Home | History | Annotate | Download | only in doc
      1 .. Copyright 2010 Nicolas Palix <npalix (a] diku.dk>
      2 .. Copyright 2010 Julia Lawall <julia (a] diku.dk>
      3 .. Copyright 2010 Gilles Muller <Gilles.Muller (a] lip6.fr>
      4 
      5 .. highlight:: none
      6 
      7 Coccinelle
      8 ==========
      9 
     10 Coccinelle is a tool for pattern matching and text transformation that has
     11 many uses in kernel development, including the application of complex,
     12 tree-wide patches and detection of problematic programming patterns.
     13 
     14 Getting Coccinelle
     15 -------------------
     16 
     17 The semantic patches included in the kernel use features and options
     18 which are provided by Coccinelle version 1.0.0-rc11 and above.
     19 Using earlier versions will fail as the option names used by
     20 the Coccinelle files and coccicheck have been updated.
     21 
     22 Coccinelle is available through the package manager
     23 of many distributions, e.g. :
     24 
     25  - Debian
     26  - Fedora
     27  - Ubuntu
     28  - OpenSUSE
     29  - Arch Linux
     30  - NetBSD
     31  - FreeBSD
     32 
     33 You can get the latest version released from the Coccinelle homepage at
     34 http://coccinelle.lip6.fr/
     35 
     36 Information and tips about Coccinelle are also provided on the wiki
     37 pages at http://cocci.ekstranet.diku.dk/wiki/doku.php
     38 
     39 Once you have it, run the following command::
     40 
     41      	./configure
     42         make
     43 
     44 as a regular user, and install it with::
     45 
     46         sudo make install
     47 
     48 Supplemental documentation
     49 ---------------------------
     50 
     51 For supplemental documentation refer to the wiki:
     52 
     53 https://bottest.wiki.kernel.org/coccicheck
     54 
     55 The wiki documentation always refers to the linux-next version of the script.
     56 
     57 Using Coccinelle on the Linux kernel
     58 ------------------------------------
     59 
     60 A Coccinelle-specific target is defined in the top level
     61 Makefile. This target is named ``coccicheck`` and calls the ``coccicheck``
     62 front-end in the ``scripts`` directory.
     63 
     64 Four basic modes are defined: ``patch``, ``report``, ``context``, and
     65 ``org``. The mode to use is specified by setting the MODE variable with
     66 ``MODE=<mode>``.
     67 
     68 - ``patch`` proposes a fix, when possible.
     69 
     70 - ``report`` generates a list in the following format:
     71   file:line:column-column: message
     72 
     73 - ``context`` highlights lines of interest and their context in a
     74   diff-like style.Lines of interest are indicated with ``-``.
     75 
     76 - ``org`` generates a report in the Org mode format of Emacs.
     77 
     78 Note that not all semantic patches implement all modes. For easy use
     79 of Coccinelle, the default mode is "report".
     80 
     81 Two other modes provide some common combinations of these modes.
     82 
     83 - ``chain`` tries the previous modes in the order above until one succeeds.
     84 
     85 - ``rep+ctxt`` runs successively the report mode and the context mode.
     86   It should be used with the C option (described later)
     87   which checks the code on a file basis.
     88 
     89 Examples
     90 ~~~~~~~~
     91 
     92 To make a report for every semantic patch, run the following command::
     93 
     94 		make coccicheck MODE=report
     95 
     96 To produce patches, run::
     97 
     98 		make coccicheck MODE=patch
     99 
    100 
    101 The coccicheck target applies every semantic patch available in the
    102 sub-directories of ``scripts/coccinelle`` to the entire Linux kernel.
    103 
    104 For each semantic patch, a commit message is proposed.  It gives a
    105 description of the problem being checked by the semantic patch, and
    106 includes a reference to Coccinelle.
    107 
    108 As any static code analyzer, Coccinelle produces false
    109 positives. Thus, reports must be carefully checked, and patches
    110 reviewed.
