3 Linux Kernel patch submission checklist
4 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
6 Here are some basic things that developers should do if they want to see their
7 kernel patch submissions accepted more quickly.
9 These are all above and beyond the documentation that is provided in
10 :ref:`Documentation/SubmittingPatches <submittingpatches>`
11 and elsewhere regarding submitting Linux kernel patches.
14 1) If you use a facility then #include the file that defines/declares
15 that facility. Don't depend on other header files pulling in ones
20 a) with applicable or modified ``CONFIG`` options ``=y``, ``=m``, and
21 ``=n``. No ``gcc`` warnings/errors, no linker warnings/errors.
23 b) Passes ``allnoconfig``, ``allmodconfig``
25 c) Builds successfully when using ``O=builddir``
27 3) Builds on multiple CPU architectures by using local cross-compile tools
28 or some other build farm.
30 4) ppc64 is a good architecture for cross-compilation checking because it
31 tends to use ``unsigned long`` for 64-bit quantities.
33 5) Check your patch for general style as detailed in
34 :ref:`Documentation/CodingStyle <codingstyle>`.
35 Check for trivial violations with the patch style checker prior to
36 submission (``scripts/checkpatch.pl``).
37 You should be able to justify all violations that remain in
40 6) Any new or modified ``CONFIG`` options don't muck up the config menu.
42 7) All new ``Kconfig`` options have help text.
44 8) Has been carefully reviewed with respect to relevant ``Kconfig``
45 combinations. This is very hard to get right with testing -- brainpower
48 9) Check cleanly with sparse.
50 10) Use ``make checkstack`` and ``make namespacecheck`` and fix any problems
55 ``checkstack`` does not point out problems explicitly,
56 but any one function that uses more than 512 bytes on the stack is a
59 11) Include :ref:`kernel-doc <kernel_doc>` to document global kernel APIs.
60 (Not required for static functions, but OK there also.) Use
61 ``make htmldocs`` or ``make pdfdocs`` to check the
62 :ref:`kernel-doc <kernel_doc>` and fix any issues.
64 12) Has been tested with ``CONFIG_PREEMPT``, ``CONFIG_DEBUG_PREEMPT``,
65 ``CONFIG_DEBUG_SLAB``, ``CONFIG_DEBUG_PAGEALLOC``, ``CONFIG_DEBUG_MUTEXES``,
66 ``CONFIG_DEBUG_SPINLOCK``, ``CONFIG_DEBUG_ATOMIC_SLEEP``,
67 ``CONFIG_PROVE_RCU`` and ``CONFIG_DEBUG_OBJECTS_RCU_HEAD`` all
68 simultaneously enabled.
70 13) Has been build- and runtime tested with and without ``CONFIG_SMP`` and
73 14) If the patch affects IO/Disk, etc: has been tested with and without
76 15) All codepaths have been exercised with all lockdep features enabled.
78 16) All new ``/proc`` entries are documented under ``Documentation/``
80 17) All new kernel boot parameters are documented in
81 ``Documentation/kernel-parameters.txt``.
83 18) All new module parameters are documented with ``MODULE_PARM_DESC()``
85 19) All new userspace interfaces are documented in ``Documentation/ABI/``.
86 See ``Documentation/ABI/README`` for more information.
87 Patches that change userspace interfaces should be CCed to
88 linux-api@vger.kernel.org.
90 20) Check that it all passes ``make headers_check``.
92 21) Has been checked with injection of at least slab and page-allocation
93 failures. See ``Documentation/fault-injection/``.
95 If the new code is substantial, addition of subsystem-specific fault
96 injection might be appropriate.
98 22) Newly-added code has been compiled with ``gcc -W`` (use
99 ``make EXTRA_CFLAGS=-W``). This will generate lots of noise, but is good
100 for finding bugs like "warning: comparison between signed and unsigned".
102 23) Tested after it has been merged into the -mm patchset to make sure
103 that it still works with all of the other queued patches and various
104 changes in the VM, VFS, and other subsystems.
106 24) All memory barriers {e.g., ``barrier()``, ``rmb()``, ``wmb()``} need a
107 comment in the source code that explains the logic of what they are doing
110 25) If any ioctl's are added by the patch, then also update
111 ``Documentation/ioctl/ioctl-number.txt``.
113 26) If your modified source code depends on or uses any of the kernel
114 APIs or features that are related to the following ``Kconfig`` symbols,
115 then test multiple builds with the related ``Kconfig`` symbols disabled
116 and/or ``=m`` (if that option is available) [not all of these at the
117 same time, just various/random combinations of them]:
119 ``CONFIG_SMP``, ``CONFIG_SYSFS``, ``CONFIG_PROC_FS``, ``CONFIG_INPUT``, ``CONFIG_PCI``, ``CONFIG_BLOCK``, ``CONFIG_PM``, ``CONFIG_MAGIC_SYSRQ``,
120 ``CONFIG_NET``, ``CONFIG_INET=n`` (but latter with ``CONFIG_NET=y``).