CODING_STYLE

   1 - 8ch indent, no tabs, except for files in man/ which are 2ch indent,
   2   and still no tabs
   3
   4 - We prefer /* comments */ over // comments, please. This is not C++, after
   5   all. (Yes we know that C99 supports both kinds of comments, but still,
   6   please!)
   7
   8 - Don't break code lines too eagerly. We do *not* force line breaks at
   9   80ch, all of today's screens should be much larger than that. But
  10   then again, don't overdo it, ~140ch should be enough really.
  11
  12 - Variables and functions *must* be static, unless they have a
  13   prototype, and are supposed to be exported.
  14
  15 - structs in MixedCase (with exceptions, such as public API structs),
  16   variables + functions in lower_case.
  17
  18 - The destructors always unregister the object from the next bigger
  19   object, not the other way around
  20
  21 - To minimize strict aliasing violations, we prefer unions over casting
  22
  23 - For robustness reasons, destructors should be able to destruct
  24   half-initialized objects, too
  25
  26 - Error codes are returned as negative Exxx. e.g. return -EINVAL. There
  27   are some exceptions: for constructors, it is OK to return NULL on
  28   OOM. For lookup functions, NULL is fine too for "not found".
  29
  30   Be strict with this. When you write a function that can fail due to
  31   more than one cause, it *really* should have "int" as return value
  32   for the error code.
  33
  34 - Do not bother with error checking whether writing to stdout/stderr
  35   worked.
  36
  37 - Do not log errors from "library" code, only do so from "main
  38   program" code. (With one exception: it is OK to log with DEBUG level
  39   from any code, with the exception of maybe inner loops).
  40
  41 - Always check OOM. There is no excuse. In program code, you can use
  42   "log_oom()" for then printing a short message, but not in "library" code.
  43
  44 - Do not issue NSS requests (that includes user name and host name
  45   lookups) from PID 1 as this might trigger deadlocks when those
  46   lookups involve synchronously talking to services that we would need
  47   to start up
  48
  49 - Do not synchronously talk to any other service from PID 1, due to
  50   risk of deadlocks
  51
  52 - Avoid fixed-size string buffers, unless you really know the maximum
  53   size and that maximum size is small. They are a source of errors,
  54   since they possibly result in truncated strings. It is often nicer
  55   to use dynamic memory, alloca() or VLAs. If you do allocate fixed-size
  56   strings on the stack, then it is probably only OK if you either
  57   use a maximum size such as LINE_MAX, or count in detail the maximum
  58   size a string can have. (DECIMAL_STR_MAX and DECIMAL_STR_WIDTH
  59   macros are your friends for this!)
  60
  61   Or in other words, if you use "char buf[256]" then you are likely
  62   doing something wrong!
  63
  64 - Stay uniform. For example, always use "usec_t" for time
  65   values. Do not mix usec and msec, and usec and whatnot.
  66
  67 - Make use of _cleanup_free_ and friends. It makes your code much
  68   nicer to read!
  69
  70 - Be exceptionally careful when formatting and parsing floating point
  71   numbers. Their syntax is locale dependent (i.e. "5.000" in en_US is
  72   generally understood as 5, while on de_DE as 5000.).
  73
  74 - Try to use this:
  75
  76       void foo() {
  77       }
  78
  79   instead of this:
  80
  81       void foo()
  82       {
  83       }
  84
  85   But it is OK if you do not.
  86
  87 - Single-line "if" blocks should not be enclosed in {}. Use this:
  88
  89   if (foobar)
  90           waldo();
  91
  92   instead of this:
  93
  94   if (foobar) {
  95           waldo();
  96   }
  97
  98 - Do not write "foo ()", write "foo()".
  99
 100 - Please use streq() and strneq() instead of strcmp(), strncmp() where applicable.
 101
 102 - Please do not allocate variables on the stack in the middle of code,
 103   even if C99 allows it. Wrong:
 104
 105   {
 106           a = 5;
 107           int b;
 108           b = a;
 109   }
 110
 111   Right:
 112
 113   {
 114           int b;
 115           a = 5;
 116           b = a;
 117   }
 118
 119 - Unless you allocate an array, "double" is always the better choice
 120   than "float". Processors speak "double" natively anyway, so this is
 121   no speed benefit, and on calls like printf() "float"s get promoted
 122   to "double"s anyway, so there is no point.
