Ravenbrook / Projects / Memory Pool System / Version 1.110 Product Sources / Product Manuals
This is the reference manual for the Memory Pool System.
This document is quite incomplete. At present it consists simply of reference descriptions of a number of MPS symbols (section 3). Many MPS symbols are not described here (see section 4 for a list). There are also no overview or protocol-oriented sections.
mps_fix MPS_FIX1 MPS_FIX12 MPS_FIX2 MPS_FIX_CALL MPS_RES_LIMIT MPS_RES_MEMORY MPS_RES_PARAM MPS_RM_CONST MPS_RM_PROT mps_sac_alloc MPS_SAC_ALLOC_FAST MPS_SAC_CLASS_LIMIT mps_sac_free MPS_SAC_FREE_FAST MPS_SCAN_BEGIN MPS_SCAN_END MPS_T_WORD MPS_WORD_SHIFT MPS_WORD_WIDTH mps_addr_t mps_align_t mps_alloc mps_alloc_pattern_ramp mps_alloc_pattern_ramp_collect_all mps_amc_apply mps_ap_alloc_pattern_begin mps_ap_alloc_pattern_end mps_ap_alloc_pattern_reset mps_ap_frame_pop mps_ap_frame_push mps_arena_clamp mps_arena_class_cl mps_arena_class_t mps_arena_class_vm mps_arena_class_vmnz mps_arena_collect mps_arena_commit_limit mps_arena_commit_limit_set mps_arena_committed mps_arena_create mps_arena_create_v mps_arena_expose mps_arena_formatted_objects_walk mps_arena_park mps_arena_release mps_arena_roots_walk mps_arena_spare_commit_limit mps_arena_spare_commit_limit_set mps_arena_spare_committed mps_arena_unsafe_expose_remember_protection mps_arena_unsafe_restore_protection mps_bool_t mps_class_amc mps_class_mvff mps_class_snc mps_class_mvt mps_class_t mps_finalize mps_fmt_A_s mps_fmt_A_t mps_fmt_B_s mps_fmt_B_t mps_fmt_auto_header_s mps_fmt_class_t mps_fmt_copy_t mps_fmt_create_A mps_fmt_create_B mps_fmt_create_auto_header mps_fmt_fwd_t mps_fmt_isfwd_t mps_fmt_pad_t mps_fmt_scan_t mps_fmt_skip_t mps_fmt_t mps_formatted_objects_stepper_t mps_free mps_lib_memcmp mps_lib_memcpy mps_lib_memset mps_lib_telemetry_control mps_message_clock mps_message_discard mps_message_finalization_ref mps_message_gc_condemned_size mps_message_gc_live_size mps_message_gc_not_condemned_size mps_message_gc_start_why mps_message_get mps_message_poll mps_message_queue_type mps_message_t mps_message_type mps_message_type_disable mps_message_type_enable mps_message_type_finalization mps_message_type_gc mps_message_type_gc_start mps_message_type_t mps_pool_check_fenceposts mps_pool_debug_option_s mps_rank_ambig mps_rank_exact mps_rank_t mps_rank_weak mps_reg_scan_t mps_res_t mps_root_create mps_root_create_fmt mps_root_create_reg mps_root_create_table mps_root_create_table_masked mps_root_scan_t mps_roots_stepper_t mps_sac_class_s mps_sac_create mps_sac_destroy mps_sac_flush mps_sac_t mps_stack_scan_ambig mps_telemetry_control mps_telemetry_flush mps_telemetry_intern mps_telemetry_label mps_thr_t mps_fixThe function mps_fix is the
part of the scanning protocol used to indicate references to the
MPS. It may only be called from within a scanning function.
Scanning.
mps_res_t mps_fix(mps_ss_t mps_ss, mps_addr_t *ref_io)
mps_ss the scan state argument that was passed to the
scanning function
ref_io a pointer to a reference within the object
being scanned
Returns a result code, see ERROR HANDLING.
If the reference rank of the object being scanned is not MPS_RANK_AMBIG then the reference
pointed to by ref_io may be modified by mps_fix.
mps.h
This function is the part of the scanning protocol used to indicate
references. Scanning functions apply it, or MPS_FIX12, or MPS_FIX1 and MPS_FIX2 to the references in the object
being scanned.
It may only be called from within a scanning function. If it is
called within a MPS_SCAN_BEGIN block, MPS_FIX_CALL must be used (yes,
really).
This function does not perform any particular operation. The MPS
may call scanning functions for a number of reasons, and mps_fix may take different actions
depending on those reasons.
mps_res_t scan_array(mps_ss_t ss, mps_addr_t object, size_t length)
{
size_t i;
mps_res_t res;
mps_addr_t *array = (mps_addr_t *)object;
for(i = 0; i < length; ++i) {
res = mps_fix(ss, &array[i]);
if(res != MPS_RES_OK) return res;
}
return res;
}
The function returns MPS_RES_OK if it was successful, in
which case the scanning function should continue to scan the rest of
the object, applying mps_fix to
the remaining references. If mps_fix returns a value other than MPS_RES_OK, the scanning function must
return that value, and may return without scanning further
references. Generally, it is better if it returns as soon
as possible.
mps_ss_t,
mps_root_scan_t,
mps_fmt_scan_t,
mps_reg_scan_t,
MPS_SCAN_BEGIN,
MPS_SCAN_END,
MPS_FIX12,
MPS_FIX1,
MPS_FIX2,
MPS_FIX_CALL
MPS_FIX1The macro MPS_FIX1 is the part
of the scanning protocol used to indicate references to the MPS. It
may only be used from within MPS_SCAN_BEGIN and MPS_SCAN_END.
Format.
MPS_FIX1(mps_ss, ref)
mps_ss the scan state argument
that was passed to the scanning function
ref a reference within the object being scanned, type
mps_addr_t
Returns a truth value (type mps_bool_t) indicating whether the
reference is likely to be interesting to the MPS.
mps.h.
MPS_FIX1 and MPS_FIX2 are a trick to speed up scanning
by splitting MPS_FIX12 into two
macros. MPS_FIX1 is a fast test
to see if the reference is likely to be interesting to the MPS; if it
returns false, the scanner can proceed to the next reference. If it
returns true, the scan method must invoke MPS_FIX2, which does the actual
fixing.
This macro may only be used in code textually between MPS_SCAN_BEGIN and MPS_SCAN_END.
mps_res_t scan_array(mps_ss_t ss, Array object, size_t length)
{
size_t i;
mps_res_t res;
mps_addr_t *array = (mps_addr_t *)object;
MPS_SCAN_BEGIN(ss)
for(i = 0; i < length; ++i) {
mps_addr_t ref = array[i];
if(MPS_FIX1(ss, ref)) {
/* if(((Object*)ref)->type == ScannableType) { */
/* You can do something here, but in the end, you must call MPS_FIX2. */
res = MPS_FIX2(ss, &array[i]);
if(res != MPS_RES_OK)
return res;
/* } */
}
}
MPS_SCAN_END(ss);
return res;
}
MPS_FIX12,
MPS_FIX2,
mps_fix,
mps_ss_t,
mps_root_scan_t,
mps_fmt_scan_t,
mps_reg_scan_t,
MPS_SCAN_BEGIN,
MPS_SCAN_END,
MPS_FIX_CALL
MPS_FIX12The macro MPS_FIX12 is the
part of the scanning protocol used to indicate references to the
MPS. It may only be used from within MPS_SCAN_BEGIN and MPS_SCAN_END.
Scanning.
MPS_FIX12(mps_ss, ref_io);
mps_ss the scan state argument that was passed to the scanning function
ref_io a pointer to a reference within the object
being scanned, type mps_addr_t *
Returns a result code, see ERROR HANDLING.
If the reference rank of the object being scanned is not MPS_RANK_AMBIG then the reference
pointed to by ref_io may be modified by MPS_FIX12.
mps.h
This macro is used in the scanning protocol to indicate
references. Scanning functions apply it or mps_fix or MPS_FIX1 and MPS_FIX2 to the references in the object
being scanned.
It may only be used in code textually between MPS_SCAN_BEGIN and MPS_SCAN_END.
It is permitted for the reference (*ref_io) to point
outside the MPS arena being scanned, or to be NULL; in that case, it
is simply ignored.
This macro does not perform any particular operation. The MPS may
call scanning functions for a number of reasons, and MPS_FIX may take different actions
depending on those reasons.
mps_res_t scan_array(mps_ss_t ss, mps_addr_t object, size_t length) {
size_t i;
mps_res_t res;
mps_addr_t *array = (mps_addr_t *)object;
MPS_SCAN_BEGIN(ss)
for(i = 0; i < length; ++i) {
res = MPS_FIX(ss, &array[i]);
if(res != MPS_RES_OK)
return res;
}
MPS_SCAN_END(ss);
return res;
}
The macro returns MPS_RES_OK
if it was successful, in which case the scanning function should
continue to scan the rest of the object, fixing the remaining
references. If MPS_FIX12 returns
a value other than MPS_RES_OK,
the scanning function must return that value, and may return without
scanning further references. Generally, it is better if it returns as
soon as possible.
mps_fix,
mps_ss_t,
mps_root_scan_t,
mps_fmt_scan_t,
mps_reg_scan_t,
MPS_SCAN_BEGIN,
MPS_SCAN_END,
MPS_FIX1,
MPS_FIX2,
MPS_FIX_CALL
MPS_FIX12 is so called, as it basically performs the work of both MPS_FIX1 and MPS_FIX2.
MPS_FIX2The macro MPS_FIX2, together with MPS_FIX1, is the part of the scanning protocol used to indicate references to the MPS. It may only be used from within MPS_SCAN_BEGIN and MPS_SCAN_END.
Scanning.
MPS_FIX2(mps_ss, ref_io);
mps_ss the scan state argument
that was passed to the scanning function
ref_io a pointer to a reference within the object
being scanned, type mps_addr_t *
Returns a result code, see ERROR HANDLING.
If the reference rank of the object being scanned is not MPS_RANK_AMBIG then the reference pointed to by ref_io may be modified by MPS_FIX2.
mps.h.
MPS_FIX1 and MPS_FIX2 are a trick to speed up scanning
by splitting MPS_FIX12 into two
macros. MPS_FIX1 is a fast test
to see if the reference is likely to be interesting to the MPS; if it
returns false, the scanner can proceed to the next reference. If it
returns true, the scan method must invoke MPS_FIX2, which does the actual
fixing.
This macro may only be used in code textually between MPS_SCAN_BEGIN and MPS_SCAN_END.
This macro does not perform any particular operation. The MPS may call scanning functions for a number of reasons, and MPS_FIX2 may take different actions depending on those reasons.
mps_res_t scan_array(mps_ss_t ss, Array object, size_t length) {
size_t i;
mps_res_t res;
mps_addr_t *array = (mps_addr_t *)object;
MPS_SCAN_BEGIN(ss)
for(i = 0; i < length; ++i) {
mps_addr_t ref = array[i];
if(MPS_FIX1(ss, ref)) {
/* if(((Object*)ref)->type == ScannableType) { */
/* You can do something here, but in the end, you must call MPS_FIX2. */
res = MPS_FIX2(ss, &array[i]);
if(res != MPS_RES_OK)
return res;
/* } */
}
}
MPS_SCAN_END(ss);
return res;
}
The macro returns MPS_RES_OK
if it was successful, in which case the scanning function should
continue to scan the rest of the object, fixing the remaining
references. IfMPS_FIX2 returns a
value other than MPS_RES_OK,
the scanning function must return that value, and may return without
scanning further references. Generally, it is better if it returns as
soon as possible.
MPS_FIX12,
MPS_FIX1,
mps_fix,
mps_ss_t,
mps_root_scan_t,
mps_fmt_scan_t,
mps_reg_scan_t,
MPS_SCAN_BEGIN,
MPS_SCAN_END,
MPS_FIX_CALL
MPS_FIX_CALLMPS_FIX_CALL is used to call a scanning function from within MPS_SCAN_BEGIN and MPS_SCAN_END.
Scanning.
MPS_FIX_CALL(ss, call);
mps_ss the scan state argument that was passed to the scanning function
call an expression (containing a call to a scanning function)
None.
mps.h.
When using the MPS_SCAN_BEGIN and MPS_SCAN_END macros, you can't
directly call a separate function to do part of the scanning, because
between MPS_SCAN_BEGIN and
MPS_SCAN_END, the
scan_state parameter is in a strange state, so you
shouldn't pass it as an argument to a function. However, if you
really want to do it (say, because you have an embedded structure
shared between two scan methods), you can pass the scan state
correctly using MPS_FIX_CALL.
Note that you must receive the return value of the scanning function called, and pass it on as described in the ERROR HANDLING section.
mps_res_t foo_scan(mps_ss_t scan_state, mps_addr_t base, mps_addr_t limit)
{
Object *obj;
Object *obj_limit;
mps_res_t res;
obj_limit = limit;
MPS_SCAN_BEGIN(scan_state)
for(obj = base; obj < obj_limit; obj++) {
if(MPS_FIX12(scan_state, &obj->left) != MPS_RES_OK)
return res;
MPS_FIX_CALL(scan_state,
res = scan_data(scan_state, &obj->data));
if(res != MPS_RES_OK) return res;
if(MPS_FIX12(scan_state, &obj->right) != MPS_RES_OK)
return res;
}
MPS_SCAN_END(scan_state);
return MPS_RES_OK;
}
You must receive the return value of the function called. Like all
scanning functions, it returns MPS_RES_OK if it was successful, in
which case the caller should continue to scan the rest of the object,
fixing the remaining references. If it returns a value other
thanMPS_RES_OK, the calling
scanning function must return that value, and may return without
scanning further references. Generally, it is better if it returns as
soon as possible.
mps_fix,
mps_ss_t,
mps_root_scan_t,
mps_fmt_scan_t,
mps_reg_scan_t,
MPS_SCAN_BEGIN,
MPS_SCAN_END,
MPS_FIX12,
MPS_FIX1,
MPS_FIX2
MPS_RES_LIMITMPS_RES_LIMIT is a result
code, indicating that an operation failed because an internal limit
was reached.
mps.h.
This result code is returned if an operation could not be completed as requested because of an internal limitation of the MPS. The precise meaning depends on the function that returned the code. Refer to the documentation of that function for details.
switch(mps_alloc(&(mps_addr_t)object, pool, size)) {
case MPS_RES_LIMIT:
bomb("The MPS has reached an internal limit");
break;
/* ... */
}
MPS_RES_MEMORYMPS_RES_MEMORY is a result code, indicating that an operation failed because it ran out of memory.
All.
mps.h
This result code is returned if an operation could not be completed because there wasn't enough memory available. You need to deallocate something or allow the garbage collector to reclaim something to free enough memory, or expand the arena (if you're using an arena for which that does not happen automatically).
Note that failing to acquire enough memory because the arena commit
limit would have been exceeded is indicated by returning MPS_RES_COMMIT_LIMIT, not
MPS_RES_MEMORY.
Note that running out of address space (as might happen in virtual
memory systems) is indicated by returning MPS_RES_RESOURCE, not MPS_RES_MEMORY.
mps_res_t,
MPS_RES_RESOURCE,
MPS_RES_COMMIT_LIMIT
MPS_RES_PARAMMPS_RES_PARAM is a result
code, indicating that an operation failed because an invalid parameter
was specified for the operation.
All.
mps.h.
This result code is returned if an operation could not be completed as requested because an invalid parameter was specified for the operation. The precise meaning depends on the function that returned the code. Refer to the documentation of that function for details.
switch( res = mps_pool_create_v(&pool, arena, class, params) ) {
case MPS_RES_PARAM:
bomb("Can't make a pool with those specifications");
break;
/* ... */
}
MPS_RM_CONSTMPS_RM_CONST is a constant used in root mode arguments to indicate constant roots.
Root.
Integral constant.
mps.h
MPS_RM_CONST is a
preprocessor macro defining a constant that can be OR'ed with other
MPS_RM_* constants, and passed as the root mode argument
to certain root creation functions (mps_root_create, mps_root_create_fmt, mps_root_create_table,
mps_root_create_table_masked,
mps_root_create_reg
).
Passing MPS_RM_CONST means
that the client program will not change the root after it is
declared. I.e., scanning the root will produce the same set of
references every time. Furthermore, for formatted and table roots, the
client program may not write to the root at all.
res = mps_root_create_table(&mmRoot,
arena,
MPS_RANK_EXACT,
MPS_RM_CONST,
(mps_addr_t)&Objects,
rootCOUNT);
mps_root_create,
mps_root_create_fmt,
mps_root_create_table,
mps_root_create_table_masked,
mps_root_create_reg
Currently ignored. -- drj 1997-12-18.
MPS_RM_PROTMPS_RM_PROT is a constant used in root mode arguments to indicate protectable roots.
Root.
Integral constant.
mps.h
MPS_RM_PROT is a
preprocessor macro defining a constant that can be OR'ed with other
MPS_RM_* constants, and passed as the root mode argument
to certain root creation functions (mps_root_create_fmt, mps_root_create_table,
mps_root_create_table_masked).
Passing MPS_RM_PROT means
that the MPS may place a hardware write barrier on any pages which any
part of the root covers. Format methods and any scanning function
(except for the one for this root) may not write data in this
root. They may read it.
You mustn't specify MPS_RM_PROT on a root allocated from
the MPS.
res = mps_root_create_table(&mmRoot,
arena,
MPS_RANK_EXACT,
MPS_RM_PROT,
(mps_addr_t)&Objects,
rootCOUNT);
mps_root_create_fmt,
mps_root_create_table,
mps_root_create_table_masked.
No page may contain parts of two or more roots with MPS_RM_PROT [how does one prevent
that?]. You mustn't specify MPS_RM_PROT if the client program or
anything other than (this instance of) the MPS is going to protect or
unprotect the relevant pages.
Future meaning: The MPS may place a hardware read and/or write
barrier on any pages which any part of the root covers. Format methods
and scanning functions (except for the one for this root) may not read
or write data in this root. You may specify MPS_RM_PROT on a root allocated from
the MPS, as long as it's not from a GCd pool. - drj 1997-12-18
This feature is far too technical for most of our clients: we should think about producing some guidelines on how to use it. - pekka 1998-01-27
There may be problems if the client wants the OS to access the root. Lots of OSes can't cope with writing to protected pages. So we'll need to document that caveat too. drj 1998-05-20
mps_sac_allocThis function allocates a block using the segregated allocation cache given.
Allocation cache
mps_res_t mps_sac_alloc(mps_addr_t *p_o, mps_sac_t sac, size_t size, mps_bool_t has_reservoir_permit);
p_o a pointer to a variable to hold the address of the new block
sac the segregated allocation cache
size the size of the block requested
has_reservoir_permit regulates access to the reservoir
If the return value is MPS_RES_OK, the address of a new block
is *p_o.
mps.h
This function allocates a block using the cache given. If no
suitable block exists in the cache, it will ask for more memory from
the associated pool. size does not have to be one of the
class sizes of the cache; it does not have to be aligned.
The client is responsible for synchronising the access to the cache, but if the cache decides to access the pool, the MPS will properly synchronize with any other threads that might be accessing the same pool.
has_reservoir_permit regulates whether the pool has
permission to get more memory from the reservoir to satisfy this
request.
void *p;
Foo *foo;
res = mps_sac_alloc(&p, sac, FooSIZE, is_in_panic);
if (res != MPS_RES_OK) {
printf("Failed to alloc foo!\n");
exit(1);
}
foo = p;
/* use foo */
mps_sac_free(sac, p, FooSIZE);
mps_sac_alloc returns
MPS_RES_MEMORY when it
fails to find enough memory; see the documentation for this return
code for recovery options. It returns MPS_RES_COMMIT_LIMIT if it
can't allocate without exceeding the arena commit limit; Free
something to make more space or increase the limit using mps_arena_commit_limit_set. It
returns MPS_RES_RESOURCE
if it has run out of swap space; Free something or terminate other
processes on the same machine.
