US 7,028,136 B1Grant
Managing idle time and performing lookup operations to adapt to refresh requirements or operational rates of the particular associative memory or other devices used to implement the system
Issue Date:2006-04-11
•24 Claims
•25 Drawing Sheets
Abstract
Methods and apparatus are disclosed for, inter alia, managing idle time and performing lookup operations to adapt to refresh requirements and/or operational rates of the particular associative memory or other devices used to implement the system. A system typically a controller coupled to a circuit including includes multiple associative memories. The controller receives performance information from at least one of the associative memories and generates idle time corresponding to the performance information. This performance information may include an actual refresh requirement for each of the associate memories or the worst-case requirement of one of the associative memories, so that the controller can intermittently provide the actual refresh cycles required, if any, by a particular implementation. This performance information may indicate for some or all instructions the actual time required by a particular instruction, so that the controller can operate based on the actual time required, rather than inefficiently operating on a worst-case time defined in a specification or data sheet.
Metadata
Assignee
- Cisco Technology, Inc.
Inventors
- Bangalore L. Priyadarshan
- Jaushin Lee
- William N. Eatherton
Application Information
Application Number:US 10/355,694
Filing Date:2003-01-31
Priority Date:2002-08-10
Art Unit:2187
Classifications
IPC:
G06F12/00
Patent Drawings (25 sheets)
Description
Cross-Reference to Related Applications
[0001] This is a continuation-in-part of application Ser. No. 10/215,719, filed Aug. 10, 2002, and is hereby incorporated by reference.
Field of the Invention
[0002] This invention especially relates to computer and communications systems, especially network routers and switches; and more particularly, the invention relates to methods and apparatus for managing idle time and performing lookup operations to adapt to refresh requirements or operational rates of the particular associative memory or other devices used to implement the system.
Background of the Invention
[0003] The communications industry is rapidly changing to adjust to emerging technologies and ever increasing customer demand. This customer demand for new applications and increased performance of existing applications is driving communications network and system providers to employ networks and systems having greater speed and capacity (e.g., greater bandwidth). In trying to achieve these goals, a common approach taken by many communications providers is to use packet switching technology. Increasingly, public and private communications networks are being built and expanded using various packet technologies, such as Internet Protocol (IP).
[0004] A network device, such as a switch or router, typically receives, processes, and forwards or discards a packet based on one or more criteria, including the type of protocol used by the packet, addresses of the packet (e.g., source, destination, group), and type or quality of service requested. Additionally, one or more security operations are typically performed on each packet. But before these operations can be performed, a packet classification operation must typically be performed on the packet.
[0005] Packet classification as required for, inter alia, access control lists (ACLs) and forwarding decisions, is a demanding part of switch and router design. The packet classification of a received packet is increasingly becoming more difficult due to ever increasing packet rates and number of packet classifications. For example, ACLs require matching packets on a subset of fields of the packet flow label, with the semantics of a sequential search through the ACL rules. IP forwarding requires a longest prefix match.
[0006] Known approaches of packet classification include using custom application-specific integrated circuits (ASICs), custom circuitry, software or firmware controlled processors, and associative memories, including, but not limited to binary content-addressable memories (binary CAMs) and ternary content-addressable memories (ternary CAMs or TCAMs). Each entry of a binary CAM typically includes a value for matching against, while each TCAM entry typically includes a value and a mask. The associative memory compares a lookup word against all of the entries in parallel, and typically generates an indication of the highest priority entry that matches the lookup word. An entry matches the lookup word in a binary CAM if the lookup word and the entry value are identical, while an entry matches the lookup word in a TCAM if the lookup word and the entry value are identical in the bits that are not indicated by the mask as being irrelevant to the comparison operations.
[0007] Associative memories are very useful in performing packet classification operations. In performing a packet classification, it is not uncommon for multiple lookup operations to be performed in parallel or in series using multiple as sociative memories basically based on a same search key or variant thereof, as one lookup operation might be related to packet forwarding while another related to quality of service determination. These associative memories may be built by different vendors, albeit typically to the same specification. However, some vendors might use different technologies or implement associative memories which operate faster than the specification. If a system is built to handle the worst case or specification values, it might not operate as fast or as efficient as it could if it was tailored to the specific parameters of the implemented devices.
[0008] For example, some vendors might use DRAM or other memory technology which requires idle time (e.g., idle or dead cycles, time delay, etc.) to refresh its memory. A known system always provides a worst case maximum number of idle cycles at minimal timing intervals to ensure that each associative memory has enough time to refresh its memory. However, even though the associative memory does not require refreshing or can refresh faster than the specification requires, the system still provides these idle cycles. Thus, the system does not always operate as fast nor as efficiently as its devices would allow it to operate. Similarly, some instructions or lookup operations may be performed by different implementations of associative memories in different amounts of time. Desired are new functionality, features, and mechanisms in associative memories and other devices to adapt to the capabilities of the devices used to implement a system.
Summary of the Invention
[0009] Methods and apparatus are disclosed for, inter alia, managing idle time and performing lookup operations to adapt to refresh requirements and/or operational rates of the particular associative memory or other devices used to implement the system. Note, the term “idle time” is an extensible term, and is used to describe any time delay or cycles generated or provided, such as, but not limited to idle or dead cycles, time delays, etc. One embodiment includes a controller coupled to a circuit, the circuit includes multiple associative memories. The controller typically receives performance information from at least one of the associative memories and generates idle time corresponding to the performance information.
[0010] In one embodiment, a first associative memory of the multiple associative memory collects associative memory performance information from each of the other associative memories and relays this information or a worst case set of the associative memory performance information to the controller.
[0011] In one embodiment, the performance information includes a refresh requirement and the idle time includes a refresh idle time. The controller intermittently generates the refresh idle time according to the refresh requirement so that at least one of the multiple associative memories can refresh some of its memory during this refresh idle time.
[0012] In one embodiment, the controller identifies a particular instruction to be performed and communicates the particular instruction to at least one particular associative memory A portion of the performance indication indicates a particular idle requirement for the particular instruction, and the controller generates idle time corresponding to the particular idle time requirement.
[0013] In one embodiment, the performance information includes a refresh indication received from a particular associative memory of the multiple associative memories that the one of the multiple associative memories requires time to refresh. In response, idle time is generated to allow the associative memory to refresh its memory.
[0014] In one embodiment, the controller identifies a particular instruction to be performed and communicates the particular instruction and a particular instruction validation pattern field to at least one particular associative memory of the plurality of associative memories. The particular associative memory generates a lookup result based on said particular instruction and copies the validation pattern field into a field of the lookup result. The controller performs a validation check by comparing the validation pattern field of the particular instruction with the validation pattern field received with the result. If they do not match, then there typically is an error in the synchronization between the instructions generated by the controller and results determined by one or more of the associative memories. The controller typically then flushes the lookup circuitry to resynchronize the results with the instructions.
Brief Description of the Drawings
[0015] The appended claims set forth the features of the invention with particularity. The invention, together with its advantages, may be best understood from the following detailed description taken in conjunction with the accompanying drawings of which:
[0016] FIGS. 1A–G are block diagrams of various exemplary systems including one or more associative memories;
[0017] FIG. 2 is a block diagram of an exemplary associative memory used in one embodiment;
[0018] FIGS. 3A–D illustrate various aspects of a control used in one embodiment;
[0019] FIGS. 4A–G illustrate various aspects of an associative memory block used in one embodiment;
[0020] FIGS. 5A–C illustrate various aspects of an output selector used in one embodiment;
[0021] FIGS. 6A , 6C–6D are flow diagrams illustrating processes to validate the receipt order of results used in one embodiment;
[0022] FIG. 6B is a block diagram illustrating a message format used in one embodiment;
[0023] FIGS. 7A–B are flow diagrams illustrating processes to request and/or generate idle time to allow an associative memory to refresh at least a portion of its memory; and
[0024] FIGS. 8A–C are flow diagrams illustrating processes used in one embodiment to adaptively provide idle time required for an instruction.
Detailed Description
[0025] Methods and apparatus are disclosed for, inter alia, managing idle time and performing lookup operations to adapt to refresh requirements and/or operational rates of the particular associative memory or other devices used to implement the system. Embodiments described herein include various elements and limitations, with no one element or limitation contemplated as being a critical element or limitation. Each of the claims individually recites an aspect of the invention in its entirety. Moreover, some embodiments described may include, but are not limited to, inter alia, systems, networks, integrated circuit chips, embedded processors, ASICs, methods, and computer-readable medium containing instructions. The embodiments described hereinafter embody various aspects and configurations within the scope and spirit of the invention, with the figures illustrating exemplary and non-limiting configurations.
