1Kernel Memory Leak Detector
7Kmemleak provides a way of detecting possible kernel memory leaks in a
8way similar to a tracing garbage collector
10with the difference that the orphan objects are not freed but only
11reported via /sys/kernel/debug/kmemleak. A similar method is used by the
12Valgrind tool (memcheck --leak-check) to detect the memory leaks in
13user-space applications.
15Please check DEBUG_KMEMLEAK dependencies in lib/Kconfig.debug for supported
21CONFIG_DEBUG_KMEMLEAK in "Kernel hacking" has to be enabled. A kernel
22thread scans the memory every 10 minutes (by default) and prints the
23number of new unreferenced objects found. To display the details of all
24the possible memory leaks:
26  # mount -t debugfs nodev /sys/kernel/debug/
27  # cat /sys/kernel/debug/kmemleak
29To trigger an intermediate memory scan:
31  # echo scan > /sys/kernel/debug/kmemleak
33To clear the list of all current possible memory leaks:
35  # echo clear > /sys/kernel/debug/kmemleak
37New leaks will then come up upon reading /sys/kernel/debug/kmemleak
40Note that the orphan objects are listed in the order they were allocated
41and one object at the beginning of the list may cause other subsequent
42objects to be reported as orphan.
44Memory scanning parameters can be modified at run-time by writing to the
45/sys/kernel/debug/kmemleak file. The following parameters are supported:
47  off - disable kmemleak (irreversible)
48  stack=on - enable the task stacks scanning (default)
49  stack=off - disable the tasks stacks scanning
50  scan=on - start the automatic memory scanning thread (default)
51  scan=off - stop the automatic memory scanning thread
52  scan=<secs> - set the automatic memory scanning period in seconds
53          (default 600, 0 to stop the automatic scanning)
54  scan - trigger a memory scan
55  clear - clear list of current memory leak suspects, done by
56          marking all current reported unreferenced objects grey
57  dump=<addr> - dump information about the object found at <addr>
59Kmemleak can also be disabled at boot-time by passing "kmemleak=off" on
60the kernel command line.
62Memory may be allocated or freed before kmemleak is initialised and
63these actions are stored in an early log buffer. The size of this buffer
64is configured via the CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE option.
66Basic Algorithm
69The memory allocations via kmalloc, vmalloc, kmem_cache_alloc and
70friends are traced and the pointers, together with additional
71information like size and stack trace, are stored in a prio search tree.
72The corresponding freeing function calls are tracked and the pointers
73removed from the kmemleak data structures.
75An allocated block of memory is considered orphan if no pointer to its
76start address or to any location inside the block can be found by
77scanning the memory (including saved registers). This means that there
78might be no way for the kernel to pass the address of the allocated
79block to a freeing function and therefore the block is considered a
80memory leak.
82The scanning algorithm steps:
84  1. mark all objects as white (remaining white objects will later be
85     considered orphan)
86  2. scan the memory starting with the data section and stacks, checking
87     the values against the addresses stored in the prio search tree. If
88     a pointer to a white object is found, the object is added to the
89     gray list
90  3. scan the gray objects for matching addresses (some white objects
91     can become gray and added at the end of the gray list) until the
92     gray set is finished
93  4. the remaining white objects are considered orphan and reported via
94     /sys/kernel/debug/kmemleak
96Some allocated memory blocks have pointers stored in the kernel's
97internal data structures and they cannot be detected as orphans. To
98avoid this, kmemleak can also store the number of values pointing to an
99address inside the block address range that need to be found so that the
100block is not considered a leak. One example is __vmalloc().
102Testing specific sections with kmemleak
105Upon initial bootup your /sys/kernel/debug/kmemleak output page may be
106quite extensive. This can also be the case if you have very buggy code
107when doing development. To work around these situations you can use the
108'clear' command to clear all reported unreferenced objects from the
109/sys/kernel/debug/kmemleak output. By issuing a 'scan' after a 'clear'
110you can find new unreferenced objects; this should help with testing
111specific sections of code.
113To test a critical section on demand with a clean kmemleak do:
115  # echo clear > /sys/kernel/debug/kmemleak
116  ... test your kernel or modules ...
117  # echo scan > /sys/kernel/debug/kmemleak
119Then as usual to get your report with:
121  # cat /sys/kernel/debug/kmemleak
123Kmemleak API
126See the include/linux/kmemleak.h header for the functions prototype.
128kmemleak_init - initialize kmemleak
129kmemleak_alloc - notify of a memory block allocation
130kmemleak_free - notify of a memory block freeing
131kmemleak_not_leak - mark an object as not a leak
132kmemleak_ignore - do not scan or report an object as leak
133kmemleak_scan_area - add scan areas inside a memory block
134kmemleak_no_scan - do not scan a memory block
135kmemleak_erase - erase an old value in a pointer variable
136kmemleak_alloc_recursive - as kmemleak_alloc but checks the recursiveness
137kmemleak_free_recursive - as kmemleak_free but checks the recursiveness
139Dealing with false positives/negatives
142The false negatives are real memory leaks (orphan objects) but not
143reported by kmemleak because values found during the memory scanning
144point to such objects. To reduce the number of false negatives, kmemleak
145provides the kmemleak_ignore, kmemleak_scan_area, kmemleak_no_scan and
146kmemleak_erase functions (see above). The task stacks also increase the
147amount of false negatives and their scanning is not enabled by default.
149The false positives are objects wrongly reported as being memory leaks
150(orphan). For objects known not to be leaks, kmemleak provides the
151kmemleak_not_leak function. The kmemleak_ignore could also be used if
152the memory block is known not to contain other pointers and it will no
153longer be scanned.
155Some of the reported leaks are only transient, especially on SMP
156systems, because of pointers temporarily stored in CPU registers or
157stacks. Kmemleak defines MSECS_MIN_AGE (defaulting to 1000) representing
158the minimum age of an object to be reported as a memory leak.
160Limitations and Drawbacks
163The main drawback is the reduced performance of memory allocation and
164freeing. To avoid other penalties, the memory scanning is only performed
165when the /sys/kernel/debug/kmemleak file is read. Anyway, this tool is
166intended for debugging purposes where the performance might not be the
167most important requirement.
169To keep the algorithm simple, kmemleak scans for values pointing to any
170address inside a block's address range. This may lead to an increased
171number of false negatives. However, it is likely that a real memory leak
172will eventually become visible.
174Another source of false negatives is the data stored in non-pointer
175values. In a future version, kmemleak could only scan the pointer
176members in the allocated structures. This feature would solve many of
177the false negative cases described above.
179The tool can report false positives. These are cases where an allocated
180block doesn't need to be freed (some cases in the init_call functions),
181the pointer is calculated by other methods than the usual container_of
182macro or the pointer is stored in a location not scanned by kmemleak.
184Page allocations and ioremap are not tracked.

Archive Download this file