US20120290628A1

US20120290628A1 - Memory Management for Garbage Collection of Critical Real Time Threads

Info

Publication number: US20120290628A1
Application number: US13/558,405
Authority: US
Inventors: Andrew Johnson
Original assignee: International Business Machines Corp
Current assignee: International Business Machines Corp
Priority date: 2007-09-25
Filing date: 2012-07-26
Publication date: 2012-11-15
Also published as: US20090083349A1; US8266190B2

Abstract

Mechanisms of memory management in a real time runtime environment having a garbage collected memory heap are provided. The runtime environment includes a critical real time thread that is precluded from accessing the garbage collected memory heap. The runtime environment further includes a scoped memory area for the allocation of objects therein for access by the critical real time thread. The mechanisms determine whether the critical real time thread currently accessing the scoped memory area is in a defined state so as to accommodate garbage collection of the scoped memory area. The mechanisms further initiate, in response to a determination that the critical real time thread currently accessing the scoped memory area is in the defined state, a process of garbage collection of the scoped memory area.

Description

BACKGROUND

The present invention relates to memory management in a computer system. More particularly it relates to garbage collection of scoped memory accessed by critical real time threads.
Memory management in runtime environments is often devised so as to provide convenience for the software engineer. For this reason, runtime environments such as Java (Java is a Registered Trademark of Sun Microsystems Inc.) include heap memory having a garbage collector. A garbage collector is a runtime facility for automatically identifying and discarding unused data from memory, such as objects, so as to free up storage.
Garbage collection is a luxury afforded by the efficiencies of modern computer systems that serves to liberate software engineers from the task of programmatically discarding each and every unused object. It does, however, present some drawbacks. One such drawback is that the process of garbage collection generally requires the suspension of execution of software threads accessing the heap. This is because the process of identifying and discarding unused objects cannot normally be undertaken while software threads actively allocate and access existing or new objects. For the garbage collector to operate effectively, the state of the heap is to be preferably static and determinate for the duration of the garbage collection process. The duration of suspension of threads for garbage collection depends on the number and variety of data objects existing in the heap and can range from an imperceptibly short duration to periods long enough to be perceived by a user. Thus the process of garbage collection can lead to delays of indeterminate frequency and duration. These downsides render garbage collected runtime environments inherently unsuitable for software depending on continuity of execution and predictability. Such software includes critical real time applications such as those used in monitoring and control applications, and these applications are normally precluded from accessing a garbage collected memory in order to ensure their consistency and stability.
To address these downsides of garbage collected heap memory, runtime environments can be adapted to use scoped memory areas. Scoped memory areas are areas of memory suitable for the allocation and de-allocation of objects without some of the disadvantages of garbage collected heaps. A software thread can enter a scoped memory area and all object allocations by the thread will then be made within the scoped memory area. The thread can also enter multiple scoped memory areas in a nested fashion. For example, the thread can enter a second scoped memory area which becomes the default allocation area. Scoped memory areas are organised in a scope graph with each scoped memory area having a parent being the previously entered scoped memory area. A first scoped memory area has a primordial scoped memory area as its parent.
A thread count is maintained for each scoped memory area of a number of active threads in the area. The thread count coupled with the scope graph indicates when a scoped memory area, including any objects allocated in the scoped memory area, is discardable. For example a scoped memory area may be discardable when an associated thread count is decremented from one to zero and the scoped memory area is not a parent for any other scoped memory area. In this way, scoped memory areas offer alternative allocation spaces for objects in an otherwise garbage collected runtime environment, such as Java. The use of such scoped memory areas can address the problem of delays and uncertainty in application execution caused by garbage collection algorithms since all objects in scoped memory areas become discardable at the same time defined in terms of thread usage of the memory area and the organisation of a scope graph.
A first object in a first scoped memory area can reference a second object in a second scoped memory area. However, since individual scoped memory areas can become completely discardable it is necessary to enforce strict rules regarding valid and invalid references between objects in different scoped memory areas. These rules are necessary to guarantee that a parent scoped memory area has a lifetime that is at least that of its child scoped memory areas so that objects in descendant memory areas can reliably reference objects in ancestor memory areas. In this way, dangling references can be avoided. This is described in “Efficient Memory-Reference Checks for Real-time Java” (Corsaro and Cytron, ACM SIGPLAN Volume 38, Issue 7 Jul. 2003).
