for more information | |
135 | on how to build with sanitizers enabled in mkosi. | |
136 | ||
5a8a9dee FA |
137 | ## Fuzzers |
138 | ||
139 | systemd includes fuzzers in `src/fuzz/` that use libFuzzer and are automatically | |
135a1add | 140 | run by [OSS-Fuzz](https://github.com/google/oss-fuzz) with sanitizers. |
53a42e62 | 141 | To add a fuzz target, create a new `src/fuzz/fuzz-foo.c` file with a `LLVMFuzzerTestOneInput` |
5a8a9dee FA |
142 | function and add it to the list in `src/fuzz/meson.build`. |
143 | ||
144 | Whenever possible, a seed corpus and a dictionary should also be added with new | |
145 | fuzz targets. The dictionary should be named `src/fuzz/fuzz-foo.dict` and the seed | |
146 | corpus should be built and exported as `$OUT/fuzz-foo_seed_corpus.zip` in | |
147 | `tools/oss-fuzz.sh`. | |
148 | ||
149 | The fuzzers can be built locally if you have libFuzzer installed by running | |
d04af6aa FS |
150 | `tools/oss-fuzz.sh`, or by running: |
151 | ||
152 | ``` | |
153 | CC=clang CXX=clang++ \ | |
154 | meson setup build-libfuzz -Dllvm-fuzz=true -Db_sanitize=address,undefined -Db_lundef=false \ | |
155 | -Dc_args='-fno-omit-frame-pointer -DFUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION' | |
156 | ninja -C build-libfuzz fuzzers | |
157 | ``` | |
158 | ||
159 | Each fuzzer then can be then run manually together with a directory containing | |
160 | the initial corpus: | |
161 | ||
162 | ``` | |
163 | export UBSAN_OPTIONS=print_stacktrace=1:print_summary=1:halt_on_error=1 | |
164 | build-libfuzz/fuzz-varlink-idl test/fuzz/fuzz-varlink-idl/ | |
165 | ``` | |
166 | ||
167 | Note: the `halt_on_error=1` UBSan option is especially important, otherwise | |
168 | the fuzzer won't crash when undefined behavior is triggered. | |
169 | ||
170 | You should also confirm that the fuzzers can be built and run using | |
39e63b44 | 171 | [the OSS-Fuzz toolchain](https://google.github.io/oss-fuzz/advanced-topics/reproducing/#building-using-docker): |
5a8a9dee FA |
172 | |
173 | ``` | |
39e63b44 EV |
174 | path_to_systemd=... |
175 | ||
176 | git clone --depth=1 https://github.com/google/oss-fuzz | |
177 | cd oss-fuzz | |
178 | ||
179 | for sanitizer in address undefined memory; do | |
180 | for engine in libfuzzer afl honggfuzz; do | |
181 | ./infra/helper.py build_fuzzers --sanitizer "$sanitizer" --engine "$engine" \ | |
182 | --clean systemd "$path_to_systemd" | |
183 | ||
184 | ./infra/helper.py check_build --sanitizer "$sanitizer" --engine "$engine" \ | |
185 | -e ALLOWED_BROKEN_TARGETS_PERCENTAGE=0 systemd | |
186 | done | |
187 | done | |
188 | ||
61ad4f25 EV |
189 | ./infra/helper.py build_fuzzers --clean --architecture i386 systemd "$path_to_systemd" |
190 | ./infra/helper.py check_build --architecture i386 -e ALLOWED_BROKEN_TARGETS_PERCENTAGE=0 systemd | |
191 | ||
39e63b44 EV |
192 | ./infra/helper.py build_fuzzers --clean --sanitizer coverage systemd "$path_to_systemd" |
193 | ./infra/helper.py coverage --no-corpus-download systemd | |
5a8a9dee FA |
194 | ``` |
195 | ||
196 | If you find a bug that impacts the security of systemd, please follow the | |
197 | guidance in [CONTRIBUTING.md](CONTRIBUTING.md) on how to report a security vulnerability. | |
198 | ||
199 | For more details on building fuzzers and integrating with OSS-Fuzz, visit: | |
200 | ||
6cec69fc LK |
201 | - [Setting up a new project - OSS-Fuzz](https://google.github.io/oss-fuzz/getting-started/new-project-guide/) |
202 | - [Tutorials - OSS-Fuzz](https://google.github.io/oss-fuzz/reference/useful-links/#tutorials) | |
4cc06b80 | 203 | |
2d92c35b DDM |
204 | ## Debugging binaries that need to run as root in vscode |
205 | ||
206 | When trying to debug binaries that need to run as root, we need to do some custom configuration in vscode to | |
207 | have it try to run the applications as root and to ask the user for the root password when trying to start | |