MPS_SAC_ALLOC_FAST,
mps_sac_free,
MPS_SAC_FREE_FAST,
mps_sac_t,
mps_reservoir_limit_set,
mps_arena_commit_limit_set,
MPS_RES_MEMORY,
MPS_RES_COMMIT_LIMIT,
MPS_RES_RESOURCE
There's also a macro called MPS_SAC_ALLOC_FAST, that does
the same thing. The macro is faster, but generates more code and does
less checking.
The block allocated can be larger than requested. Blocks not matching any class size are allocated from the next largest class, and blocks larger than the largest class size are simply allocated at the requested size (rounded up to alignment, as usual).
Objects allocated through a segregated allocation cache should only
be freed through a segregated allocation cache with the same class
structure. Using mps_free on
them can cause memory leaks, because the size of the block might be
larger than you think. Naturally, the cache must also be attached to
the same pool.
MPS_SAC_ALLOC_FASTThis macro allocates a block using the segregated allocation cache given.
Allocation cache
MPS_SAC_ALLOC_FAST(res_o, p_o, sac, size, has_reservoir_permit)
res_o |
mps_res_t |
an lvalue to hold the result code |
p_o |
mps_addr_t |
an lvalue to hold the address of the new block |
sac |
mps_sac_t |
the segregated allocation cache |
size |
size_t |
the size of the block requested |
has_reservoir_permit |
mps_bool_t |
regulates access to the reservoir |
res_o will be set to the return code. If this is
MPS_RES_OK, the address of the
new block is in p_o.
mps.h
This macro allocates a block using the cache given. If no suitable
block exists in the cache, it will ask for more memory from the
associated pool. size does not have to be one of the
class sizes of the cache; it does not have to be aligned.
The client is responsible for synchronizing the access to the cache, but if the cache decides to access the pool, the MPS will properly synchronize with any other threads that might be accessing the same pool.
has_reservoir_permit regulates whether the pool has
permission to get more memory from the reservoir to satisfy this
request.
void *p;
Foo *foo;
mps_res_t res;
MPS_SAC_ALLOC_FAST(res, p, sac, FooSIZE, is_in_panic);
if (res != MPS_RES_OK) {
printf("Failed to alloc foo!\n");
exit(1);
}
foo = p;
/* use foo */
MPS_SAC_FREE_FAST(sac, p, FooSIZE);
MPS_SAC_ALLOC_FAST
returns MPS_RES_MEMORY when
it fails to find enough memory; see the documentation for this return
code for recovery options. It returns MPS_RES_COMMIT_LIMIT if it
can't allocate without exceeding the arena commit limit; free something
to make more space or increase the limit using mps_arena_commit_limit_set. It
returns MPS_RES_RESOURCE
if it has run out of swap space; free something or terminate other
processes on the same machine.
mps_sac_alloc,
MPS_SAC_FREE_FAST,
mps_sac_free,
mps_sac_t,
mps_reservoir_limit_set,
mps_arena_commit_limit_set,
MPS_RES_MEMORY,
MPS_RES_COMMIT_LIMIT,
MPS_RES_RESOURCE
There's also a function called mps_sac_alloc, that does the same
thing.
The block allocated can be larger than requested. Blocks not matching any class size are allocated from the next largest class, and blocks larger than the largest class size are simply allocated at the requested size (rounded up to alignment, as usual).
Objects allocated through a segregated allocation cache should only
be freed through a segregated allocation cache with the same class
structure. Using mps_free on them
can cause memory leaks or assertions, because the size of the block
might be larger than you think. Naturally, the cache must also be
attached to the same pool.
The macro doesn't evaluate has_reservoir_permit,
unless it decides to access the pool.
MPS_SAC_CLASS_LIMITMPS_SAC_CLASS_LIMIT specifies how many classes mps_sac_create is guaranteed to accept.
size_t
Allocation cache
mps.h
MPS_SAC_CLASS_LIMIT specifies a lower limit on the maximum number of classes that can be described in a call to mps_sac_create, i.e., the MPS guarantees to accept at least this many classes. More might be accepted -- in fact, there might not be any limit in the implementation on the maximum number of classes, but if you specify more than this, you should be prepared to handle the error.
MPS_SAC_CLASS_LIMIT is a macro suitable for use in a constant expression, both in a #if directive and wherever else constant expressions may be used.
mps_sac_t sac;
mps_sac_class_s classes[3] = { {8, 38, 1}, {136, 19, 3}, {512, 4, 1} };
#if (MPS_SAC_CLASS_LIMIT < 3)
# error "Too many classes!"
#endif
res = mps_sac_create(&sac, pool, 3, classes);
if (res != MPS_RES_OK) {
printf("Failed to create the allocation cache!");
exit(1);
}
If you ask for too many size classes, mps_sac_create returns MPS_RES_LIMIT; you can recover by combining some small adjacent classes.
mps_sac_freeThis function frees an object using the segregated allocation cache given.
Allocation cache
void mps_sac_free(mps_sac_t sac, mps_addr_t p, size_t size);
sac the segregated allocation cache
p a pointer to the block being freed
size the size of the block being freed
None.
mps.h
This function frees an object using the cache given. If the cache would become too full,some blocks are returned to the associated pool.
size
should be the size that was specified when the object was allocated (the cache knows what the real size of the block is). The object must have been allocated through a segregated allocation cache with the same class structure,attached to the same pool.
The client is responsible for synchronising the access to the cache, but if the cache decides to access the pool, the MPS will properly synchronize with any other threads that might be accessing the same pool.
void *p;
Foo *foo;
res = mps_sac_alloc(&p, sac, FooSIZE, is_in_panic);
if (res != MPS_RES_OK) {
printf("Failed to alloc foo!\n");
exit(1);
}
foo = p;
/* use foo */
mps_sac_free(sac, p, FooSIZE);
MPS_SAC_FREE_FAST,
mps_sac_alloc,
MPS_SAC_ALLOC_FAST,
mps_sac_t
Usually, you'd use the same cache to allocate and deallocate an object.
There's also a macro called MPS_SAC_FREE_FAST, that does the same thing. The macro is faster, but generates more code and does no checking.
Note that mps_sac_free does very little checking; it's optimized for speed.Double frees and other mistakes will only be detected when the cache is flushed (which can happen by demand through mps_sac_flush or automatically), unless intervening operations have obscured symptom.
MPS_SAC_FREE_FASTMPS_SAC_FREE_FAST frees an object using the segregated allocation cache given.
Allocation cache
MPS_SAC_FREE_FAST(sac, p, size)
sac
|
mps_sac_t
|
the segregated allocation cache |
p
|
mps_addt_t
|
the address of the object to be freed |
size
|
size_t
|
the size of the object |
None.
mps.h
This macro frees an object using the cache given. If the cache would become too full, some blocks are returned to the associated pool.
size
should be the size that was specified when the object was allocated (the cache knows what the real size of the block is). The objects must have been allocated through a segregated allocation cache with the same class structure, attached to the same pool.
The client is responsible for synchronizing the access to the cache, but if the cache decides to access the pool, the MPS will properly synchronize with any other threads that might be accessing the same pool.
void *p;
Foo *foo;
mps_res_t res;
MPS_SAC_ALLOC_FAST(res, p, sac, FooSIZE, is_in_panic);
if (res != MPS_RES_OK) {
printf("Failed to alloc foo!\n");
exit(1);
}
foo = p;
/* use foo */
MPS_SAC_FREE_FAST(sac, p, FooSIZE);
mps_sac_free,
MPS_SAC_ALLOC_FAST,
mps_sac_alloc,
mps_sac_t
Usually, you'd use the same cache to allocate and deallocate an object.
There's also a function called mps_sac_free, that does the same thing. Themacro is faster, but generates more code and does no checking.
Note that MPS_SAC_FREE_FAST doesn't do any checking; it's optimized for speed. Double frees and other mistakes will only be detected when the cache is flushed (which can happen by demand through mps_sac_flush or automatically), unless intervening operations have obscured symptom.
MPS_SCAN_BEGINThe macro MPS_SCAN_BEGIN is part of the scanning protocol; together with MPS_SCAN_END, it sets up local information used by MPS_FIX*.
Scanning.
MPS_SCAN_BEGIN(ss)
ss the scan state argument that was passed to the scanning function
mps.h
This macro is used in the scanning protocol. Together with MPS_SCAN_END, it sets up local information used by the fast MPS_FIX* macros.
mps_res_t scan_array(mps_ss_t ss, mps_addr_t object, size_t length)
{
size_t i;
mps_res_t res;
mps_addr_t *array = (mps_addr_t *)object;
MPS_SCAN_BEGIN(ss)
for(i = 0; i < length; ++i) {
res = MPS_FIX12(ss, &array[i]);
if(res != MPS_RES_OK)
return res;
}
MPS_SCAN_END(ss);
return res;
}
MPS_SCAN_END,
MPS_FIX12,
MPS_FIX1,
MPS_FIX2,
MPS_FIX_CALL
Between MPS_SCAN_BEGIN and MPS_SCAN_END, you may not call another scanning function directly, because the scan state parameter is in a strange state, so you shouldn't pass it as an argument to a function. However, you can pass the scan state using MPS_FIX_CALL. You also cannot nest MPS_SCAN_BEGIN textually within another MPS_SCAN_BEGIN -- MPS_SCAN_END pair.
MPS_SCAN_ENDThe macro MPS_SCAN_END is part of the scanning protocol; it terminates a block started by MPS_SCAN_BEGIN.
Scanning.
MPS_SCAN_END(ss);
ss the scan state argument that was passed to the scanning function
mps.h
This macro is used in the scanning protocol. Together with MPS_SCAN_BEGIN, it sets up local information used by the fast MPS_FIX* macros. Note that MPS_SCAN_END completes the scanning, so successful termination of the scanning must invoke it (error branches can return without passing through).
mps_res_t scan_array(mps_ss_t ss, mps_addr_t object, size_t length) {
size_t i;
mps_res_t res;
mps_addr_t *array = (mps_addr_t *)object;
MPS_SCAN_BEGIN(ss)
for(i = 0; i < length; ++i) {
res = MPS_FIX12(ss, &array[i]);
if(res != MPS_RES_OK)
return res;
}
MPS_SCAN_END(ss);
return res;
}
MPS_SCAN_BEGIN,
MPS_FIX12,
MPS_FIX1,
MPS_FIX2,
MPS_FIX_CALL
Between MPS_SCAN_BEGIN and MPS_SCAN_END, you may not call another scanning function directly, because the scan state parameter is in a strange state, so you shouldn't pass it as an argument to a function. However, you can pass the scan state using MPS_FIX_CALL.
MPS_T_WORDMPS_T_WORD an unsigned integral type that is the same size as an object pointer.
mpstd.h.
MPS_T_WORD is a preprocessor macro defined in "mpstd.h". It is the name of an unsigned integral type that is the same size as an object pointer (so sizeof(MPS_T_WORD) == sizeof(void*)).The exact identity of the type is platform-dependent.
MPS_WORD_SHIFT,
MPS_WORD_WIDTH
MPS_WORD_SHIFTMPS_WORD_SHIFT is log base 2 of MPS_WORD_WIDTH.
Integral constant.
mpstd.h.
MPS_WORD_SHIFT is a preprocessor macro defined in "mpstd.h". It is the logarithm in base 2 of MPS_WORD_WIDTH (so 1 << MPS_WORD_SHIFT == MPS_WORD_WIDTH). The value of MPS_WORD_SHIFT is platform-dependent. Typical values are 5 and 6.
MPS_WORD_WIDTHMPS_WORD_WIDTH is the width in bits of the type MPS_T_WORD.
Integral constant.
mpstd.h.
MPS_WORD_WIDTH is a preprocessor macro defined in "mpstd.h" to be the width in bits of the type MPS_T_WORD (so MPS_WORD_WIDTH == sizeof(MPS_T_WORD) * CHAR_BIT).
This value is required for the use of the MPS C interface and the interpretation of "mps.h".It is platform-dependent. It is a power of 2; typical values are 32 and 64. It may be defined by including "mpstd.h" on a supported platform, or by defining it to be the width of MPS_T_WORD in bits.
#define MPS_WORD_WIDTH 32
mps_addr_tmps_addr_t is the type of addresses managed by the MPS, and also the type of references to objects.
Allocation, allocation point, format, root, location dependency.
typedef void *mps_addr_t;
mps.h
mps_addr_t is the type of addresses managed by the MPS, and also the type of references to objects. It is used in the MPS C interface where the MPS needs to pass a pointer to memory that is under the control of the MPS.
In accordance with standard C practice, the value NULL of type mps_addr_t will never be used to represent the address of an object.
{
mps_addr_t new_block;
mps_res_t res;
thingy *tp;
res = mps_pool_alloc(&new_block, pool, sizeof(thingy));
if(res != MPS_RES_OK) return res;
tp = new_block;
/* ... */
}
Not applicable.
mps_align_tmps_align_t is the type of an alignment.
Format, pool, allocation.
typedef size_t mps_align_t;
mps.h
An alignment specifies the address modulus to which all objects in an object format must be aligned. That is, if an alignment of 4 is specified, then the address of any object in that format modulo 4 will always be 0.
mps_align_t is a transparent type equivalent to the C type "size_t" and must be a positive power of 2.
Some pools and allocation protocols accept an alignment as an option that can be used to ensure that objects in the pool or objects allocated observe a stricter alignment than that of the object format.
mps_align_t floatAlign = 4; mps_align_t doubleFloatAlign = 8;
mps_allocmps_alloc allocates a block of memory in a pool.
Allocation.
mps_res_t mps_alloc(mps_addr_t *p, mps_pool_t pool, size_t size, ...);
p output parameter for a pointer to the block allocated
pool the pool to allocate in
size the size of the block to allocate in bytes
... (some pools can take additional arguments)
A return code.
mps.h
mps_alloc allocates a block of memory in the given pool.
mps_res_t res;
mps_addr_t p;
res = mps_alloc(&p, pool, size);
if(res != MPS_RES_OK) {
/* p hasn't been touched in this case. */
handle error;
}
/* p now contains the result, which is the address of the new block */
/* in this case. */
mps_alloc_pattern_rampReturns a allocation pattern type indicating that allocation will follow a ramp pattern.
alloc-pattern-ramp
mps_alloc_pattern_t mps_alloc_pattern_ramp();
none
Returns the allocation pattern type for ramps.
When declaring an allocation pattern for an AP, if the calls to mps_ap_alloc_pattern_begin use ramp allocation patterns (such as the result of mps_alloc_pattern_ramp), then the MPS will take this as an indication that most of the objects allocated after the call to mps_ap_alloc_pattern_begin are likely to be dead by the corresponding call to mps_ap_alloc_pattern_end.
This permits the client to indicate useful points for GC with minimal perturbation of the GCstrategy.
{
mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
do_lots_of_work();
mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
}
Cannot fail.
mps_alloc_pattern_ramp_collect_all,
mps_ap_alloc_pattern_begin
mps_alloc_pattern_ramp_collect_allmps_alloc_pattern_ramp_collect_all
Returns a ramp allocation pattern type indicating that a full GC should be done.
GC, AP, ramps
mps_alloc_pattern_t mps_alloc_pattern_ramp_collect_all();
none
Returns the allocation pattern type for full collection ramps.
This yields an allocation pattern for an AP that is similar to that returned by mps_alloc_pattern_ramp, in that it declares a ramp allocation pattern, but additionally indicates to the MPS that the next collection following the ramp should be a full GC.
This permits the client to indicate useful points for a full GC, either because most of the heap is likely to be dead, or because accurate statistics are required, with minimal perturbation of the GC strategy.
As usual, this allocation pattern should be used in matching mps_ap_alloc_pattern_begin and mps_ap_alloc_pattern_end pairs. It may nest with, but should not otherwise overlap with allocation patterns of type mps_alloc_pattern_ramp. In this case, the MPS may defer the full GC until after all ramp allocation patterns have ended.
{
mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp_collect_all());
do_lots_of_work();
mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp_collect_all());
wait_for_collection_statistics_while_doing_other_allocation();
}
Cannot fail.
mps_alloc_pattern_ramp,
mps_ap_alloc_pattern_begin
mps_amc_applymps_amc_apply is used to inspect objects in an AMC pool. You may only call it when the
arena is parked (for example, after mps_arena_collect).
Arena.
extern void mps_amc_apply(mps_pool_t, void(*f)(mps_addr_t, void *, size_t), void *, size_t);
mps_pool_t the pool whose objects you want to inspect
f a supplied function
mps_addr_t the address of the object you want to inspect
*
size_t
Not applicable.
Not applicable.
mps_amc_apply is used to inspect objects in an AMC pool. You may only call it when the arena is parked (for example, after mps_arena_collect). When called, mps_amc_apply calls the supplied function "f" once for each object in the pool, with the address of the object as its first argument. "f" is given as its second and third arguments whatever values were given as the third and fourth arguments to mps_amc_apply. (This is intended to make it easy to pass, for example, an array and its size as parameters.)
"f" will be called on both data and pad objects, and it is "f"'s job to distinguish, if required, between the two. Note (c.f. mps_arena_collect, above) that it may be called on unreachable objects that the collector has not recycled or has not been able to recycle.
The function "f" may not allocate memory or access any automatically-managed memory except the object at which it is pointed and, in the case of objects in Dylan Container Format, that object's wrapper.
Functions that park the arena:
mps_arena_park,
mps_arena_collect.
A more general heap walker which inspects all formatted objects in
the arena:
mps_arena_formatted_objects_walk
There is no equivalent function for other pool classes, but there is a more general heapwalker: mps_arena_formatted_objects_walk.
Does "You must call it when the arena is parked" mean that (a) parking an arena requires that you call this function, or (b) you can only call this function when the arena is in the parked state? LMB
(b). Changed "must call" to "may only call" drj 1998-08-25
mps_ap_alloc_pattern_beginIndicates the start of allocation following a particular pattern.
AP, Allocation, Alloc-pattern-ramp.
mps_res_t mps_ap_alloc_pattern_begin(mps_ap_t ap, mps_alloc_pattern_t alloc_pattern)
ap The allocation point in which the patterned allocation will occur
alloc_pattern The pattern of the allocation
Result code indicating whether start of the allocation pattern was successfully registered.
mps.h
This function is used, together with mps_ap_alloc_pattern_end, to indicate periods of allocation in an allocation point that follow some pattern of lifetime.
The nesting/overlapping restrictions on allocation patterns may vary depending on the particular allocation pattern type, but in general, if mps_ap_alloc_pattern_begin is used multiple times on the same allocation point without intervening calls to mps_ap_alloc_pattern_end, the calls match in a stack-like way, outermost and innermost; that is, allocation patterns may nest, but not otherwise overlap.
{
res = mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
do_some_work(); /* Leaves stuff lying around */
res = mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
do_some_more_work(); /* Tidies up after itself */
res = mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
tidy_up_first_work();
res = mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
}
Currently doesn't fail, but may in future if certain allocation patterns are inappropriate for that allocation point at that point in time.
mps_alloc_pattern_ramp,
mps_alloc_pattern_ramp_collect_all,
mps_ap_alloc_pattern_end,
mps_ap_alloc_pattern_reset
mps_ap_alloc_pattern_endIndicates the end of allocation following a particular pattern.
AP, Allocation, Alloc-pattern-ramp.
mps_res_t mps_ap_alloc_pattern_end(mps_ap_t ap, mps_alloc_pattern_t alloc_pattern)
ap The allocation point in which the patterned allocation occurred
alloc_pattern The pattern of the allocation
Result code indicating whether end of the allocation pattern was successfully registered.
This function is used, together with mps_ap_alloc_pattern_begin, to indicate periods of allocation in an allocation point that follow some pattern of lifetime.