 123
 124 - Do not mix function invocations with variable definitions in one
 125   line. Wrong:
 126
 127   {
 128           int a = foobar();
 129           uint64_t x = 7;
 130   }
 131
 132   Right:
 133
 134   {
 135           int a;
 136           uint64_t x = 7;
 137
 138           a = foobar();
 139   }
 140
 141 - Use "goto" for cleaning up, and only use it for that. i.e. you may
 142   only jump to the end of a function, and little else. Never jump
 143   backwards!
 144
 145 - Think about the types you use. If a value cannot sensibly be
 146   negative, do not use "int", but use "unsigned".
 147
 148 - Do not use types like "short". They *never* make sense. Use ints,
 149   longs, long longs, all in unsigned+signed fashion, and the fixed
 150   size types uint32_t and so on, as well as size_t, but nothing
 151   else. Do not use kernel types like u32 and so on, leave that to the
 152   kernel.
 153
 154 - Public API calls (i.e. functions exported by our shared libraries)
 155   must be marked "_public_" and need to be prefixed with "sd_". No
 156   other functions should be prefixed like that.
 157
 158 - In public API calls, you *must* validate all your input arguments for
 159   programming error with assert_return() and return a sensible return
 160   code. In all other calls, it is recommended to check for programming
 161   errors with a more brutal assert(). We are more forgiving to public
 162   users then for ourselves! Note that assert() and assert_return()
 163   really only should be used for detecting programming errors, not for
 164   runtime errors. assert() and assert_return() by usage of _likely_()
 165   inform the compiler that he should not expect these checks to fail,
 166   and they inform fellow programmers about the expected validity and
 167   range of parameters.
 168
 169 - Never use strtol(), atoi() and similar calls. Use safe_atoli(),
 170   safe_atou32() and suchlike instead. They are much nicer to use in
 171   most cases and correctly check for parsing errors.
 172
 173 - For every function you add, think about whether it is a "logging"
 174   function or a "non-logging" function. "Logging" functions do logging
 175   on their own, "non-logging" function never log on their own and
 176   expect their callers to log. All functions in "library" code,
 177   i.e. in src/shared/ and suchlike must be "non-logging". Every time a
 178   "logging" function calls a "non-logging" function, it should log
 179   about the resulting errors. If a "logging" function calls another
 180   "logging" function, then it should not generate log messages, so
 181   that log messages are not generated twice for the same errors.
 182
 183 - Avoid static variables, except for caches and very few other
 184   cases. Think about thread-safety! While most of our code is never
 185   used in threaded environments, at least the library code should make
 186   sure it works correctly in them. Instead of doing a lot of locking
 187   for that, we tend to prefer using TLS to do per-thread caching (which
 188   only works for small, fixed-size cache objects), or we disable
 189   caching for any thread that is not the main thread. Use
 190   is_main_thread() to detect whether the calling thread is the main
 191   thread.
 192
 193 - Command line option parsing:
 194   - Do not print full help() on error, be specific about the error.
 195   - Do not print messages to stdout on error.
 196   - Do not POSIX_ME_HARDER unless necessary, i.e. avoid "+" in option string.
 197
 198 - Do not write functions that clobber call-by-reference variables on
 199   failure. Use temporary variables for these cases and change the
 200   passed in variables only on success.
 201
 202 - When you allocate a file descriptor, it should be made O_CLOEXEC
 203   right from the beginning, as none of our files should leak to forked
 204   binaries by default. Hence, whenever you open a file, O_CLOEXEC must
 205   be specified, right from the beginning. This also applies to
 206   sockets. Effectively this means that all invocations to:
 207
 208   a) open() must get O_CLOEXEC passed
 209   b) socket() and socketpair() must get SOCK_CLOEXEC passed
 210   c) recvmsg() must get MSG_CMSG_CLOEXEC set
 211   d) F_DUPFD_CLOEXEC should be used instead of F_DUPFD, and so on
 212
 213 - We never use the POSIX version of basename() (which glibc defines it in
 214   libgen.h), only the GNU version (which glibc defines in string.h).
 215   The only reason to include libgen.h is because dirname()
 216   is needed. Everytime you need that please immediately undefine
 217   basename(), and add a comment about it, so that no code ever ends up
 218   using the POSIX version!
 219
 220 - Use the bool type for booleans, not integers. One exception: in public
 221   headers (i.e those in src/systemd/sd-*.h) use integers after all, as "bool"
 222   is C99 and in our public APIs we try to stick to C89 (with a few extension).