208 | the binary. To achieve this, we'll use a custom debugger path which points to a script that starts `gdb` as | |
209 | root using `pkexec`. pkexec will prompt the user for their root password via a graphical interface. This | |
210 | guide assumes the C/C++ extension is used for debugging. | |
211 | ||
212 | First, create a file `sgdb` in the root of the systemd repository with the following contents and make it | |
213 | executable: | |
214 | ||
215 | ``` | |
216 | #!/bin/sh | |
217 | exec pkexec gdb "$@" | |
218 | ``` | |
219 | ||
220 | Then, open launch.json in vscode, and set `miDebuggerPath` to `${workspaceFolder}/sgdb` for the corresponding | |
221 | debug configuration. Now, whenever you try to debug the application, vscode will try to start gdb as root via | |
222 | pkexec which will prompt you for your password via a graphical interface. After entering your password, | |
223 | vscode should be able to start debugging the application. | |
224 | ||
225 | For more information on how to set up a debug configuration for C binaries, please refer to the official | |
226 | vscode documentation [here](https://code.visualstudio.com/docs/cpp/launch-json-reference) | |
227 | ||
66dc9b46 DDM |
228 | ## Debugging systemd with mkosi + vscode |
229 | ||
230 | To simplify debugging systemd when testing changes using mkosi, we're going to show how to attach | |
c84d14c5 DDM |
231 | [VSCode](https://code.visualstudio.com/)'s debugger to an instance of systemd running in a mkosi image using |
232 | QEMU. | |
66dc9b46 DDM |
233 | |
234 | To allow VSCode's debugger to attach to systemd running in a mkosi image, we have to make sure it can access | |
c84d14c5 DDM |
235 | the virtual machine spawned by mkosi where systemd is running. mkosi makes this possible via a handy SSH |
236 | option that makes the generated image accessible via SSH when booted. Thus you must build the image with | |
ee7304df DDM |
237 | `mkosi --ssh`. The easiest way to set the option is to create a file `mkosi.conf` in the root of the |
238 | repository and add the following contents: | |
66dc9b46 DDM |
239 | |
240 | ``` | |
241 | [Host] | |
242 | Ssh=yes | |
ee7304df | 243 | RuntimeTrees=. |
66dc9b46 DDM |
244 | ``` |
245 | ||
c84d14c5 | 246 | Also make sure that the SSH agent is running on your system and that you've added your SSH key to it with |
ee7304df | 247 | `ssh-add`. Also make sure that `virtiofsd` is installed. |
c84d14c5 DDM |
248 | |
249 | After rebuilding the image and booting it with `mkosi qemu`, you should now be able to connect to it by | |
250 | running `mkosi ssh` from the same directory in another terminal window. | |
66dc9b46 DDM |
251 | |
252 | Now we need to configure VSCode. First, make sure the C/C++ extension is installed. If you're already using | |
253 | a different extension for code completion and other IDE features for C in VSCode, make sure to disable the | |
254 | corresponding parts of the C/C++ extension in your VSCode user settings by adding the following entries: | |
255 | ||
256 | ```json | |
257 | "C_Cpp.formatting": "Disabled", | |
258 | "C_Cpp.intelliSenseEngine": "Disabled", | |
259 | "C_Cpp.enhancedColorization": "Disabled", | |
260 | "C_Cpp.suggestSnippets": false, | |
261 | ``` | |
262 | ||
263 | With the extension set up, we can create the launch.json file in the .vscode/ directory to tell the VSCode | |
026d2499 WR |
264 | debugger how to attach to the systemd instance running in our mkosi container/VM. Create the file, and possibly |
265 | the directory, and add the following contents: | |
66dc9b46 DDM |
266 | |
267 | ```json | |
268 | { | |
269 | "version": "0.2.0", | |
270 | "configurations": [ | |
271 | { | |
272 | "type": "cppdbg", | |
273 | "program": "/usr/lib/systemd/systemd", | |
3f3bc1f2 | 274 | "processId": "${command:pickRemoteProcess}", |
66dc9b46 DDM |
275 | "request": "attach", |
276 | "name": "systemd", | |
277 | "pipeTransport": { | |
278 | "pipeProgram": "mkosi", | |
279 | "pipeArgs": [ | |
280 | "-C", | |
