Go back –> Main Linux bot tests page
'coccicheck' is one of the targets of the Linux kernel, this page documents how you can use it, what it provides and how to help extend it.
'coccicheck' provides a series of semantic patches written in Semantic Patch Language (SmPL) and makes use of the Coccinelle engine to interpret and complete these tests. These tests can help you vet for correctness and avoid certain bugs which have been able to be expressed semantically.
If you're a developer you are encouraged to vet your patch using cocciecheck to ensure you are not introducing issues which are known semantically. If you're a maintainer you are encouraged to be using coccicheck with M=your/maintained-code-path so you can make use of the SmPL patches to avoid known bugs in your maintained code. How maintainers can use this is explained below. Automatic code bot setups which test patches for inclusion upstream are encouraged to consider using 'make coccicheck' as well.
SmPL patches exist for a few different types of categories, documented below.
Eeach SmPL patch has a
'confidence' tag associated with it, to help highlight how confident the devlelopers of the SmPL patch are that the reports / fixes generated by the SmPL patch will be accurate and/or relevant. If an SmPL patch has a
'High' confidence annotated on it, it means that if you use coccicheck with it and it generates a report for a file there is a high degree of confidence that the report should be valid. Likewise if the confidence annotated is
'Low' you should take good care to review the report/patch generated and vet the validity of it before sending it to maintainers.
There are four different modes you can use to use the semantic patches:
'default' mode if one is not specified.
You specify the mode you want to operate. Note that not all semantic patches implement all modes. For easy use of Coccinelle, the default mode is “report”.
Two other modes provide some common combinations of these modes.
By default all SmPL files will be used to run tests using the
'report' mode. To run all tests you can run:
make coccicheck MODE=report
To produce patches for the entire kernel, run:
make coccicheck MODE=patch
For each semantic patch, a commit message is proposed. It gives a description of the problem being checked by the semantic patch, and includes a reference to Coccinelle.
As any static code analyzer, Coccinelle produces false positives. Thus, reports must be carefully checked, and patches reviewed.
If you are a maintainer you may want to use M=your/maintained-path/ to limit the scope to only the files you maintain. For instance the following will generate a report for all files in drivers/net/ only.
make coccicheck MODE=report M=drivers/net/
To only run a specific cocci test or a set of tests set the COCCI environment variable with the list of tests you want to run. For instance to only run the scripts/coccinelle/locks/double_lock.cocci you would use:
make coccicheck COCCI=scripts/coccinelle/locks/double_lock.cocci MODE=report
make coccicheck COCCI=scripts/coccinelle/locks/double_lock.cocci MODE=patch
By default the entire kernel source tree is checked. To apply Coccinelle to a specific directory, M= can be used. For example, to check drivers/net/wireless/ one may write:
make coccicheck MODE=report M=drivers/net/wireless/
To apply Coccinelle on a file basis, instead of a directory basis, the following command may be used:
make C=1 CHECK="scripts/coccicheck"
To check only newly edited code, use the value 2 for the C flag, i.e.
make C=2 CHECK="scripts/coccicheck"
In these modes, which work on a file basis, there is no information about semantic patches displayed, and no commit message proposed.
This runs every semantic patch in scripts/coccinelle by default. The COCCI variable may additionally be used to only apply a single semantic patch as shown in the previous section.
The “report” mode is the default. You can select another one with the MODE variable explained above.
Additional flags can be passed to spatch through the SPFLAGS variable.
make SPFLAGS=--use-idutils coccicheck
See spatch –help to learn more about spatch options.
New semantic patches can be proposed and submitted by kernel developers. For sake of clarity, they should be organized in the sub-directories of 'scripts/coccinelle/'.
Below we document each type of test category which as been devised.
Help vet for correctness when using certain APIs. The following tests exist:
Remove casting the values returned by memory allocation functions like kmalloc, kzalloc, kmem_cache_alloc, kmem_cache_zalloc etc. This makes an effort to find cases of casting of values returned by kmalloc, kzalloc, kcalloc, kmem_cache_alloc, kmem_cache_zalloc, kmem_cache_alloc_node, kmalloc_node and kzalloc_node and removes the casting as it is not required. The result in the patch case may need some reformatting.
Use zeroing memory allocator function rather than allocator followed by memset with 0. This considers some simple cases that are common and easy to validate. Note in particular that there are no …s in the rule, so all of the matched code has to be contiguous.
Avoid freeing bugs.
Help vet for correctness when using kernel iterators.
Vet for correct locking behaviour and avoid bugs.
Things not covered in any of the other categories.
Avoid null bugs.
Certain tests which are known to help with consistent kernel use.
Linux kernel maintainers are encouraged to make use of the coccicheck to vet for issues on their tree. This can be used to help review quality of patches before they get merged into your tree as well.
These days coccicheck is run every now and then by Julia Lawall, false positives are reviewed (in particular cocci files where confidence is annotated as low), and then once issues are confirmed patches typically are sent to maintainers.
Average developers can and should use coccicheck as well but care should be taken to ensure the
'Confidence' tag is reviewed and proper diligence is done before sending any reports / fixes to kernel maintainers.
It seems the zero day bot test infrastructure also makes use of 'make coccicheck' — this section needs to be extended to document in what capacity this is done, how often, etc.
The point of merging these SmPL patches upstream was to enable developers and maintainers to vet code themselves. Ideally developers would run 'make coccicheck M=path/to-their-code/' prior to submitting a patch upstream, and maintainers would use it regularly, perhaps after applying new patches, to avoid introducing new issues.
Time permitting Julia will also run these tests on her own when possible as linux-next moves on. There is no specific regular interval in which these tests are run. The zero day bot test infrastructure also seems to use 'make coccicheck' – but this section should be extended to document this.