    111 
    112 To enable verbose messages set the V= variable, for example::
    113 
    114    make coccicheck MODE=report V=1
    115 
    116 Coccinelle parallelization
    117 ---------------------------
    118 
    119 By default, coccicheck tries to run as parallel as possible. To change
    120 the parallelism, set the J= variable. For example, to run across 4 CPUs::
    121 
    122    make coccicheck MODE=report J=4
    123 
    124 As of Coccinelle 1.0.2 Coccinelle uses Ocaml parmap for parallelization,
    125 if support for this is detected you will benefit from parmap parallelization.
    126 
    127 When parmap is enabled coccicheck will enable dynamic load balancing by using
    128 ``--chunksize 1`` argument, this ensures we keep feeding threads with work
    129 one by one, so that we avoid the situation where most work gets done by only
    130 a few threads. With dynamic load balancing, if a thread finishes early we keep
    131 feeding it more work.
    132 
    133 When parmap is enabled, if an error occurs in Coccinelle, this error
    134 value is propagated back, the return value of the ``make coccicheck``
    135 captures this return value.
    136 
    137 Using Coccinelle with a single semantic patch
    138 ---------------------------------------------
    139 
    140 The optional make variable COCCI can be used to check a single
    141 semantic patch. In that case, the variable must be initialized with
    142 the name of the semantic patch to apply.
    143 
    144 For instance::
    145 
    146 	make coccicheck COCCI=<my_SP.cocci> MODE=patch
    147 
    148 or::
    149 
    150 	make coccicheck COCCI=<my_SP.cocci> MODE=report
    151 
    152 
    153 Controlling Which Files are Processed by Coccinelle
    154 ---------------------------------------------------
    155 
    156 By default the entire kernel source tree is checked.
    157 
    158 To apply Coccinelle to a specific directory, ``M=`` can be used.
    159 For example, to check drivers/net/wireless/ one may write::
    160 
    161     make coccicheck M=drivers/net/wireless/
    162 
    163 To apply Coccinelle on a file basis, instead of a directory basis, the
    164 following command may be used::
    165 
    166     make C=1 CHECK="scripts/coccicheck"
    167 
    168 To check only newly edited code, use the value 2 for the C flag, i.e.::
    169 
    170     make C=2 CHECK="scripts/coccicheck"
    171 
    172 In these modes, which works on a file basis, there is no information
    173 about semantic patches displayed, and no commit message proposed.
    174 
    175 This runs every semantic patch in scripts/coccinelle by default. The
    176 COCCI variable may additionally be used to only apply a single
    177 semantic patch as shown in the previous section.
    178 
    179 The "report" mode is the default. You can select another one with the
    180 MODE variable explained above.
    181 
    182 Debugging Coccinelle SmPL patches
    183 ---------------------------------
    184 
    185 Using coccicheck is best as it provides in the spatch command line
    186 include options matching the options used when we compile the kernel.
    187 You can learn what these options are by using V=1, you could then
    188 manually run Coccinelle with debug options added.
    189 
    190 Alternatively you can debug running Coccinelle against SmPL patches
    191 by asking for stderr to be redirected to stderr, by default stderr
    192 is redirected to /dev/null, if you'd like to capture stderr you
    193 can specify the ``DEBUG_FILE="file.txt"`` option to coccicheck. For
    194 instance::
    195 
    196     rm -f cocci.err
    197     make coccicheck COCCI=scripts/coccinelle/free/kfree.cocci MODE=report DEBUG_FILE=cocci.err
    198     cat cocci.err
    199 
    200 You can use SPFLAGS to add debugging flags, for instance you may want to
    201 add both --profile --show-trying to SPFLAGS when debugging. For instance
    202 you may want to use::
    203 
    204     rm -f err.log
    205     export COCCI=scripts/coccinelle/misc/irqf_oneshot.cocci
    206     make coccicheck DEBUG_FILE="err.log" MODE=report SPFLAGS="--profile --show-trying" M=./drivers/mfd/arizona-irq.c
    207 
    208 err.log will now have the profiling information, while stdout will
    209 provide some progress information as Coccinelle moves forward with
    210 work.