[0026] As used herein, the term “packet” refers to packets of all types or any other units of information or data, including, but not limited to, fixed length cells and variable length packets, each of which may or may not be divisible into smaller packets or cells. The term “packet” as used herein also refers to both the packet itself or a packet indication, such as, but not limited to all or part of a packet or packet header, a data structure value, pointer or index, or any other part or identification of a packet. Moreover, these packets may contain one or more types of information, including, but not limited to, voice, data, video, and audio information. The term “item” is used herein to refer to a packet or any other unit or piece of information or data. The phrases “processing a packet” and “packet processing” typically refer to performing some steps or actions based on the packet contents (e.g., packet header or other fields), and such steps or action may or may not include modifying and/or forwarding the packet and/or associated data.
[0027] The term “system” is used generically herein to describe any number of components, elements, sub-systems, devices, packet switch elements, packet switches, routers, networks, computer and/or communication devices or mechanisms, or combinations of components thereof. The term “computer” is used generically herein to describe any number of computers, including, but not limited to personal computers, embedded processing elements and systems, control logic, ASICs, chips, workstations, mainframes, etc. The term “processing element” is used generically herein to describe any type of processing mechanism or device, such as a processor, ASIC, field programmable gate array, computer, etc. The term “device” is used generically herein to describe any type of mechanism, including a computer or system or component thereof. The terms “task” and “process” are used generically herein to describe any type of running program, including, but not limited to a computer process, task, thread, executing application, operating system, user process, device driver, native code, machine or other language, etc., and can be interactive and/or non-interactive, executing locally and/or remotely, executing in foreground and/or background, executing in the user and/or operating system address spaces, a routine of a library and/or standalone application, and is not limited to any particular memory partitioning technique. The steps, connections, and processing of signals and information illustrated in the figures, including, but not limited to any block and flow diagrams and message sequence charts, may be performed in the same or in a different serial or parallel ordering and/or by different components and/or processes, threads, etc., and/or over different connections and be combined with other functions in other embodiments in keeping within the scope and spirit of the invention. Furthermore, the term “identify” is used generically to describe any manner or mechanism for directly or indirectly ascertaining something, which may include, but is not limited to receiving, retrieving from memory, determining, calculating, generating, etc.
[0028] Moreover, the terms “network” and “communications mechanism” are used generically herein to describe one or more networks, communications mediums or communications systems, including, but not limited to the Internet, private or public telephone, cellular, wireless, satellite, cable, local area, metropolitan area and/or wide area networks, a cable, electrical connection, bus, etc., and internal communications mechanisms such as message passing, interprocess communications, shared memory, etc. The term “message” is used generically herein to describe a piece of information which may or may not be, but is typically communicated via one or more communication mechanisms of any type.
[0029] The term “storage mechanism” includes any type of memory, storage device or other mechanism for maintaining instructions or data in any format. “Computer-readable medium” is an extensible term including any memory, storage device, storage mechanism, and other storage and signaling mechanisms including interfaces and devices such as network interface cards and buffers therein, as well as any communications devices and signals received and transmitted, and other current and evolving technologies that a computerized system can interpret, receive, and/or transmit. The term “memory” includes any random access memory (RAM), read only memory (ROM), flash memory, integrated circuits, and/or other memory components or elements. The term “storage device” includes any solid state storage media, disk drives, diskettes, networked services, tape drives, and other storage devices. Memories and storage devices may store computer-executable instructions to be executed by a processing element and/or control logic, and data which is manipulated by a processing element and/or control logic. The term “data structure” is an extensible term referring to any data element, variable, data structure, database, and/or one or more organizational schemes that can be applied to data to facilitate interpreting the data or performing operations on it, such as, but not limited to memory locations or devices, sets, queues, trees, heaps, lists, linked lists, arrays, tables, pointers, etc. A data structure is typically maintained in a storage mechanism. The terms “pointer” and “link” are used generically herein to identify some mechanism for referencing or identifying another element, component, or other entity, and these may include, but are not limited to a reference to a memory or other storage mechanism or location therein, an index in a data structure, a value, etc. The term “associative memory” is an extensible term, and refers to all types of known or future developed associative memories, including, but not limited to binary and ternary content-addressable memories, hash tables, TRIE and other data structures, etc. Additionally, the term “associative memory unit” may include, but is not limited to one or more associative memory devices or parts thereof, including, but not limited to regions, segments, banks, pages, blocks, sets of entries, etc.
[0030] The term “one embodiment” is used herein to reference a particular embodiment, wherein each reference to “one embodiment” may refer to a different embodiment, and the use of the term repeatedly herein in describing associated features, elements and/or limitations does not establish a cumulative set of associated features, elements and/or limitations that each and every embodiment must include, although an embodiment typically may include all these features, elements and/or limitations. In addition, the phrase “means for xxx” typically includes computer-readable medium containing computer-executable instructions for performing xxx.
[0031] In addition, the terms “first,” “second,” etc. are typically used herein to denote different units (e.g., a first element, a second element). The use of these terms herein does not necessarily connote an ordering such as one unit or event occurring or coming before another, but rather provides a mechanism to distinguish between particular units. Additionally, the use of a singular tense of a noun is non-limiting, with its use typically including one or more of the particular item rather than just one (e.g., the use of the word “memory” typically refers to one or more memories without having to specify “memory or memories,” or “one or more memories” or “at least one memory”, etc.). Moreover, the phrases “based on x” and “in response to x” are used to indicate a minimum set of items x from which something is derived or caused, wherein “x” is extensible and does not necessarily describe a complete list of items on which the operation is performed, etc. Additionally, the phrase “coupled to” is used to indicate some level of direct or indirect connection between two elements or devices, with the coupling device or devices modifying or not modifying the coupled signal or communicated information. The term “subset” is used to indicate a group of all or less than all of the elements of a set. Moreover, the term “or” is used herein to identify a selection of one or more, including all, of the conjunctive items.
[0032] Methods and apparatus are disclosed for, inter alia, managing idle time and performing lookup operations to adapt to refresh requirements and/or operational rates of the particular associative memory or other devices used to implement the system. A system typically a controller coupled to a circuit including includes multiple associative memories. The controller receives performance information from at least one of the associative memories and generates idle time corresponding to the performance information. This performance information may include an actual refresh requirement for each of the associate memories or the worst-case requirement of one of the associative memories, so that the controller can intermittently provide the actual refresh cycles required, if any, by a particular implementation. This performance information may indicate for some or all instructions the actual time required by a particular instruction, so that the controller can operate based on the actual time required, rather than inefficiently operating on a worst-case time defined in a specification or data sheet.
[0033] In one embodiment, the performance information includes a refresh indication received from a particular associative memory of the multiple associative memories that the one of the multiple associative memories requires time to refresh. In response, idle time is generated to allow the associative memory to refresh its memory.
[0034] In one embodiment, a first associative memory of the multiple associative memory collects associative memory performance information from each of the other associative memories and relays this information or a worst case set of the associative memory performance information to the controller.
[0035] In one embodiment, the controller identifies a particular instruction to be performed and communicates the particular instruction and a particular instruction validation pattern field to at least one particular associative memory of the plurality of associative memories. The particular associative memory generates a lookup result based on said particular instruction and copies the validation pattern field into a field of the lookup result. The controller performs a validation check by comparing the validation pattern field of the particular instruction with the validation pattern field received with the result. If they do not match, then there typically is an error in the synchronization between the instructions generated by the controller and results determined by one or more of the associative memories. The controller typically then flushes the lookup circuitry to resynchronize the results with the instructions.
[0036] One embodiment performs lookup operations using associative memories, including, but not limited to modifying search keys within an associative memory based on modification mappings, forcing a no-hit condition in response to a highest-priority matching entry including a force no-hit indication, selecting among various associative memory blocks or sets or banks of associative memory entries in determining a lookup result, and detecting and propagating error conditions. In one embodiment, each block retrieves a modification mapping from a local memory and modifies a received search key based on the mapping and received modification data. In one embodiment, each of the associative memory entries includes a field for indicating that a successful match on the entry should or should not force a no-hit result. In one embodiment, an indication of which associative memory sets or banks or entries to use in a particular lookup operation is retrieved from a memory.
[0037] One embodiment performs error detection and handling by identifying, handling and communication errors, which may include, but is not limited to array parity errors in associative memory entries and communications errors such as protocol errors and interface errors on input ports. Array parity errors can occur as a result of failure-in-time errors which are typical of semiconductor devices. One embodiment includes a mechanism to scan associative memory entries in background, and to identify any detected errors back to a control processor for re-writing or updating the flawed entry. In one embodiment, certain identified errors or received error conditions are of a fatal nature in which no processing should be performed. For example, in one embodiment, a fatal error causes an abort condition. In response, the device stops an in-progress lookup operation and just forwards error and possibly no-hit signals. Typically, these signals are generated at the time the in-progress lookup operation would have generated its result had it not been aborted so as to maintain timing among devices in a system including the associative memory.
