malloc(), alloca(), calloc(), free(), mallinfo(), mallopt(), memorymap(), realloc(), valloc() - main memory allocator
System V Synopsis
The functionality in the old package has been incorporated into The library corresponding to the linker option is now an empty library on
PA-RISC based systems only; it does not exist on Makefiles that reference this library on PA-RISC based systems will continue to work, but
they will not work on
has been deprecated at HP-UX 11i Version 1 and is now obsolete. The function is more useful for statistics.
The functions described here provide a simple, general purpose memory allocation package:
Allocates space for a block of at least
size bytes, but does not initialize the space.
Allocates space for an array of
nelem elements, each of size elsize bytes, and initializes the space to zeros. Actual amount of space allocated will
be at least nelem * elsize bytes.
Changes the size of the block pointed to by
ptr to size bytes and returns a pointer to the (possibly moved) block. Existing contents are unchanged up to the
lesser of the new and old sizes. If ptr is a NULL pointer, behaves like for the specified size. If size is zero and
ptr is not a NULL pointer, the object it points to is freed and NULL is returned. of blocks with special alignments,
such as those created by is not supported.
Allocates space for a block of at least
size bytes starting on a boundary aligned to a multiple of the value returned by (__SC_PAGESIZE). This space is
Deallocates the space pointed to by
ptr (a pointer to a block previously allocated by or and makes the space available for further allocation. If ptr is
a NULL pointer, no action occurs.
Provides for control over the allocation algorithm
and other options in the package. The available values for cmd are:
Set maxfast to value. The algorithm allocates all blocks below the size of maxfast in large groups, then
doles them out very quickly. The default value for maxfast is zero.
Set numlblks to value. The above mentioned ``large groups'' each contain numlblks blocks. numlblks must
be greater than 1. The default value for numlblks is
Set grain to value. The sizes of all blocks smaller than maxfast are considered to be rounded up to the
nearest multiple of grain. grain must be greater than zero. The default value of grain is the small-
est number of bytes that can accommodate alignment of any data type. value is rounded up to a multi-
ple of the default when grain is set.
Block all blockable signals in
and This option is provided for those who need to write signal handlers that allocate memory. When
set, the routines can be called from within signal handlers (they become re-entrant). Default action
is to block all blockable signals.
NOTE: performance of will be impacted considerably when the option is set.
Do not block all blockable signals in
and This option cancels signal blocking initiated by the option.
Enable the feature of
This option is provided for 32-bit applications with one arena. Release of last block happens only
when size of the free block is greater than the threshold.
The threshold value is defined as The feature is disabled if the value is greater than or equal to
threshold value. Default action is not to release last block.
These values are defined in the header file.
can be called repeatedly; but once the first small block is allocated, it is not possible to change the and values.
Provides instrumentation describing space usage,
but cannot be called until the first small block is allocated. It returns the structure
The structure is defined in the header file.
Each of the allocation routines returns a pointer to space suitably aligned (after possible pointer coercion) for storage of any type of
Displays the contents of the memory allocator
for HP-UX 32-bit operating systems only. A list of addresses and block descriptions is written (using to standard
output. If the value of the show_stats parameter is 1, statistics concerning number of blocks and sizes used will
also be written. If the value is zero, only the memory map will be written.
The addresses and sizes displayed by may not correspond to those requested by an application. The size of a block
(as viewed by the allocator) includes header information and padding to properly align the block. The address is
also offset by a certain amount to accommodate the header information.
has been deprecated at HP-UX 11i Version 1 and is now obsolete.
Allocates space from the stack of the caller
for a block of at least size bytes, but does not initialize the space. The space is automatically freed when the
calling routine exits.
Memory returned by is not related to memory allocated by other memory allocation functions. Behavior of addresses
returned by as parameters to other memory functions is undefined.
The implementation of this routine is system dependent and its use is discouraged.
Upon successful completion, and return a pointer to space suitably aligned (after possible pointer coercion) for storage of any type of
object. Otherwise, they return a NULL pointer. If returns a NULL pointer, the memory pointed to by the original pointer is left intact.
returns zero for success and nonzero for failure.
and return a NULL pointer if there is no available memory, or if the memory managed by has been detectably corrupted. This memory may
become corrupted if data is stored outside the bounds of a block, or if an invalid pointer (a pointer not generated by or is passed as an
argument to or
If is called after any allocation of a small block and cmd is not set to or or if cmd or value is invalid, nonzero is returned. Otherwise,
it returns zero.
and set to and return a NULL pointer when an out-of-memory condition arises.
and set to and return a NULL pointer when the memory being managed by has been detectably corrupted.
For PA-RISC based systems, the performance of the family of APIs can be tuned via the environment variables and
For systems, in addition to and three global variables can be used for performance tuning: and
For threaded applications, uses multiple arenas. Memory requests from different threads are handled by different arenas. can be used to
adjust the number of arenas and how many pages each time an arena expands itself (the expansion factor), assuming that the page size is
4096 bytes. In general, the more threads in an application, the more arenas should be used for better performance. The number of arenas
can be from 1 to 64 for threaded applications.