One such rule is that a scoped memory area has at most one parent. This is known as the “single parent rule”. The single parent rule provides that the order of entry into scoped memory areas by different threads is always the same because the order of entry of threads into scoped memory areas defines the structure of the scoped memory areas in the scope graph. Further, a source object allocated in a source scoped memory area can only reference target objects in the same or ancestral scoped memory areas. An ancestral scoped memory area is a memory area which is entered by a thread before the source scoped memory area. The source scoped memory area is said to have inner scope with respect to that of the target object. This criterion ensures that the source object will always be de-allocated before the target object. Attempts to create references which do not satisfy the criterion must fail according to the Real-Time Specification for Java (Bollella et al, Addison-Wesley, 2000). Thus the criterion must be checked in an efficient manner whenever a new reference to a target object is created.
One way to determine whether this criterion is satisfied for the creation of a reference between objects is to iterate through scoped memory areas in a scope graph. Iteration starts at the source scoped memory area and progresses through each parent scoped memory area checking for the target scoped memory area. If the target scoped memory area is encountered within the scope graph during the iterations then it can be concluded that the criterion is satisfied because the target area is an ancestor of the source area. If the iterations reach the primordial scoped memory area (the “root” of the scope graph) then it can be concluded that the criterion is not satisfied because the target scoped memory area is not an ancestor of the source scoped memory area.
An alternative approach is to determine an index of the distance between the scoped memory area of the target object and the primordial scoped memory area. This distance can be considered to be a “depth” of the target scoped memory area. Additionally, an index of the distance between the scoped memory area of the source object and the primordial scoped memory area can be determined as a “depth” of the source scope. The term depth here is used to indicate that scoped memory areas exist at levels which are some distance from the primordial scoped memory area. Deeper scoped memory areas are located at a level in the scope graph which is further from the primordial scoped memory area. A comparison of these indices of depth will indicate whether the target scoped memory area is potentially an ancestor of the source scoped memory area because descendent areas will always be deeper (further from the primordial scoped memory area) than ancestor areas.
FIG. 1 is a block diagram of an exemplary scoped memory system 100 in the prior art. The scoped memory system is an arrangement of a part of a storage of a computer system such as the scoped memory system described in the Real-Time Specification for Java (Bollella et al, Addison-Wesley, 2000). In use, the scoped memory system 100 includes a scope graph 102 which is a data structure arrangement of zero or more scoped memory areas such as scoped memory area 106. Scope graph 102 is commonly referred to in the art as a “scope stack” although the data structure of scope graph 102 is not necessarily a stack data structure. The scope graph 102 can be implemented as a directed graph data structure, a stack data structure, a tree data structure, a linked list or any data structure suitable for the storage of zero or more scoped memory areas.
Scoped memory area 106 is a defined area of memory in storage suitable for the allocation and reference of software objects. Objects can be allocated and accessed in an allocated objects 112 part of the scoped memory area by software threads which execute in the context of the scoped memory area 106. The scoped memory area 106 further includes a thread count 110 as a count of a number of software threads executing in the context of the scoped memory area 106. A thread executes in the context of the scoped memory area 106 by explicitly entering the scoped memory area, such as by calling of an application programming interface function. Once a thread has entered the scoped memory area 106, allocations of objects will be undertaken within the scoped memory area 106 unless specifically allocated elsewhere by the thread, such as in the general heap or another scoped memory area.
Scoped memory area 106 further includes a parent field 108 which includes a reference to a parent scoped memory area. As described above, a scoped memory area has at most one parent. An arrangement of scoped memory areas defined by way of a parental relationships there-between defines the scope graph 102. An initial scoped memory area having no particular parent will refer to a primordial scoped memory area 104 as its parent by way of a placeholder.
Critical software depending on continuity of execution and predictability can use scoped memory areas for the allocation of objects. Since scoped memory areas are not garbage collected, delays and uncertainty associated with garbage collection processes do not affect scoped memory or software threads using scoped memory. While this provides certainty for critical software threads, it presents the disadvantage that individual objects allocated in scoped memory are not discarded until all objects in the scoped memory are discardable. Thus scoped memory can become consumed by numerous unused objects.
It would therefore be advantageous to provide the advantages of garbage collection with the certainty and consistency of scoped memory.