[0038] In one embodiment, including cascaded or connected associative memory devices, error status messages indicating any error type and its corresponding source are propagated to indicate the error status to the next device and/or a control processor. In addition, the communicated signal may indicate and generate an abort condition in the receiving device. In one embodiment, the receiving device does not perform its next operation or the received instruction, or it may abort its current operation or instruction. Moreover, the receiving device may or may not delay a time amount corresponding to that which its processing would have required in performing or completing the operation or instruction so as to possibly maintain the timing of a transactional sequence of operations.
[0039] FIGS. 1A–G are block diagrams of various exemplary systems and configurations thereof, with these exemplary systems including one or more embodiments for managing idle time and performing lookup operations using associative memories. First, FIG. 1A illustrates one embodiment of a system, which may be part of a router or other communications or computer system, for performing lookup operations to produce results which can be used in the processing of packets. In one embodiment, control logic 110, via signals 111, programs and updates associative memory or memories 115, such as, but not limited to one or more associative memory devices, banks, and/or sets of associative memory entries which may or may not be part of the same associative memory device and/or bank. In one embodiment, control logic 110 also programs memory 120 via signals 123. In one embodiment, control logic 110 includes custom circuitry, such as, but not limited to discrete circuitry, ASICs, memory devices, processors, etc.
[0040] In one embodiment, packets 101 are received by packet processor 105. In addition to other operations (e.g., packet routing, security, etc.), packet processor 105 typically generates one or more items, including, but not limited to one or more packet flow identifiers based on one or more fields of one or more of the received packets 101 and possibly from information stored in data structures or acquired from other sources. Packet processor 105 typically generates a lookup value 103 which is provided to control logic 110 for providing control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memory or memories 115, which perform lookup operations and generate one or more results 117. In one embodiment, a result 117 is used is by memory 120 to produce a result 125. Control logic 110 then relays result 107, based on result 117 and/or result 125, to packet processor 105. In response, one or more of the received packets are manipulated and forwarded by packet processor 105 as indicated by packets 109. Note, results 117, 125 and 107 may include indications of error conditions.
[0041] FIG. 1B illustrates one embodiment for managing idle time and performing lookup operations using associative memories, including, but not limited to modifying search keys within an associative memory based on modification mappings, forcing a no-hit condition in response to a highest-priority matching entry including a force no-hit indication, selecting among various associative memory blocks or sets or banks of associative memory entries in determining a lookup result, and detecting and propagating error conditions. Control logic 130, via signals 132, programs associative memory or memories 136. In addition, control logic 130 provides control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memory or memories 136, which perform lookup operations to generate results, error, and control signals 134, which are received by control logic 130. Note, as used herein “control signals” is an expansive term which may include, but is not limited to performance information, refresh or idle requirements or requests, flow control information (e.g., external or internal such as controlling queues, etc.), acknowledgements, etc.
[0042] FIG. 1C illustrates one embodiment for managing idle time and performing lookup operations using associative memories, including, but not limited to modifying search keys within an associative memory based on modification mappings, forcing a no-hit condition in response to a highest-priority matching entry including a force no-hit indication, selecting among various associative memory blocks or sets or banks of associative memory entries in determining a lookup result, and detecting and propagating error conditions. Control logic 140, via signals 141–143, programs associative memories 146–148. In addition, control logic 140 provides control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memories 146–148, which perform lookup operations to generate results, error, and control signals 144–145. As shown each progressive stage forwards error messages to a next associative memory stage or to control logic 140. For example, associative memory 148 relays received error indications via signals 144 via signals 145 to control logic 140. Control signals 144–145 may also communicate performance information and/or refresh requests.
[0043] Moreover, in one embodiment, a synchronization bit or other validation pattern field is included in messages 141–145 sent between devices 140 and 146–148, with the value being set or changed at predetermined periodic intervals such that each device 140, 146–148 expects the change. One embodiment uses a single synchronization bit, and if this bit is set in the request or input data 141–145 to a device 146–148, then the device 146–148 will set this bit in the corresponding reply or output data 143–145. For example, in one embodiment, control processor or logic 140 sets the sync bit in its request data 141 periodically, say once in every eight requests. Control processor or logic 140 also monitors the sync bit in the reply data 145. If any kind of error altered the request-reply association (or transaction timing) between the control processor or logic 140 and the associative memories 146–148, then control processor or logic 140 can detect it and recover from that error (by flushing the pipeline, etc.)
[0044] In this manner, devices, especially those as part of a transactional sequence, can synchronize themselves with each other. Resynchronization of devices may become important, for example, should an error condition occur, such as an undetected parity error in a communicated instruction signal (e.g., the number of parity errors exceed the error detection mechanism). There is a possibility that a parity error in an instruction goes undetected and that completely changes the transaction timing. Also, there could be other types of “unknown” errors that can put the control processor or logic and the associative memory chain out of synchronization.
[0045] FIG. 1D illustrates one embodiment for managing idle time and performing lookup operations using associative memories, including, but not limited to modifying search keys within an associative memory based on modification mappings, forcing a no-hit condition in response to a highest-priority matching entry including a force no-hit indication, selecting among various associative memory blocks or sets or banks of associative memory entries in determining a lookup result, and detecting and propagating error conditions. Control logic 150, via signals 151–153, programs associative memories 156–158. In addition, control logic 150 provides control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memories 156–158, which perform lookup operations to generate results, error, and control signals 154–155. Control signals 154–155 may also communicate performance information and/or refresh requests. As shown each progressive stage forwards error messages to a next associative memory stage or to control logic 150. For example, associative memory 158 relays received error indications via signals 154 via signals 155 to control logic 150. Moreover, in one embodiment, a synchronization bit or other validation pattern field is included in messages 151–155.
[0046] FIG. 1E illustrates one embodiment for managing idle time and performing lookup operations using associative memories, including, but not limited to modifying search keys within an associative memory based on modification mappings, forcing a no-hit condition in response to a highest-priority matching entry including a force no-hit indication, selecting among various associative memory blocks or sets or banks of associative memory entries in determining a lookup result, and detecting and propagating error conditions. Control logic 160, via signals 161–163, programs associative memories 166–168. In addition, control logic 160 provides control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memories 166–168, which perform lookup operations to generate results and error signals 164–165 which are communicated to control logic 160. Control signals 164–165 may also communicate performance information and/or refresh requests.
[0047] FIG. 1F illustrates one embodiment for managing idle time and performing lookup operations using associative memories. Control logic 180, via signals 181–183, programs associative memories 186–188. In addition, control logic 180 provides control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memories 186–188, which perform lookup operations to generate results and error signals 184–185 which are communicated to control logic 180. Control signals 1886–187 may communicate performance information and/or refresh requests. In one embodiment, control logic 180 receives performance information from each associative memory 186–188. In one embodiment, associative memory 186 collects performance information from each downstream associative memory 188, and then relays all performance information of merely the worst-case performance information based on its own and the received performance information. In one embodiment, associative memories 186–188 indicated flow control signals 186–187 to indicate when it is ready to perform a next operation.
[0048] FIG. 1G illustrates a system 190, which may be part of a router or other communications or computer system, and which manages idle time and performs lookup operations using associative memories. Also, in one embodiment, system 190 distributes distributing entries among associative memory units and selectively enables less than all of the associative memory units when performing a lookup operation. In one embodiment, system 190 includes a processing element 191, memory 192, storage devices 193, one or more associative memories 194, and an interface 195 for connecting to other devices, which are coupled via one or more communications mechanisms 199 (shown as a bus for illustrative purposes).
[0049] Various embodiments of system 190 may include more or less elements. The operation of system 190 is typically controlled by processing element 191 using memory 192 and storage devices 193 to perform one or more tasks or processes, such as programming and performing lookup operations using associative memory or memories 194. Memory 192 is one type of computer-readable medium, and typically comprises random access memory (RAM), read only memory (ROM), flash memory, integrated circuits, and/or other memory components. Memory 192 typically stores computer-executable instructions to be executed by processing element 191 and/or data which is manipulated by processing element 191 for implementing functionality in accordance with one embodiment of the invention. Storage devices 193 are another type of computer-readable medium, and typically comprise solid state storage media, disk drives, diskettes, networked services, tape drives, and other storage devices. Storage devices 193 typically store computer-executable instructions to be executed by processing element 191 and/or data which is manipulated by processing element 191 for implementing functionality in accordance with one embodiment of the invention.
[0050] In one embodiment, processing element 191 receives performance information from associative memory or memories 194. In one embodiment, processing element 191 provides control and data information (e.g., lookup words, modification data, profile IDs, etc.) to associative memory or memories 194, which perform lookup operations to generate lookup results and possibly error indications, which are received and used by processing element 191 and/or communicated to other devices via interface 195.
