3While there is much to be said for a solid and community-oriented design
   4process, the proof of any kernel development project is in the resulting
   5code.  It is the code which will be examined by other developers and merged
   6(or not) into the mainline tree.  So it is the quality of this code which
   7will determine the ultimate success of the project.
   9This section will examine the coding process.  We'll start with a look at a
  10number of ways in which kernel developers can go wrong.  Then the focus
  11will shift toward doing things right and the tools which can help in that
  154.1: PITFALLS
  17* Coding style
  19The kernel has long had a standard coding style, described in
  20Documentation/CodingStyle.  For much of that time, the policies described
  21in that file were taken as being, at most, advisory.  As a result, there is
  22a substantial amount of code in the kernel which does not meet the coding
  23style guidelines.  The presence of that code leads to two independent
  24hazards for kernel developers.
  26The first of these is to believe that the kernel coding standards do not
  27matter and are not enforced.  The truth of the matter is that adding new
  28code to the kernel is very difficult if that code is not coded according to
  29the standard; many developers will request that the code be reformatted
  30before they will even review it.  A code base as large as the kernel
  31requires some uniformity of code to make it possible for developers to
  32quickly understand any part of it.  So there is no longer room for
  33strangely-formatted code.
  35Occasionally, the kernel's coding style will run into conflict with an
  36employer's mandated style.  In such cases, the kernel's style will have to
  37win before the code can be merged.  Putting code into the kernel means
  38giving up a degree of control in a number of ways - including control over
  39how the code is formatted.
  41The other trap is to assume that code which is already in the kernel is
  42urgently in need of coding style fixes.  Developers may start to generate
  43reformatting patches as a way of gaining familiarity with the process, or
  44as a way of getting their name into the kernel changelogs - or both.  But
  45pure coding style fixes are seen as noise by the development community;
  46they tend to get a chilly reception.  So this type of patch is best
  47avoided.  It is natural to fix the style of a piece of code while working
  48on it for other reasons, but coding style changes should not be made for
  49their own sake.
  51The coding style document also should not be read as an absolute law which
  52can never be transgressed.  If there is a good reason to go against the
  53style (a line which becomes far less readable if split to fit within the
  5480-column limit, for example), just do it.
  57* Abstraction layers
  59Computer Science professors teach students to make extensive use of
  60abstraction layers in the name of flexibility and information hiding.
  61Certainly the kernel makes extensive use of abstraction; no project
  62involving several million lines of code could do otherwise and survive.
  63But experience has shown that excessive or premature abstraction can be
  64just as harmful as premature optimization.  Abstraction should be used to
  65the level required and no further.
  67At a simple level, consider a function which has an argument which is
  68always passed as zero by all callers.  One could retain that argument just
  69in case somebody eventually needs to use the extra flexibility that it
  70provides.  By that time, though, chances are good that the code which
  71implements this extra argument has been broken in some subtle way which was
  72never noticed - because it has never been used.  Or, when the need for
  73extra flexibility arises, it does not do so in a way which matches the
  74programmer's early expectation.  Kernel developers will routinely submit
  75patches to remove unused arguments; they should, in general, not be added
  76in the first place.
  78Abstraction layers which hide access to hardware - often to allow the bulk
  79of a driver to be used with multiple operating systems - are especially
  80frowned upon.  Such layers obscure the code and may impose a performance
  81penalty; they do not belong in the Linux kernel.
  83On the other hand, if you find yourself copying significant amounts of code
  84from another kernel subsystem, it is time to ask whether it would, in fact,
  85make sense to pull out some of that code into a separate library or to
  86implement that functionality at a higher level.  There is no value in
  87replicating the same code throughout the kernel.
  90* #ifdef and preprocessor use in general
  92The C preprocessor seems to present a powerful temptation to some C
  93programmers, who see it as a way to efficiently encode a great deal of
  94flexibility into a source file.  But the preprocessor is not C, and heavy
  95use of it results in code which is much harder for others to read and
  96harder for the compiler to check for correctness.  Heavy preprocessor use
  97is almost always a sign of code which needs some cleanup work.
  99Conditional compilation with #ifdef is, indeed, a powerful feature, and it
 100is used within the kernel.  But there is little desire to see code which is
 101sprinkled liberally with #ifdef blocks.  As a general rule, #ifdef use
 102should be confined to header files whenever possible.
