3 A testsuite covering functionality of btrfs-progs, ie. the checker, image, mkfs
4 and similar tools. There are no additional requirements on kernel features
5 (other than `CONFIG_BTRFS_FS` built-in or module), the
6 tests build on top of the core functionality like snapshots and device
7 management. In some cases optional features are turned on by mkfs and the
8 filesystem image could be mounted, such tests might fail if there's lack of
13 Run the tests from the top directory:
21 or selectively from the `tests/` directory:
28 The verbose output of the tests is logged into a file named after the test
29 category, eg. `fsck-tests-results.txt`.
33 The tests are prefixed by a number for ordering and uniqueness. To run a
40 where `MASK` is a glob expression that will execute only tests
41 that match the MASK. Here the test number comes handy:
44 $ make TEST=001\* test-fsck
45 $ TEST=001\* ./fsck-tests.sh
48 will run the first test in fsck-tests subdirectory.
55 * tests targeted at bugs that are fixable by fsck
57 *tests/convert-tests/:*
59 * coverage tests of ext2/3/4 and btrfs-convert options
63 * collection of fuzzed or crafted images
64 * tests that are supposed to run various utilities on the images and not
69 * tests for command line interface, option coverage, weird option combinations that should not work
70 * not necessary to do any functional testing, could be rather lightweight
71 * functional tests should go to to other test dirs
72 * the driver script will only execute `./test.sh` in the test directory
76 * anything that does not fit to the above, the test driver script will only
77 execute `./test.sh` in the test directory
79 *tests/common, tests/common.convert:*
81 * script with shell helpers, separated by functionality
85 * default testing image, the file is never deleted by the scripts but
86 truncated to 0 bytes, so it keeps it's permissions. It's eg. possible to
87 host it on NFS, make it `chmod a+w` for root.
90 ## Other tuning, environment variables
94 It's possible to wrap the tested commands to utilities that might do more
95 checking or catch failures at runtime. This can be done by setting the
96 `INSTRUMENT` environment variable:
99 INSTRUMENT=valgrind ./fuzz-tests.sh # in tests/
100 make INSTRUMENT=valgrind test-fuzz # in the top directory
103 The variable is prepended to the command *unquoted*, all sorts of shell tricks
106 Note: instrumentation is not applied to privileged commands (anything that uses
109 ### Verbosity, test tuning
111 * `TEST_LOG=tty` -- setting the variable will print all commands executed by
112 some of the wrappers (`run_check` etc), other commands are not printed to the
113 terminal (but the full output is in the log)
115 * `TEST_LOG=dump` -- dump the entire testing log when a test fails
117 * `TEST_ENABLE_OVERRIDE` -- defined either as make arguments or via
118 `tests/common.local` to enable additional arguments to some commands, using
119 the variable(s) below (default: false, enable by setting to 'true')
121 * `TEST_ARGS_CHECK` -- user-defined arguments to `btrfs check`, before the
122 test-specific arguments
124 Multiple values can be separated by `,`.
128 Some commands require root privileges (to mount/umount, access loop devices).
129 It is assumed that `sudo` will work in some way (no password, password asked
130 and cached). Note that instrumentation is not applied in this case, for safety
131 reasons. You need to modify the test script instead.
135 The tests are supposed to cleanup after themselves if they pass. In case of
136 failure, the rest of the tests are skipped and intermediate files, mounts and
137 loop devices are kept. This should help to investigate the test failure but at
138 least the mounts and loop devices need to be cleaned before the next run.
140 This is partially done by the script `clean-tests.sh`, you may want to check
141 the loop devices as they are managed on a per-test basis.
143 ### Prototyping tests, quick tests
145 There's a script `test-console.sh` that will run shell commands in a loop and
146 logs the output with the testing environment set up.
148 ### Runtime dependencies
150 The tests use some common system utilities like `find`, `rm`, `dd`. Additionally,
151 specific tests need the following packages installed: `acl`, `attr`,
152 `e2fsprogs`, `reiserfsprogs`
157 1. Pick the category for the new test or fallback to `misc-tests` if not sure. For
158 an easy start copy an existing `test.sh` script from some test that might be
159 close to the purpose of your new test. The environment setup includes the
160 common scripts and/or prepares the test devices. Other scripts contain examples
161 how to do mkfs, mount, unmount, check, etc.
163 2. Use the highest unused number in the sequence, write a short descriptive title
164 and join by dashes `-`. This will become the directory name, eg. `012-subvolume-sync-must-wait`.
166 3. Write a short description of the bug and how it's tested to the comment at the
167 begining of `test.sh`. You don't need to add the file to git yet.
169 4. Write the test commands, comment anything that's not obvious.
171 5. **Test your test.** Use the `TEST` variable to jump right to your test:
173 $ make TEST=012\* tests-misc # from top directory
174 $ TEST=012\* ./misc-tests.sh # from tests/
177 6. The commit changelog should reference a commit that either introduced or
178 fixed the bug (or both). Subject line of the shall mention the name of the
179 new directory for ease of search, eg. `btrfs-progs: tests: add 012-subvolume-sync-must-wait`
182 ### Crafted/fuzzed images
184 Images that are create by fuzzing or specially crafted to trigger some error
185 conditions should be added to the directory *fuzz-tests/images*, accompanied by
186 a textual description of the source (bugzilla, mail), the reporter, brief
187 description of the problem or the stack trace.
189 If you have a fix for the problem, please submit it prior to the test image, so
190 the fuzz tests always succeed when run on random checked out. This helps
194 # Coding style, best practices
198 * quote all variables by default, any path, even the TOP could need that, and
200 * there are exceptions:
201 * `$SUDO_HELPER` as it might be intentionally unset
202 * the variable is obviously set to a value that does not require it
203 * use `#!/bin/bash` explicitly
204 * check for all external dependencies (`check_prereq_global`)
205 * check for internal dependencies (`check_prereq`), though the basic set is
206 always built when the tests are started through make
207 * use functions instead of repeating code
208 * generic helpers could be factored to the `common` script
209 * cleanup after successful test
210 * use common helpers and variables
214 * pull external dependencies if we can find a way to replace them: example is
215 `xfs_io` that's conveniently used in fstests but we'd require `xfsprogs`,
217 * throw away (redirect to */dev/null*) output of commands unless it's justified
218 (ie. really too much text, unnecessary slowdown) -- the test output log is
219 regenerated all the time and we need to be able to analyze test failures or
220 just observe how the tests progress
221 * cleanup after failed test -- the testsuite stops on first failure and the
222 developer can eg. access the environment that the test created and do further
224 * this might change in the future so the tests cover as much as possible, but
225 this would require to enhance all tests with a cleanup phase