[0051] FIG. 2 illustrates an associative memory 200 used in one embodiment for performing lookup operations using associative memories, including, but not limited to modifying search keys within an associative memory based on modification mappings, forcing a no-hit condition in response to a highest-priority matching entry including a force no-hit indication, selecting among various associative memory blocks or sets or banks of associative memory entries in determining a lookup result, and detecting and propagating error conditions. As shown, control logic 210 receives input control signals 202 which may include programming information. In turn, control logic 210 may update information and data structures within itself, program/update associative memory blocks 218–219, and/or output selectors 231–232. Note, in one embodiment, each of the associative memory blocks 218–219 include one or more associative memory sets or banks of associative memories entries, and logic or circuitry for performing lookup operations. In one embodiment, control logic 210 receives flow and/or other (e.g., performance information) control signals 243 and generates flow and/or other (e.g., performance information) control signals 245.
[0052] In one embodiment, input data 201, which may include, but is not limited to search keys and modification data, is received by associative memory 200 and distributed to associative memory blocks 218–219, and possibly forwarded to other downstream associative memories in a cascaded configuration. In addition, input control information 202, which may include, but is not limited to profile IDs (e.g., a value), instructions, programming information, is received by control logic 210, and possibly forwarded to other downstream associative memories in a cascaded configuration. In addition, in one embodiment, previous stage lookup results and/or error indications are received from previous stage associative memories in a cascaded configuration or from other devices by control logic 210. Note, in one embodiment, input data 201, input control 202, previous stage results and errors 203, and/or portions thereof are communicated directly to associative memory blocks 218–219 and/or output selectors 231–232. Note, in one embodiment (and as shown in FIG. 2 ), input data 201, input control 202, previous stage results and errors 203 are respectively buffered in queues 205–207 to synchronize their arrival. Control logic receives an indication when the data is available and releases the buffered information via sync signal 209.
[0053] Control logic 210 possibly processes and/or forwards the received information via block control signals 211–212 to associative memory blocks 218–219 and via selector control signals and previous stage results 215 (which typically includes the received profile ID) to output selectors 231–232. In addition, control logic 210 may generate error signals 216 based on a detected error in the received information or in response to received error condition indications. Note, in one embodiment, control logic 210 merely splits or regenerates a portion of or the entire received input control 202 and optional previous stage results and errors 203 signals as selector control signals and previous stage results signals 215 and/or error signals 216. In addition, control logic 210 could initiate an abort operation wherein a lookup operation will not occur because of a detected or received notification of an error condition.
[0054] In one embodiment, control logic 210 identifies data representing which associative memory blocks 218–219 to enable, which associative memory blocks 218–219 each output selector 231–232 should consider in determining its lookup result, and/or modification mappings each associative memory block 218–219 should use in modifying an input search key. In one embodiment, this data is retrieved, based on received input control information 202 (e.g., a profile ID or other indication), from one or more memories, data structures, and/or other storage mechanisms. This information is then communicated as appropriate to associative memory blocks 218–219 via block control signals 211–212, and/or output selectors 231–232 via selector control signals and previous stage results signals 215.
[0055] In one embodiment, associative memory blocks 218–219 each receive a search key and possibly modification data via signal 201, and possibly control information via block control signals 211–212. Each enabled associative memory block 218–219 then performs a lookup operation based on the received search key, which may include generating a lookup word by modifying certain portions of the search key based on received modification data and/or modification mappings. Each associative memory 218–219 typically generates a result 228–229 which are each communicated to each of the output selectors 231–232. In one embodiment, each associative memory block 218–219 that is not enabled generates a no-hit signal as its corresponding result 228–229. In one embodiment, output selectors 231–232 receive an indication of the associative memory blocks 218–219 that is not enabled.
[0056] Output selectors 231 evaluate associative memory results 228–229 to produce results 240. In one embodiment, each output selector has a corresponding identified static or dynamic subset of the associate memory results 228–229 to evaluate in determining results 240. In one embodiment, an identification of this corresponding subset is provided to each output selector 231–232 via selector control signals 215. In one embodiment, each of the output selectors 231–232 receives a profile ID via selector control signals 215 and performs a memory lookup operation based on the received profile ID to retrieve an indication of the particular associate memory results 228–229 to evaluate in determining results 240.
[0057] Moreover, in one embodiment, results 240 are exported over one or more output buses 240, each typically connected to a different set of one or more pins of a chip of the associative memory. In one embodiment, the number of output buses used and their connectivity to outputs selectors 231–232 are static, while in one embodiment the number of output buses used and their connectivity to outputs selectors 231–232 are configurable, for example, at initialization or on a per or multiple lookup basis. In one embodiment, an output bus indication is received by an output selector 231–232, which uses the output bus indication to determine which output bus or buses to use. For example, this determination could include, but is not limited to a direct interpretation of the received output bus indication, performing a memory read operation based on the received output bus indication, etc. In one embodiment, an output selector 231–232 performs a memory access operation based on a profile ID to determine which output bus or buses to use for a particular lookup operation. Thus, depending on the configuration, a single or multiple output buses/pins can selectively be used to communicate results 240, with this decision possibly being made based on the tradeoff of receiving multiple results simultaneously versus the number of pins required.
[0058] Associative memory 200 provides many powerful capabilities for simultaneously producing one or more results 240. For example, in one embodiment, based on a received profile ID, control logic 210 identifies which of the one or more associative memory blocks 218–219 to enable and then enables them, and provides the profile ID to output selectors 231 for selecting a lookup result among the multiple associative memory blocks 218–219. Each of the associative memory blocks 218–219 may receive/identify a modification mapping based on the profile ID, with this modification mapping possibly being unique to itself. This modification mapping can then be used in connection with received modification data to change a portion of a received search key to produce the actual lookup word to be used in the lookup operation. Also, certain entries may be programmed with force no-hit indications to generate a no-hit result for the corresponding associative memory block 218–219 should a corresponding entry be identified as the highest priority entry matching the lookup word. Each of these enabled associative memories 218–219 typically generate a result (e.g., no-hit, hit with highest priority matching entry or location thereof identified) which is typically communicated to each of the output selectors 231–232. Note, in one embodiment, the results are only communicated to the particular output selectors 231–232 which are to consider the particular result in selecting their respective highest priority result received from associative memory blocks 218–219 and possibly other lookup results from previous stage associative memories. Additionally, in certain configurations, multiple associative memories 200 are cascaded or coupled in other methods so that results from one or more stages may depend on previous stage results, such that a lookup can be programmed to be performed across multiple associative memories 200. These and other constructs provided by associative memory 200 and configurations thereof provide powerful programmable lookup search capabilities and result selection mechanisms using one or more stages of associative memories 200, each including N associative memories blocks 218–219 and M output selectors 231–232. In one embodiment, the actual values of N and M may vary among associative memories 200.
[0059] FIG. 3A illustrates a control 300 (which may or may not correspond to control logic 210 of FIG. 2 ) of an associative memory used in one embodiment. As shown, control 300 includes control logic 310 and memory 311. In one embodiment, programming signals 303 are received, and in response, one or more data structures in memory 311 are updated. In addition, control logic generates programming signals 318. In one embodiment, programming 318 is the same as programming signals 303 and thus a physical connection can be used rather than passing through control logic 310. One embodiment of a programming process is illustrated in FIG. 3C , in which processing begins with process block 380. Processing then proceeds to process block 382, wherein programming signals are received. Next, in process block 384, data structures and other elements (e.g., associative memory blocks, output selectors, etc.) are updated. Processing is completed as indicated by process block 386.
[0060] Returning to FIG. 3A , in performing a lookup operation, input data 301, input control 302, and optionally previous stage results and errors 304 (such as in a cascaded associative memory configuration) are received by control logic 310. In response, one or more data structures in memory 311 are referenced. Control logic 310 generates input data 314, block control signals 315, output selector control signals and (optionally) previous stage results 316, and possibly an error signal 319 indicating a detected error condition or a received error indicator. In one embodiment, input data 314 is the same as input data 301 and thus a physical connection can be used rather than passing through control logic 310.
[0061] FIG. 3B illustrates one set of data structures used in one embodiment. Enable array 320 is programmed with an associative memory block enable indicator 325 for each profile ID 321 to be used. Each associative memory block enable indicator 325 identifies which associative memory blocks are to be enabled for a given lookup operation. Thus, based on a profile ID 321 received via input control 302 (FIG. 3A ), enable array 320 can be retrieved from memory 311 (FIG. 3A ), which can then be used to generate associative memory block enable signals included in block control signals 315 (FIG. 3A ). In one embodiment, associative memory block enable indicator 325 is a bitmap data structure, while in one embodiment, associative memory block enable indicator 325 is a list, set, array, or any other data structure.
[0062] Output selector array 330 is programmed with an output selector ID 335 identifying which output selector, such as, but not limited to output selectors 231–232 (FIG. 2 ) for each tuple (profile ID 331, associative memory block ID 332). Thus, based on a profile ID 331 received over via input control 302 (FIG. 3A ), an output selector ID 335 can be identified for each associative memory block ID 332. In one embodiment, output selector ID 335 is a numeric identifier, while in one embodiment, output selector ID 335 is any value or data structure.