 103Conditionally-compiled code can be confined to functions which, if the code
 104is not to be present, simply become empty.  The compiler will then quietly
 105optimize out the call to the empty function.  The result is far cleaner
 106code which is easier to follow.
 108C preprocessor macros present a number of hazards, including possible
 109multiple evaluation of expressions with side effects and no type safety.
 110If you are tempted to define a macro, consider creating an inline function
 111instead.  The code which results will be the same, but inline functions are
 112easier to read, do not evaluate their arguments multiple times, and allow
 113the compiler to perform type checking on the arguments and return value.
 116* Inline functions
 118Inline functions present a hazard of their own, though.  Programmers can
 119become enamored of the perceived efficiency inherent in avoiding a function
 120call and fill a source file with inline functions.  Those functions,
 121however, can actually reduce performance.  Since their code is replicated
 122at each call site, they end up bloating the size of the compiled kernel.
 123That, in turn, creates pressure on the processor's memory caches, which can
 124slow execution dramatically.  Inline functions, as a rule, should be quite
 125small and relatively rare.  The cost of a function call, after all, is not
 126that high; the creation of large numbers of inline functions is a classic
 127example of premature optimization.
 129In general, kernel programmers ignore cache effects at their peril.  The
 130classic time/space tradeoff taught in beginning data structures classes
 131often does not apply to contemporary hardware.  Space *is* time, in that a
 132larger program will run slower than one which is more compact.
 134More recent compilers take an increasingly active role in deciding whether
 135a given function should actually be inlined or not.  So the liberal
 136placement of "inline" keywords may not just be excessive; it could also be
 140* Locking
 142In May, 2006, the "Devicescape" networking stack was, with great
 143fanfare, released under the GPL and made available for inclusion in the
 144mainline kernel.  This donation was welcome news; support for wireless
 145networking in Linux was considered substandard at best, and the Devicescape
 146stack offered the promise of fixing that situation.  Yet, this code did not
 147actually make it into the mainline until June, 2007 (2.6.22).  What
 150This code showed a number of signs of having been developed behind
 151corporate doors.  But one large problem in particular was that it was not
 152designed to work on multiprocessor systems.  Before this networking stack
 153(now called mac80211) could be merged, a locking scheme needed to be
 154retrofitted onto it.  
 156Once upon a time, Linux kernel code could be developed without thinking
 157about the concurrency issues presented by multiprocessor systems.  Now,
 158however, this document is being written on a dual-core laptop.  Even on
 159single-processor systems, work being done to improve responsiveness will
 160raise the level of concurrency within the kernel.  The days when kernel
 161code could be written without thinking about locking are long past.
 163Any resource (data structures, hardware registers, etc.) which could be
 164accessed concurrently by more than one thread must be protected by a lock.
 165New code should be written with this requirement in mind; retrofitting
 166locking after the fact is a rather more difficult task.  Kernel developers
 167should take the time to understand the available locking primitives well
 168enough to pick the right tool for the job.  Code which shows a lack of
 169attention to concurrency will have a difficult path into the mainline.
 172* Regressions
 174One final hazard worth mentioning is this: it can be tempting to make a
 175change (which may bring big improvements) which causes something to break
 176for existing users.  This kind of change is called a "regression," and
 177regressions have become most unwelcome in the mainline kernel.  With few
 178exceptions, changes which cause regressions will be backed out if the
 179regression cannot be fixed in a timely manner.  Far better to avoid the
 180regression in the first place.
 182It is often argued that a regression can be justified if it causes things
 183to work for more people than it creates problems for.  Why not make a
 184change if it brings new functionality to ten systems for each one it
 185breaks?  The best answer to this question was expressed by Linus in July,
 188        So we don't fix bugs by introducing new problems.  That way lies
 189        madness, and nobody ever knows if you actually make any real
 190        progress at all. Is it two steps forwards, one step back, or one
 191        step forward and two steps back?
 195An especially unwelcome type of regression is any sort of change to the
 196user-space ABI.  Once an interface has been exported to user space, it must
 197be supported indefinitely.  This fact makes the creation of user-space
 198interfaces particularly challenging: since they cannot be changed in
 199incompatible ways, they must be done right the first time.  For this
 200reason, a great deal of thought, clear documentation, and wide review for
 201user-space interfaces is always required.
 207For now, at least, the writing of error-free code remains an ideal that few
 208of us can reach.  What we can hope to do, though, is to catch and fix as
 209many of those errors as possible before our code goes into the mainline
 210kernel.  To that end, the kernel developers have put together an impressive
 211array of tools which can catch a wide variety of obscure problems in an
 212automated way.  Any problem caught by the computer is a problem which will
 213not afflict a user later on, so it stands to reason that the automated
 214tools should be used whenever possible.
 216The first step is simply to heed the warnings produced by the compiler.
 217Contemporary versions of gcc can detect (and warn about) a large number of
 218potential errors.  Quite often, these warnings point to real problems.
 219Code submitted for review should, as a rule, not produce any compiler
 220warnings.  When silencing warnings, take care to understand the real cause
 221and try to avoid "fixes" which make the warning go away without addressing
 222its cause.
 224Note that not all compiler warnings are enabled by default.  Build the
 225kernel with "make EXTRA_CFLAGS=-W" to get the full set.
 227The kernel provides several configuration options which turn on debugging
 228features; most of these are found in the "kernel hacking" submenu.  Several
 229of these options should be turned on for any kernel used for development or
 230testing purposes.  In particular, you should turn on:
 233   extra set of warnings for problems like the use of deprecated interfaces
 234   or ignoring an important return value from a function.  The output
 235   generated by these warnings can be verbose, but one need not worry about
 236   warnings from other parts of the kernel.
 238 - DEBUG_OBJECTS will add code to track the lifetime of various objects
 239   created by the kernel and warn when things are done out of order.  If
 240   you are adding a subsystem which creates (and exports) complex objects
 241   of its own, consider adding support for the object debugging
 242   infrastructure.
 244 - DEBUG_SLAB can find a variety of memory allocation and use errors; it
 245   should be used on most development kernels.
 248   number of common locking errors.
 250There are quite a few other debugging options, some of which will be
 251discussed below.  Some of them have a significant performance impact and
 252should not be used all of the time.  But some time spent learning the
 253available options will likely be paid back many times over in short order. 
 255One of the heavier debugging tools is the locking checker, or "lockdep."
 256This tool will track the acquisition and release of every lock (spinlock or
 257mutex) in the system, the order in which locks are acquired relative to
 258each other, the current interrupt environment, and more.  It can then
 259ensure that locks are always acquired in the same order, that the same
 260interrupt assumptions apply in all situations, and so on.  In other words,
 261lockdep can find a number of scenarios in which the system could, on rare
 262occasion, deadlock.  This kind of problem can be painful (for both
 263developers and users) in a deployed system; lockdep allows them to be found
 264in an automated manner ahead of time.  Code with any sort of non-trivial
 265locking should be run with lockdep enabled before being submitted for
 268As a diligent kernel programmer, you will, beyond doubt, check the return
 269status of any operation (such as a memory allocation) which can fail.  The
 270fact of the matter, though, is that the resulting failure recovery paths
 271are, probably, completely untested.  Untested code tends to be broken code;
 272you could be much more confident of your code if all those error-handling
 273paths had been exercised a few times.
 275The kernel provides a fault injection framework which can do exactly that,
 276especially where memory allocations are involved.  With fault injection
 277enabled, a configurable percentage of memory allocations will be made to
 278fail; these failures can be restricted to a specific range of code.
 279Running with fault injection enabled allows the programmer to see how the
 280code responds when things go badly.  See
 281Documentation/fault-injection/fault-injection.txt for more information on
 282how to use this facility.
 284Other kinds of errors can be found with the "sparse" static analysis tool.
 285With sparse, the programmer can be warned about confusion between
 286user-space and kernel-space addresses, mixture of big-endian and
 287small-endian quantities, the passing of integer values where a set of bit
 288flags is expected, and so on.  Sparse must be installed separately (it can
 289be found at if your
 290distributor does not package it); it can then be run on the code by adding
 291"C=1" to your make command.
 293The "Coccinelle" tool ( is able to find a wide
 294variety of potential coding problems; it can also propose fixes for those
 295problems.  Quite a few "semantic patches" for the kernel have been packaged
 296under the scripts/coccinelle directory; running "make coccicheck" will run
 297through those semantic patches and report on any problems found.  See
 298Documentation/coccinelle.txt for more information.