{
res = mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
do_some_work(); /* Leaves stuff lying around */
res = mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
do_some_more_work(); /* Tidies up after itself */
res = mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
tidy_up_first_work();
res = mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
}
Will fail if there is no extant allocation pattern of that type. May fail in future if certain allocation patterns are inappropriate for that allocation point at that point in time.
mps_alloc_pattern_ramp,
mps_alloc_pattern_ramp_collect_all,
mps_ap_alloc_pattern_begin,
mps_ap_alloc_pattern_reset
mps_ap_alloc_pattern_resetIndicates the end of allocation pattern on an allocation point.
AP, Alloc-pattern-ramp
mps_res_t mps_ap_alloc_pattern_reset(mps_ap_t ap);
ap The allocation point in which the patterned allocation occurred
Result code indicating whether end of the allocation patterns was successfully registered.
This function may be used in place of mps_ap_alloc_pattern_end to end all extant allocation patterns on an allocation point. It is anticipated that this may be used to recover from error conditions.
{
res = mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
do_some_work(); /* Leaves stuff lying around */
res = mps_ap_alloc_pattern_begin(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
res = do_some_more_work(); /* Tidies up after itself */
if(res != mps_res_ok) {
res = mps_ap_alloc_pattern_reset(ap);
assert(res == mps_res_ok);
return;
}
res = mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
tidy_up_first_work();
res = mps_ap_alloc_pattern_end(ap, mps_alloc_pattern_ramp());
assert(res == mps_res_ok);
}
Cannot fail at present. May fail in future if certain allocation patterns cannot be ended for that allocation point at that point in time.
mps_alloc_pattern_ramp,
mps_alloc_pattern_ramp_collect_all,
mps_ap_alloc_pattern_begin,
mps_ap_alloc_pattern_end
mps_ap_frame_popDeclares that a set of objects in a particular frame are dead or likely to be dead.
AP Stack Protocol
mps_res_t (mps_ap_frame_pop)(mps_ap_t /* ap */, mps_frame_t /* frame */)
mps_ap_t ap
The allocation point in which the frame was pushed.
frame
The frame.
A result code in the usual way.
This function pops the specified frame making its parent the current frame (frames are implicitly created using the push operation, see mps_ap_frame_push). Popping invalidates the specified frame and all frames pushed since the specified frame. Popping the frame makes a declaration about the set of objects which were allocated in the specified frame and also all frames which were pushed since the specified frame. It can be used to declare a set of objects dead or likely to be mostly dead; the exact interpretation of the declaration depends on pool class that the allocation point is in (the same pool class that that objects are in). Typically pool classes which are mostly manually managed will use this declaration to mean that the objects are dead and their space can be reclaimed immediately, whereas pool classes which are mostly automatically managed will use this declaration to mean that the objects are likely to be mostly dead (the pool class may use this declaration to alter its collection decisions). Consult the pool class documentation for details.
In general a frame other than the current frame can be popped (all frames pushed more recently will be invalidated as well, as described above), but a particular pool class may impose the restriction that only the current frame may be popped. This restriction means that every push must have a corresponding pop. Consult the pool class documentation for details.
It is illegal to pass invalid frames to any MPS function. In particular it is not possible to pop frames out of order (so the sequence "A = push, B = push, pop A, pop B" is illegal) or to pop to the same frame twice (so the sequence "A = push, pop A, pop A" is illegal).
More comprehensive documentation is available in the protocol document (AP Stack Protocol).
<example of how to use the symbol>
<how the client program should handle errors that the symbol returns, if applicable>
mps.protocol.alloc-point.stack,
mps_ap_frame_push
None.
mps_ap_frame_pushDeclares a new frame as part of the AP stack protocol.
AP Stack Protocol
mps_res_t (mps_ap_frame_push)(mps_frame_t * /* frameReturn */, mps_ap_t /* ap */);
mps_frame_t *frameReturn The frame return parameter. A new frame (declared by this function) is stored in this location if this function is successful.
mps_ap_t ap The allocation point in which the new frame is declared.
A result code in the usual way. The creation of new frame objects (which is implicit in the action of this function) can consume resources, so this function can fail because there are insufficient resources. This function may fail if the correct protocol is not followed by the client.
This function declares a new frame in the specified allocation point, makes that new frame a child of the current frame, changes the current frame to be the newly created frame, and returns a handle to the frame. Frames have two important features: A single frame identifies a set of objects(those objects that are "allocated in the frame") which can be destroyed (or declared dead) in a pop operation (see mps_ap_frame_pop); They are arranged in a partially ordered sequence (this is important when the pop operation is used). A fuller and more useful description is found in the APstack protocol document (protocol.mps.alloc-point.stack).
[missing]
Errors can either be because the client hasn't followed the correct protocol in which case there isn't much that we can recommend or else because some needed resource isn't available. The usual course of actions when short of resources is recommended.
mps_ap_frame_pop,
protocol.mps.alloc-point.stack
mps_arena_clampmps_arena_clamp puts the specified arena into the clamped state.
Arena.
extern void mps_arena_clamp(mps_arena_t);
arena -- the arena to be put into the clamped state
None.
mps.h
mps_arena_clamp puts the specified arena into the clamped state. In the clamped state, no
object motion will occur and the staleness of location dependencies will not change. All references
to objects loaded while the arena is clamped will keep the same binary representation until after it is released.
In a clamped arena, incremental collection may still occur, but it will not be visible to the mutator and no new collections will begin. Space used by unreachable objects will not be recycled until the arena becomes unclamped.
mps_arena_park,
mps_arena_release
mps_arena_class_clmps_arena_class_cl returns the client arena class.
Arena.
mps_arena_class_t mps_arena_class_cl(void)
None.
Returns the client arena class.
mpsacl.h
This function is used to get hold of the client arena class, for the purpose of passing it to mps_arena_create.
mps_arena_t arena;
int main(void)
{
void *block;
mps_res_t res;
block = malloc(ARENA_SIZE);
if(block == NULL) {
printf("Not enough memory!");
exit(1);
}
res = mps_arena_create(&arena, mps_arena_class_cl(), ARENA_SIZE, block);
if(res != MPS_RES_OK) {
printf("ARENA_SIZE too small");
exit(2);
}
/* rest of program */
}
None.
A client arena gets its managed memory from the client. This memory block is passed when the arena is created. When creating a client arena, mps_arena_create takes two extra arguments:
mps_res_t mps_arena_create(mps_arena_t *mps_arena_o, mps_arena_class_t mps_arena_class_cl, size_t size, void *block)
block is the address of the memory block managed by the arena, andsize is its size in bytes. If mps_arena_create returns MPS_RES_MEMORY, then the block was too small to hold the internal arena structures.Allocate a (much) larger one, and try again. mps_arena_create returns MPS_RES_FAIL, if the MPS library is copy-protected by a security device, such as a dongle, and a valid security device cannot be found.
mps_arena_class_t"m ps_arena_class_t " is the type of arena classes.
Arena.
typedef struct mps_arena_s *mps_arena_t;
mps_arena_class_s is an incomplete structure type used only to declare the opaque type mps_arena_class_t.
mps.h
mps_arena_class_t is the type of arena classes. It is opaque.
The definition of the client arena class in the "mpsacl.h" header:
extern mps_arena_class_t mps_arena_class_cl(void);
None.
mps_arena_class_vmmps_arena_class_vm returns the virtual memory arena class.
Arena.
mps_arena_class_t mps_arena_class_vm(void)
None.
Returns the virtual memory arena class.
mpsavm.h
This function is used to get hold of the virtual memory arena class, for the purpose of passing it to mps_arena_create. The VM arenas use the OS virtual memory interfaces to allocate memory. The chief consequence of this is that the arena can manage many more virtual addresses than it needs to commit memory to. This gives it flexibility as to where to place objects, which reduces fragmentation and helps make garbage collection more efficient.
This class is similar to mps_arena_class_vmnz but uses a more complex placement policy, which is more suited to copying garbage collection.
mps_arena_t arena;
int main(void)
{
mps_res_t res;
res = mps_arena_create(&arena, mps_arena_class_vm(), ARENA_SIZE);
if(res != MPS_RES_OK) {
printf("Not enough memory!");
exit(1);
}
/* rest of program */
}
None.
mps_arena_create,
mps_arena_class_vmnz
A virtual memory arena gets its managed memory from the operating system's virtual memory services. An initial address space size is passed when the arena is created. When creating a virtual memory arena, mps_arena_create takes one extra argument:
mps_res_t mps_arena_create(mps_arena_t *arena_o,
mps_arena_class_t arena_class_vm,
size_t size)
size is the initial amount of virtual address space, in bytes, that the arena will reserve (this space is initially reserved so that the arena can subsequently use it without interference from other parts of the program, but most of it is not committed, so it don't require any RAM or backing store). The arena may allocate more virtual address space beyond this initial reservation as and when it deems it necessary. The MPS is most efficient if you reserve an address space that is several times larger than your peak memory usage.
mps_arena_create returns MPS_RES_RESOURCE if it fails to reserveadequate address space to place the arena in; possibly other parts of the program are reserving too much virtual memory. It returns MPS_RES_MEMORY when it fails to allocate memory for the internal arena structures; either size was far too small or you ran out of swap space.It returns MPS_RES_FAIL, if the library is copy-protected by a security device, such as a dongle, and a valid security device cannot be found.
Virtual memory arenas are not available on the Mac platforms, other than MacOS X. You will get a linking error, if you attempt to use this function.
mps_arena_class_vmnzAn arena class like mps_arena_class_vm but with a different placement policy.
Arena.
mps_arena_class_t mps_arena_class_vmnz(void);
None.
Returns the VMNZ arena class.
mpsavm.h
Returns the VMNZ arena class (stands for Virtual Memory No Zones, if you really care.) This class can be passed to mps_arena_create in order to create a VMNZ arena. The VMNZ arenas use the OS virtual memory interfaces to allocate memory. The chief consequence of this is that the arena can manage many more virtual addresses than it needs to commit memory to. This gives it flexibility as to where to place objects.
This class is similar to mps_arena_class_vm but uses a simpler placement policy, that makes it slightly faster.
mps_arena_t arena;
int main(void)
{
mps_res_t res;
res = mps_arena_create(&arena, mps_arena_class_vmnz(), ARENA_SIZE);
if(res != MPS_RES_OK) {
printf("Not enough memory!");
exit(1);
}
/* rest of program */
}
No errors.
mps_arena_create,
mps_arena_class_vm
This class takes an extra argument when used in mps_arena_create (see example).The extra parameter should be of type size_t. It specifies the amount of virtual address space, in bytes, that this arena should use. The arena will reserve this amount of virtual address space from the OS during initialization. It will not subsequently use any more address space(compare with mps_arena_class_vm which can grow).
mps_arena_create returns MPS_RES_RESOURCE if it fails to reserve adequate address space to place the arena in; possibly other parts of the program are reserving too much virtual memory. It returns MPS_RES_MEMORY when it fails to allocate memory for the internal arena structures; either size was far too small or you ran out of swap space.It returns MPS_RES_FAIL, if the library is copy-protected by a security device, such as a dongle, and a valid security device cannot be found.
Virtual memory arenas are not available on the Mac platforms, other than MacOS X. You will get a linking error, if you attempt to use this function.
mps_arena_collectmps_arena_collect collects the arena and puts it in the parked state.
Arena.
void mps_arena_collect(mps_arena_t arena);
arena the arena to collect
mps.h
mps_arena_collect collects the arena and puts it in the parked state. Collecting the arena attempts to recycle as many unreachable objects as possible and reduce the size of the arena as much as possible (though in some cases it may increase because it becomes more fragmented). If you do not want the arena to be in the parked state, you must explicitly call mps_arena_release aftermps_arena_collect.
Note that the collector may not be able to recycle some objects (such as those near the destination of ambiguous references) even though they are not reachable.
[missing]
No errors.
mps_arena_park,
mps_arena_release
None.
mps_arena_commit_limitReturns the current commit limit associated with the arena in bytes.
Arena
size_t mps_arena_commit_limit(mps_arena_t arena)
arena -- the arena
Returns the current commit limit as a number of bytes in a size_t
mps.h
Returns the current commit limit associated with the arena in bytes. The commit limit can be changed using the function mps_commit_limit_set. The commit limit is used to control how much memory the MPS can obtain from the OS. See Arena Protocol for details.
limit = mps_arena_commit_limit(arena);
No errors.
mps_arena_committed,
mps_arena_commit_limit_set
None.
mps_arena_commit_limit_setChanges the current commit limit associated with the arena.
Arena
mps_res_t mps_arena_commit_limit_set(mps_arena_t arena, size_t limit)
arena -- the arena
limit -- the new commit limit in bytes
Returns a result code.
mps.h
The commit limit of the arena is set to the limit given. The commit limit controls how much memory the MPS will obtain from the OS. See Arena Protocol for details. The commit limit cannot beset to a value that is lower than the number of bytes that the MPS is using. If an attempt is made to set the commit limit to a value greater than or equal to that returned bymps_arena_committed then it will succeed. If an attempt is made to set the commit limit to a value less than that returned by mps_arena_committed then it will succeed only if the amount committed by the MPS can be reduced by reducing the amount of spare committed memory; in such a case the spare committed memory will be reduced appropriately and the attempt will succeed.
do {
res = mps_arena_commit_limit_set(arena, limit - 100 * 1024);
if(res != MPS_RES_OK)
flush_caches();
} while(res != MPS_RES_OK);
Returns MPS_RES_OK when successful, and some other result code when not.
mps_arena_committed,
mps_arena_commit_limit,
mps_arena_spare_commit_limit_set
mps_arena_commit_limit_set puts a limit on all memory committed by the MPS. The"spare committed" memory can be limited separately with mps_arena_spare_commit_limit_set. Note that "spare committed" memory is subject to both limits; there cannot be more spare committed memory than the spare commit limit, and there can't be so much spare committed memory that there is more committed memory than the commit limit.
mps_arena_committedmps_arena_committed returns the amount of memory (backing store) in use by the arena, both for storing client objects and for its own data structures.
Arena.
extern size_t mps_arena_committed(mps_arena_t arena)
arena -- the arena
Returns a number of bytes (the amount of committed memory) as a size_t.
mps.h
mps_arena_committed returns the amount of memory (backing store) in use by the arena (also known as "committed memory"). The value returned is a number of bytes.
Committed memory may be used both for storing client objects and for storing MPS datastructures. In addition the MPS maintains committed memory which is not being used (for either of the above purposes). This memory is known as "spare committed" memory (see mps_arena_spare_committed). The amount of "spare committed" memory can change atany time, in particular in will be reduced as appropriate in order meet client requests.
The reasons that the committed memory (as return by this function) might be large than the sum of the sizes of client allocated objects are:
some memory is used internally by the MPS to manage its own data structures and to record information about client objects (such as free lists, page tables, colour tables, statistics, etc).
operating systems (and hardware) typically restrict programs to requesting and releasing memory with a certain granularity (for example, pages), so extra memory is committed when this rounding is necessary.
there might be "spare committed" memory.
The amount of committed memory is a good measure of how much virtual memory resource ("swapspace") the MPS is using from the OS.
This function may be called whether the arena is unclamped, clamped or parked, if called when the arena in unclamped then the value may change after this function returns. A possible use might be to call it just after mps_arena_collect to (over-)estimate the size of the heap.
If you want to know how much memory the MPS is using then you're probably interested in the value mps_arena_committed() - mps_arena_spare_committed().
The amount of committed memory can be limited with the function mps_arena_commit_limit.
mps_arena_collect,
mps_arena_clamp,
mps_arena_park,
mps_arena_release
-
mps_arena_createmps_arena_create is used to create an arena.
Arena.
mps_res_t mps_arena_create(mps_arena_t *mps_arena_o, mps_arena_class_t mps_arena_class, ...)
mps_arena_o pointer to a variable to store the new arena in
mps_arena_class the arena class
... initialization arguments for the arena class
Different for each arena class. See mps_arena_class_*.
If the return value is MPS_RES_OK, the new arena is in *mps_arena_o.
mps.h
mps_arena_create is used to create an arena.
mps_arena_t arena;
int main(void)
{
mps_res_t res;
res = mps_arena_create(&arena, mps_arena_class_vm(), ARENA_SIZE);
if(res != MPS_ RES_OK) {
printf("Not enough memory!");
exit(1);
}
/* rest of program */
}
mps_arena_create returns MPS_RES_FAIL, if the MPS library is copy-protected by a security device, such as a dongle, and a valid security device cannot be found.Other error codes are specific to each arena class. See mps_arena_class_*.
mps_arena_create_v,
mps_arena_class_*,
mps_arena_destroy
mps_arena_create_vmps_arena_create_v is used to create an arena.
Arena.
mps_res_t mps_arena_create_v(mps_arena_t *mps_arena_o, mps_arena_class_t mps_arena_class, va_list args)
mps_arena_o pointer to a variable to store the new arena in
mps_arena_class the arena class
args initialization arguments for the arena class
Different for each arena class. See mps_arena_class_*.
If the return value is MPS_RES_OK, the new arena is in *mps_arena_o.
mps.h
mps_arena_create_v is used to create an arena. It is exactly the same as mps_arena_create, except that it takes the arena class initialization arguments in a va_list.
mps_arena_create_v returns MPS_RES_FAIL, if the MPS library is copy-protected by a security device, such as a dongle, and a valid security device cannot be found. Other error codes are specific to each arena class. See mps_arena_class_*.
mps_arena_create,
mps_arena_class_*,
mps_arena_destroy
mps_arena_exposemps_arena_expose ensures
that the MPS is not protecting any pages in the arena with read- or
write-memory protection barriers.
Arena, clamp, park, protection
mps_arena_expose(mps_arena);
extern void mps_arena_expose(mps_arena_t);
(mps_arena_t mps_arena)
mps_arena is an MPS arena object.
None.
mps.h
This function will ensure that the MPS is not protecting (with memory
read/write barriers) any page in the arena.
This is expected to only be useful for debugging.
The arena is left in the clamped state (see mps_arena_clamp).
Since barriers are used during a collection, calling this function has
the same effect as calling mps_arena_park; all collections are
run to completion and the arena is clamped so that no new collections
begin. The MPS also uses barriers to maintain remembered sets (an
optimisation to help avoid scanning work); calling
this function will effectively destroy the remembered sets and any
optimisation gains.
Calling this function will introduce a slow down, primarily for two reasons: any active collections will be run to completion before this function returns; the next collection will have to recompute all the remembered sets by scanning the entire heap.
The second aspect of the slow down, having the next collection recompute
the remembered sets, can be avoided by using mps_arena_unsafe_expose_remember_protection
instead of mps_arena_expose, and calling
mps_arena_unsafe_restore_protection
before calling mps_arena_release.
Those functions have unsafe aspects and place restrictions on what the
client can do (basically no exposed data can be changed).
None.
There can be no errors.
mps_arena_clamp,
mps_arena_park,
mps_arena_release,
mps_arena_unsafe_expose_remember_protection,
mps_arena_unsafe_restore_protection
mps_arena_formatted_objects_walkmps_arena_formatted_objects_walk
mps_arena_formatted_objects_walk is used to iterate over all formatted objects in the MPS heap.
None.
mps_arena_formatted_objects_walk(mps_arena, client_step_function, client_step_closure_p,client_step_closure_s);
extern void mps_arena_formatted_objects_walk(mps_arena_t, mps_formatted_objects_stepper_t, void *, size_t);
(mps_arena_t mps_arena, mps_formatted_objects_stepper_t stepper, void *p, size_t s)
mps_arena is an MPS arena object.
stepper is a client-supplied function (pointer) of
the right type (see mps_formatted_objects_stepper_t).
This function is applied to every object in all formatted pools.
This function should take the argument list (mps_addr_t object,
mps_fmt_t format,mps_pool_t pool, void *p, size_t s) and return
void.
object is the object to which the function is being
applied.
format is the format (an MPS format object)
of the object. pool is the pool in which the object
resides. p and s are copies of the
corresponding values that the client passed into mps_arena_formatted_objects_walk
originally.
p and s are passed into the function specified by the stepper argument whenever the MPS calls that function. See mps_formatted_objects_stepper_t.
