36. General MPS types¶
36.2. Rationale¶
Some types are declared to resolve a point of design, such as the best type to use for array indexing.
Some types are declared so that the intention of code is clearer. For
example, Byte
is necessarily unsigned char
, but it’s better to
say Byte
in your code if it’s what you mean.
36.3. Concrete types¶
-
unsigned
AccessSet
¶
.access-set: An AccessSet
is a bitset of Access
modes,
which are AccessREAD
and AccessWRITE
. AccessSetEMPTY
is
the empty AccessSet
.
-
struct AddrStruct *
Addr
¶
.addr: Addr
is the type used for “managed addresses”, that is,
addresses of objects managed by the MPS.
.addr.def: Addr
is defined as struct AddrStruct *
, but
AddrStruct
is never defined. This means that Addr
is always an
incomplete type, which prevents accidental dereferencing, arithmetic,
or assignment to other pointer types.
.addr.use: Addr
should be used whenever the code needs to deal
with addresses. It should not be used for the addresses of memory
manager data structures themselves, so that the memory manager remains
amenable to working in a separate address space. Be careful not to
confuse Addr
with void *
.
.addr.ops: Limited arithmetic is allowed on addresses using
AddrAdd()
and AddrOffset()
(impl.c.mpm). Addresses may also be
compared using the relational operators ==
, !=
, <
, <=
,
>
, and >=
.
.addr.ops.mem: We need efficient operators similar to memset()
,
memcpy()
, and memcmp()
on Addr
; these are called AddrSet()
,
AddrCopy()
, and AddrComp()
. When Addr
is compatible with
void *
, these are implemented through the functions
mps_lib_memset()
, mps_lib_memcpy()
, and mps_lib_memcmp()
functions in the plinth (impl.h.mpm).
.addr.conv.c: Addr
is converted to mps_addr_t
in the MPS C
Interface. mps_addr_t
is defined to be the same as void *
, so
using the MPS C Interface confines the memory manager to the same
address space as the client data.
.addr.readonly: For read-only addresses, see .readonlyaddr.
.align: Align
is an unsigned integral type which is used to
represent the alignment of managed addresses. All alignments are
positive powers of two. Align
is large enough to hold the maximum
possible alignment.
.align.use: Align
should be used whenever the code needs to
deal with the alignment of a managed address.
.align.conv.c: Align
is converted to mps_align_t
in the MPS
C Interface.
-
unsigned
Attr
¶
.attr: Pool attributes. A bitset of pool class attributes, which are:
Attribute |
Description |
---|---|
|
Is garbage collecting, that is, parts may be reclaimed. Used to decide which segments are condemned. |
|
Is moving, that is, objects may move in memory. Used to update the set of zones that might have moved and so implement location dependency. |
There is an attribute field in the pool class (PoolClassStruct
)
which declares the attributes of that class. See
design.mps.pool.field.attr.
-
int
Bool
¶
.bool: The Bool
type is mostly defined so that the intention of
code is clearer. In C, Boolean expressions evaluate to int
, so
Bool
is in fact an alias for int
.
.bool.value: Bool
has two values, TRUE
and FALSE
. These
are defined to be 1
and 0
respectively, for compatibility with
C Boolean expressions (so one may set a Bool
to the result of a C
Boolean expression).
.bool.use: Bool
is a type which should be used when a Boolean
value is intended, for example, as the result of a function. Using a
Boolean type in C is a tricky thing. Non-zero values are “true” (when
used as control conditions) but are not all equal to TRUE
. Use
with care.
.bool.check: BoolCheck()
simply checks whether the argument is
TRUE
(1
) or FALSE
(0
).
.bool.check.inline: The inline macro version of BoolCheck
casts
the int
to unsigned
and checks that it is <= 1
. This is
safe, well-defined, uses the argument exactly once, and generates
reasonable code.
.bool.check.inline.smaller: In fact we can expect that the “inline”
version of BoolCheck()
to be smaller than the equivalent function
call. On IA-32 for example, a function call will be 3 instructions
(total 9 bytes), the inline code for BoolCheck()
will be 1
instruction (total 3 bytes) (both sequences not including the test
which is the same length in either case).
.bool.check.inline.why: As well as being smaller (see
.bool.check.inline.smaller) it is faster. On 1998-11-16 drj
compared w3i3mv\hi\amcss.exe
running with and without the macro
for BoolCheck
on the PC Aaron. “With” ran in 97.7% of the time
(averaged over 3 runs).
.bool.bitfield: When a Boolean needs to be stored in a bitfield,
the type of the bitfield must be unsigned:1
, not Bool:1
.
(That’s because the two values of the type Bool:1
are 0
and
-1
, which means that assigning TRUE
would require a sign
conversion.) To make it clear why this is done, misc.h
provides
the BOOLFIELD
macro.
.bool.bitfield.assign: To avoid warnings about loss of data from
GCC with the -Wconversion
option, misc.h
provides the
BOOLOF
macro for coercing a value to an unsigned single-bit field.
.bool.bitfield.check: A Boolean bitfield cannot have an incorrect
value, and if you call BoolCheck()
on such a bitfield then GCC 4.2
issues the warning “comparison is always true due to limited range of
data type”. When avoiding such a warning, reference this tag.
-
unsigned
BufferMode
¶
.buffermode: BufferMode
is a bitset of buffer attributes. See
design.mps.buffer. It is a sum of the following:
Mode |
Description |
---|---|
|
Buffer is attached to a region of memory. |
|
Buffer has been flipped. |
|
Buffer remains permanently trapped, so that all reserve and commit events can be logged. |
|
Buffer is in the process of being detached. |
-
unsigned char
Byte
¶
.byte: Byte
is an unsigned integral type corresponding to the
unit in which most sizes are measured, and also the units of
sizeof
.
.byte.use: Byte
should be used in preference to char
or
unsigned char
wherever it is necessary to deal with bytes
directly.
.byte.source: Byte
is a just pedagogic version of unsigned
char
, since char
is the unit of sizeof
.
.clock: Clock
is an unsigned integral type representing clock
time since some epoch.
.clock.use: A Clock
value is returned by the plinth function
mps_clock
. It is used to make collection scheduling decisions and
to calibrate the time stamps on events in the telemetry stream.
.clock.units: The plinth function mps_clocks_per_sec
defines
the units of a Clock
value.
.clock.conv.c: Clock
is converted to mps_clock_t
in the MPS
C Interface.
-
unsigned
Compare
¶
.compare: Compare
is the type of tri-state comparison
values.
Value |
Description |
---|---|
|
A value compares less than another value. |
|
Two values compare the same. |
|
A value compares greater than another value. |
.count: Count
is an unsigned integral type which is large
enough to hold the size of any collection of objects in the MPS.
.count.use: Count
should be used for a number of objects
(control or managed) where the maximum number of objects cannot be
statically determined. If the maximum number can be statically
determined then the smallest unsigned integer with a large enough
range may be used instead (although Count
may be preferable for
clarity).
.count.use.other: Count
may also be used to count things that
aren’t represented by objects (for example, levels), but only where it
can be determined that the maximum count is less than the number of
objects.
.epoch: An Epoch
is a count of the number of flips that have
occurred, in which objects may have moved. It is used in the
implementation of location dependencies.
Epoch
is converted to mps_word_t
in the MPS C Interface, as a
field of mps_ld_s
.
-
unsigned
FindDelete
¶
.finddelete: FindDelete
represents an instruction to one of the
find methods of a Land
as to what it should do if it finds a
suitable block. See design.mps.land. It takes one of the following
values:
Value |
Description |
---|---|
|
Don’t delete after finding. |
|
Delete from low end of block. |
|
Delete from high end of block. |
|
Delete entire block. |
-
void
(*Fun)
(void)¶
.fun: Fun
is the type of a pointer to a function about which
nothing more is known.
.fun.use: Fun
should be used where it’s necessary to handle a
function in a polymorphic way without calling it. For example, if you
need to write a function g
which passes another function f
through to a third function h
, where h
knows the real type of
f
but g
doesn’t.
.index: Index
is an unsigned integral type which is large
enough to hold any array index.
.index.use: Index
should be used where the maximum size of the
array cannot be statically determined. If the maximum size can be
determined then the smallest unsigned integer with a large enough
range may be used instead.
-
unsigned
LocusPrefKind
¶
.locusprefkind: The type LocusPrefKind
expresses a preference for
addresses within an address space. It takes one of the following
values:
Kind |
Description |
---|---|
|
Prefer high addresses. |
|
Prefer low addresses. |
|
Prefer addresses in specified zones. |
-
unsigned
MessageType
¶
.messagetype: MessageType
is the type of a message. See
design.mps.message. It takes one of the following values:
Message type |
Description |
---|---|
|
A block is finalizable. |
|
A garbage collection finished. |
|
A garbage collection started. |
-
unsigned
Rank
¶
.rank: Rank
is an enumeration which represents the rank of a
reference. The ranks are:
Rank |
Index |
Description |
---|---|---|
|
0 |
The reference is ambiguous. That is, it must be assumed to be a reference, but not updated in case it isn’t. |
|
1 |
The reference is exact, and refers to an object. |
|
2 |
The reference is exact and final, so special action is required if only final or weak references remain to the object. |
|
3 |
The reference is exact and weak, so should be deleted if only weak references remain to the object. |
Rank
is stored with segments and roots, and passed around.
Rank
is converted to mps_rank_t
in the MPS C Interface.
The ordering of the ranks is important. It is the order in which the references must be scanned in order to respect the properties of references of the ranks. Therefore they are declared explicitly with their integer values.
Note
Could Rank
be an unsigned short
or unsigned char
?
Note
This documentation should be expanded and moved to its own document, then referenced from the implementation more thoroughly.
-
unsigned
RankSet
¶
.rankset: RankSet
is a set of ranks, represented as a bitset.
-
const struct AddrStruct *
ReadonlyAddr
¶
.readonlyaddr: ReadonlyAddr
is the type used for managed
addresses that an interface promises it will only read through, never
write. Otherwise it is identical to Addr
.
.ref: Ref
is a reference to a managed object (as opposed to any
old managed address). Ref
should be used where a reference is
intended.
Note
This isn’t too clear – richard
.refset: RefSet
is a conservative approximation to a set of
references. See design.mps.refset.
-
int
Res
¶
.res: Res
is the type of result codes. 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. These codes are just broad categories with
mnemonic names for various sorts of problems.
Result code |
Description |
---|---|
|
The operation succeeded. Return parameters may only be updated if OK is returned, otherwise they must be left untouched. |
|
The arena’s commit limit would have been exceeded as a result of allocation. |
|
Something went wrong which doesn’t fall into any of the other categories. The exact meaning depends on the call. See documentation. |
|
An I/O error occurred. Exactly what depends on the function. |
|
An internal limitation was reached. For example, the maximum number of somethings was reached. We should avoid returning this by not including static limitations in our code, as far as possible. (See rule.impl.constrain and rule.impl.limits.) |
|
Needed memory (committed memory, not address space) could not be obtained. |
|
An invalid parameter was passed. Normally reserved for parameters passed from the client. |
|
A needed resource could not be obtained. Which
resource depends on the call. See also
|
|
The operation, or some vital part of it, is unimplemented. This might be returned by functions which are no longer supported, or by operations which are included for future expansion, but not yet supported. |
.res.use: Res
should be returned from any function which might
fail. Any other results of the function should be passed back in
“return” parameters (pointers to locations to fill in with the
results).
Note
This is documented elsewhere, I think – richard
.res.use.spec: The most specific code should be returned.
-
unsigned
RootMode
¶
.rootmode: RootMode
is an unsigned integral type which is used
to represent an attribute of a root:
Root mode |
Description |
---|---|
|
Client program will not change the root after it is registered. |
|
Root is protectable: the MPS may place a barrier on any page containing any part of the root. |
|
Root is protectable: the MPS may place a barrier on any page completely covered by part of the root. |
.rootmode.const.unused: RootModeCONSTANT
has no effect. This
mode was introduced in the hope of being able to maintain a remembered
set for the root without needing a write barrier, but it can’t work as
described, since you can’t reliably create a valid registered constant
root that contains any references. (If you add the references before
registering the root, they may have become invalid; but you can’t add
them afterwards because the root is supposed to be constant.)
.rootmode.conv.c: RootMode
is converted to mps_rm_t
in the
MPS C Interface.
-
unsigned
RootVar
¶
.rootvar: The type RootVar
is the type of the discriminator for
the union within RootStruct
.
-
unsigned
Serial
¶
.serial: A Serial
is a number which is assigned to a structure
when it is initialized. The serial number is taken from a field in the
parent structure, which is incremented. Thus, every instance of a
structure has a unique “name” which is a path of structures from the
global root. For example, “the third arena’s fifth pool’s second
buffer”.
Why? Consistency checking, debugging, and logging. Not well thought out.
-
unsigned
Shift
¶
.shift: Shift
is an unsigned integral type which can hold the
amount by which a Word
can be shifted. It is therefore large
enough to hold the word width (in bits).
.shift.use: Shift
should be used whenever a shift value (the
right-hand operand of the <<
or >>
operators) is intended, to
make the code clear. It should also be used for structure fields which
have this use.
Note
Could Shift
be an unsigned short
or unsigned char
?
-
unsigned long
Sig
¶
.sig: Sig
is the type of signatures, which are written into
structures when they are created, and invalidated when they are
destroyed. They provide a limited form of run-time type checking and
dynamic scope checking. See design.mps.sig.
.size: Size
is an unsigned integral type large enough to
hold the size of any object which the MPS might manage.
.size.byte: Size
should hold a size calculated in bytes.
Warning
This is violated by GenParams.capacity
(which is measured in
kilobytes).
.size.use: Size
should be used whenever the code needs to deal
with the size of managed memory or client objects. It should not be
used for the sizes of the memory manager’s own data structures, so
that the memory manager is amenable to working in a separate address
space. Be careful not to confuse it with size_t
.
.size.ops: SizeIsAligned()
, SizeAlignUp()
,
SizeAlignDown()
and SizeRoundUp()
.
.size.conv.c: Size
is converted to size_t
in the MPS C
Interface. This constrains the memory manager to the same address
space as the client data.
-
unsigned
TraceId
¶
.traceid: A TraceId
is an unsigned integer which is less than
TraceLIMIT
. Each running trace has a different TraceId
which
is used to index into the tables and bitfields that record the state
of that trace. See design.mps.trace.instance.limit.
-
unsigned
TraceSet
¶
.traceset: A TraceSet
is a bitset of TraceId
,
represented in the obvious way:
member(ti, ts) ⇔ ((1<<ti) & ts) != 0
In the multiple-traces design, each region of memory may be a
different colour for each trace. Thus the set of traces for which a
segment is grey (say) is represented by a TraceSet
. See
design.mps.trace.
-
unsigned
TraceStartWhy
¶
.tracestartwhy: TraceStartWhy
represents the reason that a
trace was started. It takes one of the following values:
Reason |
Description |
---|---|
|
Generation zero of a chain reached capacity. |
|
Need to start full collection now, or there won’t be enough memory to complete it. |
|
Client had idle time available. |
|
Client requested incremental collection. |
|
Client requested full collection. |
|
Walking references. |
|
Request by MPS extension. |
-
unsigned
TraceState
¶
.tracestate: TraceState
represents the current state in a
trace’s lifecycle. See design.mps.trace. It takes one of the
following values:
State |
Description |
---|---|
|
Nothing happened yet. |
|
Segments condemned (made white); initial grey set created; scanning rate calculated. |
|
Buffers flipped; roots scanned; location dependencies made stale; grey segments protected. |
|
There are no outstanding grey segments. |
|
All segments reclaimed. |
-
MPS_T_ULONGEST
ULongest
¶
.ulongest: ULongest
is the longest unsigned integer on the
platform. (We used to use unsigned long
but this assumption is
violated by 64-bit Windows.) This type should be used for calculations
where any integer might be passed. Notably, it is used in WriteF()
to print any integer.
-
MPS_T_WORD
Word
¶
.word: Word
is an unsigned integral type which matches the size
of the machine word, that is, the natural size of the machine
registers and addresses.
.word.use: Word
should be used where an unsigned integer is
required that might range as large as the machine word.
.word.source: Word
is derived from the macro MPS_T_WORD
which is declared in impl.h.mpstd according to the target platform
(design.mps.config.pf.word).
.word.conv.c: Word
is converted to mps_word_t
in the MPS C
Interface.
.word.ops: WordIsAligned()
, WordAlignUp()
,
WordAlignDown()
and WordRoundUp()
.
-
MPS_T_WORD
Work
¶
.work: Work
is an unsigned integral type representing
accumulated work done by the collector.
.work.impl: Work is implemented as a count of the bytes scanned by the collector in segments and roots. This is a very crude measure, because it depends on the scanning functions supplied by the mutator, which we know very little about.
.zoneset: ZoneSet
is a conservative approximation to a set of
zones. See design.mps.refset.
36.4. Abstract types¶
.adts: The following types are abstract data types, implemented as
pointers to structures. For example, Ring
is a pointer to a
RingStruct
. They are described elsewhere.
AllocFrame
, AllocPattern
, AP
, Arena
, BootBlock
,
Buffer
, Chain
, Chunk
, Format
, Globals
, Land
,
LD
, Lock
, LocusPref
, MutatorContext
, PoolClass
,
Page
, Pool
, PoolDebugMixin
, Range
, Ring
, Root
,
ScanState
, Seg
, SegBuf
, StackContext
, Thread
,
Trace
, VM
.
-
void *
Pointer
¶
.pointer: The type Pointer
is the same as void *
, and
exists to sanctify functions such as PointerAdd()
.