SUMMARY

In one illustrative embodiment, a method of memory management in a real time runtime environment having a garbage collected memory heap is provided. The runtime environment includes a critical real time thread that is precluded from accessing the garbage collected memory heap. The runtime environment further includes a scoped memory area for the allocation of objects therein for access by the critical real time thread.
The method comprises determining whether the critical real time thread currently accessing the scoped memory area is in a defined state so as to accommodate garbage collection of the scoped memory area. The method further comprises initiating, in response to a determination that the critical real time thread currently accessing the scoped memory area is in the defined state, a process of garbage collection of the scoped memory area.
In other illustrative embodiments, a computer program product comprising a computer useable or readable medium having a computer readable program is provided. The computer readable program, when executed on a computing device, causes the computing device to perform various ones, and combinations of, the operations outlined above with regard to the method illustrative embodiment.
In yet another illustrative embodiment, a system/apparatus is provided. The system/apparatus may comprise one or more processors and a memory coupled to the one or more processors. The memory may comprise instructions which, when executed by the one or more processors, cause the one or more processors to perform various ones, and combinations of, the operations outlined above with regard to the method illustrative embodiment.
These and other features and advantages of the present invention will be described in, or will become apparent to those of ordinary skill in the art in view of, the following detailed description of the example embodiments of the present invention.

BRIEF DESCRIPTION OF SEVERAL VIEWS OF THE DRAWINGS

The invention, as well as a preferred mode of use and further objectives and advantages thereof, will best be understood by reference to the following detailed description of illustrative embodiments when read in conjunction with the accompanying drawings, wherein:

FIG. 1 is a block diagram of an exemplary scoped memory system in the prior art;

FIG. 2 is an example block diagram of a computer system suitable for the operation of illustrative embodiments of the present invention;

FIG. 3 is an example block diagram of a software application in execution in a software runtime environment in accordance with one illustrative embodiment of the present invention;

FIG. 4 is an example flowchart of a method of the scoped memory garbage collector in accordance with one illustrative embodiment of the present invention; and

FIG. 5 is an example block diagram of a scoped memory area in use by a thread in accordance with one illustrative embodiment of the present invention;

FIG. 6 is an example block diagram of nested scoped memory areas in accordance with one illustrative embodiment of the present invention;

FIG. 7 a is an example flowchart of a method for identifying root references to objects allocated in a scoped memory area for garbage collection in accordance with one illustrative embodiment of the present invention;

FIG. 7 b is an example flowchart of a method for identifying a set of potential child scoped memory areas for the garbage collection of FIG. 6 in accordance with one illustrative embodiment of the present invention;

FIG. 8 is an example flowchart of a method of memory management in a real time runtime environment in accordance with one illustrative embodiment of the present invention;

FIG. 9 a is an example block diagram of a critical real time thread in a software application in accordance with one illustrative embodiment of the present invention; and

FIG. 9 b is an example block diagram of a critical real time thread in a software application in accordance with an alternative illustrative embodiment of the present invention.