    211 
    212 DEBUG_FILE support is only supported when using coccinelle >= 1.2.
    213 
    214 .cocciconfig support
    215 --------------------
    216 
    217 Coccinelle supports reading .cocciconfig for default Coccinelle options that
    218 should be used every time spatch is spawned, the order of precedence for
    219 variables for .cocciconfig is as follows:
    220 
    221 - Your current user's home directory is processed first
    222 - Your directory from which spatch is called is processed next
    223 - The directory provided with the --dir option is processed last, if used
    224 
    225 Since coccicheck runs through make, it naturally runs from the kernel
    226 proper dir, as such the second rule above would be implied for picking up a
    227 .cocciconfig when using ``make coccicheck``.
    228 
    229 ``make coccicheck`` also supports using M= targets.If you do not supply
    230 any M= target, it is assumed you want to target the entire kernel.
    231 The kernel coccicheck script has::
    232 
    233     if [ "$KBUILD_EXTMOD" = "" ] ; then
    234         OPTIONS="--dir $srctree $COCCIINCLUDE"
    235     else
    236         OPTIONS="--dir $KBUILD_EXTMOD $COCCIINCLUDE"
    237     fi
    238 
    239 KBUILD_EXTMOD is set when an explicit target with M= is used. For both cases
    240 the spatch --dir argument is used, as such third rule applies when whether M=
    241 is used or not, and when M= is used the target directory can have its own
    242 .cocciconfig file. When M= is not passed as an argument to coccicheck the
    243 target directory is the same as the directory from where spatch was called.
    244 
    245 If not using the kernel's coccicheck target, keep the above precedence
    246 order logic of .cocciconfig reading. If using the kernel's coccicheck target,
    247 override any of the kernel's .coccicheck's settings using SPFLAGS.
    248 
    249 We help Coccinelle when used against Linux with a set of sensible defaults
    250 options for Linux with our own Linux .cocciconfig. This hints to coccinelle
    251 git can be used for ``git grep`` queries over coccigrep. A timeout of 200
    252 seconds should suffice for now.
    253 
    254 The options picked up by coccinelle when reading a .cocciconfig do not appear
    255 as arguments to spatch processes running on your system, to confirm what
    256 options will be used by Coccinelle run::
    257 
    258       spatch --print-options-only
    259 
    260 You can override with your own preferred index option by using SPFLAGS. Take
    261 note that when there are conflicting options Coccinelle takes precedence for
    262 the last options passed. Using .cocciconfig is possible to use idutils, however
    263 given the order of precedence followed by Coccinelle, since the kernel now
    264 carries its own .cocciconfig, you will need to use SPFLAGS to use idutils if
    265 desired. See below section "Additional flags" for more details on how to use
    266 idutils.
    267 
    268 Additional flags
    269 ----------------
    270 
    271 Additional flags can be passed to spatch through the SPFLAGS
    272 variable. This works as Coccinelle respects the last flags
    273 given to it when options are in conflict. ::
    274 
    275     make SPFLAGS=--use-glimpse coccicheck
    276 
    277 Coccinelle supports idutils as well but requires coccinelle >= 1.0.6.
    278 When no ID file is specified coccinelle assumes your ID database file
    279 is in the file .id-utils.index on the top level of the kernel, coccinelle
    280 carries a script scripts/idutils_index.sh which creates the database with::
    281 
    282     mkid -i C --output .id-utils.index
    283 
    284 If you have another database filename you can also just symlink with this
    285 name. ::
    286 
    287     make SPFLAGS=--use-idutils coccicheck
    288 
    289 Alternatively you can specify the database filename explicitly, for
    290 instance::
    291 
    292     make SPFLAGS="--use-idutils /full-path/to/ID" coccicheck
    293 
    294 See ``spatch --help`` to learn more about spatch options.
    295 
    296 Note that the ``--use-glimpse`` and ``--use-idutils`` options
    297 require external tools for indexing the code. None of them is
    298 thus active by default. However, by indexing the code with
    299 one of these tools, and according to the cocci file used,
    300 spatch could proceed the entire code base more quickly.