None.
mps.h
mps_arena_formatted_objects_walk
is used to iterate over all formatted objects in the MPS heap. A
client-supplied function is called for every object in all formatted
pools; the object, the format, and the pool are passed to the user
supplied function, as well as user supplied closure variables.
Applies stepper function to a pool-class-specific collection of objects (that is, the pool class determines which objects in its instances get walked). Typically pool classes will arrange that all validly formatted objects are walked. During a trace this will in general be only the black objects, though the leaf pool class (LO), for example, will walk all objects since they are validly formatted whether they are black or white. Padding objects may be walked at the pool classes discretion, the client should handle this case.
The user supplied stepper function is called in a restricted context. It may not in general call any MPS function.
[not yet]
There are none.
mps_amc_apply (the
historical walker),
mps_formatted_objects_stepper_t
mps_arena_has_addrmps_arena_has_addr tests
whether an address is managed by a particular arena.
Arena.
extern mps_bool_t mps_arena_has_addr(mps_arena_t arena,
mps_addr_t addr);
arena an arena
addr an address
A boolean. Returns true if the address is managed by the arena, false otherwise.
mps.h
mps_arena_has_addr determines
whether a particular address is managed by a particular arena. An arena
manages a portion of total address space available on the hardware
architecture. No two arenas overlap so for any particular address this
function will return true for at most one arena. In general not all the
architecture addresses are managed by some arena; some addresses will not
be managed by any arena. This is what allows the MPS to cooperate with
other memory managers, shared object loaders, memory mapped file I/O,
and such like - it does not steal the whole address space.
The results from this function are true only for the instant at which the function returned. In some circumstances the results may immediately become invalidated (for example, a garbage collection may occur, the address in question may become free, the arena may choose to unmap the address and return storage to the operating system); for reliable results call this function whilst the arena is parked.
Can't fail.
mps_arena_park used to park an
arena
None.
mps_arena_parkmps_arena_park puts the specified arena into the parked state.
Arena.
extern void mps_arena_park(mps_arena_t arena);
arena the arena to park
None.
mps.h
mps_arena_park puts the specified arena into the parked state. While an arena is parked,no object motion will occur and the staleness of location dependencies will not change. All references to objects loaded while the arena is parked will keep the same binary representation until after it is released.
Any current collection is run to completion before the arena is parked, and no new collections will start. When an arena is in the parked state, it is necessarily not in the middle of a collection.
Can't fail.
mps_arena_clamp,
mps_arena_release
None.
mps_arena_releasemps_arena_release puts the specified arena into the unclamped state.
Arena.
extern void mps_arena_release(mps_arena_t);
None.
mps.h
mps_arena_release puts the specified arena into the unclamped state. While an arena is unclamped, garbage collection, object motion, and other background activity can take place.
Can't fail.
mps_arena_clamp,
mps_arena_park
None.
mps_arena_roots_walkmps_arena_roots_walk is used to iterate over all roots of the MPS heap.
None.
mps_arena_roots_walk(mps_arena, client_step_function, client_step_closure_p,client_step_closure_s);
extern void mps_arena_roots_walk(mps_arena_t, mps_roots_stepper_t, void *, size_t);
(mps_arena_t mps_arena, mps_roots_stepper_t stepper, void *p, size_t s)
mps_arena in an MPS arena object.
stepper is a client-supplied function (pointer) of the right type (see mps_roots_stepper_t). This function is applied to every reference to the heap from every root object registered with the arena. This function should take the argument list (mps_addr_t *ref, mps_root_t root, void *p, size_t s). ref is the address of a root which references an object in the arena. root is the registered root (an MPS root object) of which ref is a single reference, p and s are copies of the corresponding values that the client passed into mps_arena_roots_walk originally.
p and s are passed into the function specified by the stepper argument whenever the MPS calls that function. See mps_roots_stepper_t
None.
mps.h
mps_arena_roots_walk is used to iterate over all roots of the MPS heap. A client-supplied function is called for every root reference which points to an object in any automatically managed pools; the address of the root reference and the MPS root object are passed to the user supplied function, as well as some closure variables.
May only be called when the arena is in the parked state.
Applies stepper to each reference in any roots registered with the arena and which point to objects in automatically managed pools. If the root has rank MPS_RANK_AMBIG then the reference might not be to the start of an object; the client should handle this case. There is no guarantee that the reference corresponds to the actual location that holds the pointer to the object (since this might be a register, for example) - but the actual location will be passed if possible. This may aid analysis of roots via a debugger.
[not yet]
There are none.
mps_arena_formatted_objects_walk
mps_arena_spare_commit_limitRetrieves the value of the spare commit limit (previously set with mps_arena_spare_commit_limit_set).
Arena.
extern size_t mps_arena_spare_commit_limit(mps_arena_t arena)
mps_arena_t arena
Specifies the arena to retrieve the spare commit limit of.
Returns, as a size_t, the value of the spare commit limit.
mps.h
Returns the current value of the spare commit limit which is the value most recently set with mps_arena_spare_commit_limit_set. (See mps_arena_spare_commit_limit_set fordetails).
[missing]
There are no errors.
mps_arena_spare_commit_limit_set
None.
mps_arena_spare_commit_limit_setmps_arena_spare_commit_limit_set
Sets the limit of the amount of spare committed memory.
Arena.
extern void mps_arena_spare_commit_limit_set(mps_arena_t arena, size_t limit)
mps_arena_t arena
The arena to which the new limit should apply.
size_t limit
The value of the new limit (specified in bytes).
mps.h
The limit argument specifies a new "spare commit limit". The spare commit limit specifies the maximum amount of bytes of "spare committed" memory the MPS is allowed to have. Setting it to a value lower than the current amount of spare committed memory would immediately cause sufficient spare committed memory to be uncommitted so as to bring the value under the limit. In particular setting to 0 will mean that the MPS will have no "spare committed" memory.
"spare committed" memory is the term for describing memory which the arena is managing as free memory (so not in use by any pool and not otherwise in use for obscure internal reasons) but which remains committed (mapped from the OS). It is used by the arena to (attempt to) avoid calling the OS to repeatedly unmap and map areas of VM. "spare committed" memory is counted as committed memory as counted by mps_arena_committed and restricted by mps_arena_commit_limit.
Non-VM arenas do not have this concept, but they support the two functions mps_arena_spare_commit_limit and mps_arena_spare_commit_limit_set. The functions simply get and retrieve a value but do nothing else in that case.
Initially the value is some configuration-dependent value.
The value of the limit can be retrieved with mps_arena_spare_commit_limit.
[missing]
There are no errors.
None.
mps_arena_spare_committedReturns the number of bytes of spare committed memory.
Memory
size_t mps_arena_spare_committed(mps_arena_t);
The arena to which the query applies.
Returns the number of bytes of spare committed memory.
mps.h
"Spare committed" memory is the term for describing memory which the arena is committed from the OS but which is free (so not in use by any pool and not otherwise in use for obscure internal reasons). It is used by the arena to (attempt to) avoid calling the OS to repeatedly uncommit and commit areas of VM (because calling the OS to commit and uncommit memory is typically expensive)."Spare committed" memory can be used for grant client requests; if this is done when the MPS would otherwise have had to call the OS to commit more memory then the MPS has avoid some OS calls.
"spare committed" memory is counted as part of committed memory. The amount of committed memory can be retrieved with mps_arena_committed (see mps_arena_committed).
The amount of "spare committed" memory can be limited by using mps_arena_spare_commit_limit_set (see mps_arena_spare_commit_limit_set ), and the value of that limit can be retrieved with mps_arena_spare_commit_limit (see mps_arena_spare_commit_limit ). This is analogous to the functions for limiting the amount of committed memory.
[missing]
[missing]
mps_arena_spare_commit_limit_set,
mps_arena_spare_commit_limit
None.
mps_arena_unsafe_expose_remember_protectionmps_arena_unsafe_expose_remember_protection
mps_arena_unsafe_expose_remember_protection
is like mps_arena_expose but
additionally indicates that the MPS should remember some internal state
which can be used later to avoid slow down. This function is
potentially unsafe and must be used carefully.
Arena, clamp, park, protection
mps_arena_unsafe_expose_remember_protection(mps_arena);
extern void mps_arena_unsafe_expose_remember_protection(mps_arena_t);
(mps_arena_t mps_arena)
mps_arena is an MPS arena object.
None.
mps.h
This function does the same as mps_arena_expose in that it ensures
the MPS is not protecting any page in the arena and also clamps the
arena.
Additionally, using this function indicates to the MPS that it should
remember the protection state internally.
Later on the client should indicate that the remembered protection state
should be restored by using the mps_arena_unsafe_restore_protection
function.
Restore the remembered protections is only safe if the contents of the
exposed pages have not been changed; therefore this function should only
be used if you do not intend changing the pages, and the remembered
protection must only be restored if the pages have not been changed.
Releasing the arena from the clamped state, by calling mps_arena_release, will cause the MPS to
forget the remembered state. Restoring the remembered protection state,
using mps_arena_unsafe_restore_protection,
will also cause the MPS to forget the remembered state.
The MPS will remember the protection state if resources (memory) are available. If memory is low then only some or possibly none of the protection state will be remembered, with a corresponding inability to avoid slow down later. It is not possible for the client to tell whether the MPS has in fact remembered the protection state.
None.
There can be no errors.
mps_arena_clamp,
mps_arena_unsafe_restore_protection
mps_arena_unsafe_restore_protectionmps_arena_unsafe_restore_protection
mps_arena_unsafe_restore_protection
restores the protection state that the MPS remembered when the client
called mps_arena_unsafe_expose_remember_protection.
If used correctly this should avoid any slow down that would otherwise
occur.
Arena, clamp, park, protection
mps_arena_unsafe_restore_protection(mps_arena);
extern void mps_arena_unsafe_restore_protection(mps_arena_t);
(mps_arena_t mps_arena)
mps_arena is an MPS arena object.
None.
mps.h
This function restores the protection that the MPS has remembered
(during a period when the arena is exposed).
The client can cause the MPS to remember the protection state by using
the mps_arena_unsafe_expose_remember_protection
function.
The point of remember and restoring the protection state is to
avoid the slow down that happens when mps_arena_expose is used. Normally
when this function is used the next garbage collection will be slow
because the MPS has to do a lot of work to recover remembered sets;
normally the remembered sets are preserved by the MPS protecting the
relevant pages, but if the protection is removed then the remembered
sets have to be discarded and recomputed.
This recomputation of remembered sets can be avoided by using mps_arena_unsafe_expose_remeber_protection
instead, and using mps_arena_unsafe_restore_protection
to restore the remembered protections instead of recomputing them.
This function has unsafe aspects.
In order for it to be used safely the client must not have changed the
exposed data between the call to mps_arena_unsafe_expose_remember_protection
and mps_arena_unsafe_restore_protection.
If the client has changed the exposed data then
mps_arena_unsafe_restore_protection
must not be called - simply call mps_arena_release to continue normal
collections.
Note that this function does not release the arena from the clamped
state;
mps_arena_release should be called to continue normal
collections.
Calling this function causes the MPS to forget the remember protection state; as a consequence the same remembered state cannot be restored more than once.
None.
There can be no errors.
mps_arena_release,
mps_arena_unsafe_expose_remember_protection
mps_bool_tmps_bool_t is a transparent type, equivalent to int, that is used in the MPS C interfaceto indicate that a boolean value is intended.
Not applicable.
Not applicable.
Not applicable.
typedef int mps_bool_t;
mps.h
When used as an input parameter to the MPS, a value of 0 indicates "false" and any other value indicates "true". As an output parameter or function return from the MPS, 0 indicates "false",and 1 indicates "true". Note that an mps_bool_t value can be used in a conditional context, such as in an "if" statement.
if(mps_ld_isstale(&ld, space, obj)) {
mps_ld_reset(&ld, space);
mps_ld_add(&ld, space, obj);
}
None.
mps_class_amcmps_class_amc returns the pool class object for the Automatic Mostly Copying pool class.
Pool
mps_class_t mps_class_amc(void)
No arguments.
Returns a pool class object.
mpscamc.h
This function returns an object of type mps_class_t which represents the Automatic MostlyCopying pool class.
This pool class requires an extra argument when used in mps_pool_create:
res = mps_pool_create(&pool, arena, mps_class_amc(), format);
The extra argument, format, should be of type mps_fmt_t and specifies the format of the objects allocated in the pool.
An AMC pool is both scannable and collectable. Objects may contain exact references to other objects that will preserve such other objects. Objects may be reclaimed if they are not reachable from a root. Objects may move during collection, unless reachable via a (direct) ambiguous reference. Objects in an AMC pool may be registered for finalization. Exact (that is, non-ambiguous)references into an object in an AMC pool must be to the start of the object.
The AMC pool class exploits assumptions about object lifetimes and inter-connection variously referred to as "the generational hypothesis". In particular, the following tendencies will be efficiently exploited by such a pool:
- Most objects die young;
- Objects that don't die young will live a long time;
- Most references are backwards in time.
mps_ap_frame_push and mps_ap_frame_pop may be used on an allocation point in an AMC pool.They do not declare the affected objects to be definitely dead (compare with the SNC pool class),but have an undefined effect on the collection strategy.
If an allocation point is created in an AMC pool, the call to mps_ap_create will take no additional parameters.
mps_ap_frame_pop,
mps_ap_frame_push,
mps_ap_create
mps_class_mvffUsed as a parameter to mps_pool_create to create an MVFF pool.
Pool, Allocation Points.
mps_class_t mps_class_mvff(void)
None.
The function returns a class object that can be passed to mps_pool_create.
mpscmvff.h
MVFF pools implement a first-fit policy. The pool requires six parameters to pool creation:
mps_size_t extendBy -- The size of segment to allocate by default;
mps_size_t avgSize -- The average size of objects to be allocated;
mps_align_t alignment -- The alignment of addresses for allocation (and freeing) in thepool;
mps_bool_t slotHigh
mps_bool_t arenaHigh
mps_bool_t firstFit
The alignment is the alignment of ranges that can be allocated and freed. If an unaligned size is passed to mps_alloc or mps_free, it will be rounded up to the pool's alignment. The minimum alignment supported by pools of this class is
sizeof(void *).
The three boolean parameters may be set to (0, 0, 1) or (1, 1, 1). No other settings of these parameters is currently recommended.
Buffered allocation (mps_reserve and mps_commit) is also supported, but in that case, the policy is rather different: buffers are filled worst-fit, and allocation is always upwards from the base. The arenaHigh parameter regulates whether new segments are acquired at high or low addresses;the slotHigh and firstFit parameters do not affect buffered allocation. Buffered and unbuffered allocation can be used at the same time, but in that case, the first allocation point must be created before any call to mps_alloc.
Cached allocation ( MPS_SAC_ALLOC and MPS_SAC_FREE ) is also supported, but in that case,the policy is a little different: allocation from the cache follows its own policy (typicallyfirst-fit), and only when the cache needs to acquire more blocks from the underlying MVFF pool does it use the usual algorithm to choose blocks for the cache.
if(mps_pool_create(&pool, arena, mps_class_mvff(), 8 * 1024, 135, 4, 0, 0, 1)
!= MPS_RES_OK) {
printf("Error creating pool!");
exit(2);
}
mps_pool_create,
mps_reserve,
mps_commit.
It is usually not advisable to use buffered and unbuffered allocation at the same time,because the worst-fit policy of buffer filling will grab all the large blocks, leading to severe fragmentation. Use two separate pools instead.
Note that using buffered allocation prevents (for obscure technical reasons) the pool from allocating across segment boundaries. This can cause added external fragmentation if objects are allocated that are a significant fraction of the segment size. (This quirk will disappear in a future version.)
mps_class_sncReturns the pool class object (of type mps_class_t) for the Stack No Check pool class.
Pool.
mps_class_t mps_class_snc(void)
No arguments.
Returns a pool class object.
mpscsnc.h
This function returns an object of type mps_class_t which represents the Stack No Check pool class.
This pool class requires an extra argument when used in mps_pool_create:
res = mps_pool_create(&pool, arena, mps_class_snc(), format);
The extra argument, format, should be of type mps_fmt_t and specifies the format of the objects allocated in the pool (in a similar way to mps_class_amc). The format should provide at least the methods: scan, skip, pad.
An SNC pool is scannable, in that objects may contain references to objects in other pools that will keep those objects alive (depending on rank). In this sense, an SNC pool is a de-facto root.
Exact references may point to (the start of) objects in an SNC pool, but will have no effect on whether those objects are either scanned or kept alive.
If mps_ap_frame_pop is used on an allocation point in an SNC pool (after a corresponding call to mps_ap_frame_push), then the objects affected by the pop are effectively declared dead, and may be reclaimed by the collector. Extant references to such objects from reachable or de facto alive objects are safe, but such other objects should be dead; that is, such references must never be used.
If an allocation point is created in an SNC pool, then the call to mps_ap_create will take as an additional parameter the rank (of type mps_rank_t) of references in the objects to be created in that allocation point. Currently, only rank exact (mps_rank_exact) is supported.
Objects in an SNC pool may not be registered for finalization.
Objects in an SNC pool will not move.
Cannot fail.
mps_class_amc,
mps_ap_frame_pop,
mps_ap_frame_push,
mps_ap_create
mps_class_mvtmps_class_mvt is a function that returns the MVT pool class object.
Allocation point.
mps_class_t mps_class_mvt(void);
C function
None.
The MVT pool class object.
mpscmv2.h
The function mps_class_mvt returns the MVT pool
class object, which can be used to create an MVT pool instance by
passing the class object as the mps_class_t (third) argument to
mps_pool_create.
The MVT pool class manually manages variable-sized, unformatted objects. The MVT pool uses an allocation policy termed "temporal fit". Temporal fit attempts to place consecutive allocations next to each other. It relies on delaying reuse as long as possible to permit freed blocks to coalesce, thus maximizing the number of consecutive allocations that can be co-located. Temporal fit permits a very fast allocator and a deallocator competitive in speed with all other known policies.
Temporal fit is intended to take advantage of knowledge of object lifetimes, either apriori knowledge or knowledge acquired by profiling. The best performance of the MVT pool will be achieved by allocating objects with similar expected deathtimes together.
A simple policy can be implemented to take advantage of MVT: Object size is typically well-correlated with object life-expectancy, and birthtime plus lifetime gives deathtime, so allocating objects of similar size sequentially from the same pool instance should result in objects allocated close to each other dying at about the same time.
An application that has several classes of objects of widely differing life expectancy will best be served by creating a different MVT pool instance for each life-expectancy class. A more sophisticated policy can use either the programmer's knowledge of the expected lifetime of an objector any characteristic of objects that correlates with lifetime to choose an appropriate pool instance to allocate in.
Allocating objects with unknown or very different deathtimes together will pessimize the space performance of MVT.
if(mps_pool_create(&pool, arena, mps_class_mvt(), 8, 32, 256, 70, 20)
!= MPS_RES_OK) {
printf("Error creating pool!");
exit(2);
}
mps_class_mvt cannot result in an error.
Creation
The MVT pool class has five creation parameters:
mps_res_t mps_pool_create(mps_pool_t * pool, mps_arena_t arena, mps_class_t mvt_class, size_t minimum_size, size_t mean_size, size_t maximum_size, mps_count_t reserve_depth mps_count_t fragmentation_limit);
Sizes
minimum_size, mean_size, and maximum_size are the minimum, mean, and maximum (typical) size in bytes of objects expected to be allocated in the pool. Objects smaller than minimum size may be allocated, but the pool is not guaranteed to manage them space-efficiently. Objects larger than maximum_size may be allocated, but the pool is not guaranteed to manage them space-efficiently.Furthermore, partial freeing is not supported for objects larger than maximum size; doing so will result in the storage of the object never being reused. Mean_size need not be an accurate mean,although the pool will manage mean_size objects more efficiently.