DETAILED DESCRIPTION

FIG. 2 is a block diagram of a computer system suitable for the operation of embodiments of the present invention. A central processor unit (CPU) 202 is communicatively connected to a storage 204 and an input/output (I/O) interface 206 via a data bus 208. The storage 204 can be any read/write storage device such as a random access memory (RAM) or a non-volatile storage device. An example of a non-volatile storage device includes a disk or tape storage device. The I/O interface 206 is an interface to devices for the input or output of data, or for both input and output of data. Examples of I/O devices connectable to I/O interface 206 include a keyboard, a mouse, a display (such as a monitor) and a network connection.
FIG. 3 is an example block diagram of a software application in execution in a software runtime environment in accordance with one illustrative embodiment of the present invention. The software application includes threads of execution 302 to 308. Each thread is a separate stream of software execution and includes software instructions and thread data normally stored in an area of storage known as a memory stack. Threads 302 to 308 include references 310 to 320 into allocated heap memory including a garbage collected heap 332 and a scoped memory area 330. The garbage collected heap 332 is a memory allocation space for threads in execution in the runtime environment. Software objects 340, 338 are allocated in the garbage collected heap 332. Periodically a garbage collector is operable to process the garbage collected heap 332 to discard unused objects as is well known in the art. The scoped memory area 330 is also a memory allocation space for threads in execution, and software objects 334, 336 are allocated in the scoped memory area 330. Scoped memory area 330 includes a thread count 350 of a number of threads configured to access the scoped memory area 330. The thread count is useful to determine when the scoped memory area 330 is discardable, such as when the thread count 350 indicates that no threads access the scoped memory area 330. Scoped memory area further includes a reference to a parent scoped memory area, such as a primordial scoped memory area, as is well known in the art.
References 310 to 320 are object references within threads 302 to 308. For example, each such reference can include a memory address, base and offset, or pointer to a software object. Thread 302 includes a reference 310 to object 340 in the garbage collected heap 332. Thread 302 further includes a reference 312 to object 338 in the garbage collected heap 332. Thread 304 includes a reference 313 to object 338 in the garbage collected heap and a reference 314 to object 336 in the scoped memory area 330. Thread 306 includes references 316 and 318 to objects 336 and 334 respectively in the scoped memory area. Thread 308 includes a reference 320 to object 334 in the scoped memory area. Thus, thread 302 includes only references to objects in the garbage collected heap. Threads 306 and 308 include only references to objects in the scoped memory area 330. Thread 304 includes reference to objects in both the garbage collected heap 332 and the scoped memory area 330. In execution threads 302 and 304 with references to objects allocated in the garbage collected heap 332 will be affected by the periodic processing of the garbage collector in that heap 332. On the other hand, threads 306 and 308 are not affected by the garbage collector in the garbage collected heap 332 since they have no references to objects in that heap 332. In contrast, threads 306 and 308 employ scoped memory area 330 that is discardable in dependence on the thread count 350.
In accordance with one illustrative embodiment of the present invention, the techniques of garbage collection are further applied to the scoped memory area 330 by way of a scoped memory garbage collector 360. The scoped memory garbage collector 360 is a software or hardware component of the runtime environment operable to identify and discard unused objects in the scoped memory area 330. In this way, the advantages of garbage collection are provided for the scoped memory area 330, and threads that are restricted to using only scoped memory, such as critical real time threads, are able to benefit from garbage collection.
FIG. 4 is an example flowchart of a method of the scoped memory garbage collector 360 of FIG. 3 in accordance with one illustrative embodiment of the present invention. At step 402 the method identifies a set of root references for objects allocated in the scoped memory area 330. Root references are references leading directly or indirectly to objects allocated in the scoped memory area 330. Such references are normally at the root of a chain of references, such as stored in processor registers, on a software program stack (including local variables and temporaries), and global variables. The identification of a set of root references is inherent in methods of garbage collection and is well known in the art. For example, such an identification will include identifying all software threads having direct or indirect references to objects allocated in the scoped memory area 330. Thus, an appropriate set of root references for a scoped memory area 330 may include all references in threads having entered the scoped memory area 330.
