| POOL(9) | Kernel Developer's Manual | POOL(9) |
pool_init,
pool_destroy, pool_get,
pool_put, pool_prime,
pool_sethiwat,
pool_setlowat,
pool_sethardlimit —
resource-pool manager
#include
<sys/pool.h>
void
pool_init(struct pool *pp,
size_t size, u_int align,
u_int align_offset, int flags,
const char *wchan, struct
pool_allocator *palloc, int ipl);
void
pool_destroy(struct
pool *pp);
void *
pool_get(struct
pool *pp, int
flags);
void
pool_put(struct
pool *pp, void
*item);
void
pool_prime(struct
pool *pp, int
n);
void
pool_sethiwat(struct
pool *pp, int
n);
void
pool_setlowat(struct
pool *pp, int
n);
void
pool_sethardlimit(struct
pool *pp, int n,
const char *warnmess,
int ratecap);
These utility routines provide management of pools of fixed-sized areas of memory. Resource pools set aside an amount of memory for exclusive use by the resource pool owner. This can be used by applications to guarantee the availability of a minimum amount of memory needed to continue operation independent of the memory resources currently available from the system-wide memory allocator (malloc(9)).
The function
pool_init()
initializes a resource pool. The arguments are:
pool_get().
This argument must be a power of two. If zero, the alignment defaults to
an architecture-specific natural alignment.PR_NOTOUCH, or
PR_PSERIALIZE. If
PR_NOTOUCH is given, free items are never used to
keep internal state so that the pool can be used for non memory backed
objects. If PR_PSERIALIZE is given, then the
allocator guarantees that a passive serialization barrier equivalent to
“xc_barrier(0)” will be performed before the object's
backing store is returned to the system.
PR_PSERIALIZE implies
PR_NOTOUCH. Because of the guarantees provided by
PR_PSERIALIZE, objects muste never be freed to a
pool using this option from either hard or soft interrupt context, as
doing so may block.pool_get() must wait for items to be returned to
the pool.NULL or
pool_allocator_kmem, in which case the default
kernel memory allocator will be used. It can also be set to
pool_allocator_nointr when the pool will never be
accessed from interrupt context.The function pool_destroy() destroys a
resource pool. It takes a single argument pp
identifying the pool resource instance.
pool_get() allocates an item from the pool
and returns a pointer to it. The arguments are:
PR_NOWAIT is given,
pool_get() returns NULL.
If PR_WAITOK is given and allocation is attempted
with no resources available, the function will sleep until items are
returned to the pool. If both PR_LIMITFAIL and
PR_WAITOK are specified, and the pool has reached
its hard limit, pool_get() will return
NULL without waiting, allowing the caller to do
its own garbage collection; however, it will still wait if the pool is not
yet at its hard limit. If the PR_ZERO flag is
specified, then the memory returned will be zeroed first using
memset(3).pool_put()
returns the pool item pointed at by item to the
resource pool identified by the pool handle pp. If the
number of available items in the pool exceeds the maximum pool size set by
pool_sethiwat() and there are no outstanding
requests for pool items, the excess items will be returned to the system.
The arguments to pool_put() are:
pool_get().A pool will attempt to increase its resource usage to keep up with the demand for its items. Conversely, it will return unused memory to the system should the number of accumulated unused items in the pool exceed a programmable limit.
The targets for the minimum and maximum number of
free items which a pool should try to keep available are known as the high
and low
watermarks.
The functions
pool_sethiwat()
and pool_setlowat() set a pool's high and low
watermarks, respectively.
The limits for the minimum and maximum number of
total items (both free and allocated) that the pool can have at any time are
specified by the functions
pool_prime()
and pool_sethardlimit(), respectively. The defaults
for these limits are 0 and
UINT_MAX, respectively. Changing these limits will
cause memory to be immediately allocated to the pool or freed from the pool
as needed.
Note that undefined behaviour results when mixing the storage providing methods supported by the pool resource routines.
The pool resource code uses a per-pool lock to
protect its internal state. If any pool functions are called in an interrupt
context, the caller must block all interrupts that might cause the code to
be reentered. Additionally, the functions
pool_init()
and
pool_destroy()
should never be called in interrupt context.
Pool usage logs can be enabled by defining the compile-time option
POOL_DIAGNOSTIC.
The pool manager is implemented in the file sys/kern/subr_pool.c.
free(9), malloc(9), memoryallocators(9), pool_cache(9), uvm(9)
The NetBSD pool manager appeared in NetBSD 1.4.
| April 12, 2020 | NetBSD 11.0 |