[0063] Modification mapping array 340 is programmed with a modification mapping 345 for each tuple (profile ID 341, output selector ID 342). Thus, based on a profile ID 341 received over via input control 302 (FIG. 3A ), a modification mapping 345 can be identified for each output selector ID 342. In one embodiment, each modification mapping is a data structure identifying how to modify a received search key with received modification data.
[0064] FIG. 3D illustrates a process used in one embodiment for initiating a lookup operation. Processing begins with process block 360, and proceeds to process block 362, wherein input data and control signals are received. Next, in process block 364, any previous stage results and error indications are received. As determined in process block 366, if an abort operation should be performed, such as, but not limited to in response to a received fatal error indication or an identified fatal error condition, then processing proceeds to process block 374 (discussed hereinafter). Otherwise, in process block 368, the enable bitmap, output selector configuration, and modification mappings are received based on the profile ID. Next, in process block 370, data and control signals based on the retrieved and received information are forwarded to the associative memory blocks and output selectors. As determined in process block 372, if an error condition is identified or has been received, then in process block 374, an error indication, typically including an indication of the error type and its source is generated or forwarded. Processing is complete as indicated by process block 376.
[0065] FIG. 4A illustrates an associative memory block 400 used in one embodiment. Associative memory block 400 typically includes control logic 410 and associative memory entries, global mask registers, operation logic and priority encoder 412 (e.g., elements for performing the associative memory match operation on a received lookup word). In one embodiment, sets of associative memory entries are grouped into banks of associative memory entries. In one embodiment, programming signals 401 are received, and in response, one or more associative memory entries and/or global mask registers in block 412 are updated. In one embodiment, an associative memory block 400 corresponds to a set or bank of associative memory entries and a mechanism for performing a lookup operation on the set or bank of associative memory entries to produce one or more results. In one embodiment, no mask register is included in associative memory block 400.
[0066] Moreover, one embodiment of associative memory block 400 includes a memory 413 for storing configuration information, which may allow an associative memory block 400 to retrieve the information from memory 413 rather than receive it from another source. For example, in one embodiment, modification mapping data (e.g., modification mapping 345 of FIG. 3B ) or other information is programmed into memory 413. Then, associative memory block 400 retrieves the modification mapping information, such as based on a received profile ID (e.g., rather than receiving the modification mapping signal 404).
[0067] Additionally, in one embodiment, a search key 402, modification data 403, modification mapping 404, an enable signal 405, a global mask enable signal 406, and a global mask select signal 407 are received. In response to performing a lookup operation and/or detecting an error condition, such as a parity fault in one of the associative memory entries, result and error indications 411 are generated. In one embodiment, associative memory entries are checked for parity errors in background. The use of these signals and information in one embodiment are further described in relation to FIGS. 4B–4G .
[0068] Turning to FIG. 4B , one embodiment includes multiple global mask registers 415 for use in a lookup operation on associative memory entries 416. Global mask enable signal 406 enables the use of a global mask register, while global mask select 407 identifies which of multiple masks to apply to each of the associative memory entries. Lookup word 414 is applied to associative memory entries 416, with possibly using one or more of global masks stored in global mask registers 415, to generate hit/no hit indication 417 and possibly hit location 418 and/or error indication 419, which are incorporated directly or indirectly into result and error indications 411 (FIG. 4A ).
[0069] FIG. 4C illustrates an error indication 420 used in one embodiment. As shown, error indication 420 includes an error indication 421 for identifying if any or possibly the number of error indications included therein. For any identified error condition or received error indication, an encoded description of each error is included in one or more of the error descriptors 422–423. In one embodiment, a bitmap is used in one or more of error descriptors 422–423, wherein each bit represents a possible error condition, and the value of the bit indicates whether or not a corresponding error has been identified (including received from a prior component or stage.) In one embodiment, each error descriptor 422–423 corresponds to a different component, interface, or previous stage. In one embodiment, error indication 420 is used by other components in communicating error conditions or lack thereof.
[0070] FIG. 4D illustrates an associative memory entry 430 used in one embodiment. As shown, associative memory entry 430 includes a value 431, an optional mask 432, force no hit indication 433, valid/invalid flag 434, and an error detection value 435. Error detection value 435 may be one or more parity bits, a cyclic redundancy checksum value, or a value corresponding to any other mechanism used for detecting data corruption errors. In one embodiment, value 431 is of a configurable width. In one embodiment, this configurable width includes 80 bits, 160 bits and 320 bits. In one embodiment, such as that of a binary content-addressable memory, no mask field 432 is included. In one embodiment, the width of mask field 432 is variable, and typically, although not required, matches the width of value field 431. In one embodiment, fields 431–435 are stored in a single physical memory; while in one embodiment, fields 431–435 are stored in multiple physical memories.
[0071] FIG. 4E illustrates a mechanism to modify a search key based on modification mapping and modification information used in one embodiment. As shown, a modification mapping bit 443 is used to control selector 440 which selects either search key unit (e.g., one or more bits, bytes, etc.) 441 or modification data unit 442 as the value for lookup unit 445, which is typically a portion of the actual lookup word to be used in matching associative memory entries in a lookup operation.
[0072] FIG. 4F illustrates a mechanism to modify a search key 456 based on modification mapping 450 and modification data 454 used in one embodiment. In one embodiment, modification mapping 450 corresponds to a modification mapping 345 (FIG. 3B ). As shown in FIG. 4F , modification mapping 450 includes a source portion 451 and a destination portion 452. Referring to the lower portion of FIG. 4F , modification data 454 includes four bytes and search key 456 includes eight bytes. The source portion 451 of modification mapping 450 identifies which bytes of modification data 454 are to be used in generating lookup word 458, and the destination portion 452 of modification mapping 450 identifies where the corresponding bytes to be used of modification data 454 are to be placed in lookup word 458, with the remaining bytes coming from search key 456. In other words, modification mapping 450 and modification data 454 are used to replace certain specified data units in search key 456 in producing the value which will be used in matching the associative memory entries. Of course, various embodiments use different numbers of bits and bytes for modification mapping 450 and modification data 454. In one embodiment, modification mapping 450 includes an indication of the portion of search key 456 to modify (e.g., the value of J in one embodiment, the high-order bytes, the low order bytes, etc.).
[0073] FIG. 4G illustrates an associative memory process used in one embodiment in performing a lookup operation. Processing begins with process block 470, and proceeds to process block 472. If the associative memory is not enabled, then processing proceeds to process block 490 wherein a result with a no hit indication is generated, and processing continues to process block 484. Otherwise, in process block 474, the lookup word is determined typically based on the search key, modification mapping, and modification data. Note, in one embodiment, the search key is used as the lookup word and there is no concept of a modification mapping or modification data. Next, in process block 476, the lookup word is used to match the associative memory entries with consideration of a selected and enabled global mask, if any. Note, in one embodiment, there is no concept of a global mask.
[0074] As determined in process block 478, if at least one match has been identified, then processing proceeds to process block 480, otherwise to process block 490, wherein a result with a no hit indication is generated and processing proceeds to process block 484. Otherwise, as determined in process block 480, if the highest priority matching entry includes a force no hit indication, then processing proceeds to process block 490, wherein a result with a no hit indication is generated and processing proceeds to process block 484. Otherwise, in process block 482, a result indicating a hit (i.e., successful match) with the highest priority matching entry identified is generated.
[0075] In process block 484, the result is communicated to at least the identified output selector or selectors. In one embodiment, the output selector to which to communicate the result is identified by output selector ID 335 (FIG. 3B ). As determined in process block 486, if an error condition has been identified or received, then in process block 492, a signal is generated indicating the type and location of the error. In one embodiment, error indication 420 (FIG. 4C ) is used. Processing is complete as indicated by process block 499.
[0076] FIG. 5A illustrates of an output selector 500 (which may or may not correspond to an output selector 231–232 of FIG. 2 ) used in one embodiment. As shown, output selector 500 includes control logic 510 and memory 511. In one embodiment, programming signals 504 are received, and in response, one or more data structures in memory 511 are updated.
[0077] FIG. 5B illustrates one data structure used in one embodiment. Available array 520 is programmed with an associative memory blocks and optionally previous stage results available for use indicator 525 for each profile ID 521 to be used. Each indicator 525 identifies which, if any, associative memory blocks, sets of entries or associative memory banks are to be considered in determining which matching associative entry to select for the ultimate highest-priority matching associative memory entry. In one embodiment, indicator 525 further identifies which previous stage results to consider. Thus, based on a profile ID 521 received over via selector control signal 501 (FIG. 5A ), available array 520 can be retrieved from memory 511 (FIG. 5A ). In one embodiment, there is an implied priority ordering of associative memory blocks and any previous stage results, while in one embodiment this priority ordering for determining the ultimate highest-priority matching entry is programmable and/or variable per lookup operation. In one embodiment, associative memory blocks available for use indicator 525 is a bitmap data structure, while in one embodiment, associative memory blocks available for use indicator 525 is a list, set, array, or any other data structure.