At step 404, each reference in the set of root references is traversed recursively to identify objects that are allocated in the scoped memory area 330. Recursive traversal of a reference involves accessing all references stored at a referenced location in a recursive fashion as is undertaken, for example, when fully traversing a linked list or tree data structure. The purpose of the traversal is to identify all objects allocated in the scoped memory area 330 that are referenced from a root reference. When such an object is identified, it is marked. Marking of objects can be undertaken by any suitable fashion, such as by setting a flag within the object itself or by generating or amending an entry corresponding to the object in a separate data structure, such as a table. In this way, all objects in the scoped memory area 330 referenced by the software application in execution (such as via the threads 302 to 308) are marked.
At step 406 the method identifies objects in the scoped memory area 330 that are not marked. The objects identified at step 406 are unused objects—that is, objects to which there exists no identified reference in the software application. The objects identified at step 406 are discarded. Discarding an object can include, inter alia, marking the object as discardable, marking the memory consumed by an object as free memory, or actively deleting the object from memory.
FIG. 5 is an example block diagram of a scoped memory area 503 in use by a thread 502 in accordance with one illustrative embodiment of the present invention. Thread 502 enters the scoped memory area 503 for the allocation of objects B and A. Other objects C, D, E and F are also allocated in the scoped memory area 503 and are variously referenced and unreferenced. The method of FIG. 4 will now be applied to the arrangement of FIG. 5 to illustrate a method of the preferred embodiment of the present invention.
Initially, at step 402, a set of root references for all objects allocated in the scoped memory area 503 is identified. This set includes only references in thread 502 since thread 502 is the only thread having references to objects in the scoped memory area 503. Thus, the set of root references comprises references 510 and 504. At step 404 the method traverses each of the root reference 510 and 504 and marks objects in the scoped memory area 503. Thus, traversing root reference 510, object B is marked since it is allocated in the scoped memory area 503. The marking of object B is indicated by a tick. Traversing root reference 504 object A is marked since it is allocated in the scoped memory area 503. The traversal of root reference 504 is recursively performed for reference 506 to object E which is also in the scoped memory area 503 and so is also marked. Further, the traversal of reference 508 to object D which is also in the scoped memory area 503 and so is also marked. Thus, on completion of step 404, objects B, A, E and D are marked in the scoped memory area 503 since these objects are referenced directly or indirectly from the root references 510 and 504. Finally, at step 406, the method identifies objects allocated in the scoped memory area 503 that are not marked. Objects C and F are therefore identified and these objects are discarded. In this way, a process of garbage collection can be applied to the scoped memory area 503.
In order to undertake the operations of FIG. 4 described above, it may be necessary to prevent changes to the references to, and objects within, a scoped memory in order that the scoped memory garbage collector 360 can operate on the scoped memory area 330 in a static state. To achieve this, all threads accessing the scoped memory area 330 are preferably transitioned to a suspended state in which execution of the software threads is stopped. Correspondingly, on completion of the garbage collection process, all threads accessing the scoped memory area 330 are preferably transitioned from the suspend state to an active state.
Further, on completion of a garbage collection operation it can be preferable to undertake a tidy-up operation on the scoped memory area 330 using compaction. Compaction is a process of moving objects within a memory space such that the memory they occupy is substantially contiguous, leaving relatively large areas of free memory. In contrast, in uncompacted memory, many allocated objects can be widely dispersed leaving many small areas of free memory there-between. Compaction will involve moving objects allocated in the scoped memory area 330 to contiguous (or substantially contiguous) memory locations in the scoped memory area 330. Additionally, references to each moved object will need to be updated to correspond to a new memory location for the object. This is known as a process of fixing up or modifying references.
The identification of root references has so far been described in terms of direct or indirect references to a scoped memory area in software threads. However, the rules of scoped memory described above further provide for references from inner scopes to outer scopes. Such references can also be used to indicate that an object allocated in a scoped memory area is used, and these references are therefore important to include in the set of roots.