Reserve Depth
reserve_depth is the expected hysteresis of the object population. When pool objects are freed, the pool will retain sufficient storage to allocate reserve_depth objects of mean_size for near term allocations (rather than immediately making that storage available to other pools).
If a pool has a stable object population, one which only grows over the lifetime of the pool, or one which grows steadily and then shrinks steadily, use a reserve_depth of 0.
It is always safe to use a reserve depth of 0, but if the object population typically fluctuates in a range (e.g., the client program may repeatedly create and destroy a subset of objects in a loop), it is more efficient for the pool to retain enough storage to satisfy that fluctuation. For example, if a pool has an object population that typically fluctuates between 8,000and 10,000, use a reserve_depth of 2,000.
The reserve will not normally be available to other pools for allocation, even when it is not used by the pool. If this is undesirable, a reserve depth of 0 may be used for a pool whose object population does vary, at a slight cost in efficiency. The reserve does not guarantee any particular amount of allocation.
Fragmentation Limit
fragmentation_limit is a percentage in (0, 100] that can be used to set an upper limit on the space overhead of MVT in case object deathtimes and allocations do not correlate well.
If the free space managed by the pool as a ratio of all the space managed by the pool exceeds the specified percentage, the pool will fall back to a first fit allocation policy,exploiting space more efficiently at a cost in time efficiency.
A fragmentation_limit of 0 would cause the pool to operate as a first-fit pool, at a significant cost in time-efficiency, therefore is not permitted.
A fragmentation_limit of 100 will cause the pool to use temporal fit (unless resources are exhausted). If the objects allocated in the pool have similar lifetime expectancies, this mode will have the best time- and space-efficiency. If the objects have widely varying lifetime expectancies,this mode will be time-efficient, but may be space-inefficient. An intermediate setting can be used to limit the space-inefficiency of temporal fit due to varying object life expectancies.
Allocation
The MVT pool class only supports allocation through allocation points. See mps_ap_create.
Deallocation
The MVT pool class supports explicit freeing. See mps_pool_free.
Need a life-expectancy parameter! How else will different instances choose their Loci?
Need an alignment parameter. Perhaps this is embedded in a format parameter (when all pools have at least a null format).
It is conceivable that a client would want to mix manual and automatic pools with the manual pool being able to be a root for the automatic. To do so, MVT would need to support formatted objects and scanning. This may be added someday.
Eventually the MM product will include profiling tools that will help determine object characteristics that correlate with object lifetime and suggest how to configure the appropriate number of MVT pool instances and what characteristics to dispatch on when choosing which instance to allocate from.
[From mail.ptw.1998-08-19.02-33(0) ]
Remember Wilson's statement that the goal of a memory manager is to exploit the regularities in allocation patterns? My intent in the interface parameters is to accept measurable regularities in object populations, then the implementation can exploit them.
Perhaps the pool should accept some description of the mean and deviation of the object sizes, object population, and object lifetimes. Is that what you are getting at? [Reserve_depth is in some sense a deviation.]
mps_class_tmps_class_t is the type of pool classes.
Pool.
mps.h
mps_class_t is the abstract type of pool classes. It is opaque. A pool class may be obtained by calling the class function for the appropriate class, such as mps_class_amc for the AMC class. A pool class is used when creating a pool with mps_pool_create or mps_pool_create_v.
mps_pool_create,
mps_pool_create_v
mps_class_s is an incomplete structure type used only to define mps_class_t.
mps_finalizeRegisters an object for finalization.
Finalization, message.
mps_res_t mps_finalize(mps_arena_t arena, mps_addr_t *object_ref)
arena
-- the arena in which the object lives
object_ref
-- a pointer to a reference to the object to be finalized
A result code.
mps.h
This function registers the specified object for finalization. This object must be an object allocated from a pool in the specified arena. Violations of this constraint may not be checked by the MPS, and may be unsafe (cause the MPS to crash in undefined ways).
An object becomes finalizable if it is registered for finalization and the collector observes that it would otherwise be reclaimable. Once an object is finalizable the MPS may choose to finalize it (by posting a finalization message, see below) at any future time. Note that the subsequent creation of strong references to the object (from, say, weak references) may cause finalization to occur when an object is not otherwise reclaimable.
When an object is finalizable, it may be finalized up to N times, where N is the number of times it has been registered for finalization. When an object is finalized, it is also deregistered for finalization (so that it will not be finalized again from the same registration).
Finalization is performed by passing a finalization message to the client, containing an exact reference to the object. See the message protocol, mps_message_type_finalization, and mps_message_finalization_ref for details.
If an object is registered for finalization multiple times, then there may be multiple finalization messages on the queue at the same time. On the other hand it may be necessary to discard previous finalization messages for an object before all such messages are posted on the message queue. In other words a finalization message may prevent other finalizations of the same object from occurring until the message is deleted; or, it may not. We don't provide any guarantees either way. Clients performing multiple registrations must cope with both behaviors. In any case we expect it to be unusual for clients to register the same object multiple times.
Note that there is no guarantee that finalization will be prompt.
Weak references do not prevent objects from being finalized. At the point that an object is finalized, weak references will still validly refer to the object. The fact that an object is registered for finalization prevents weak references to that object from being deleted.
Note that there will be no attempt to finalize objects in the context of mps_arena_destroy or mps_pool_destroy. mps_pool_destroy should therefore not be invoked on pools containing objects registered for finalization.
Not all pool classes support finalization of objects. In general only pools that manage objects whose liveness is determined by garbage collection will support finalization of objects. For more information, see the Pool Class Catalog.
[missing]
[missing]
mps_message_type_finalization,
mps_rank_weak,
mps_arena_destroy,
mps_pool_destroy
This function receives a pointer to a reference. This is to avoid placing the restriction on the client that the C call stack be a root.
mps_fmt_A_smps_fmt_A_s is a structure used to create object formats of variant A.
Format.
typedef struct mps_fmt_A_s {
mps_align_t align;
mps_fmt_scan_t scan;
mps_fmt_skip_t skip;
mps_fmt_copy_t copy;
mps_fmt_fwd_t fwd;
mps_fmt_isfwd_t isfwd;
mps_fmt_pad_t pad;
} mps_fmt_A_s;
mps.h
Objects of this type are intended to be used in the creation of object formats. Object formats describe the layout of client objects.
mps_fmt_A_s is a structure that represents the particular collection of methods and values that describes an object format of variant A.
Broadly speaking, the object formats of this variant are suitable for use in copying or moving memory managers.
mps_fmt_A_s has the
following methods: scan, skip, copy, fwd, isfwd, pad, and the following value:align.
align is an integer value defines the alignment of objects allocated with this format. It should be large enough to satisfy the alignment requirements of any field in the objects,and it cannot be larger than the arena alignment. For details of the methods, consult the reference pages for the type of each method.
mps_fmt_t create_format(mps_arena_t arena)
{
mps_fmt my_format;
mps_res_t res;
mps_fmt_A_s my_format_A = { my_alignment, &my_scan, &my_skip, &my_copy, &my_fwd,
&my_isfwd, &my_pad };
res = mps_fmt_create_A(&my_format, arena, &my_format_A);
assert(res != MPS_RES_OK);
return my_format;
}
mps_fmt_create_A,
mps_fmt_scan_t,
mps_fmt_skip_t
mps_fmt_copy_t,
mps_fmt_fwd_t,
mps_isfwd_t,
mps_pad_t,
mps_align_t,
mps_fmt_B_s
mps_fmt_A_tmps_fmt_A_t is the type pointer to mps_fmt_A_s.
Format.
typedef struct mps_fmt_A_s {
mps_align_t align;
mps_fmt_scan_t scan;
mps_fmt_skip_t skip;
mps_fmt_copy_t copy;
mps_fmt_fwd_t fwd;
mps_fmt_isfwd_t isfwd;
mps_fmt_pad_t pad;
} mps_fmt_A_s;
typedef struct mps_fmt_A_s *mps_fmt_A_t;
mps.h
mps_fmt_A_t is the type pointer to mps_fmt_A_s. A value of this type represents a collection of methods and values that can be used to create a format object of type mps_fmt_t. This type represents a particular collection of methods and values; other collections are represented by other types.
Objects of type mps_fmt_A_t are intended to be used in the creation of object formats.Object formats describe the layout of client objects. The function mps_fmt_create_A takes an mps_fmt_A_t as one of its arguments and creates an object of type mps_fmt_t (an object format).
See the documentation of mps_fmt_A_s for further details.
mps_fmt_A_s,
mps_fmt_t,
mps_fmt_create_A
mps_fmt_B_smps_fmt_B_s is a transparent structure used to create object formats of variantB.
Format.
typedef struct mps_fmt_B_s {
mps_align_t align;
mps_fmt_scan_t scan;
mps_fmt_skip_t skip;
mps_fmt_copy_t copy;
mps_fmt_fwd_t fwd;
mps_fmt_isfwd_t isfwd;
mps_fmt_pad_t pad;
mps_fmt_class_t mps_class;
} mps_fmt_B_s;
mps.h
Objects of this type are intended to be used in the creation of object formats. Object formats describe the layout of client objects. mps_fmt_B_s is a structure that represents the particular collection of methods and values that describes an object format of variant B.
mps_fmt_B_s is the same as mps_fmt_A_s except for the addition of the mps_class method. Broadly speaking, the object formats of variety B are suitable for use in copying or moving memory managers (just like variety A); the addition of the class method allows more information to be passed to various support tools (such as graphical browsers).
mps_fmt_B_s has the following methods: scan, skip, copy, fwd, isfwd, pad, mps_class, and the following value: align.
align is an integer value defines the alignment of objects allocated with this format. It should be large enough to satisfy the alignment requirements of any field in the objects, and it cannot be larger than the arena alignment. For details of the methods, consult the reference pages for the type of each method.
mps_fmt_t create_format(mps_arena_t arena)
{
mps_fmt_B_s my_format_B = { my_alignment, &my_scan, &my_skip, &my_copy,
&my_fwd, &my_isfwd, &my_pad, &my_class };
mps_fmt my_format;
mps_res_t res;
res = mps_fmt_create_B(&my_format, arena, &my_format_B);
assert(res != MPS_RES_OK);
return my_format;
}
mps_fmt_create_B,
mps_fmt_scan_t,
mps_fmt_skip_t,
mps_fmt_copy_t,
mps_fmt_fwd_t,
mps_isfwd_t,
mps_pad_t,
mps_align_t,
mps_class_t,
mps_fmt_A_s
The mps_class field used to be called "class", but that was problematic for C++, so we changed it.
mps_fmt_B_tmps_fmt_B_t is a type passed to mps_fmt_create_B. It represents the collection of methods and values used to create a mps_fmt_t. You are expected to declare and create structures of this type if you require an object of type mps_fmt_B_t.
Format
typedef struct mps_fmt_B_s {
mps_align_t align;
mps_fmt_scan_t scan;
mps_fmt_skip_t skip;
mps_fmt_copy_t copy;
mps_fmt_fwd_t fwd;
mps_fmt_isfwd_t isfwd;
mps_fmt_pad_t pad;
mps_fmt_class_t class;
} mps_fmt_B_s;
typedef struct mps_fmt_B_s *mps_fmt_B_t;
mps.h
mps_fmt_B_t is the equivalent to mps_fmt_A_t that should be passed tomps_fmt_create_B. It is suitable for format variety A collectors that need to use tools that useclass information.
See the documentation for the symbol mps_fmt_B_s for further details.
mps_fmt_B_s,
mps_fmt_t,
mps_fmt_create_B,
mps_fmt_A_t
None.
mps_fmt_auto_header_smps_fmt_auto_header_s is a structure used to create object formats of variant auto_header.
Format.
typedef struct mps_fmt_auto_header_s {
mps_align_t align;
mps_fmt_scan_t scan;
mps_fmt_skip_t skip;
mps_fmt_fwd_t fwd;
mps_fmt_isfwd_t isfwd;
mps_fmt_pad_t pad;
size_t mps_headerSize;
} mps_fmt_auto_header_s;
mps.h
Objects of this type are intended to be used in the creation of object formats. Object formats describe the layout of client objects. mps_fmt_auto_header_s isa structure that represents the particular collection of methods and values that describes an object format of variant auto_header.
Broadly speaking, the object formats of this variant are suitable for use in automatic memory management for objects with headers (hence the name). More precisely, this variant is intended for formats where the client's pointers point some distance into the memory block containing the object. This typically happens when the objects have a common header used for memory management or class system purposes, but this situation also arises when the low bits of a pointer are used for a tag. The MPS does not care what the reason is, only about the offset of the pointer in relation to the memory block.
mps_fmt_auto_header_shas the following methods: scan, skip, fwd, isfwd, pad, and the following values: align and mps_headerSize.
align is an integer value defines the alignment of objects allocated with this format. It should be large enough to satisfy the alignment requirements of any field in the objects, and it cannot be larger than the arena alignment.
mps_headerSize is the size of the header, i.e., the offset of a client pointer from the base the memory block. For details of the methods, consult the reference pages for the type of each method.
mps_fmt_t create_format(mps_arena_t arena)
{
mps_fmt format;
mps_res_t res;
mps_fmt_auto_header_s format_desc = { my_alignment, &my_scan, &my_skip, &my_fwd,
&my_isfwd, &my_pad, HEADER_SIZE };
res = mps_fmt_create_auto_header(&format, arena, &format_desc);
assert(res != MPS_RES_OK);
return format;
}
mps_fmt_create_auto_header,
mps_fmt_scan_t,
mps_fmt_skip_t,
mps_fmt_fwd_t,
mps_isfwd_t,
mps_pad_t,
mps_align_t,
mps_fmt_A_s
For technical reasons, client objects must be longer than the header, i.e., objects consisting of only a header are not supported. However, if the header size is larger than or equal to alignment, the pad method must still be able to create padding objects down to alignment size.
At the moment, this format only works with pool classes AMC and AMCZ.
mps_fmt_class_tmps_fmt_class_t is a function pointer type for the class method of a format.
Format. Telemetry.
typedef mps_addr_t (*mps_fmt_class_t)(mps_addr_t addr);
addr the address of the object whose class is of interest
Returns an address that the client associates with the class or type of the object.
mps.h
mps_fmt_class_t is t he type of a format's class method. A class method returns an address that is related to the class of the object, for passing on to various support tools (such as graphical browsers).
A class method is provided by the client as part of a format (see Format Protocol).
The exact meaning of the return value is up to the client, but it would typically bear some relation to class or type in the client program. The client may have objects that represent classes or types. These may be associated with strings via mps_telemetry_intern and mps_telemetry_label.
mps_addr_t my_class_method(mps_addr_t object) {
my_object_generic_t generic_object = object;
return (mps_addr_t)(generic_object.class);
}
A class method is not allowed to fail, but may return NULL.
It is recommended that NULL be returned for padding objects and forwarded objects.
mps_fmt_copy_tmps_fmt_copy_t is a function pointer type for the copy method of a format. [Obsolete. RHSK 2006-06-06]
Format.
typedef void (*mps_fmt_copy_t)(mps_addr_t old, mps_addr_t new);
old -- the address of the object
new -- the address to which the object should be copied
mps.h
[Note: mps_fmt_copy_t is obsolete: the MPS does not call this format method. The MPS simply copies all the bytes to the new location (using the length reported by the skip format method). RHSK 2006-06-06]
mps_fmt_copy_t is a function pointer type for the copy method of a format. A copy method copies an object to a new location. It may be called by the MPS as part of copying garbage collection, for example.
A copy method is required in some formats (in particular formats A and B (see mps_fmt_A_t and mps_fmt_B_t)). A copy method takes the address of an object and another address, and copies the object to the new address. The new and the old locations are guaranteed not to overlap.
void my_copy_method(mps_addr_t old, mps_addr_t new)
{
size_t length = (char*)my_skip_method(old) - (char *)old;
memcpy(new, old, length);
}
A copy method is not allowed to fail.
mps_fmt_t,
mps_fmt_create_A,
mps_fmt_A_t,
mps_fmt_B_t,
mps_fmt_create_B
Most pools will just ignore Copy methods, and do the copy themselves.
mps_fmt_create_AFunction for create a format of variety A.
Format.
mps_res_t mps_fmt_create_A(mps_fmt_t *fmt_o, mps_arena_t arena, mps_fmt_A_s *fmt_A);
fmt_o
- the address of a variable to hold the new format
arena
- the arena in which to create the format
fmt_A
- format description of variety A
Result status. If the return value is MPS_RES_OK, the new format is in *fmt_o.
mps.h
This function creates a format from a user format specification of variety A.
mps_fmt_t create_format(mps_arena_t arena)
{
mps_fmt_A_s my_format_A = { my_alignment, &my_scan, &my_skip, &my_copy,&my_fwd,
&my_isfwd, &my_pad };
mps_fmt my_format;
mps_res_t res;
res = mps_fmt_create_A(&my_format, arena, &my_format_A);
if(res != MPS_RES_OK) {
fprintf(stderr, "Couldn't create format.\n");
exit(1);
}
return my_format;
}
The MPS may exhaust some resource in the course of mps_fmt_create_A and will return an appropriate error code in such circumstances.
mps_fmt_A_s,
mps_fmt_t,
mps_fmt_create_B
mps_fmt_create_BFunction for create a format of variety B.
Format.
mps_res_t mps_fmt_create_B(mps_fmt_t *fmt_o, mps_arena_t arena, mps_fmt_B_s *fmt_B);
arena - the arena in which to create the format
fmt_B - format description of variety B
Result status. If the return value is MPS_RES_OK, the new format is in*fmt_o.
mps.h
This function creates a format from a user format specification of variety B. It is very similar to mps_fmt_create_A.
mps_fmt_t create_format(mps_arena_t arena)
{
mps_fmt_B_s my_format_B = { my_alignment, &my_scan, &my_skip, &my_copy,
&my_fwd, &my_isfwd, &my_pad, &my_class };
mps_fmt my_format;
mps_res_t res;
res = mps_fmt_create_B(&my_format, arena, &my_format_B);
assert(res != MPS_RES_OK);
return my_format;
}
The MPS may exhaust some resource in the course of mps_fmt_create_B and will return an appropriate error code in such circumstances.
mps_fmt_B_s,
mps_fmt_t,
mps_fmt_create_A
mps_fmt_create_auto_headerFunction for create a format of variety auto_header.
Format.
mps_res_t mps_fmt_create_auto_header(mps_fmt_t *fmt_o, mps_arena_t arena, mps_fmt_auto_header_s *fmt_st);
fmt_o
- the address of a variable to hold the new format
arena
- the arena in which to create the format
fmt_st
- format description of variety auto_header
Result status. If the return value is MPS_RES_OK, the new format is in *fmt_o.
mps.h
This function creates a format from a user format specification of variety auto_header.
mps_fmt_t create_format(mps_arena_t arena)
{
mps_fmt_auto_header_s format_desc = { my_alignment, &my_scan, &my_skip, &my_fwd,
&my_isfwd, &my_pad, HEADER_SIZE };
mps_fmt format;
mps_res_t res;
res = mps_fmt_create_auto_header(&format, arena, &format_desc);
assert(res != MPS_RES_OK);
return format;
}
The MPS may exhaust some resource in the course of mps_fmt_create_auto_headerand will return an appropriate error code in such circumstances.
mps_fmt_auto_header_s,
mps_fmt_t,
mps_fmt_create_A
mps_fmt_fwd_tThe type of a format's forward method.
Format.
typedef void (*mps_fmt_fwd_t)(mps_addr_t old, mps_addr_t new);
old
the address of an object
new
the address where the object has been moved
None.
mps_fmt_fwd_t is the type of a format's forward method. A forward method is used to store relocation information in a heap. It may be called by the MPS as part of copying garbage collection.