For non-threaded applications, only one arena is used. If the environment variable is not set, or the number of arenas is set to be out of
the range, the default number of 8 will be used. The expansion factor is from 1 to 4096, default value is 32. Again, if the factor is out
of the range, the default value will be used.
Here is an example of how to use
This means that the number of arenas is 16, and the expansion size is 8*4096 bytes. In general, the more arenas you use, the smaller the
expansion factor should be, and vice versa.
is used to turn on the small block allocator, and to set up parameters for the small block allocator, namely, maxfast, grain, and numlblks.
Applications with small block allocator turned on usually run faster than with it turned off. Small block allocator can be turned on
through however, it is not early enough for C++/Java applications. The environment variable turns it on before the application starts.
The call can still be used the same way. If the environment variable is set, and no small block allocator has been used, the subsequent
calls can still overwrite whatever is set through If the environment variable is set, and small block allocator has been used, then will
have no effect.
To use this environment variable,
This means that the maxfast size is 512, the number of small blocks is 100, and the grain size is 16. You have to supply all 3 values, and
in that order. If not, the default values will be used instead.
Three new global variables, and are introduced for Itanium-based systems to over-ride the environment option. When these three variables
are initialized within an application, has no effect. This way, a finely tuned application can lock in performance across different user
environments. However, as a subsequent call to before any block of memory was allocated will change the behavior. By default, these three
variables will be initialized to zero at start up. It is the same as setting them to
By default, SBA (Small Block Allocation) is turned on for Itanium-based systems. This may contribute to better application performance. A
user can set
This will turn off SBA.
For all other possible values, please refer to
is used to turn on the thread local cache. Turning this option on sets up a private cache for each thread to which access is effectively
non-threaded, so there is less contention on the arenas. For some multi-threaded applications this can give a significant performance
The thread local cache saves blocks of sizes that have previously been used, and thus may be requested again. The size of the cache is
configurable. The cache is organized in buckets of sizes that are powers of two; that is, there is a bucket for all blocks in the size
range 64-127 bytes, another for 128-255 bytes, and so on.
Thread Local Cache can be tuned by setting the environment variable as follows:
The values must be supplied in the exact order indicated. The first three parameters are mandatory, and the last two are optional.
denotes the number of pointers cached per bucket. If is 0, then thread local cache is disabled. The maximum value for is 32768.
is an indication of the number of buckets. The maximum block size that will be cached is can range between 8 and 32.
is an indication in minutes of after how long blocks in an unused cache will be released to the arena. is only a hint, so caches may or
may not be retired after the specified time period. If is 0, retirement is disabled. The maximum value for is 1440 (that is, 24 hours).
enables cached blocks to be exchanged among threads. A thread that heavily allocates blocks of a certain size is soon bound to run out of
blocks in its private cache. The exchange allows the thread to borrow blocks from a global pool if available. This may be more efficient
than going back to the arena. is a hint to the caching algorithm indicating a particular number of cache misses after which it will search
the global pool for appropriate sized blocks. The algorithm also releases unused cache blocks to the global pool. If is 0 or not set,
cache exchange is turned off.
is an indication of the size of the global cache pool. This parameter is valid only if is turned on. The default value for is 8.
The maximum number of blocks that will be cached for each thread is
Here are examples of how to use
This means that is 1024, is 32, is 20 minutes, is 4, and is 8.
This is a valid configuration where the global pool is not activated.
and have no effect on non-threaded applications, while has.
NOTE: Modifying these variables increases the chances of surfacing existing user memory defects such as buffer overrun.
functions use and (see brk(2)) to increase the address space of a process. Therefore, an application program that uses or must not use
them to decrease the address space, because this confuses the functions.
and do not check their pointer argument for validity.
The following actions are considered bad programming practices and should not be done. The results are unpredictable, probably undesirable
and not supported. Examples of undesirable results are loss of data, memory fault, bus error or infinite loop.
o Attempting to or a pointer not generated as the result of a call to or
o Reading or writing data outside the boundaries of an allocated block.
o Attempting to an aligned block such as the result of
The following actions are strongly discouraged and may be unsupported in a future version of
o Attempting to the same block twice.
o Depending on unmodified contents of a block after it has been freed.
o Attempting to a block after it is freed.
Undocumented features of earlier memory allocators have not been duplicated. Applications which used any of the above bad programming
practices or discouraged practices are not guaranteed to continue functioning at future releases.
The only external difference between the old allocator and the allocator is that the old allocator would return a NULL pointer for a
request of zero bytes. The allocator returns a valid memory address. This is not a concern for most applications.
By default, SBA (Small Block Allocation) is turned on for Itanium-based systems and is turned off for PA-RISC systems. This was due to
performance concerns. Please refer to the section of this manpage for details.
brk(2), errno(2), thread_safety(5).