FIG. 6 is an example block diagram of nested scoped memory areas 602, 604, 606 and 608 in accordance with one illustrative embodiment of the present invention. Scoped memory area 602 is a scoped memory area for garbage collection and includes allocated objects G, H, and I. The scoped memory area 602 has further two inner scoped memory areas 604 and 606. Scoped memory area 604 includes allocated objects J, K, and L. Object J includes a reference to object G and object K includes a reference to object H. Scoped memory area 606 includes allocated object M. Scoped memory area 606 further includes an inner scoped memory area 608 having a single object N. Object N includes a reference to object I in scoped memory area 602. Thus, the scoped memory area for garbage collection 602 has three inner scoped memory areas: scoped memory area 604; scoped memory area 606; and scoped memory area 608. Such inner scoped memory areas are also known as child scoped memory areas since they have parental relationships 620, 622 and 624, with their immediately outer scoped memory areas. In the method of garbage collection of FIG. 4, references from objects within these child scoped memory areas are preferably to be included in the set of root references to ensure objects are not discarded from the scoped memory area for garbage collection 602 that are referenced by inner scoped memory areas.
FIG. 7 a is an example flowchart of a method for identifying root references to objects allocated in the scoped memory area for garbage collection 602 in accordance with one illustrative embodiment of the present invention. Initially, at step 702, the method identifies a set of child scoped memory areas of the scoped memory area for garbage collection 602. The set of child scoped memory therefore comprises scoped memory areas 604, 606, and 608. Subsequently, at step 704, the method identifies, in each child scoped memory area of the set, references to objects allocated in the scoped memory area for garbage collection 602. Thus, references 632, 634, and 636 are identified as references to objects allocated in the scoped memory area for garbage collection 602 and these references are identified as root references at step 706. In this way, references from inner (child) scopes to outer (parent) scopes can be accommodated during a garbage collection process.
The step 702 of the method of FIG. 7 a requires the identification of a set of child scoped memory areas for the scoped memory area for garbage collection 602. Where it is impractical or inefficient to identify strictly only, and all, child scoped memory areas, an alternative approximation is suitable for identifying a set of potential child scoped memory areas.
FIG. 7 b is an example flowchart of a method for identifying a set of potential child scoped memory areas for the garbage collection 602 of FIG. 6 in accordance with one illustrative embodiment of the present invention. At step 722 all threads using the scoped memory area for garbage collection 602 are identified. This can be achieved by parsing all threads in an application and, for each thread, searching for a reference to the scoped memory area for garbage collection 602 in a thread scope stack. At step 724, an identification is made of all scoped memory areas entered by all the identified threads. Subsequently, at step 726, scoped memory areas that are potential child scoped memory areas of the scoped memory area for garbage collection 602 are identified. It may not be practical or efficient to determine exactly which of the scoped memory areas identified at step 724 is a child of the scoped memory area for garbage collection 602, but a comparison of the relative depths of each identified scoped memory area and the scoped memory area for garbage collection will give an indication of potential children. This is because child scoped memory areas will always have a depth that is deeper than their parent. In this way, a set of potential child scoped memory areas can be identified for processing in accordance with steps 704 and 706 of FIG. 7 a.
It will be appreciated by those skilled in the art that the set of root references identified by the methods of FIGS. 7 a and 7 b will be supplemented by software threads having references to objects in the scoped memory area since these are also root references. Subsequently, the complete set of root references can be applied to the method of FIG. 4 for the garbage collection of a scoped memory area.
The aforementioned preferred embodiments provide the advantages of garbage collected memory to a scoped memory area. The garbage collection process can only affect the threads accessing the scoped memory area. The extent of any unfavourable effects of such garbage collection is thus constrained only to those threads accessing the scoped memory area. Since the number of such threads accessing the scoped memory area is typically small (in comparison with all threads in an application), the significance of such effects is also constrained. In spite of this, even these constrained implications of garbage collection in scoped memory areas may not be satisfactory for certain critical real time threads in critical software applications. For example, threads involved in the monitoring or control of critical processes in real time—the delay to which could pose a threat to the successful completion of a process, or to the safety of resources, equipment, or even animal or human life—may not be able to accommodate even the least significant impact of such garbage collection in scoped memory areas. An example of such a thread is a ‘NoHeapRealtime’ thread in the Java runtime environment.