A forward method is provided by the client as part of a format (see Format Protocol ). TheMPS calls a forward method when it has relocated an object. The forward method when called must replace the object at 'old' with a forwarding marker that points to the address 'new'. The forwarding marker must meet the following requirements:
it must be possible for the MPS to call other format methods with the address of a forwarding marker as the argument.
he forwarding marker must not be bigger than the original object.
t must be possible to distinguish the forwarding marker from ordinary objects using the isfwd method (see mps_fmt_isfwd_t), and the isfwd method must return the address'new'.
/* define the function */
void example_fwd(mps_addr_t old, mps_addr_t new)
{
/* ... */
}
/* also define example_scan, example_skip, etc */
/* store pointer to function in the format variant struct */
struct mps_fmt_B_s example_fmt_B = {
4, /* align */
example_scan,
example_skip,
example_copy,
example_fwd,
example_isfwd,
example_pad,
example_class
};
/* The (address of the) example_fmt_B object can now be passed to */
/* mps_fmt_create_B to create a format. */
mps_fmt_A_s,
mps_fmt_B_s,
mps_fmt_auto_header_s,
mps_fmt_isfwd_t
This method is never invoked by the GC on an object in a non-moving pool.
mps_fmt_isfwd_tThe type of a format's isfwd ("is forwarded") method.
Format.
typedef mps_addr_t (*mps_fmt_isfwd_t)(mps_addr_t addr);
addr
the address of a candidate object
Either a null pointer to indicate the object at
addr
has not been relocated, orthe new location of the object if there is a forwarding marker at
addr
indicating thatthe object has been relocated.
The type of a format's isfwd ("is forwarded") method. An isfwd method is used to test whether an object has been relocated using the format's forward method.
An isfwd method is provided by the client as part of a format (see protocol.mps.format(0) ).The MPS calls the isfwd method to determine whether an object in the heap has been relocated or not.Objects in the heap are relocated using the format's forward method (see mps_fmt_fwd_t). When the isfwd method is called the parameter addr will be the address of either an object or a forwarding marker created with the forward method. If it is an object (so it has not been relocated)the method should return a null pointer; otherwise it is a forward marker indicating the address of the relocated object, the address of the relocated object should be returned (this should be the same as the 'new' parameter that was passed to the forward method that created the forwarding marker).
<example of how to use the symbol>
mps_fmt_A_s,
mps_fmt_B_s,
mps_fmt_auto_header_s,
mps_fmt_fwd_t
This method is never invoked by the GC on an object in a non-moving pool.
mps_fmt_pad_tThe type of a format's pad method.
Format.
typedef void (*mps_fmt_pad_t)(mps_addr_t addr, size_t size);
addr
The address at which to create a padding object.
size
The size (in bytes) of the padding object to be created.
None.
The type of a format's pad method. A pad method is used to create padding objects.
A pad method is provided by the client as part of a format (see Format Protocol ). The MPS calls a pad method when it wants to create a padding object. Typically the MPS creates padding objects to fill in otherwise unused gaps in memory; they allow the MPS to pack objects in fixed-size units (such as OS pages). The pad method should create a padding object of the specified size at the specified address. The size can be any aligned (to the format alignment) size. A padding object should be acceptable to other methods in the format (scan, skip, isfwd, etc.).
<example of how to use the symbol>
mps_fmt_scan_tType of the scan method of a format.
Format, Scanning.
typedef mps_res_t (*mps_fmt_scan_t)(mps_ss_t scan_state, mps_addr_t base, mps_addr_t limit)
scan_state
a scan state
base
a client pointer to the first object in the block to be scanned
limit
a client pointer to the object just beyond the end of the block
A result code.
mps.h
This is the type of scanning functions provided by the client in some format variants and mps_root_create_fmt. When the MPS needs to scan objects in an area of memory that this scanning function has been registered for, it will be called with a scan state and the limits of the block of objects to scan. It must then indicate references within the objects by usingmps_fix or one of the alternatives.
The base and limit arguments are client pointers, as usual. Note that there might not be any object at the location indicated by limit.
/* Scanner for a simple Scheme-like language with just two interesting types */
mps_res_t scan_objs(mps_ss_t ss, mps_addr_t base, mps_addr_t limit)
{
mps_res_t res;
mps_addr_t obj;
MPS_SCAN_BEGIN(ss)
for(obj = base; obj < limit;) { /* obj maps over the objects to scan */
switch(((Object*)obj)->type) {
case ArrayType:
{
size_t i;
Array *array = (Array *)obj;
for(i = 0; i < array->length; ++i) { /* fix each element */
res = MPS_FIX12(ss, &array->contents[i]);
if(res != MPS_RES_OK) return res;
}
obj = AddrAdd(obj, ArraySize(array)); /* move to next object */
break;
}
case StackFrameType:
{
StackFrame *frame = (StackFrame *)obj;
for(i = frame->size; i > 0; --i) { /* fix each local var */
res = MPS_FIX12(ss, &frame->locals[i]);
if(res != MPS_RES_OK) return res;
}
res = MPS_FIX12(ss, &frame->next);
if(res != MPS_RES_OK) return res;
obj = AddrAdd(obj, StackFrameSize(frame));
break;
}
default: /* other types don't contain references */
obj = AddrAdd(obj, DefaultSize(obj));
break;
}
}
MPS_SCAN_END(ss);
return res;
}
If a fixing operation returns a value other than MPS_RES_OK, the scanning function must
return that value, and may return without scanning further
references. Generally, itis better if it returns as soon as
possible. If the scanning is completed successfully, the
function should return MPS_RES_OK.
mps_fmt_A_s,
mps_fmt_B_s,
mps_fmt_auto_header_s,
mps_root_create_fmt,
mps_fix,
MPS_FIX12,
MPS_FIX1,
MPS_FIX2,
MPS_FIX_CALL,
MPS_SCAN_BEGIN,
MPS_SCAN_END
mps_fmt_skip_tmps_fmt_skip_t is a function pointer type for the skip method of a format.
Format.
typedef mps_addr_t (*mps_fmt_skip_t)(mps_addr_t obj);
obj
the client pointer to the object to be skipped
The skip method should return the address of the next object.
mps.h
mps_fmt_skip_t is a function pointer type for the skip method of a format.
These methods are provided by the client as part of a format and invoked by the MPS (seeFormat Protocol). The skip method takes the client pointer to the object. The method should return the client pointer to the next object, whether there is one or not. With no headers, this is the address just past the end of this object; with headers, it's the address just past where the header of next object would be. It is always the case that the difference between the argument and the return value is the size of the block containing the object.
mps_addr_t my_skip_method(mps_addr_t object)
{
char *p = (char *)object;
my_object_t my_object = (my_object_t)object;
return((mps_addr_t)(p + my_object->length));
}
A skip method is not allowed to fail.
mps_fmt_A_s,
mps_fmt_B_s,
mps_fmt_auto_header_s
mps_fmt_tmps_fmt_t is the type of object formats.
Format.
typedef struct mps_fmt_s *mps_fmt_t;
mps_fmt_s is an incomplete structure type used only to declare the opaque type mps_fmt_t.
mps.h
mps_fmt_t is the opaque type of object formats. An object format is a way for the MPS and client programs to communicate regarding the layout of client objects. For more information, seeFormat Protocol.
#include "mps.h"
#include "mpscamc.h"
#include <stdlib.h>
struct mps_fmt_A_s fmt_A_s = {
(mps_align_t)4,
scan, skip, copy, move, isMoved, pad
};
void go(mps_space_t space)
{
mps_fmt_t format;
mps_res_t res;
mps_pool_t pool;
res = mps_fmt_create_A(&format, space, &mps_fmt_A_s);
if(res != MPS_RES_OK)
abort();
res = mps_pool_create(&pool, space, mps_class_amc(), format);
if(res != MPS_RES_OK)
abort();
/* do some stuff here */
mps_pool_destroy(pool);
mps_format_destroy(format);
}
mps_fmt_create_A,
mps_fmt_create_B,
mps_fmt_destroy,
mps_fmt_A_t
mps_formatted_objects_stepper_tmps_formatted_objects_stepper_t
Type of the client supplied heap walker function.
Heap walking.
typedef void (*mps_formatted_objects_stepper_t)(mps_addr_t, mps_fmt_t, mps_pool_t, void *,size_t )
mps_formatted_objects_stepper
is a type not a function so it doesn't take any arguments; however the
function pointed to by an object of this type does. Such functions
take the following argument list:
(mps_addr_t object, mps_fmt_t format, mps_pool_t pool, void *p, size_t s)
object is a pointer to the (client) object.
format is the MPS format of the client object.
pool in the MPS pool in which the client object resides.
p and s are two closure values which are copies of the corresponding values which the client
passed into the heap walking function, mps_arena_formatted_objects_walk.
The function pointed to by an object of type mps_formatted_objects_stepper_t returns no
arguments.
mps.h
This symbol describe the type of pointers passed into the heap
walking function
mps_arena_formatted_objects_walk.
The heap walker arranges to apply this
function to all objects on the heap, see
mps_arena_formatted_objects_walk
for details.
<example of how to use the symbol>
The function pointed to by an object of type mps_formatted_objects_stepper_t have no way to
return an error code to the caller.
mps_arena_formatted_objects_arena_walk
mps_freeFrees a block of memory to a pool.
Allocation
void mps_free(mps_pool_t pool, mps_addr_t p, size_t size);
pool the pool of the object to be freed
p a pointer to the object to the freed
size the size of the object to the freed in bytes
None.
mps.h
Frees an object of memory, returning the memory block to the pool it was allocated from.The pool might then decide to make it available to other pools, but the way this happens depends onthe pool class and the current situation.
mps_free takes a size argument, because it is most efficient to do so. In practical programs, the type of an object is usually known at the point in the code that calls the deallocation function, and hence the size is trivially available. In such cases. storing the size on the MPS side would cost time and memory, and make it hard to get good virtual memory behaviour (as it is, the deallocation code doesn't have to touch the dead object at all).
Undoubtedly, one day, we'll get around to writing a pool that stores the size of each object.
mps_lib_memcmpA plinth function similar to C's "memcmp".
Plinth
int mps_lib_memcmp(const void *s1, const void *s2, size_t n);
s1, s2 pointers to memory blocks to be compared
n length of the blocks, in bytes
An integer that is greater than, equal to, or less than zero, accordingly as the block pointed to by "s1" is greater than, equal to, or less than the block pointer to by "s2".
mpslib.h
This function is intended to have the same semantics as the "memcmp" function of the ANSI C standard (section 7.11.4.1).
Like other plinth features, it is used by the MPS and provided by the client (possibly using the ANSI plinth, mpsliban.c).
None, clients don't use it.
None.
mps_lib_memset,
mps_lib_memcpy,
mpsliban.c
None.
mps_lib_memcpyA plinth function similar to C's "memcpy".
Plinth
void *mps_lib_memcpy(void *dest, const void *source, size_t n);
dest destination of copy
source source of copy
n length of the blocks, in bytes
Returns the value of the dest argument.
mpslib.h
This function is intended to have the same semantics as the "memcpy" function of the ANSI C standard (section 7.11.2.1).
Like other plinth features, it is used by the MPS and provided by the client (possibly using the ANSI plinth, mpsliban.c).
None, clients don't use it.
None.
mps_lib_memset,
mps_lib_memcmp,
mpsliban.c
None.
mps_lib_memsetA plinth function similar to C's "memset".
Plinth
void *mps_lib_memset(void *s, int c, size_t n);
s destination of copy
c byte (when converted to an unsigned char) to copy
n length of the block, in bytes
Returns the value of s.
mpslib.h
This function is intended to have the same semantics as the "memset" function of the ANSI C standard (section 7.11.6.1).
Like other plinth features, it is used by the MPS and provided by the client (possibly using the ANSI plinth, mpsliban.c).
None, clients don't use it.
None.
mps_lib_memcpy,
mps_lib_memcmp,
mpsliban.c
None.
mps_lib_telemetry_controlPlinth function to supply a default value for telemetry filters from environment.
Telemetry
unsigned long mps_lib_telemetry_control();
None.
In the absence of environmental data, a default of zero is recommended.
The default value of the telemetry filter, as derived from the environment. It is recommended that the environment be consulted for a symbol analogous to MPS_TELEMETRY_CONTROL, subject to local restrictions.
Depends on access to the environment.
See mps_telemetry_control for more information on the significant of the values.
See the supplied ANSI plinth for an example implementation.
mps_message_clockmps_message_clock
returns the time at which the MPS posted the message (only for certain message types).
Message.
mps_clock_t mps_message_clock(mps_arena_t arena, mps_message_t message)
arena
-- the arena
message
-- any message retrieved with mps_message_get and not yet discarded
For supported message types: the time at which the MPS posted the message. For other message types: zero.
mps.h
Messages are asynchronous: they are posted by the MPS, wait on a queue, and are later collected by the client. Each message (of supported types) records the time that it was posted, and this is what mps_message_clock returns.
The time returned is the mps_clock_t value returned by the library function mps_clock at the time the message was posted. You can subtract one clock value from another to get the time interval between the posting of two messages.
Only the following supported message types record the time of posting:
For other message types, the value returned is always zero.
mps_message_t message;
mps_clock_t posted_at;
if(mps_message_get(&message, arena, mps_message_type_gc_start())) {
posted_at = mps_message_clock(arena, message);
printf("Collection started at %ul.\n", (unsigned long)posted_at);
}
Can't fail.
The example ANSI plinth, mpsliban.c, implements mps_clock by calling ISO C time.h's clock(). The difference between two of these clock values may be converted to seconds by dividing by ISO C time.h's CLOCKS_PER_SEC conversion factor.
mps_message_discardmps_message_discard
is used to indicate that the client is done with the message and the MPS
can now reclaim any storage associated with the message.
Message.
void mps_message_discard(mps_arena_t arena, mps_message_t message)
arena
-- the arena
message
-- the message
None.
mps.h
mps_message_discard
is used to indicate that the client has no further use for the specified
message in the specified arena. After this call, the message is invalid
and should not be passed as argument to any message functions.
Messages are essentially manually managed. This call allows the MPS to reclaim storage associated with messages. If the client does not discard their messages then the resources used may grow without bound.
As well as consuming resources, messages may have other visible effects that require them to be tidied by calling this function. In particular finalization messages refer to their finalized object, and will prevent the object from being reclaimed (subject to the usual garbage collection liveness analysis). A finalized object cannot possibly be reclaimed until its corresponding finalization messages have been discarded (all such messages in the case of multiple messages for the same object).
[missing]
Can't fail.
None.
mps_message_finalization_refmps_message_finalization_ref returns the "finalization reference" property of the specified message in the specified arena.
Message, finalization.
void mps_message_finalization_ref(mps_addr_t *object_ref, mps_arena_t arena, mps_message_tmessage)
object_ref -- a pointer to a reference to the object to which the message pertains
arena -- the arena that the message is in
message -- a message of a message type that supports this method
None.
mps.h
This method returns the "finalization reference" property of the specified message in the specified arena. The message must be of a message type that supports this method; currently, the only such type is that of finalization messages, as returned by mps_message_type_finalization.
The reference returned by this method is a reference to the object
that was originally registered for finalization (by a call to mps_finalize).
Note that the reference returned is subject to the normal constraints, such as might be imposed by a moving collection, if appropriate. For this reason, it is returned indirectly via "object_ref" to enable the client to place it directly into scanned memory, without imposing the restriction that the C stack be a root.
The message itself is not affected by invoking this method. Until
the client calls mps_message_discard to discard
the message it will refer to the object and prevent its reclamation.
mps_message_*,
mps_message_type_finalization,
mps_finalize
mps_message_gc_condemned_sizemps_message_gc_condemned_size returns the "condemned size" property of the specified message in the specified arena.
Message, GC.
size_t mps_message_gc_condemned_size(mps_arena_t arena, mps_message_tmessage)
arena-- the arena
message-- a message of a message type that supports this method
An approximate size for the set of objects condemned in the collection that generated the message.
mps.h
Currently, the only type of message that supports this property is
mps_message_type_gc,
such messages are generated whenever a garbage collection completes. This
method returns an approximation to the size of the set of objects that
were condemned in that collection.
mps_message_*
mps_message_gc_live_sizemps_message_gc_live_size returns the "live size" property of the specified message in the specified arena.
Message, GC.
size_t mps_message_gc_live_size(mps_arena_t arena, mps_message_t message)
arena -- the arena;
message -- a message of a message type that supports this method.
The total size of the condemned objects that survived the collection that generated the message.
mps.h
Currently, the only type of message that supports this property is mps_message_type_gc, such messages are generated whenever a garbage collection completes. This method returns the size of the set of objects that were condemned in that collection, but survived.
mps_message_*
mps_message_gc_not_condemned_sizemps_message_gc_not_condemned_size
mps_message_gc_not_condemned_size returns the "not condemned size" property of the specified message in the specified arena.
Message, GC.
size_t mps_message_gc_not_condemned_size(mps_arena_t arena, mps_message_t message)
arena -- the arena
message -- a message of a message type that supports this method
An approximate size for the set of objects that were in collected pools, but were not condemned in the collection that generated the message.
mps.h
Currently, the only type of message that supports this property is mps_message_type_gc; such messages are generated whenever a garbage collection completes. This method returns an approximation to the size of the set of objects that were in collected pools (so potentially subject to garbage collection), but were not condemned in that collection.
mps_message_*
mps_message_gc_start_whymps_message_gc_start_why returns the a
string that describes why a particular collection started.
Message, GC.
const char * mps_message_gc_start_why(mps_arena_t arena, mps_message_t message)
arena -- the arena
message -- a message of a message type that supports
this method (mps_message_type_gc_start())
A pointer to a string that is a textual explanation of why this collection is starting.
mps.h
Currently, the only type of message that supports this property is
mps_message_type_gc_start;
such messages are generated whenever a garbage collection starts.
This method returns a string describing why the collection started.
The contents of the string must not be modified by the client. The
string and the pointer are only valid as long as the message has not
been discarded (with mps_message_discard).
mps_message_*
mps_message_getGets a message of the specified type from a message queue.
Message.
mps_bool_t mps_message_get(mps_message_t *message_return, mps_arena_t arena, mps_message_type_tmessage_type)
message_return
-- the handle to the message that was removed from the queue
arena
-- the arena
message_type
-- the type of message
Returns true if a message has been removed from the queue, false if not.
mps.h
If there is a message of the specified type on the message queue of the specified arena,then this function removes one such message from the queue, returns a handle to it via themessage_return argument, and returns true. Otherwise it returns false.
mps_message_*
mps_message_pollmps_message_poll determines whether there are currently any messages on a message queue.
Message.
mps_bool_t mps_message_poll(mps_arena_t arena)
arena
-- the arena whose message queue you are interested in
A flag to indicate whether there are any messages on the queue.
mps.h
mps_message_poll is used to determine whether there are currently any messages on the message queue of the specified arena.
[missing]
Can't fail.
If you expect a particular type of message, it is usually more practical to just call mps_message_get.
mps_message_queue_typemps_message_queue_type returns the type of the first message on a message queue.
Message.
mps_bool_t mps_message_queue_type(mps_message_type_t *message_type_return, mps_arena_t arena)
message_type_return -- the type of the first message on the queue of the specified arena
arena -- the arena
"True" if there are any messages on the queue of the specified arena, "false" if not.
mps.h
If there are any messages on the queue of the specified arena, then this function returns"true", and also returns the type of the first message via "message_type_return". Otherwise it returns "false".
mps_message_*
mps_message_tmps_message_t is used as a handle on an individual message.
Message.
typedef struct mps_message_s *mps_message_t
mps_message_s is an incomplete structure type used only to declare the opaque type mps_message_t.
mps.h
The opaque type mps_message_t is used as a handle on an individual message. Messages are manually managed. They are created at the instigation of the MPS (but see mps_message_type_enable), and are deleted by the client.