281 | "/path/to/systemd/repo/directory/on/host/system/", | |
282 | "ssh" | |
283 | ], | |
284 | "debuggerPath": "/usr/bin/gdb" | |
285 | }, | |
286 | "MIMode": "gdb", | |
287 | "sourceFileMap": { | |
ee7304df | 288 | "/root/src/systemd": { |
66dc9b46 DDM |
289 | "editorPath": "${workspaceFolder}", |
290 | "useForBreakpoints": false | |
291 | }, | |
66dc9b46 DDM |
292 | } |
293 | } | |
294 | ] | |
295 | } | |
296 | ``` | |
297 | ||
298 | Now that the debugger knows how to connect to our process in the container/VM and we've set up the necessary | |
299 | source mappings, go to the "Run and Debug" window and run the "systemd" debug configuration. If everything | |
300 | goes well, the debugger should now be attached to the systemd instance running in the container/VM. You can | |
301 | attach breakpoints from the editor and enjoy all the other features of VSCode's debugger. | |
302 | ||
303 | To debug systemd components other than PID 1, set "program" to the full path of the component you want to | |
304 | debug and set "processId" to "${command:pickProcess}". Now, when starting the debugger, VSCode will ask you | |
305 | the PID of the process you want to debug. Run `systemctl show --property MainPID --value <component>` in the | |
306 | container to figure out the PID and enter it when asked and VSCode will attach to that process instead. | |
948d085e | 307 | |
818e46ae | 308 | ## Debugging systemd-boot |
948d085e | 309 | |
badea0d6 ZJS |
310 | During boot, systemd-boot and the stub loader will output messages like |
311 | `systemd-boot@0x0A` and `systemd-stub@0x0B`, providing the base of the loaded | |
312 | code. This location can then be used to attach to a QEMU session (provided it | |
313 | was run with `-s`). See `debug-sd-boot.sh` script in the tools folder which | |
314 | automates this processes. | |
315 | ||
316 | If the debugger is too slow to attach to examine an early boot code passage, | |
317 | the call to `DEFINE_EFI_MAIN_FUNCTION()` can be modified to enable waiting. As | |
318 | soon as the debugger has control, we can then run `set variable wait = 0` or | |
319 | `return` to continue. Once the debugger has attached, setting breakpoints will | |
320 | work like usual. | |
948d085e JJ |
321 | |
322 | To debug systemd-boot in an IDE such as VSCode we can use a launch configuration like this: | |
323 | ```json | |
324 | { | |
325 | "name": "systemd-boot", | |
326 | "type": "cppdbg", | |
327 | "request": "launch", | |
328 | "program": "${workspaceFolder}/build/src/boot/efi/systemd-bootx64.efi", | |
329 | "cwd": "${workspaceFolder}", | |
330 | "MIMode": "gdb", | |
331 | "miDebuggerServerAddress": ":1234", | |
332 | "setupCommands": [ | |
333 | { "text": "shell mkfifo /tmp/sdboot.{in,out}" }, | |
334 | { "text": "shell qemu-system-x86_64 [...] -s -serial pipe:/tmp/sdboot" }, | |
335 | { "text": "shell ${workspaceFolder}/tools/debug-sd-boot.sh ${workspaceFolder}/build/src/boot/efi/systemd-bootx64.efi /tmp/sdboot.out systemd-boot.gdb" }, | |
336 | { "text": "source /tmp/systemd-boot.gdb" }, | |
337 | ] | |
338 | } | |
339 | ``` | |
d12e9bdc DDM |
340 | |
341 | ## Hacking on the kernel + systemd | |
342 | ||
343 | If you're hacking on the kernel in tandem with systemd, you can clone a kernel repository in mkosi.kernel/ in | |
344 | the systemd repository, and mkosi will automatically build that kernel and install it into the final image. | |
345 | To prevent the distribution's kernel from being installed (which isn't necessary since we're building our | |
94c9855a | 346 | own kernel), you can add the following snippets to mkosi.conf.d/20-local.conf: |
d12e9bdc DDM |
347 | |
348 | (This snippet is for Fedora, the list of packages will need to be changed for other distributions) | |
349 | ||
350 | ``` | |
351 | [Distribution] | |
352 | CacheInitrd=no | |
353 | ||
354 | [Content] | |
355 | BasePackages=conditional | |
356 | Packages=systemd | |
357 | util-linux | |
358 | dracut | |
359 | ``` |