FIG. 8 is an example flowchart of a method of memory management in a real time runtime environment in accordance with one illustrative embodiment of the present invention. The method of FIG. 8 seeks to provide the advantages of garbage collection for such critical real time threads.
Referring to FIGS. 8, 9 a, and 9 b, a critical real time thread 912 is precluded from using a garbage collected heap 332 due to its critical real time nature. Instead, the critical real time thread allocates objects in a scoped memory area 503. The critical real time thread 912 has associated a garbage collection identifier 914 as a software or hardware identifier for indicating when the critical real time thread 912 can reasonably accommodate a garbage collection process of the scoped memory area 503. Such a garbage collection process can be carried out in accordance with the methods of FIGS. 4, 7 a, and 7 b, as discussed above. The garbage collection identifier 914 can form part of the critical real time thread 912, or can be stored separately such as within application data for a runtime execution environment. Further, the garbage collection identifier 914 can be actively set and unset by the critical real time thread 912.
Alternatively, a separate thread monitor 902 can be configured to passively monitor the state of the critical real time thread 912 in order to set and unset the garbage collection identifier 914 (see FIG. 9 b). For example, the thread monitor 902 can be a software or hardware component of a runtime environment operable to examine the execution profile of the critical real time thread 912 to identify states or phases of execution of the thread when the thread is more able to accommodate a garbage collection process.
Referring to the method of FIG. 8, initially, at step 802, the critical real time thread 912 is in execution using a scoped memory area 503. At step 804 the method determines if the garbage collection identifier 914 indicates that the critical real time thread 912 is able to accommodate a garbage collection process. If the garbage collection identifier 914 indicates that the critical real time thread 912 is unable to accommodate a garbage collection process, the method returns to step 802. Alternatively, if the garbage collection identifier 914 indicates that the critical real time thread 912 is able to accommodate a garbage collection process, the method proceeds to step 806 where the garbage collection process is undertaken for the scoped memory area 503 (such as in accordance with the method of FIG. 4). In this way, the undertaking of a garbage collection process of the scoped memory area 503 is constrained to only those times when the critical real time thread 912 is able to accommodate such a process as indicated by the garbage collection identifier 914.
Insofar as embodiments of the invention described are implementable, at least in part, using a software-controlled programmable processing device, such as a microprocessor, digital signal processor or other processing device, data processing apparatus or system, it will be appreciated that a computer program for configuring a programmable device, apparatus or system to implement the foregoing described methods is envisaged as an aspect of the present invention. The computer program may be embodied as source code or undergo compilation for implementation on a processing device, apparatus or system or may be embodied as object code, for example.
Suitably, the computer program is stored on a carrier medium in machine or device readable form, for example in solid-state memory, magnetic memory such as disk or tape, optically or magneto-optically readable memory such as compact disk or digital versatile disk etc., and the processing device utilises the program or a part thereof to configure it for operation. The computer program may be supplied from a remote source embodied in a communications medium such as an electronic signal, radio frequency carrier wave or optical carrier wave. Such carrier media are also envisaged as aspects of the present invention.
It will be understood by those skilled in the art that, although the present invention has been described in relation to the above described example embodiments, the invention is not limited thereto and that there are many possible variations and modifications which fall within the scope of the invention.
The scope of the present invention includes any novel features or combination of features disclosed herein. The applicant hereby gives notice that new claims may be formulated to such features or combination of features during prosecution of this application or of any such further applications derived therefrom. In particular, with reference to the appended claims, features from dependent claims may be combined with those of the independent claims and features from respective independent claims may be combined in any appropriate manner and not merely in the specific combinations enumerated in the claims.