[0078] Returning to FIG. 5A , in the performance of a lookup operation, output selector 500 receives selector control signal 501, which may include a profile ID. In addition, output selector 500 receives any relevant previous stage results 502 and results 503 from zero or more of the associative memory blocks from which the highest-priority entry will be selected, and which, if any, will be identified in generated result 515.
[0079] Moreover, in one embodiment, selector control signal 501 including an enable indication, the enable indication including an enabled or not enabled value, such that in when a not enable value is received, output selector 500 is not enabled and does not select among results from blocks 1–N 503 or optional previous stage results 502. In one embodiment, when not enabled, output selector 500 generates a result signal 515 indicting a no hit, not enabled, or some other predetermined or floating value.
[0080] Additionally, in one embodiment, result 515 is communicated over a fixed output bus, which may or may not be multiplexed with other results 515 generated by other output selectors 500. In one embodiment, the associative memory may include one or more output buses, each typically connected to a single pin of a chip of the associative memory, with the selection of a particular output bus possibly being hardwired or configurable, with the configuration possibly being on a per lookup basis, such as that determined from a received value or configuration information retrieved from a memory (e.g., based on the current profile ID.) In such a configuration, control logic 510 (or other mechanism) typically selects which output bus (and the timing of sending result 515) to use for a particular or all results 515.
[0081] A process used in one embodiment for receiving and selecting a highest-priority associative memory entry, if any, is illustrated in FIG. 5C . Processing begins with process block 540, and proceeds to process block 542, wherein the results from the associative memory blocks and the profile ID are received. In process block 544, the set of associative memory blocks to consider in determining the result is retrieved from a data structure/memory based on the profile ID. In process block 546, any relevant previous stage results are received from coupled associative memories. Next, in process block 548, the highest priority match from the available associative memory block and previous stage results is identified, if any. Then, in process block 550, the result is communicated over a fixed or identified output bus/pin or to some other destination, with the result typically including a no hit indication or a hit indication and an identification of the ultimate highest-priority matching associative memory entry. Processing is complete as indicated by process block 552.
[0082] Next, FIG. 6A illustrates a process used in one embodiment by a controller to generate a validation pattern which can be used to validate/synchronize instructions with received results. Processing begins with process block 600, and proceeds to process block 602 wherein the validation pattern is initialized. This synchronization pattern can be any validation pattern, such as, but not limited to a single bit changed between a one an zero at predetermined times (or randomly if the sequence is communicated to the receiver process of the controller), multiple bits modified in a predetermined pattern (e.g., a sequence number) etc. Any validation pattern can be used which provides a mechanism to validate the receipt of results in the order corresponding to the instructions provided to the associative memory or memories.
[0083] Next, as determined in process block 604, when an error (e.g., the validation pattern of the received result does not match that generated by the corresponding instruction or instructions) has been identified (such as via the process illustrated in FIG. 6D ), then in process block 610, the associative memories are flushed by delaying to allow the results in process to be completed. These results are typically dropped. Processing then returns to process block 602.
[0084] Otherwise, in process block 606, an instruction to be performed is identified and the current value of the validation pattern is inserted in the instruction, and optionally, especially if a non-deterministic or random validation pattern is used, this current value of the validation pattern is communicated to the receiver (e.g., via a message, stored in memory or a queue, etc.). Next, in process block 608, the instruction is transmitted to one or more associative memories. In process block 610, a next current value of the validation pattern receiver is determined, and processing returns to process block 604.
[0085] FIG. 6B illustrates an exemplary message format 630 used in one embodiment. As shown, message format 630 includes an instruction, data and/or result field 631, a validation field 632, and a refresh request field 633. Note, all messages do not include all fields, and may include additional or different fields.
[0086] FIG. 6C illustrates a process used by an associative memory in one embodiment. Processing begins with process block 640, and proceeds to process block 642, wherein a message is received. Next, in process block 644, the value of the validation pattern is extracted, and the specified instruction is performed. Next, in process block 646, a message is prepared with the result and possibly other values, as well as the extracted value of the validation field. In process block 648, the message is sent. Processing returns to process block 642.
[0087] FIG. 6D illustrates a process used by a receiver process, such in a controller, used in one embodiment. Processing begins with process block 660, and proceeds to process block 662, wherein the validation pattern is initialized. Note, the sending process (e.g., that illustrated in FIG. 6A ) must use the same validation pattern as this process. In process block 664, a message is received. In process block 666, the value of the validation pattern is extracted from the message. As determined in process block 668, if the extracted validation pattern data matches that which was expected (e.g., matching that of the corresponding instruction which caused the received message to be generated), then in process block 670, the received message is processed, and processing returns to process block 672. Otherwise, in process block 672, the sending process is notified of the identified error and processing is delayed to allow for the system to flush, then processing returns to process block 662.
[0088] FIGS. 7A–B are flow diagrams illustrating processes to request and/or generate idle time to allow an associative memory to refresh at least a portion of its memory. In one embodiment, the process illustrated in FIG. 7A is performed by an associative memory or other memory device, while the process illustrated in FIG. 7B is performed by a controller.
[0089] Processing of FIG. 7A begins with process block 700, and proceeds to process block 702, wherein processing loops until the associative memory identifies it requires its memory to be refreshed. Note, as used herein, an associative memory refreshing its memory refers to refreshing some or all, but not necessarily all of its memory locations. Then in process block 704, a refresh flag is inserted or set in a message (e.g., message 630, FIG. 6B ), with the message to be communicated to the controller to provide idle time during which the associative memory can refresh its memory locations. Processing then returns to process block 702.
[0090] Processing of FIG. 7B begins with process block 730, and proceeds to process block 732, wherein a result or other message is received. As determined in process block 734, if one or more associative memories have requested time to refresh its memory, then in process block 736, idle time/cycles corresponding to the refresh request are generated. Processing returns to process block 732.
[0091] FIGS. 8A–C are flow diagrams illustrating processes used in one embodiment to adaptively provide idle time required for an instruction. Note, in one embodiment, the processes illustrated in FIGS. 8A and 8C are performed by a controller, while the process illustrated in FIG. 8B is performed by one or more associative memories.
[0092] Processing of FIG. 8A begins with process block 800, and proceeds to process block 802, wherein all associative memories are signaled to provide their performance parameters and configuration information or an associative memory coupled to the controller is signaled to collect performance parameters and configuration information from the other associative memories. Next, in process block 804, the performance and configuration information is received or retrieved from one or more associative memories. Next, in process block 806, the controller generates instructions and data at appropriate rates and intervals for identified configuration and performance of the one or more associative memories, and provides idle time/cycles for identified refresh requirements of the one or more associative memories.
[0093] Next, processing of FIG. 8B begins with process block 830, and proceeds to process block 832, wherein a request for performance and configuration information is received (or this process automatically proceeds to process block 834 upon initialization). Next, in process block 834, a request is sent/forwarded to the coupled associative memories so that all associative memories will forward their performance and configuration information (as not all associative memories are directly connected to the controller). Next, as determined in process block 836, if the associative memory is connected to the controller (or otherwise designated), then as determined in process block 850, if this associative memory should collect the performance and configuration information, then in process block 852, the associative memory collects the performance and configuration information from the other associative memories; and then in process block 854, the associative memory signals the controller with the configuration and worst-case or all performance information. Otherwise, in process block 838, the associative memory signals its performance and configuration information to another associative memory or controller, and forwards any such received information from another associative memory. Processing is complete as indicated by process block 840.
[0094] Next, processing of FIG. 8C begins with process block 870, and proceeds to process block 872, wherein the controller identifies an instruction to perform. Next, in process block 874, the identified instruction is signaled to one or more associative memories. In process block 876, based on the received performance information, the number of idle cycles or time to delay is identified for the instruction, and in process block 878, the controller delays the corresponding amount of time or cycles. Processing then returns to process block 872 to perform a next instruction.
[0095] In view of the many possible embodiments to which the principles of our invention may be applied, it will be appreciated that the embodiments and aspects thereof described herein with respect to the drawings/figures are only illustrative and should not be taken as limiting the scope of the invention. For example and as would be apparent to one skilled in the art, many of the process block operations can be re-ordered to be performed before, after, or substantially concurrent with other operations. Also, many different forms of data structures could be used in various embodiments. The invention as described herein contemplates all such embodiments as may come within the scope of the following claims and equivalents thereof.
Claims
What is claimed is:
1. A method for managing idle time in a system including a controller coupled to a circuit, the circuit including a plurality of associative memories, the method comprising:
the controller receiving performance information from at least one of the plurality of associative memories; and
the controller generating idle time corresponding to said performance information;
a first associative memory of the plurality of associative memory collecting associative memory performance information from each of the other associative memories of the plurality of associative memories; and
wherein said receiving the performance information includes receiving from the first associative memory a worst case set of said associative memory performance information.