 300Other kinds of portability errors are best found by compiling your code for
 301other architectures.  If you do not happen to have an S/390 system or a
 302Blackfin development board handy, you can still perform the compilation
 303step.  A large set of cross compilers for x86 systems can be found at 
 307Some time spent installing and using these compilers will help avoid
 308embarrassment later.
 313Documentation has often been more the exception than the rule with kernel
 314development.  Even so, adequate documentation will help to ease the merging
 315of new code into the kernel, make life easier for other developers, and
 316will be helpful for your users.  In many cases, the addition of
 317documentation has become essentially mandatory.
 319The first piece of documentation for any patch is its associated
 320changelog.  Log entries should describe the problem being solved, the form
 321of the solution, the people who worked on the patch, any relevant
 322effects on performance, and anything else that might be needed to
 323understand the patch.  Be sure that the changelog says *why* the patch is
 324worth applying; a surprising number of developers fail to provide that
 327Any code which adds a new user-space interface - including new sysfs or
 328/proc files - should include documentation of that interface which enables
 329user-space developers to know what they are working with.  See
 330Documentation/ABI/README for a description of how this documentation should
 331be formatted and what information needs to be provided.
 333The file Documentation/kernel-parameters.txt describes all of the kernel's
 334boot-time parameters.  Any patch which adds new parameters should add the
 335appropriate entries to this file.
 337Any new configuration options must be accompanied by help text which
 338clearly explains the options and when the user might want to select them.
 340Internal API information for many subsystems is documented by way of
 341specially-formatted comments; these comments can be extracted and formatted
 342in a number of ways by the "kernel-doc" script.  If you are working within
 343a subsystem which has kerneldoc comments, you should maintain them and add
 344them, as appropriate, for externally-available functions.  Even in areas
 345which have not been so documented, there is no harm in adding kerneldoc
 346comments for the future; indeed, this can be a useful activity for
 347beginning kernel developers.  The format of these comments, along with some
 348information on how to create kerneldoc templates can be found in the file
 351Anybody who reads through a significant amount of existing kernel code will
 352note that, often, comments are most notable by their absence.  Once again,
 353the expectations for new code are higher than they were in the past;
 354merging uncommented code will be harder.  That said, there is little desire
 355for verbosely-commented code.  The code should, itself, be readable, with
 356comments explaining the more subtle aspects.
 358Certain things should always be commented.  Uses of memory barriers should
 359be accompanied by a line explaining why the barrier is necessary.  The
 360locking rules for data structures generally need to be explained somewhere.
 361Major data structures need comprehensive documentation in general.
 362Non-obvious dependencies between separate bits of code should be pointed
 363out.  Anything which might tempt a code janitor to make an incorrect
 364"cleanup" needs a comment saying why it is done the way it is.  And so on.
 369The binary interface provided by the kernel to user space cannot be broken
 370except under the most severe circumstances.  The kernel's internal
 371programming interfaces, instead, are highly fluid and can be changed when
 372the need arises.  If you find yourself having to work around a kernel API,
 373or simply not using a specific functionality because it does not meet your
 374needs, that may be a sign that the API needs to change.  As a kernel
 375developer, you are empowered to make such changes.
 377There are, of course, some catches.  API changes can be made, but they need
 378to be well justified.  So any patch making an internal API change should be
 379accompanied by a description of what the change is and why it is
 380necessary.  This kind of change should also be broken out into a separate
 381patch, rather than buried within a larger patch.
 383The other catch is that a developer who changes an internal API is
 384generally charged with the task of fixing any code within the kernel tree
 385which is broken by the change.  For a widely-used function, this duty can
 386lead to literally hundreds or thousands of changes - many of which are
 387likely to conflict with work being done by other developers.  Needless to
 388say, this can be a large job, so it is best to be sure that the
 389justification is solid.  Note that the Coccinelle tool can help with
 390wide-ranging API changes.
 392When making an incompatible API change, one should, whenever possible,
 393ensure that code which has not been updated is caught by the compiler.
 394This will help you to be sure that you have found all in-tree uses of that
 395interface.  It will also alert developers of out-of-tree code that there is
 396a change that they need to respond to.  Supporting out-of-tree code is not
 397something that kernel developers need to be worried about, but we also do
 398not have to make life harder for out-of-tree developers than it needs to
 400 kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.