An mps_message_t is a reference into MPS managed memory, and can safely be stored as such in scannable memory.
Not applicable.
mps_message_*
mps_message_typemps_message_type returns the type of a message.
Message.
mps_message_type_t mps_message_type(mps_arena_t arena, mps_message_t message)
arena -- the arena containing the message
message -- a valid message; that is, one previously returned by mps_message_get, and notdiscarded via mps_message_discard
The type of the specified message.
mps.h
mps_message_type returns the type of a message.
mps_message_type_disablemps_message_type_disable restores the arena to the default state whereby messages of thespecified type are not generated.
This reverses the effect of an earlier call to "m ps_message_type_enable".
Message.
void mps_message_type_disable(mps_arena_t arena, mps_message_type_t message_type)
arena -- the arena
message_type -- the message type to be disabled
None.
mps.h
This procedure may be used by the client to specify that messages of the specified type should not created for the specified arena.
Messages are not generated by default, but the client may enable the generation of messages with mps_message_type_enable.
Any existing messages of the specified type are flushed from the message queue.
[none]
Never fails.
mps_message_*
It is permitted to call this function when the message type is already disabled. Such a call will have no effect.
mps_message_type_enablemps_message_type_enable allows messages of the specified type to be created for thespecified arena. Without such enabling, the MPS will, by default, not generate any messages of thattype.
Message.
void mps_message_type_enable(mps_arena_t arena, mps_message_type_t message_type)
arena -- the arena
message_type -- the message type to be enabled
None.
mps.h
This procedure may be used by the client to specify that messages of the specified type maybe created for the specified arena. Without such enabling, the MPS will by default not generate any messages of that type.
Note that the enabling of messages of a particular type implies that the client application will handle and discard message of that type, or the message queue may consume unbounded resources.
The client may disable message generation again by means of an equivalent call to mps_message_type_disable.
[none]
Never fails.
mps_message_*
"Message Protocol"
It is permitted to call this function when the message type is already enabled. Such a call will have no effect.
mps_message_type_finalizationmps_message_type_finalization returns the type of finalization messages.
Message, Finalization.
mps_message_type_t mps_message_type_finalization(void)
None.
The type of finalization messages.
Not applicable.
mps_message_type_finalization
returns the type of finalization messages. Finalization messages are
used by the MPS to implement finalization (see mps_finalize). When the MPS detects that an
object is finalizable, it finalizes the object by posting a message of
this type (note that there might be delays between the object becoming
finalizable, the MPS detecting that, and the message being posted).
In addition to the usual methods applicable to messages, finalization
messages support the mps_message_finalization_ref
method which returns a reference to the object that was registered for
finalization.
{
mps_message_type_t type;
if(mps_message_queue_type(&type, arena)) {
if(type == mps_message_type_finalization()) {
process_finalization_message_from_queue();
} else {
unknown_message_type();
}
}
}
mps_message_*,
mps_finalize
mps_message_type_gcmps_message_type_gc returns the type of garbage collection statistic messages.
Message.
mps_message_type_t mps_message_type_gc(void)
None.
The type of garbage collection statistic messages.
mps.h
mps_message_type_gc returns the type of garbage collection statistic messages. Garbage collection statistic messages are used by the MPS to give the client information about garbage collections that have occurred. Such information may be useful in analysing the client's memory usage over time.
The access methods specific to a message of this type are:
mps_message_gc_live_size -- gives the total size of the condemned objects that survived the collection that generated the message
mps_message_gc_condemned_size
-- gives an approximate size for the set of objects condemned in the collection that generated the message.
mps_message_gc_not_condemned_size -- gives an approximate size for the set of objects that were in collected pools, but were not condemned in the collection that generated the message.
{
mps_message_t message;
if(mps_message_get(&message, arena, mps_message_type_gc())) {
size_t live, condemned, not_condemned;
live = mps_message_gc_live_size(arena, message);
condemned = mps_message_gc_condemned_size(arena, message);
not_condemned = mps_message_gc_not_condemned_size(arena,message);
mps_message_discard(arena, message);
process_collection_stats(live, condemned, not_condemned);
}
}
Cannot fail.
mps_message_*.
mps_message_type_gc_start
mps_message_type_gc_start
returns the type of garbage collection start messages.
Message.
mps_message_type_t mps_message_type_gc_start(void)
None.
The type of garbage collection start messages.
mps.h
mps_message_type_gc_start
returns the type of garbage collection start messages.
The messages contain information about why the collection started. See
mps_message_gc_start_why.
The access methods specific to a message of this type are:
mps_message_gc_start_why --
Returns a string that is a description of why the collection started.
{
mps_message_t message;
if(mps_message_get(&message, arena, mps_message_type_gc_start())) {
printf("Collection started; reason: %s\n",
mps_message_gc_start_why(arena, message));
}
}
Cannot fail.
mps_message_*.
mps_message_type_tmps_message_type_t is the type of message types.
Message.
mps.h
mps_message_type_t is the type whose values are the various message types. It is opaque.
mps_message_*
mps_pool_check_fencepostsCheck all the fenceposts in the pool.
Debug
void mps_pool_check_fenceposts(mps_pool_t pool)
pool the pool whose fenceposts are to be checked
mps.h
This function is a debugging feature to check all the fenceposts in the pool. If a corrupted fencepost is found, an assert will fire. It is only useful to call this on a debug pool that had fenceposting turned, it does nothing on other pools.
mps_pool_check_fenceposts(gene_pool);
If a corrupted fencepost is found, an assert will fire. You will probably want to look at the problem with a debugger.
mps_class_*_debug
mps_pool_debug_option_sThis structure is used to pass debug options to mps_pool_create for debug classes.
Debug.
typedef struct mps_pool_debug_option_s {
void *fence_template;
size_t fence_size;
} mps_pool_debug_option_s;
fence_template
the template for fencepost contents
fence_size
the size of the template in bytes
mps.h
Structures of this type are used to pass debug options to mps_pool_create when creating instances of debug classes.
Fenceposting is enabled by specifying a non-zero fence_size; the size must be a multiple of the [pool/format] alignment. The content of fenceposts is given as a template that is simply copied onto each fencepost (although sometimes the MPS will create fenceposts smaller than the given size, for example, to pad out some bit that was left unused because of alignmentrequirements).
static mps_pool_debug_option_s debugOptions = { (void *)"postpost", 8 };
if(mps_pool_create(&pool, arena, mps_class_ams_debug(),
&debugOptions, 8192, 135, 8)
!= MPS_RES_OK) {
printf("Error creating pool!"); exit(2);
}
Fencepost templates allow the client to specify complicated patterns that mimic illegal datavalues, that would cause an assert to fire if read by mistake, and that would never be written by any operation that writes at the wrong address by mistake.
Another trick is to make the pattern contain an instruction sequence that would cause theprogram to error or stop if executed by mistake.
mps_rank_ambigFunction returning the value representing "rank ambig".
Allocation, Root, Scanning.
mps_rank_ambig()
mps_rank_t mps_rank_ambig(void)
None.
Returns a value of type mps_rank_t representing "rank ambig".
mps.h
Used to get a value for "rank ambig", which is used to denote that certain references (in a root, for example) are ambiguous references.
mps_rank_exactUsed to declare references which the client wishes to be exact references.
Allocation, Root, Scanning.
mps_rank_t mps_rank_exact(void);
No arguments.
Returns a rank (see mps_rank_t) which can be used to declare references to be exact references.
mps.h
Used to declare references which the client wishes to be exact, non-weak references.
[missing]
mps_rank_t,
mps_rank_ambig,
mps_rank_weak
mps_rank_tA type whose values are "reference ranks".
Allocation, Root.
typedef unsigned int mps_rank_t;
mps.h
mps_rank_t is a concrete type. It is an alias (via the C typedef mechanism) for "unsigned int" provided for convenience and clarity. An object of type mps_rank_t can store a value representing one reference rank. Reference ranks are used to conveniently express specific semantics of particular references. See "MPS Scanning Protocol" for descriptions of these semantics, and mps_rank_* for the actual ranks used to declare these semantics.
(Probably won't be used explicitly, most likely to be seen in the prototype declaration for other MPS functions. For example, mps_root_create.)
mps_rank_*
mps_rank_weakFunction to return a value used to represent "rank weak".
Allocation, Scanning.
extern mps_rank_t mps_rank_weak(void);
None.
Returns a value of type mps_rank_t that represent "rank weak".
mps.h
mps_rank_weak returns a value used to represent "rank weak".
"Rank weak" is often used to denote that certain references (in a root or in objects allocated in a pool) are weak references.
<example of how to use the symbol>
mps_reg_scan_tType of root scanning functions for mps_root_create_reg.
Root.
typedef mps_res_t (*mps_reg_scan_t)( mps_ss_t scan_state, mps_thr_t thread, void *p, size_t s)
scan_state a scan state
thread the thread
p a value passed through from root registration
s a value passed through from root registration
A result code.
mps.h
This is the type of root scanning functions the client provides to mps_root_create_reg.These functions will be called, whenever the root needs to be scanned, and passed the "p" and "s"values specified in the call to mps_root_create_reg.
mps_root_create_reg,
mps_stack_scan_ambig
Users are not expected to write any scanning functions of this type. The one function supplied with the MPS, mps_stack_scan_ambig, should be enough for most purposes.
mps_res_tmps_res_t is the type of result codes returned by operations that may fail.
typedef int mps_res_t;
mps.h
A result code indicates the success or failure of an operation, along with the reason for failure. Like UNIX error codes, the meaning of the code depends on the call that returned it. Refer to the documentation of the function for the exact meaning. This documentation describes the broad categories with mnemonic names for various sorts of problems.
MPS_RES_OK: The operation succeeded. Out and in/out parameters will only be updated if OK is returned, otherwise they will be left untouched. MPS_RES_OK is zero.
MPS_RES_FAIL: Something went wrong that does not fall into any of the other categories. The exact meaning depends on the call. See the documentation of the function.
MPS_RES_RESOURCE: A needed resource could not be obtained. Which resource, depends on the call. Compare with MPS_RES_MEMORY, which is a special case of this.
MPS_RES_MEMORY: Needed memory (committed memory, not address space) could not be obtained. (A more detailed explanation).
MPS_RES_LIMIT: An internal limitation was reached. For example, the maximum number of something was reached. (A more detailed explanation).
MPS_RES_UNIMPL: The operation, or some vital part of it, is unimplemented. This might be returned by functions that are no longer supported, or by operations that are included for future expansion, but not yet supported.
MPS_RES_IO: An I/O error occurred. Exactly what depends on the function.
MPS_RES_COMMIT_LIMIT: The arena's commit limit would have been exceeded as a result of (explicit or implicit) allocation. See protocol.arena.commit.
MPS_RES_PARAM: A parameter of the operation was invalid. (A more detailed explanation).
Any function that might fail will return a result code. Any other results of the function are passed back in "return" parameters. See MPS Interface Conventions for more information.
mps_addr_t p;
mps_res_t res;
res = mps_alloc(&p, pool, sizeof(struct spong));
if(res != MPS_RES_OK) {
handle_memory_error(res);
abort();
}
For more examples, s ee doc.mps.ref-man.if-conv.
MPS_RES_*
mps_root_createThe function mps_root_create declares a root that consists of all the references indicated by a scanning function.
Root.
mps_res_t mps_root_create(mps_root_t *root_o, mps_arena_t arena, mps_rank_t rank, mps_rm_trm, mps_root_scan_t scan, void *p, size_t s)
root_o a pointer to a variable to store the new root structure
arena the arena
rank the rank of references in the root
rm the root mode
scan the scanning function
p a value to be passed to the scanning function
s a value to be passed to the scanning function
If the return value is MPS_RES_OK, a new root structure in "*root_o".
mps.h.
The client provides a scanning function, that will be called with a scan state and "p" and"s", whenever the root needs to be scanned. See mps_root_scan_t for details.
If the rank of the root is not MPS_RANK_AMBIG, the contents of the root have to be valid whenever a GC happens, i.e., they have to be references to actual objects or "NULL". If you're using asynchronous GC, this could be right after the root is registered, so the root has to be valid when it is registered. It's OK for a root to have entries which point to memory not managed by the MPS --they will simply be ignored.
static mps_root_t mmRoot;
int main(void)
{
mps_res_t res;
/* ... */
res = mps_root_create(&mmRoot, arena, MPS_RANK_EXACT, (mps_rm_t)0,
&rootScanner, NULL, 0);
/* see doc of mps_root_scan_t for definition of rootScanner */
if(res != MPS_RES_OK)
exit(1);
/* ... */
}
mps_root_create returns MPS_RES_MEMORY when it fails to allocate memory for the internal root structure; you need to deallocate or reclaim something to make enough space, or expand the arena.
mps_root_scan_t,
mps_rm_t,
mps_rank_t,
mps_root_t,
mps_root_create_fmt,
mps_root_create_table,
MPS_RM_CONST
"p" and "s" are just arbitrary data that scanning function can use. This is needed because Clacks local functions.
mps_root_create_fmtThe function mps_root_create_fmt declares a root that consists of a block of objects, and provides a scanning function for them.
Root.
mps_res_t mps_root_create_fmt(mps_root_t *root_o, mps_arena_t arena, mps_rank_t rank, mps_rm_t rm, mps_fmt_scan_t scan, mps_addr_t base, mps_addr_t limit)
root_o a pointer to a variable to store the new root structure
arena the arena
rank the rank of references in the root
rm the root mode
scan the scanning function
base the address of the start of the root
limit the address just beyond the end of the root
If the return value is MPS_RES_OK, the new root in "*root_o".
mps.h
The client provides a scanning function, that will be called with a scan state and an area of memory, whenever the root needs to be scanned. See mps_fmt_scan_t for details.
If the rank of the root is not MPS_RANK_AMBIG, the contents of the root have to be valid whenever a GC happens, i.e., they have to be references to actual objects or "NULL". If you're using asynchronous GC, this could be right after the root is registered, so the root has to be valid when it is registered. It's OK for a root to have entries which point to memory not managed by the MPS --they will simply be ignored.
static mps_root_t mmRoot;
SegmentDescriptor DataSegment;
int main(void)
{
mps_res_t res;
/* ... */
res = mps_root_create_fmt(&mmRoot, arena, MPS_RANK_EXACT, (mps_rm_t)0,
&scan_objs,
(mps_addr_t)DataSegment.base,
(mps_addr_t) (DataSegment.base + SegmentLength) );
/* see doc of mps_fmt_scan_t for definition of scan_objs */
if(res != MPS_RES_OK)
exit( EXIT_FAILURE );
/* ... */
}
mps_root_create_fmt returns MPS_RES_MEMORY when it fails to allocate memory for the internal root structure; you need to deallocate or reclaim something to make enough space, or expand the arena.
mps_fmt_scan_t,
mps_rm_t,
mps_rank_t,
mps_root_t,
mps_root_create,
mps_root_create_table,
MPS_RM_PROT,
MPS_RM_CONST
This is like mps_root_create_table, except you get to supply your own scanning function.This is like mps_root_create, except the scanning function has a slightly different argument list(and the MPS knows where the root is).
mps_root_create_regmps_root_create_reg registers a thread as a root.
Root.
mps_res_t mps_root_create_reg(mps_root_t * root_o, mps_arena_t arena, mps_rank_t rank, mps_rm_t rm, mps_thr_t thread, mps_reg_scan_t scan, void *p, size_t s)
root_o a pointer to a variable to store the new root structure
arena the arena
rank the rank of references in the root
rm the root mode
thread the thread to the registered as a root
scan the scanning function
p a value to be passed to the scanning function
s a value to be passed to the scanning function
If the return value is MPS_RES_OK, a new root structure in "*root_o".
mps.h
mps_root_create_reg
declares the state of a thread as a root. The client provides a
scanning function that will be called and passed "p" and "s", whenever
the root needs to be scanned. See mps_reg_scan_t for details.
If the rank of the root is not MPS_RANK_AMBIG, the contents of the root have to be valid whenever a GC happens, i.e., they have to be references to actual objects or "NULL". If you're using asynchronous GC, this could be right after the root is registered, so the root has to be valid when it is registered. It's OK for a root to have entries which point to memory not managed by the MPS --they will simply be ignored.
typedef struct {
mps_root_t mmRoot;
mps_thr_t thread;
/* ... */
} ThreadLocals;
void InitThread(ThreadLocals *thr)
{
/* This is a hack to find the bottom of the stack. */
void *stackBottom=&stackBottom;
mps_thread_reg(&thr->thread, arena);
mps_root_create_reg(&thr->mmRoot, arena, MPS_RANK_AMBIG, (mps_rm_t) 0,
thr->thread, mps_stack_scan_ambig, stackBottom, 0);
/* ... */
}
mps_root_create_reg returns MPS_RES_MEMORY when it fails to allocate memory for the internal root structure; you need to deallocate or reclaim something to make enough space, or expand the arena.
mps_stack_scan_ambig,
mps_reg_scan_t
Only one suitable scanning function is supplied with the MPS, namely mps_stack_scan_ambig.
mps_root_create_tablemps_root_create_table create s a root that is a vector of references.
Root.
mps_res_t mps_root_create_table(mps_root_t *root_o, mps_arena_t arena, mps_rank_t rank, mps_rm_t rm, mps_addr_t *base, size_t size)
root_o a pointer to a variable for storing the new root structure in
arena the arena
rank the rank of the references in this root
rm the root mode
base a pointer to the vector of references that is being registered
size the number of references in the vector being registered
If the return value is MPS_RES_OK, the new root in "*root_o".
mps.h
This function declares a root that is a vector of references.
If the rank of the root is not MPS_RANK_AMBIG, the contents of the root have to be valid whenever a GC happens, i.e., they have to be references to actual objects or "NULL". If you're using asynchronous GC, this could be right after the root is registered, so the root has to be valid when it is registered. It's OK for a root to have entries which point to memory not managed by the MPS --they will simply be ignored.
static mps_root_t mmRoot;
Object *Objects[rootCOUNT];
int main(void)
{
mps_res_t res;
/* ... */
res = mps_root_create_table(&mmRoot, arena, MPS_RANK_EXACT, (mps_rm_t)0,
(mps_addr_t) &Objects, rootCOUNT );
if(res != MPS_RES_OK)
exit(1);
/* ... */
}
mps_root_create_table returns MPS_RES_MEMORY when it fails to allocate memory for the internal root structure; you need to deallocate or reclaim something to make enough space, or expand the arena.
mps_root_create_table_masked,
MPS_RM_PROT,
MPS_RM_CONST
mps_root_create_table_maskedmps_root_create_table_masked creates a root that is a vector of tagged values.
Root.
mps_res_t mps_root_create_table_masked(mps_root_t *root_o, mps_arena_t arena, mps_rank_t rank, mps_rm_t rm, mps_addr_t *base, size_t size, mps_word_t mask);
root_o a pointer to a variable for storing the new root structure in
arena the arena
rank the rank of the references in this root
rm the root mode
base a pointer to the vector of references that is being registered
size the number of references in the vector being registered
mask any element that has any of the bits in mask set is ignored
If the return value is MPS_RES_OK, the new root in "*root_o".
mps.h
mps_root_create_table_masked creates a root that is a table of tagged values. The mask parameter indicates which bits of a pointer are tag bits. References are assumed to have a tag of zero, values with other tags are ignored.
If the rank of the root is not MPS_RANK_AMBIG, the contents of the root have to be valid whenever a GC happens, i.e., they have to be references to actual objects or "NULL". If you're using asynchronous GC, this could be right after the root is registered, so the root has to be valid when it is registered. It's OK for a root to have entries which point to memory not managed by the MPS --they will simply be ignored.