2. The method of claim 1, wherein said performance information includes a refresh requirement and said idle time includes a refresh idle time;
wherein said generating idle time includes intermittently generating the refresh idle time according to the refresh requirement; and
wherein at least one of the plurality of associative memories refreshes its memory during the refresh idle time.
3. The method of claim 2, further comprising:
the controller identifying a particular instruction to be performed; and
the controller communicating the particular instruction to at least one particular associative memory of the plurality of associative memories;
wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said generating idle time includes generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
4. The method of claim 1, further comprising:
the controller identifying a particular instruction to be performed; and
the controller communicating the particular instruction to at least one particular associative memory of the plurality of associative memories;
wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said generating idle time includes generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
5. A method for managing idle time in a system including a controller coupled to a circuit, the circuit including a plurality of associative memories, the method comprising:
the controller receiving performance information from at least one of the plurality of associative memories; and
the controller generating idle time corresponding to said performance information;
wherein said performance information includes a refresh requirement and said idle time includes a refresh idle time;
wherein said generating idle time includes intermittently generating the refresh idle time according to the refresh requirement; and
wherein at least one of the plurality of associative memories refreshes its memory during the refresh idle time.
6. The method of claim 5, further comprising:
the controller identifying a particular instruction to be performed; and
the controller communicating the particular instruction to at least one particular associative memory of the plurality of associative memories;
wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said generating idle time includes generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
7. A method for managing idle time in a system including a controller coupled to a circuit, the circuit including a plurality of associative memories, the method comprising:
the controller receiving performance information from at least one of the plurality of associative memories; and
the controller generating idle time corresponding to said performance information;
the controller identifying a particular instruction to be performed; and
the controller communicating the particular instruction to at least one particular associative memory of the plurality of associative memories;
wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said generating idle time includes generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
8. The method of claim 7, wherein said performance information includes a refresh indication received from a particular associative memory of the plurality of associative memories that said one of the plurality of associative memories requires time to refresh;
wherein said idle time includes a refresh idle time;
wherein said generating idle time includes generating the refresh idle time in response to receiving the refresh indication; and
wherein the particular associative memory refreshes at least a portion of its memory during the refresh idle time.
9. A method for managing idle time in a system including a controller coupled to a circuit, the circuit including a plurality of associative memories, the method comprising:
the controller receiving performance information from at least one of the plurality of associative memories; and
the controller generating idle time corresponding to said performance information;
wherein said performance information includes a refresh indication received from a particular associative memory of the plurality of associative memories that said one of the plurality of associative memories requires time to refresh;
wherein said idle time includes a refresh idle time;
wherein said generating idle time includes generating the refresh idle time in response to receiving the refresh indication; and
wherein the particular associative memory refreshes at least a portion of its memory during the refresh idle time.
10. A method for managing idle time in a system including a controller coupled to a circuit, the circuit including a plurality of associative memories, the method comprising:
the controller receiving performance information from at least one of the plurality of associative memories; and
the controller generating idle time corresponding to said performance information;
the controller identifying a particular instruction to be performed; and
the controller communicating the particular instruction and a particular instruction sync field to at least one particular associative memory of the plurality of associative memories;
said at least one particular associative memory generating a lookup result based on said particular instruction;
the controller receiving a result message including the lookup result and a result sync field; and
the controller validating the lookup result, said validating including comparing the particular instruction sync field with the result sync field.
11. The method of claim 10, wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said generating idle time includes generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
12. The method of claim 10, wherein the particular instruction sync field comprises a single bit.
13. The method of claim 10, wherein the particular instruction sync field includes a plurality of bits.
14. The method of claim 10, further comprising generating a flush amount of idle time in response to determining that the particular instruction sync field does not equal the result sync field.
15. The method of claim 10, wherein said at least one particular associative memory of the plurality of associative memories copies the particular instruction sync field into the result message.
16. The method of claim 10, wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said generating idle time includes generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
17. The method of claim 10, wherein the result message includes a refresh indication;
wherein said idle time includes a refresh idle time;
wherein said generating idle time includes generating the refresh idle time in response to receiving the result message; and
wherein the particular associative memory refreshes at least a portion of its memory during the refresh idle time.
18. The method of claim 10, further comprising:
a first associative memory of the plurality of associative memory collecting associative memory performance information from each of the other associative memories of the plurality of associative memories; and
wherein said receiving the performance information includes receiving from the first associative memory a worst case set of said associative memory performance information.
19. The method of claim 10, wherein said performance information includes a refresh requirement and said idle time includes a refresh idle time;
wherein said generating idle time includes intermittently generating the refresh idle time according to the refresh requirement; and
wherein at least one of the plurality of associative memories refreshes its memory during the refresh idle time.
20. An apparatus for managing idle time, the apparatus comprising:
a controller; and
a circuit, coupled to the controller, including a plurality of associative memories;
wherein the controller includes:
means for receiving performance information from at least one of the plurality of associative memories; and
means for generating idle time corresponding to said performance information;
wherein a first associative memory of the plurality of associative memory includes means for collecting associative memory performance information from each of the other associative memories of the plurality of associative memories; and
wherein said means for receiving the performance information includes means for receiving from the first associative memory a worst case set of said associative memory performance information.
21. An apparatus for managing idle time, the apparatus comprising:
a controller; and
a circuit, coupled to the controller, including a plurality of associative memories;
wherein the controller includes:
means for receiving performance information from at least one of the plurality of associative memories; and
means for generating idle time corresponding to said performance information;
wherein said performance information includes a refresh requirement and said idle time includes a refresh idle time;
wherein said means for generating idle time includes means for intermittently generating the refresh idle time according to the refresh requirement; and
wherein at least one of the plurality of associative memories refreshes its memory during the refresh idle time.
22. An apparatus for managing idle time, the apparatus comprising:
a controller; and
a circuit, coupled to the controller, including a plurality of associative memories;
wherein the controller includes:
means for receiving performance information from at least one of the plurality of associative memories; and
means for generating idle time corresponding to said performance information;
wherein the controller includes:
means for identifying a particular instruction to be performed; and
means for communicating the particular instruction to at least one particular associative memory of the plurality of associative memories;
wherein a portion of said performance indication indicates a particular idle requirement for the particular instruction; and
wherein said means for generating idle time includes means for generating, after said communicating the instruction, idle time corresponding to the particular idle time requirement.
23. An apparatus for managing idle time, the apparatus comprising:
a controller; and
a circuit, coupled to the controller, including a plurality of associative memories;
wherein the controller includes:
means for receiving performance information from at least one of the plurality of associative memories; and
means for generating idle time corresponding to said performance information;
wherein said performance information includes a refresh indication received from a particular associative memory of the plurality of associative memories that said one of the plurality of associative memories requires time to refresh;
wherein said idle time includes a refresh idle time;
wherein said means for generating idle time includes means for generating the refresh idle time in response to receiving the refresh indication; and
wherein the particular associative memory refreshes at least a portion of its memory during the refresh idle time.
24. An apparatus for managing idle time, the apparatus comprising:
a controller; and
a circuit, coupled to the controller, including a plurality of associative memories;
wherein the controller includes:
means for receiving performance information from at least one of the plurality of associative memories; and
means for generating idle time corresponding to said performance information;
wherein the controller includes:
means for identifying a particular instruction to be performed; and
means for communicating the particular instruction and a particular instruction sync field to at least one particular associative memory of the plurality of associative memories, said at least one particular associative memory generating a lookup result based on said particular instruction;
means for receiving a result message including the lookup result and a result sync field; and
means for validating the lookup result, said means for validating including means for comparing the particular instruction sync field with the result sync field.