    301 
    302 SmPL patch specific options
    303 ---------------------------
    304 
    305 SmPL patches can have their own requirements for options passed
    306 to Coccinelle. SmPL patch specific options can be provided by
    307 providing them at the top of the SmPL patch, for instance::
    308 
    309 	// Options: --no-includes --include-headers
    310 
    311 SmPL patch Coccinelle requirements
    312 ----------------------------------
    313 
    314 As Coccinelle features get added some more advanced SmPL patches
    315 may require newer versions of Coccinelle. If an SmPL patch requires
    316 at least a version of Coccinelle, this can be specified as follows,
    317 as an example if requiring at least Coccinelle >= 1.0.5::
    318 
    319 	// Requires: 1.0.5
    320 
    321 Proposing new semantic patches
    322 -------------------------------
    323 
    324 New semantic patches can be proposed and submitted by kernel
    325 developers. For sake of clarity, they should be organized in the
    326 sub-directories of ``scripts/coccinelle/``.
    327 
    328 
    329 Detailed description of the ``report`` mode
    330 -------------------------------------------
    331 
    332 ``report`` generates a list in the following format::
    333 
    334   file:line:column-column: message
    335 
    336 Example
    337 ~~~~~~~
    338 
    339 Running::
    340 
    341 	make coccicheck MODE=report COCCI=scripts/coccinelle/api/err_cast.cocci
    342 
    343 will execute the following part of the SmPL script::
    344 
    345    <smpl>
    346    @r depends on !context && !patch && (org || report)@
    347    expression x;
    348    position p;
    349    @@
    350 
    351      ERR_PTR@p(PTR_ERR(x))
    352 
    353    @script:python depends on report@
    354    p << r.p;
    355    x << r.x;
    356    @@
    357 
    358    msg="ERR_CAST can be used with %s" % (x)
    359    coccilib.report.print_report(p[0], msg)
    360    </smpl>
    361 
    362 This SmPL excerpt generates entries on the standard output, as
    363 illustrated below::
    364 
    365     /home/user/linux/crypto/ctr.c:188:9-16: ERR_CAST can be used with alg
    366     /home/user/linux/crypto/authenc.c:619:9-16: ERR_CAST can be used with auth
    367     /home/user/linux/crypto/xts.c:227:9-16: ERR_CAST can be used with alg
    368 
    369 
    370 Detailed description of the ``patch`` mode
    371 ------------------------------------------
    372 
    373 When the ``patch`` mode is available, it proposes a fix for each problem
    374 identified.
    375 
    376 Example
    377 ~~~~~~~
    378 
    379 Running::
    380 
    381 	make coccicheck MODE=patch COCCI=scripts/coccinelle/api/err_cast.cocci
    382 
    383 will execute the following part of the SmPL script::
    384 
    385     <smpl>
    386     @ depends on !context && patch && !org && !report @
    387     expression x;
    388     @@
    389 
    390     - ERR_PTR(PTR_ERR(x))
    391     + ERR_CAST(x)
    392     </smpl>
    393 
    394 This SmPL excerpt generates patch hunks on the standard output, as
    395 illustrated below::
    396 
    397     diff -u -p a/crypto/ctr.c b/crypto/ctr.c
    398     --- a/crypto/ctr.c 2010-05-26 10:49:38.000000000 +0200
    399     +++ b/crypto/ctr.c 2010-06-03 23:44:49.000000000 +0200
    400     @@ -185,7 +185,7 @@ static struct crypto_instance *crypto_ct
    401  	alg = crypto_attr_alg(tb[1], CRYPTO_ALG_TYPE_CIPHER,
    402  				  CRYPTO_ALG_TYPE_MASK);
    403  	if (IS_ERR(alg))
    404     -		return ERR_PTR(PTR_ERR(alg));
    405     +		return ERR_CAST(alg);
    406 
    407  	/* Block size must be >= 4 bytes. */
    408  	err = -EINVAL;
    409 
    410 Detailed description of the ``context`` mode
    411 --------------------------------------------
    412 
    413 ``context`` highlights lines of interest and their context
    414 in a diff-like style.