 223
 224 - When you invoke certain calls like unlink(), or mkdir_p() and you
 225   know it is safe to ignore the error it might return (because a later
 226   call would detect the failure anyway, or because the error is in an
 227   error path and you thus couldn't do anything about it anyway), then
 228   make this clear by casting the invocation explicitly to (void). Code
 229   checks like Coverity understand that, and will not complain about
 230   ignored error codes. Hence, please use this:
 231
 232       (void) unlink("/foo/bar/baz");
 233
 234   instead of just this:
 235
 236       unlink("/foo/bar/baz");
 237
 238 - Don't invoke exit(), ever. It is not replacement for proper error
 239   handling. Please escalate errors up your call chain, and use normal
 240   "return" to exit from the main function of a process. If you
 241   fork()ed off a child process, please use _exit() instead of exit(),
 242   so that the exit handlers are not run.
 243
 244 - Please never use dup(). Use fcntl(fd, F_DUPFD_CLOEXEC, 3)
 245   instead. For two reason: first, you want O_CLOEXEC set on the new fd
 246   (see above). Second, dup() will happily duplicate your fd as 0, 1,
 247   2, i.e. stdin, stdout, stderr, should those fds be closed. Given the
 248   special semantics of those fds, it's probably a good idea to avoid
 249   them. F_DUPFD_CLOEXEC with "3" as parameter avoids them.
 250
 251 - When you define a destructor or unref() call for an object, please
 252   accept a NULL object and simply treat this as NOP. This is similar
 253   to how libc free() works, which accepts NULL pointers and becomes a
 254   NOP for them. By following this scheme a lot of if checks can be
 255   removed before invoking your destructor, which makes the code
 256   substantially more readable and robust.
 257
 258 - Related to this: when you define a destructor or unref() call for an
 259   object, please make it return the same type it takes and always
 260   return NULL from it. This allows writing code like this:
 261
 262             p = foobar_unref(p);
 263
 264   which will always work regardless if p is initialized or not, and
 265   guarantees that p is NULL afterwards, all in just one line.
 266
 267 - Use alloca(), but never forget that it is not OK to invoke alloca()
 268   within a loop or within function call parameters. alloca() memory is
 269   released at the end of a function, and not at the end of a {}
 270   block. Thus, if you invoke it in a loop, you keep increasing the
 271   stack pointer without ever releasing memory again. (VLAs have better
 272   behaviour in this case, so consider using them as an alternative.)
 273   Regarding not using alloca() within function parameters, see the
 274   BUGS section of the alloca(3) man page.
 275
 276 - Use memzero() or even better zero() instead of memset(..., 0, ...)
 277
 278 - Instead of using memzero()/memset() to initialize structs allocated
 279   on the stack, please try to use c99 structure initializers. It's
 280   short, prettier and actually even faster at execution. Hence:
 281
 282           struct foobar t = {
 283                   .foo = 7,
 284                   .bar = "bazz",
 285           };
 286
 287   instead of:
 288
 289           struct foobar t;
 290           zero(t);
 291           t.foo = 7;
 292           t.bar = "bazz";
 293
 294 - When returning a return code from main(), please preferably use
 295   EXIT_FAILURE and EXIT_SUCCESS as defined by libc.
 296
 297 - The order in which header files are included doesn't matter too
 298   much. systemd-internal headers must not rely on an include order, so
 299   it is safe to include them in any order possible.
 300   However, to not clutter global includes, and to make sure internal
 301   definitions will not affect global headers, please always include the
 302   headers of external components first (these are all headers enclosed
 303   in <>), followed by our own exported headers (usually everything
 304   that's prefixed by "sd-"), and then followed by internal headers.
 305   Furthermore, in all three groups, order all includes alphabetically
 306   so duplicate includes can easily be detected.
 307
 308 - To implement an endless loop, use "for (;;)" rather than "while
 309   (1)". The latter is a bit ugly anyway, since you probably really
 310   meant "while (true)"... To avoid the discussion what the right
 311   always-true expression for an infinite while() loop is our
 312   recommendation is to simply write it without any such expression by
 313   using "for (;;)".
 314
 315 - Never use the "off_t" type, and particularly avoid it in public
 316   APIs. It's really weirdly defined, as it usually is 64bit and we
 317   don't support it any other way, but it could in theory also be
 318   32bit. Which one it is depends on a compiler switch chosen by the
 319   compiled program, which hence corrupts APIs using it unless they can
 320   also follow the program's choice. Moreover, in systemd we should
 321   parse values the same way on all architectures and cannot expose
 322   off_t values over D-Bus. To avoid any confusion regarding conversion
 323   and ABIs, always use simply uint64_t directly.