#define tagMASK 0x0003
static mps_root_t mmRoot;
Object *Objects[rootCOUNT];
int main(void)
{
mps_res_t res;
/* ... */
res = mps_root_create_table_masked(&mmRoot, arena, MPS_RANK_EXACT, (mps_rm_t)0,
(mps_addr_t)&Objects, rootCOUNT,
(mps_word_t)tagMASK);
if(res != MPS_RES_OK)
exit(1);
/* ... */
}
mps_root_create_table_masked returns MPS_RES_MEMORY when it fails to allocate memory for the internal root structure; you need to deallocate or reclaim something to make enough space,or expand the arena.
mps_root_create_table,
MPS_RM_PROT,
MPS_RM_CONST
mps_root_scan_tType of root scanning functions for mps_root_create.
Root.
typedef mps_res_t (*mps_root_scan_t)(mps_ss_t scan_state, void * p, size_t s)
scan_state a scan state
p an argument passed through from mps_root_create
s an argument passed through from mps_root_create
A result code.
mps.h
This is the type of root scanning functions the client provides tomps_root_create. The MPS will call these functions whenever the root needs to be scanned, with a scan state (of type mps_ss_t ), and the p ands values specified in the call to mps_root_create. Apart from the argument list, the scanning function works like the format scan methods: it needs to indicate all references using mps_fix or MPS_FIX*.
static StackFrame *stackBottom;
/* root scanner for an imaginary interpreter for a stack-oriented language */
static mps_res_t rootScanner(mps_ss_t ss, void * p, size_t s)
{
StackFrame *frame;
size_t i;
mps_res_t res;
UNUSED(p);
UNUSED(s);
for(frame = stackBottom; frame != NULL; frame = frame->next)
for(i = frame->size; i > 0; --i) {
res = mps_fix(ss, &frame->locals[i]);
if(res != MPS_RES_OK) return res;
}
return res;
}
If a fixing operation returns a value other than MPS_RES_OK, the scanning function must
return that value, and may return without scanning further
references. Generally, it is better if it returns as soon as
possible. If the scanning is completed successfully, the
function should return MPS_RES_OK.
mps_root_create,
mps_ss_t,
mps_fix,
MPS_SCAN_BEGIN,
MPS_SCAN_END,
MPS_FIX12,
MPS_FIX1,
MPS_FIX2,
MPS_FIX_CALL,
mps_fmt_scan_t
mps_roots_stepper_tType of the client-supplied root walker component.
None.
typedef void (*mps_roots_stepper_t)( mps_addr_t *, mps_root_t, void *, size_t )
The function pointed to by an object of type mps_roots_stepper_t takes the followingargument list:
(mps_addr_t *ref, mps_root_t root, void *p, size_t s)
ref is the address of a root which references an object in the arena. It's a pointer to a root which points to "something" in the client heap. That "something" will be an object if the root is an exact root. But it might be an interior pointer to an object if the root is an ambiguous root.
root is the MPS root object which contains ref.
p and s are two closure values which are copies of the corresponding values which the client passed into mps_arena_roots_walk.
he function pointed to by an object of type mps_roots_stepper_t returns no values.
mps.h
A pointer to a function is passed into the function mps_arena_roots_walk; the pointer has this type. The root walker arranges to apply this function to all objects which are directly referenced from the roots.
<example of how to use the symbol>
he function pointed to by an object of type mps_roots_stepper_t has no way of signalling an error to the caller.
mps_sac_class_sA structure describing a size class to be passed as an argument to mps_sac_create.
Allocation cache
typedef struct mps_sac_class_s {
size_t mps_block_size;
size_t mps_cached_count;
unsigned mps_frequency;
} mps_sac_class_s;
mps.h
mps_sac_class_s is the element type of the array passed tomps_sac_create to describe the size classes. Each element of this array describes one class by specifying block_size, the maximum size (in bytes) in this class; cached_count, the number of objects of this class to cache; and frequency, a number that describes the frequency of requests (allocation and deallocation combined ) in this class relative to all the other classes. The classes should be given in the order of ascending size.
block_size s have to be aligned to the pool alignment. All sizes must be different, and the smallest size must be large enough to hold a void *.
cached_count
is advice to the MPS on how many blocks to cache, not an absolute limit. The cache policy tries to accommodate fluctuations in the population and minimize the cost of responding to client requests; the purpose of this parameter is to limit how much memory the client is willing to set aside for this purpose. However, a
cached_count
of zero prevents any caching of blocks falling into that class.
The MPS automatically provides an "overlarge" class for arbitrarily large objects above the largest class described. Allocations falling into the overlarge class are not cached.
mps_sac_t sac;
mps_sac_class_s classes[3] = { {8, 38, 1}, {136, 19, 3}, {512, 4, 1} };
res = mps_sac_create(&sac, pool, 3, classes);
if (res != MPS_RES_OK) {
printf("Failed to create the allocation cache!");
exit(1);
}
Any blocks whose size falls between two classes are allocated from the larger class.
mps_sac_createThis function creates a segregated allocation cache.
Allocation cache
mps_res_t mps_sac_create(mps_sac_t *sac_o, mps_pool_t pool, size_t classes_count, mps_sac_class_s *classes);
sac_o a pointer to a variable to hold the cache created
pool the pool the cache is attached to
classes_count the number of the size classes
classes pointer to the first element of an array describing the size classes
If the return value is MPS_RES_OK, a new cache in *sac_o.
mps.h
This function creates an allocation cache whose free-list is segregated into the given size classes. The cache can get more memory from the given pool, or return memory to it.
Segregated allocation caches can be associated with any pool that supports mps_alloc and mps_free.
The size classes are described by an array of element type mps_sac_class_s (q.v.). This array is used to initialize the cache, and is not needed aftermps_sac_create returns. There might be a limit on how many classes can be described,but it will be no less than MPS_SAC_CLASS_LIMIT. You must specify at least one class.The MPS automatically provides an "overlarge" class for arbitrarily large objects above the largest class described. Allocations falling into the overlarge class are not cached.
mps_sac_t sac;
mps_sac_class_s classes[3] = { {8, 38, 1}, {136, 19, 3}, {512, 4, 1} };
res = mps_sac_create(&sac, pool, 3, classes);
if (res != MPS_RES_OK) {
printf("Failed to create the allocation cache!");
exit(1);
}
mps_sac_create returns MPS_RES_MEMORY orMPS_RES_COMMIT_LIMIT when it fails to allocate memory for the internal cache structure;see the documentation for those return codes for recovery options. It returnsMPS_RES_LIMIT if you ask for too many size classes; combine some small adjacent classes. It returns MPS_RES_PARAM if the pool doesn't support segregated allocation caches.
mps_sac_class_s,
MPS_SAC_CLASS_LIMIT,
mps_sac_destroy,
MPS_RES_MEMORY,
MPS_RES_COMMIT_LIMIT,
MPS_RES_LIMIT,
MPS_RES_PARAM,
mps_sac_t
Too many classes will slow down allocation; too few classes waste more space in internal fragmentation. It is assumed that overlarge allocations are rare; otherwise, you would add another class for them, or even create separate allocation caches or pools for them.
Some pools will work more efficiently with caches than others. In the future, the MPS might offer pools specially optimized for particular types of cache.
Segregated allocation caches work poorly with debug pool classes at the moment: the checking only happens when blocks are moved between the cache and the pool. This will be fixed, but the speed of allocation with a debug class will always be similar to mps_alloc, rather than cached speed.
mps_sac_destroyThis function destroys a segregated allocation cache.
Allocation cache
void mps_sac_destroy(mps_sac_t);
sac the segregated allocation cache
None.
mps.h
This function destroys a segregated allocation cache. All memory held in it is returned to the associated pool.
res = mps_sac_create(&sac, pool, 3, classes);
if (res != MPS_RES_OK) {
printf("Failed to create the allocation cache!");
exit(1);
}
/* Use sac. */
mps_sac_destroy(sac);
mps_pool_destroy(pool);
Destroying the cache might well cause the pool to return some memory to the arena, but that's up to the pool's usual policy.
Destroying the cache has no effect on objects allocated through it.
mps_sac_flushThis function flushes the segregated allocation cache given.
Allocation cache
void mps_sac_flush(mps_sac_t sac);
sac the segregated allocation cache
None.
mps.h
This function flushes the segregated allocation cache given, returning all memory held in it to the associated pool.
The client is responsible for synchronising the access to the cache, but the MPS will properly synchronize with any other threads that might be accessing the same pool.
mps_sac_t sac_small, sac_large;
res = mps_sac_create(&sac_small, pool, 3, small_classes);
if (res != MPS_RES_OK) {
printf("Failed to create the small allocation cache!");
exit(1);
}
res = mps_sac_create(&sac_large, pool, 3, large_classes);
if (res != MPS_RES_OK) {
printf("Failed to create the large allocation cache!");
exit(1);
}
/* Use sac_small. */
mps_sac_flush(sac_small);
/* Use sac_large. */
mps_sac_flush(sac_large);
/* Use sac_small. */
This is something that you'd typically do when you know you won't be using the cache for awhile, but want to hold on to the cache itself. Destroying a cache has the effect of flushing it,naturally.
Flushing the cache might well cause the pool to return some memory to the arena, but that's up to the pool's usual policy.
Note that the MPS might also decide to take memory from the cache without the client requesting a flush.
mps_sac_tType of segregated allocation caches.
Allocation cache
typedef struct mps_sac_s *mps_sac_t;
mps.h
A value of this type represents an allocation cache with segregated free lists. It is an opaque type.
mps_sac_t sac;
mps_sac_class_s classes[3] = { {8, 38, 1}, {136, 19, 3}, {512, 4, 1} };
res = mps_sac_create(&sac, pool, 3, classes);
if (res != MPS_RES_OK) {
printf("Failed to create the allocation cache!");
exit(1);
}
mps_sac_create,
mps_sac_destroy,
MPS_SAC_ALLOC,
mps_sac_alloc,
MPS_SAC_FREE,
mps_sac_free,
mps_sac_flush
None.
mps_stack_scan_ambigA scanning function for ambiguous scanning of thread states.
Root.
mps_res_t mps_stack_scan_ambig(mps_ss_t scan_state, mps_thr_t thread, void *stack_bottom, size_t ignore)
scan_state a scan state
thread the thread
stack_bottom a pointer to the bottom of the stack
ignore ignored
A result code.
mps.h
This is a root scanning function of type mps_reg_scan_t. It will scan all integer registers and everything on the stack of the thread given, and can therefore only be used with roots of rank MPS_RANK_AMBIG. It will only scan things at the given stack bottom pointer or higher on the stack (that is, more recently added). References are assumed to be represented as machine words, and are required to be 4-byte-aligned; unaligned values are ignored.
Clients don't call this function, it is used as an argument of mps_root_create_reg.
typedef struct {
mps_root_t mmRoot;
mps_thr_t thread;
/* ... */
} ThreadLocals;
void InitThread(ThreadLocals *thr)
{
/* This is a hack to find the bottom of the stack. */
void *stackBottom=&stackBottom;
mps_thread_reg(&thr->thread, arena);
mps_root_create_reg(&thr->mmRoot, arena, MPS_RANK_AMBIG, (mps_rm_t)0,
thr->thread, mps_stack_scan_ambig, stackBottom, 0)
/* ... */
}
mps_reg_scan_t,
mps_root_create_reg
The MPS provides this function because it's hard to write (it's OS- and architecture-dependent and possibly compiler-dependent).
mps_telemetry_controlThis function is used to read and change the filters on the telemetry stream.
Telemetry.
mps_word_t mps_telemetry_control(mps_word_t reset_mask, mps_word_t flip_mask);
reset_mask is a bit mask indicating the bits that should be reset, regardless of previous value.
flip_mask is a bit mask indicating the bits whose value should be flipped after the resetting.
The function returns the previous value of the telemetry filter control.
This function is used to read and change the filters on the telemetry stream. It is generally for use by developers.
The parameters reset_mask and flip_mask allow specifying any binary operation on the filter control. To use this function for typical operations, the parameters should be set as follows:
Operation reset_mask flip_mask
set(M) M M
reset(M) M 0
flip(M) 0 M
read() 0 0
The significance of the bits is liable to change, but the current values (number the least significant bit as zero) are:
0 -- per space or arena
1 -- per pool
2 -- per trace or scan
3 -- per page (segment)
4 -- per reference or fix
5 -- per allocation or object
6 -- user events (e.g., mps_telemetry_intern)
mps_telemetry_flushThis function is used to flush the internal event buffers.
Telemetry.
void mps_telemetry_flush(void);
mps.h
This function is used to flush the internal event buffers into the event stream. This function also calls mps_lib_io_flush on the event stream itself. This ensures that even the latest events are now properly recorded, should the application terminate (uncontrollably as a result of a bug, for example) or some interactive tool require access to the event data. You could even try calling this from a debugger after a problem.
mps_telemetry_flush();
mps_telemetry_internThis function registers a string with the MPS, and receives a unique identifier in return.This identifier is suitable for use with mps_telemetry_label.
Telemetry
mps_word_t mps_telemetry_intern(char *)
The function receives a name as a nul-terminated string in the usual C way. The string's length should not exceed 256 characters, including nul terminating character. In appropriate varieties this restriction is checked and will cause the MPS to issue an ASSERT. So don't do it.
The function returns a unique identifier that may be used to represent the string in future.
The intention of this function is to provide an immediate
identifier that can be used to concisely represent a string for the
purposes of mps_telemetry_label. Note that
the appropriate settings must be made to the telemetry filter (via
mps_telemetry_control) before
this function is invoked; the associate event is of the user kind.
The string's length should not exceed 256 characters, including nul terminating character.This will cause the MPS to issue an ASSERT in appropriate varieties.
mps_telemetry_labelThis function associates an identifier returned from mps_telemetry_intern, and hence
a string, with an address, in the telemetry stream.
telemetry
void mps_telemetry_label(mps_addr_t, mps_word_t);
The function receives an address and an identifier. The identifier should be one returned by mps_telemetry_intern in the same session.
This function is intended to associate the address with an identifier in the telemetry stream. Note that the user kind must be set in the telemetry filter.
Typical uses include:
- Label pools with a human-meaningful name;
- Label allocated objects with their type or class.
mps_telemetry_intern,
mps_telemetry_control,
mps_thr_t
mps_thr_tmps_thr_t is the type of thread records registered with the MPS.
Threads.
typedef mps_thr_s *mps_thr_t;
mps.h
An object of the opaque type mps_thr_t is a thread registration. In a multi-threaded environment where incremental garbage collection is used, threads must be registered with the MPS so that the MPS can examine their state.
An object of type mps_thr_t is obtained using the thread registration function mps_thread_reg.
mps_thr_t this_thread; mps_res_t res; res = mps_thread_reg(&this_thread, space); if(res != MPS_RES_OK) return res;
mps_reg_t,
mps_thread_reg,
mps_thread_dereg,
mps_reg_scan_t,
mps_root_create_reg,
mps_stack_scan_ambig
The following MPS symbols are used or defined in MPS header files, and intended for client use, but are not yet documented in this reference manual.
[This section is very out-of-date. RB 2012-08-15]
mps_arena_t mps_pool_t mps_chain_t mps_root_t mps_ap_t mps_ld_t mps_ss_t mps_alloc_pattern_t mps_frame_t mps_word_t mps_shift_t mps_rm_t MPS_RES_OK MPS_RES_FAIL MPS_RES_RESOURCE MPS_RES_UNIMPL MPS_RES_IO MPS_RES_COMMIT_LIMIT mps_ap_s mps_sac_freelist_block_s mps_sac_s mps_ld_s mps_ss_s mps_fmt_fixed_s MPS_BEGIN MPS_END mps_arena_step mps_arena_start_collect mps_arena_destroy mps_arena_reserved mps_arena_extend mps_arena_retract mps_fmt_create_fixed mps_fmt_destroy mps_addr_pool mps_addr_fmt mps_pool_create mps_pool_create_v mps_pool_destroy mps_gen_param_s mps_chain_create mps_chain_destroy mps_alloc_v mps_ap_create mps_ap_create_v mps_ap_destroy mps_reserve mps_commit mps_ap_fill mps_ap_fill_with_reservoir_permit mps_ap_trip MPS_SAC_ALLOC MPS_SAC_FREE mps_reservoir_limit_set mps_reservoir_limit mps_reservoir_available mps_reserve_with_reservoir_permit MPS_RESERVE_BLOCK MPS_RESERVE_WITH_RESERVOIRf_PERMIT_BLOCK mps_root_destroy mps_tramp_t mps_tramp mps_thread_reg mps_thread_dereg mps_ld_reset mps_ld_add mps_ld_merge mps_ld_isstale mps_collections mps_definalize mps_alert_collection_set mps_pool_check_free_space mps_lib_get_EOF mps_lib_stream_s mps_lib_get_stderr mps_lib_get_stdout mps_lib_fputc mps_lib_fputs mps_lib_assert_fail mps_clock_t mps_clock mps_clocks_per_sec mps_class_amcz mps_class_ams mps_class_ams_debug mps_class_awl mps_class_lo mps_mv_free_size mps_mv_size mps_class_mv mps_class_mv_debug mps_mvt_free_size mps_mvt_size mps_mvff_free_size mps_mvff_size mps_class_mvff_debug mps_SEH_filter mps_SEH_handler mps_io_t mps_io_create mps_io_destroy mps_io_write mps_io_flush MPS_PF_STRING MPS_PF_ALIGN MPS_ARCH_I3 MPS_ARCH_I4 MPS_ARCH_PP MPS_ARCH_S8 MPS_ARCH_S9 MPS_BUILD_GC MPS_BUILD_MV MPS_BUILD_SC MPS_OS_FR MPS_OS_LI MPS_OS_SO MPS_OS_W3 MPS_OS_XC MPS_PF_FRI3GC MPS_PF_LII3GC MPS_PF_LIPPGC MPS_PF_SOS8GC MPS_PF_SOS9SC MPS_PF_W3I3MV MPS_PF_XCPPGC
| 2002-05-27 | RB | Created from individual MPS reference pages, originally written and mainted in Lotus Notes by members of the Memory Management Group of Global Graphics (formerly Harlequin). I found many errors caused by the various conversions that this text has been through. There are probably many more. |
| 2002-06-17 | RB | Removed Global Graphics specific entries for confidential sources not included in open source release. |
| 2002-06-18 | NB | Added contents table to section 3. |
| 2002-06-20 | NB | Quite a bit of proof-reading, to insert missing spaces. Also reformatted for easier editing, including the "See Also" sections. |
| 2002-06-21 | NB | Removed obsolete symbols. |
| 2003-01-01 | DRJ | [various edits in 2003] |
| 2006-06-06 | RHSK | Marked copy format method as obsolete. |
| 2007-06-19 | DRJ | mps_finalize and finalization messages: correct, clarify, and expand descriptions |
| 2008-10-24 | RHSK | mps_arena_has_addr entry: fix wrong title; remove from undocumented symbols list. (Also: fix html tag error; add missing document history entries). |
| 2008-11-25 | RHSK | mps_message_clock: add entry. |
| 2010-03-02 | RHSK | mps_addr_pool, mps_addr_fmt, mps_alert_collection_set: as yet Undocumented. |
This document is copyright © 1997-2002, 2003, 2006, 2007, 2008, 2010 Ravenbrook Limited. All rights reserved. This is an open source license. Contact Ravenbrook for commercial licensing options.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
This software is provided by the copyright holders and contributors "as is" and any express or implied warranties, including, but not limited to, the implied warranties of merchantability, fitness for a particular purpose, or non-infringement, are disclaimed. In no event shall the copyright holders and contributors be liable for any direct, indirect, incidental, special, exemplary, or consequential damages (including, but not limited to, procurement of substitute goods or services; loss of use, data, or profits; or business interruption) however caused and on any theory of liability, whether in contract, strict liability, or tort (including negligence or otherwise) arising in any way out of the use of this software, even if advised of the possibility of such damage.
$Id: //info.ravenbrook.com/project/mps/version/1.110/manual/reference/index.html#2 $
Ravenbrook / Projects / Memory Pool System / Version 1.110 Product Sources / Product Manuals