    415 
    416       **NOTE**: The diff-like output generated is NOT an applicable patch. The
    417       intent of the ``context`` mode is to highlight the important lines
    418       (annotated with minus, ``-``) and gives some surrounding context
    419       lines around. This output can be used with the diff mode of
    420       Emacs to review the code.
    421 
    422 Example
    423 ~~~~~~~
    424 
    425 Running::
    426 
    427 	make coccicheck MODE=context COCCI=scripts/coccinelle/api/err_cast.cocci
    428 
    429 will execute the following part of the SmPL script::
    430 
    431     <smpl>
    432     @ depends on context && !patch && !org && !report@
    433     expression x;
    434     @@
    435 
    436     * ERR_PTR(PTR_ERR(x))
    437     </smpl>
    438 
    439 This SmPL excerpt generates diff hunks on the standard output, as
    440 illustrated below::
    441 
    442     diff -u -p /home/user/linux/crypto/ctr.c /tmp/nothing
    443     --- /home/user/linux/crypto/ctr.c	2010-05-26 10:49:38.000000000 +0200
    444     +++ /tmp/nothing
    445     @@ -185,7 +185,6 @@ static struct crypto_instance *crypto_ct
    446  	alg = crypto_attr_alg(tb[1], CRYPTO_ALG_TYPE_CIPHER,
    447  				  CRYPTO_ALG_TYPE_MASK);
    448  	if (IS_ERR(alg))
    449     -		return ERR_PTR(PTR_ERR(alg));
    450 
    451  	/* Block size must be >= 4 bytes. */
    452  	err = -EINVAL;
    453 
    454 Detailed description of the ``org`` mode
    455 ----------------------------------------
    456 
    457 ``org`` generates a report in the Org mode format of Emacs.
    458 
    459 Example
    460 ~~~~~~~
    461 
    462 Running::
    463 
    464 	make coccicheck MODE=org COCCI=scripts/coccinelle/api/err_cast.cocci
    465 
    466 will execute the following part of the SmPL script::
    467 
    468     <smpl>
    469     @r depends on !context && !patch && (org || report)@
    470     expression x;
    471     position p;
    472     @@
    473 
    474       ERR_PTR@p(PTR_ERR(x))
    475 
    476     @script:python depends on org@
    477     p << r.p;
    478     x << r.x;
    479     @@
    480 
    481     msg="ERR_CAST can be used with %s" % (x)
    482     msg_safe=msg.replace("[","@(").replace("]",")")
    483     coccilib.org.print_todo(p[0], msg_safe)
    484     </smpl>
    485 
    486 This SmPL excerpt generates Org entries on the standard output, as
    487 illustrated below::
    488 
    489     * TODO [[view:/home/user/linux/crypto/ctr.c::face=ovl-face1::linb=188::colb=9::cole=16][ERR_CAST can be used with alg]]
    490     * TODO [[view:/home/user/linux/crypto/authenc.c::face=ovl-face1::linb=619::colb=9::cole=16][ERR_CAST can be used with auth]]
    491     * TODO [[view:/home/user/linux/crypto/xts.c::face=ovl-face1::linb=227::colb=9::cole=16][ERR_CAST can be used with alg]]
    492