Patent Citations (84)
| Patent | Date | Inventor | Cited By |
|---|---|---|---|
| US3648254(A) | 1972-03-01 | Beausoleil | Applicant |
| US4296475(A) | 1981-10-01 | Nederlof et al. | Applicant |
| US4791606(A) | 1988-12-01 | Threewitt et al. | Applicant |
| US4996666(A) | 1991-02-01 | Duluk, Jr. | Applicant |
| US5088032(A) | 1992-02-01 | Bosack | Applicant |
| US5319763(A) | 1994-06-01 | Ho et al. | Applicant |
| US5339076(A) | 1994-08-01 | Jiang | Applicant |
| US5383146(A) | 1995-01-01 | Threewitt | Applicant |
| US5404482(A) | 1995-04-01 | Stamm et al. | Applicant |
| US5428565(A) | 1995-06-01 | Shaw | Applicant |
| US5440715(A) | 1995-08-01 | Wyland | Applicant |
| US5450351(A) | 1995-09-01 | Heddes | Applicant |
| US5481540(A) | 1996-01-01 | Huang | Applicant |
| US5515370(A) | 1996-05-01 | Rau | Applicant |
| US5684954(A) | 1997-11-01 | Kaiserswerth et al. | Applicant |
| US5740171(A) | 1998-04-01 | Mazzola et al. | Applicant |
| US5748905(A) | 1998-05-01 | Hauser et al. | Applicant |
| US5802567(A) | 1998-09-01 | Liu et al. | Applicant |
| US5841874(A) | 1998-11-01 | Kempke et al. | Applicant |
| US5842040(A) | 1998-11-01 | Hughes et al. | Applicant |
| US5852569(A) | 1998-12-01 | Srinivasan et al. | Applicant |
| US5856940(A) | 1999-01-01 | Rao | Examiner |
| US5898689(A) | 1999-04-01 | Kumar et al. | Applicant |
| US5920886(A) | 1999-07-01 | Feldmeier | Applicant |
| US5930359(A) | 1999-07-01 | Kempke et al. | Applicant |
| US5956336(A) | 1999-09-01 | Loschke et al. | Applicant |
| US5978885(A) | 1999-11-01 | Clark, II | Applicant |
| US6000008(A) | 1999-12-01 | Simcoe | Applicant |
| US6041389(A) | 2000-03-01 | Rao | Applicant |
| US6047369(A) | 2000-04-01 | Colwell et al. | Applicant |
| US6061368(A) | 2000-05-01 | Hitzelberger | Applicant |
| US6069573(A) | 2000-05-01 | Clark, II et al. | Applicant |
| US6081440(A) | 2000-06-01 | Washburn et al. | Applicant |
| US6091725(A) | 2000-07-01 | Cheriton et al. | Applicant |
| US6097724(A) | 2000-08-01 | Kartalopoulos | Applicant |
| US6134135(A) | 2000-10-01 | Andersson | Applicant |
| US6137707(A) | 2000-10-01 | Srinivasan et al. | Applicant |
| US6141738(A) | 2000-10-01 | Munter et al. | Applicant |
| US6148364(A) | 2000-11-01 | Srinivasan et al. | Applicant |
| US6154384(A) | 2000-11-01 | Nataraj et al. | Applicant |
| US6175513(B1) | 2001-01-01 | Khanna | Applicant |
| US6181698(B1) | 2001-01-01 | Hariguchi | Applicant |
| US6199140(B1) | 2001-03-01 | Srinivasan et al. | Applicant |
| US6219748(B1) | 2001-04-01 | Srinivasan et al. | Applicant |
| US6236658(B1) | 2001-05-01 | Essbaum et al. | Applicant |
| US6237061(B1) | 2001-05-01 | Srinivasan et al. | Applicant |
| US6240003(B1) | 2001-05-01 | McElroy | Applicant |
| US6240485(B1) | 2001-05-01 | Srinivasan et al. | Applicant |
| US6243667(B1) | 2001-06-01 | Kerr et al. | Applicant |
| US6246601(B1) | 2001-06-01 | Pereira | Applicant |
| US6285378(B1) | 2001-09-01 | Duluk, Jr. | Applicant |
| US6289414(B1) | 2001-09-01 | Feldmeier et al. | Applicant |
| US6295576(B1) | 2001-09-01 | Ogura et al. | Applicant |
| US6307855(B1) | 2001-10-01 | Hariguchi | Applicant |
| US6308219(B1) | 2001-10-01 | Hughes | Applicant |
| US6317350(B1) | 2001-11-01 | Pereira et al. | Applicant |
| US6331961(B1) | 2001-12-01 | Kengeri et al. | Examiner |
| US6374326(B1) | 2002-04-01 | Kansal et al. | Applicant |
| US6377577(B1) | 2002-04-01 | Bechtolsheim et al. | Applicant |
| US6389506(B1) | 2002-05-01 | Ross et al. | Applicant |
| US6430190(B1) | 2002-08-01 | Essbaum et al. | Applicant |
| US6434662(B1) | 2002-08-01 | Greene et al. | Applicant |
| US6467019(B1) | 2002-10-01 | Washburn | Applicant |
| US6526474(B1) | 2003-02-01 | Ross | Applicant |
| US6535951(B1) | 2003-03-01 | Ross | Applicant |
| US6546391(B1) | 2003-04-01 | Tsuruoka | Applicant |
| US6563754(B1) | 2003-05-01 | Lien et al. | Examiner |
| US6597595(B1) | 2003-07-01 | Ichiriu et al. | Applicant |
| US6606681(B1) | 2003-08-01 | Uzun | Applicant |
| US6651096(B1) | 2003-11-01 | Gai et al. | Applicant |
| US6658002(B1) | 2003-12-01 | Ross et al. | Applicant |
| US6707692(B1) | 2004-03-01 | Hata | Applicant |
| US6715029(B1) | 2004-03-01 | Trainin et al. | Applicant |
| US6717946(B1) | 2004-04-01 | Hariguchi et al. | Applicant |
| US6738862(B1) | 2004-05-01 | Ross et al. | Applicant |
| US6775737(B1) | 2004-08-01 | Warkede et al. | Applicant |
| US6848035(B1) | 2005-01-01 | Akiyama et al. | Examiner |
| US2001/0038554(A1) | 2001-11-01 | Takata et al. | Applicant |
| US2002/0056020(A1) | 2002-05-01 | Sartore et al. | Examiner |
| US2003/0005146(A1) | 2003-01-01 | Miller et al. | Applicant |
| US2003/0223259(A1) | 2003-12-01 | Roth | Applicant |
| US2003/0231631(A1) | 2003-12-01 | Pullela | Applicant |
| US2004/0030802(A1) | 2004-02-01 | Eatherton et al. | Applicant |
| US2004/0030803(A1) | 2004-02-01 | Eatherton et al. | Applicant |
Non-Patent Literature (15)
- Delgado-Frias et al., Mar. 22-23, 1996, pp. 246-251.Examiner
- Jon P. Wade and Charles G. Sodini, “A Ternary Content Addressable Search Engine,” IEEE Journal of Solid-State Circuits, vol. 24, No. 4, Aug. 1989, pp. 1003-1013.Applicant
- Teuvo Kohonen, Content-Addressable Memories, 1987, pp. 128-129 and 142-144, Springer-Verlang, New York.Applicant
- Brian Dipert, ed., “Special-purpose SRAMs Smooth the Ride,” EDN, Jun. 24, 1999, pp. 93-104.Applicant
- “What is a CAM (Content-Addressable Memory)?,” Application Brief AB-N6, Rev. 2a, Music Semiconductors, Milpitas, CA, Sep. 30, 1998, 4 pages.Applicant
- “Reading Out the Valid LANCAM Memory Entries,” Application Brief AB-N4, Rev. 1a, Music Semiconductors, Milpitas, CA, Sep. 30, 1998, 4 pages.Applicant
- “Extending the LANCAM Comparand,” Application Brief AB-N3, Rev. 1.0a Draft, Music Semiconductors, Milpitas, CA, Sep. 30, 1998, 4 pages.Applicant
- “Advantages of CAM in ASIC-Based Network Address Processing,” Application Brief AB-N11, Rev. 1.2a Draft, Music Semiconductors, Milpitas, CA, Sep. 30, 1998, 4 pages.Applicant
- “Virtual Memory Applications of the MU9C1480A LANCAM,” Application Note AN-N3, Rev. 1a, Music Semiconductors, Milpitas, CA, Sep. 30, 1998, 12 pages.Applicant
- “Using the MU9C1965A LANCAM MP for Data Wider than 128 Bits,” Application Note AN-N19, Rev. 1a, Music Semiconductors, Milpitas, CA, Sep. 30, 1998, 16 pages.Applicant
- “Fast IPv4 and IPv4 CIDR Address Translation and Filtering Using the MUAC Routing CoProcessor (RCP),” Application Note AN-N25, Rev. 0a, Music Semiconductors, Milpitas, CA, Oct. 1, 1998, 16 pages.Applicant
- “Using MUSIC Devices and RCPs for IP Flow Recognition,” Application Note AN-N27, Rev. 0, Music Semiconductors, Milpitas, CA, Oct. 21, 1998, 20 pages.Applicant
- “Wide Ternary Searches Using Music CAMs and RCPs,” Application Note AN-N31, Rev. 0, Music Semiconductors, Milpitas, CA, Apr. 13, 1999, 8 pages.Applicant
- Tong-Bi Pei and Charles Zukowski, “VLSI Implementation of Routing Tables: Tries and CAMS,” Networking in the Nineties, Proceedings of the Annual Joint Conference of the Computer and Communications Societies, New York, Apr. 7, 1991, pp. 515-524, vol. 2, Conf. 10.Applicant
- Anthony McAuley and Paul Francis, “Fast Routing Table Lookup Using CAMs,” Networking: Foundation for the Future, Proceedings of the Annual Joint Conference of the Computer and Communications Societies, Los Alamitos, Mar. 28, 1993, pp. 1382-1391, vol. 2, Conf 12.Applicant