Claims

1. A method, in a data processing system, of memory management in a real time runtime environment having a garbage collected memory heap, the runtime environment including a critical real time thread that is precluded from accessing the garbage collected memory heap, the runtime environment further including a scoped memory area for the allocation of objects therein for access by the critical real time thread, the method comprising:

determining whether the critical real time thread currently accessing the scoped memory area is in a defined state so as to accommodate garbage collection of the scoped memory area; and

in response to a determination that the critical real time thread currently accessing the scoped memory area is in the defined state, initiating a process of garbage collection of the scoped memory area.

2. The method of claim 1, wherein the critical real time thread includes an indicator for indicating whether or not the critical real time thread is in the defined state, and wherein determining whether the critical real time thread is in the defined state comprises determining a state of the indicator.

3. The method of claim 2, wherein the indicator is actively set and unset by the critical real time thread.

4. The method of claim 1, wherein determining whether the critical real time thread is in the defined state comprises:

passively monitoring execution of the critical real time thread to identify behaviors of the critical real time thread that indicate that the critical real time thread is in the defined state.

5. The method of claim 1, wherein the critical real time thread is a NoHeapRealtime thread of a Java runtime environment.

6. An apparatus for memory management in a real time runtime environment having a garbage collected memory heap, the runtime environment including a critical real time thread that is precluded from accessing the garbage collected memory heap, the runtime environment further including a scoped memory area for the allocation of objects therein for access by the critical real time thread, the apparatus comprising:

a processor; and

a memory coupled to the processor, wherein the memory comprises a computer program which configures the processor to:

determine whether the critical real time thread currently accessing the scoped memory area is in a defined state so as to accommodate garbage collection of the scoped memory area; and

initiate, in response to a determination that the critical real time thread currently accessing the scoped memory area is in the defined state, a process of garbage collection of the scoped memory area.

7. The apparatus of claim 6, wherein the critical real time thread includes an indicator for indicating whether or not the critical real time thread is in the defined state, and wherein the computer program configures the processor to determine whether the critical real time thread is in the defined state by determining a state of the indicator.

8. The apparatus of claim 7, wherein the indicator is actively set and unset by the critical real time thread.

9. The apparatus of claim 6, wherein the computer program configures the processor to determine whether the critical real time thread is in the defined state by:

10. The apparatus of claim 6, wherein the critical real time thread is a NoHeapRealtime thread of a Java runtime environment.

11-12. (canceled)

13. The method of claim 1, wherein determining whether the critical real time thread is in the defined state comprises determining a state of a hardware identifier associated with the critical real time thread, the hardware identifier having a first state indicating accommodation of garbage collection operations the scoped memory area, and a second state when garbage collection operations cannot be performed on the scoped memory area.

14. The method of claim 1, wherein the process of garbage collection comprises:

identifying, by a scoped memory garbage collector of the data processing system, a set of root references for objects allocated in the scoped memory area;

recursively traversing and marking, by the scoped memory garbage collector, objects that are referenced from the set of root references and that are allocated in the scoped memory area;

identifying, by the scoped memory garbage collector, objects in the scoped memory area of the scoped memory system that are not so marked; and

discarding the identified objects that are not so marked from the scoped memory system.

15. The method of claim 14, wherein identifying the set of root references further comprises:

identifying a set of one or more child scoped memory areas of the scoped memory area; and

identifying, in each scoped memory area of the set, references to the scoped memory area as root references for objects allocated in the scoped memory area, thereby generating the set of root references.

16. The apparatus of claim 6, wherein the computer program configures the processor to determine whether the critical real time thread is in the defined state by determining a state of a hardware identifier associated with the critical real time thread, the hardware identifier having a first state indicating accommodation of garbage collection operations on the scoped memory area, and a second state when garbage collection operations cannot be performed on the scoped memory area.

17. The apparatus of claim 6, wherein the process of garbage collection comprises:

18. The apparatus of claim 17, wherein identifying the set of root references further comprises:

19. A computer program product in a computer readable medium, comprising computer program code which, when loaded into a computer system and executed thereon, causes the computer system to perform a memory management operation in a real time runtime environment having a garbage collected memory heap, the runtime environment including a critical real time thread that is precluded from accessing the garbage collected memory heap, the runtime environment further including a scoped memory area for the allocation of objects therein for access by the critical real time thread, the memory management operation comprising: