APPARATUS FOR CONTROLLING A DIGITAL

SIGNAL PROCESSOR FOR RADIO ISOLATION

AND ASSOCIATED METHODS

FIELD OF THE INVENTION

[01] This patent document relates generally to real-time systems and, more particularly, to systems and methods for controlling real-time systems, such as a highly integrated RF system.

BACKGROUND OF THE INVENTION

[02] In various types of circuits, interference can cause problems with the operation of the circuits. These types of problems may be especially evident in highly integrated systems, where the operation of one portion of a device can interfere with the operation another portion of the device. For example, in a circuit containing both RF circuitry and digital circuitry, interference between the RF and digital circuitry can cause significant disturbance to the RF circuits, making the implementation on an integrated circuit very difficult.

[03] In a typical prior art RF receiver, transmitter, or transceiver, RF circuitry generally resides in a different circuit partition (e.g., integrated circuit (IC), die, etc.)

than does signal-processing circuitry (e.g., baseband), partly due to the problem of interference. RF circuitry typically includes analog circuitry that has a relatively high sensitivity to noise and interference. Furthermore, the RF circuitry in some applications, for example, in a mobile telephone apparatus, may have to detect signals as small as a few nano- volts in amplitude. The performance of a device may suffer as a result of noise and interference from sources external or even internal to. the communication apparatus.

[04] In a typical communication apparatus, such as a mobile telephone apparatus, digital circuitry produces digital signals with relatively small rise and fall times, or with fast transitions or sharp edges. Furthermore, those signals often have relatively high frequencies. As a result, these high frequency signals, and their harmonics, can interfere with, and adversely impact the performance of, the RF circuitry. As a result, typical prior art communication devices use more than one circuit partition. For example, one partition may include the RF circuitry, while a second partition includes the digital circuitry.

[05] Using more than one partition for RF circuitry and the digital circuitry, however, has several disadvantages, such as increased component count, size, and overall cost, and more potential for decreased reliability and increased manufacturing failures. Therefore, a need exists for highly integrated devices having all circuitry in one partition. For example, in the field of RF communication devices, there is a need

for a highly integrated RF apparatus that includes a complete radio in one partition, die, IC, etc.

SUMMARY OF THE INVENTION

[06] This invention contemplates highly integrated RF apparatus and associated methods, hi one embodiment, a mobile telephone apparatus includes a transceiver adapted to transmit and receive RF signals, a DSP, and an interrupt controller configured to effect the operation of the DSP such that at least a portion of the DSP is disabled while the transceiver is transmitting or receiving RF signals.

[07] In another embodiment, a method of controlling a mobile telephone apparatus includes using a DSP to process RF signals received by the apparatus during bursts, and using a sequencer to generate an interrupt to at least partially shut down the DSP to minimize noise in the mobile telephone apparatus during bursts.

[08] In another embodiment, a method of minimizing interference between a transceiver formed on an integrated circuit and a DSP formed on the same integrated circuit includes using a sequencer to generate an interrupt to disable at least a portion of the DSP to reduce interference between the DSP and the transceiver during time periods when the transceiver is transmitting or receiving signals.

[09] In another embodiment, a method of minimizing interference between a transceiver formed on an integrated circuit and a DSP formed on the same integrated circuit includes operating the DSP in a first mode of operation when the transceiver is not transmitting or receiving signals, and operating the DSP in a second mode when the transceiver is transmitting or receiving signals, wherein the transition from the first mode to the second mode is initiated by an interrupt generated by a sequencer.

[10] Other features and advantages of the present invention will be apparent from the accompanying drawings and from the detailed description that follows below.

BRIEF DESCRIPTION OF THE DRAWINGS

[11] The present invention is illustrated by way of example and not limitation in the figures of the accompanying drawings, in which like references indicate similar elements and in which:

[12] FIG. 1 is a block diagram of a mobile communication apparatus.

[13] FIG. 2 shows a set of events that occur in a general communication system according to the invention.

[14] FIG. 3 is a block diagram of an interrupt control system that may be used with the present invention.

[15] FIG. 4 is a diagram illustrating an interrupt map.

[16] FIG. 5 is a diagram illustrating an interrupt map, including two new interrupts used to trigger the stopping and starting of a processor.

[17] FIG. 6 is a flowchart illustrating the operation of the present invention.

[18] FIG. 7 illustrates an embodiment of the reception of a shut down interrupt and a shut down interrupt service routine.

[19] FIG. 8 illustrates an embodiment of the reception of a restart interrupt and a restart interrupt service routine.

[20] FIG. 9 shows an example of a GSM data frame and the relative timing of the shut down and restart interrupts.

[21] FIG. 10 is a timing diagram illustrating the system operation context for an RF time-slot.

[22] FIG. 11 is a block diagram of an interrupt control system that may be used with the present invention.

[23] FIG. 12 is a flowchart illustrating the operation of the invention shown in FIG. 11.

DETAILED DESCRIPTION

[24] This invention relates to highly integrated RF and digital systems. In one application of the invention, the concepts described below obviate the need for partitioning RF circuitry and digital circuitry (e.g., signal-processing circuitry), as well as address challenges relating to such integration. In one exemplary embodiment of the present invention, in a communication system, RF circuitry and signal- processing circuitry (e.g., digital signal processor (DSP), microprocessor, microcontroller, general -purpose logic circuitry, and the like) may reside in the same circuit partition. Of course, the present invention may be used with any other desired system or device.

[25] Generally, in one example, one aspect of the present invention relates to time- domain isolation of different parts of an apparatus (e.g., isolating RF circuitry from digital circuitry in time). In an example of a communication system having signal processing circuitry and RF circuitry, the RF circuitry generally operates when the signal-processing circuitry is inactive, and vice-versa. Note, however, that the operation of the RF circuitry and the signal -processing circuitry may overlap by some amount. As a consequence, the digital switching noise and associated harmonic content do not interfere with the performance of the RF circuitry, and vice- versa. Another aspect of the present invention relates to techniques for efficiently stopping

and starting signal processing circuitry to achieve the time-domain isolation mentioned above.

[26] In order to provide a context for understanding this description, the following description illustrates one example of an environment in which the present invention may be used. Of course, the invention may also be used in many other types of environments (e.g., a real-time operating system based system). Techniques of the present invention maybe used for any desired applications, including a wireless transmission system such as mobile or cellular communication devices or other wireless devices. Examples of systems where the present invention may be used include, but are not limited to, GSM, GPRS, EDGE, TDMA, PCS, DCS, or any similarly configured communication system.

[27] FIG. 1 is a block diagram of a mobile communication apparatus 10. Note that FIG. 1 shows the apparatus 10 generally, and that such an apparatus will include various other components, as persons of ordinary skill in the art who have the benefit of the description of the invention understand. The apparatus 10 shown in FIG. 1 includes a circuit partition 12 (e.g., an integrated circuit (IC), die, multi-chip module, package, EMI cavity, etc.), including a baseband 14 and RF front-end circuitry 16 (as well as other digital and RF circuitry ). The baseband 14 generally functions to control the operation of the apparatus 10, and may include microcontroller, digital signal processors, logic circuits, memory, etc. A processor or controller may be comprised of a multiple processors, i.e., a plurality of processing elements. The RF

front-end circuitry 16 generally provides an interface to a power amplifier 18 (to facilitate the transmission of signals) and the front-end interface 20 (for the routing of signals to and from the antenna). Of course, the circuit partition 12 includes various other digital and RF circuitry, as persons of ordinary skill in the art who have the benefit of the description of the invention understand. When transmitting signals, the power amplifier 18 provides amplified signals to the front-end interface 20, which then provides the amplified signals to the antenna 22. When receiving signals, signals are received by the antenna 22, and sent to the RF front-end circuitry 16 via the front- end interface 20.

[28] FIG. 2 illustrates a set of events that occur in a general communication system according to the present invention. The example shown in FIG. 2 relates to a system that operates according to a TDMA protocol. Broadly speaking, two alternate events take place in this example: RF reception or transmission, and signal processing. Put another way, the system arranges in time the RF reception or transmission activities and the signal -processing activities so as to avoid or reduce interference between the RF circuitry and the digital signal-processing circuitry.

[29] Referring to FIG. 2, communication systems or apparatus according to exemplary embodiments of the invention use a plurality of RF time-slots 30A, 30B, 30C, and so on. Such systems or apparatus also employ a plurality of signal- processing time-slots 32A, 32B, and so on. Generally, during RF time-slots 30A-30C, the system or apparatus (e.g., the RF front-end circuitry 16 shown in FIG. 1) may

receive RF signals or transmit RF signals, process the received signals, and perform any other desired manipulation of the data. Subsequently, during signal -processing time-slots 32A-32B, the system or apparatus (e.g., the baseband 14) may perform signal-processing tasks.

[30] Alternatively, during RF time-slots 30A-30C, the system or apparatus may transmit RF signals. In this mode of operation, during signal -processing time-slots 32A-32B, the system or apparatus (e.g., the baseband 14) performs signal -processing tasks (e.g., voice, data) and performs any other desired manipulation of the data. Subsequently, during RF time-slots 30A-30C, the system or apparatus (e.g., the RF front-end circuitry 16) may process transmit signals and perform RF operations (for example, up-conversion) and transmit an RF signal.

[31] Note that the signal-processing tasks performed during signal -processing time- slots 32A-32B constitute various signal-processing functions in an RF communication apparatus. Examples of such tasks include modulation, coding, decoding, and the like. Also note that depending on the specific protocol, architecture, and circuitry used, the system or apparatus may receive and transmit simultaneously, as desired. Typically, though, the system either transmits signals or receives signals during any of the RF time-slots, or in bursts. For example, a GSM-compliant system or apparatus, such as a mobile telephone, either receives or transmits RF signals in one or more bursts of activity during RF time-slots. Note that the RF and signal processing time-

slots can overlap or otherwise vary from that shown in FIG. 2. Also, the positions of the RF or signal processing time-slots in a GSM frame may change over time.

[32] Note that the RF time-slots 30A-30C shown in FIG. 2 may have the same or different durations, as desired. Generally, the RF time-slots 30A-30C may have unequal lengths so as to accommodate a wide variety of circuitry, systems, protocols, and specifications, as desired. Each of the RF time-slots 30A-30C may include several other time-slots or a frame, depending on the particular communication protocol or technique used. For example, in a GSM application, each RF period may include a GSM slot, multiple slots, or multiple frames used to transmit, receive, or monitor.

[33] Similarly, the signal-processing time-slots 32A-32B shown in FIG. 2 may have similar or dissimilar durations, as desired. Generally, the signal-processing time- slots may have unequal lengths so as to accommodate a broad array of signal- processing apparatus, circuitry, algorithms, and processing techniques. Each of signal- processing time-slots 32A-32B may include several other time-slots or time divisions, depending on the particular communication protocol and/or signal -processing techniques and the particular circuitry and technology used. For example, a signal- processing time-slot may include several time-slots, with a portion of a particular circuitry active or processing signals during one or more of the time-slots.

[34] Furthermore, the signal-processing tasks may be performed in a serial or multiplexed manner (e.g. , by sharing hardware to perform a variety of tasks), in a

parallel manner (e.g. , by using dedicated hardware for each signal-processing task), or in a combination of the two techniques, as desired. The choice of signal-processing hardware, firmware, and software depends on the design and performance specifications for a given desired implementation, as persons of ordinary skill in the art who have the benefit of the description of the invention understand.

[35] To accomplish the isolation illustrated in FIG. 2, the RF circuitry and the signal-processing circuitry can be activated and deactivated, in correspondence with transitions from one time-slot to another. The activation and deactivation may be accomplished in a variety of ways. As mentioned above, another aspect of the present invention relates to techniques for efficiently stopping and starting processors and other circuitry.

[36] Efficiently starting and stopping a general purpose processor and/or a digital signal processor presents several challenges. It is desirable to safely stop a processor with minimum latency, so that a maximum amount of time can be allocated for processing tasks while the radio (or other RF circuitry) is not in use. Simply stopping a clock to a processor may leave the processor (as well as the processor bus, memories, or peripheral devices) in an invalid state, or a state which may not be capable of returning to normal operations. By carefully implementing a stopping procedure (e.g., the procedures described below), the buses are well suited for data movement. In addition, with an efficient stopping procedure, time spent before the radio use can be used more effectively, where tasks desired in preparation for radio

use can occur as close to the start of the radio use as possible. In one example, the state of the processors while the radio is used is maintained, but can also be transitioned quickly back to normal operation after use of the radio. An efficient processor restarting procedure will minimize the time that the processor will spend restarting, allowing more time for processing other tasks.

[37] When implementing stopping and starting procedures, there are several considerations that can be taken into consideration. Again, as described above, in order to maintain the best possible RF performance, noise associated with digital logic switching should be controlled during critical times (e.g., when a radio is transmitting or receiving). One method to control the digital switching noise is to stop all of the switching, or at least minimize it. Ideally, the stopping and starting techniques should be safe for the overall system, such that the states of the processor pipelines are not impacted by spurious clocks, for example. It is also desirable to stop the processor at a specific time, since use of the radio is dictated by a prescribed schedule. It is desirable to stop the processor as quickly as possible in order to maximize the amount of time that the processor is available to process tasks.

[38] While the techniques of the present invention for efficiently stopping and starting one or more processors can be implemented in a number of ways, following is a description of techniques for stopping and starting a processor that is controlled using an interrupt architecture. Generally, an interrupt is a signal received by a processor that causes a temporary halt in the execution of a program while some other

task is performed. After the task is performed, control returns to the original program. If multiple interrupts are received, the interrupts are serviced based on a priority system, where the highest priority interrupts are serviced first. It is common for a real time system (e.g., a mobile telephone, etc.) to utilize interrupts to control the operation of the system. Such a system may include a processor, an interrupt handler program, an interrupt controller, and a sequencer, or interrupt generator.

[39] FIG. 3 is a block diagram of an interrupt control system 40 that may be used with the present invention. The system 40 includes a microcontroller (MCU) 42, which may be part of the baseband 14 shown in FIG. 1. The MCU 42 shown in FIG. 3 includes memory 44, cache 46, and interrupt logic 48. The memory 44 and cache 46 are used for storing information for use by the MCU 42. The MCU 42 may also make use of memory or cache coupled externally. The interrupt logic 48 is coupled to an interrupt controller 50 by line 52. The interrupt controller 50 is also coupled to an MCU bus 54, which is also coupled to the MCU 42. A system timer 56 is also coupled to the MCU bus 54, as well as to the interrupt controller 50, via line 58. The system timer 56 acts as a sequencer, or interrupt generator for generating interrupts for the interrupt controller 50. FIG. 3 also shows two representative peripheral devices 60 and 62, which are coupled to the MCU bus 54 and to the interrupt controller 50 via lines 64 and 66, respectively. Any desired number of peripherals supported by the system may be used. Examples of peripheral devices include, but are not limited to, a UART, a SIM card interface, a DMA controller, external interrupt sources, etc.

[40] In the embodiment shown in FIG. 3, interrupts can be generated by the system timer 56, peripheral devices 60 and 62, as well as other devices that may not be shown in FIG. 3. When one or more interrupts are received by the interrupt controller 50, the interrupt controller 50 determines what processing path to take next, at least partly depending on the relative priorities of received interrupts. The interrupt controller 50 then sends control signals the interrupt logic 48 of the MCU 42 via line 52.

[41] Generally, the present invention efficiently stops and starts a processor by (1) providing a way to trigger a transition; and (2) providing procedures to efficiently stop and start the processor. In one example, a transition is triggered by generating and servicing an interrupt of relatively high priority (described in more detail below). One advantage of using this triggering technique, is that preexisting, legacy code can be maintained, while still achieving the desired objectives. Upon receiving the interrupt of relatively high priority, the processor is stopped or started such that the processor can operate in a normal, processing mode, and in a low noise, low power, sleep-like state mode (described in more detail below).

[42] In one example, the present invention configures a first interrupt of relatively high priority that will be generated when the processor is to be stopped (i.e., immediately before radio use). A second interrupt of relatively high priority is configured which will be generated when the processor is to be restarted (i.e., immediately after radio use). FIG. 4 is a diagram illustrating a typical legacy interrupt map. FIG. 4 lists N interrupts (labeled 1 through N), sorted by priority, where

interrupt 1 has the highest priority, and interrupt N has the lowest priority. The priorities of the interrupts are labeled 0 through N-I, where the lowest numbers correspond to the highest priorities.

[43] As mentioned above, the present invention can be implemented without disturbing preexisting legacy code (i.e., code corresponding to the N interrupts illustrated in FIG. 4). FIG. 5 is a diagram illustrating an interrupt map, including two new interrupts used to trigger the stopping and starting of the processor. Like FIG. 4, FIG. 5 lists the N legacy interrupts (labeled 1 through N). Two new interrupts (N+l and N+2) are also shown. The new interrupts (N+l and N+2) are re-mapped such that they have the highest priorities. In this example, interrupt N+l has a priority of -1, which is higher than all of the normal, legacy interrupts. Interrupt N+l can also be thought of as the "-1 interrupt". In one example, interrupt N+l triggers a shut down interrupt routine (described below) and is generated prior to radio use. Interrupt N+2 has a priority of -2, which is higher than all of the normal, legacy interrupts, as well as interrupt N+l. Interrupt N+2 can be thought of as the "-2 interrupt". In one example, interrupt N+2 triggers a restart interrupt routine (described below) and is generated after radio use. Note that the terms "-1 interrupt" and "-2 interrupt" are merely just convenient names given to the interrupts and that other nomenclature can be used, as desired. Generally, the two new interrupts are configured to have a higher priority than all of the legacy interrupts.

[44] As mentioned, the present invention may use any desired procedures for starting and stopping a processor. FIG. 6 is a flowchart illustrating the operation of the present invention. The process begins with step 6-10, where one or more interrupts are generated and received by the interrupt controller. Generally, as interrupts are received, the interrupt controller will make a decision as to the next processing step that the processor will take. If multiple interrupts are received, the interrupt with the highest priority will be serviced first. At step 6-12, the process determines whether a shut down interrupt (e.g., the -1 interrupt described above) was received. If not (i.e., a normal, legacy interrupt was received), the process proceeds to step 6-14, where the new interrupt is processed. In the case where a previously received higher priority interrupt has not yet been completed, this processing is continued. The process then goes back to step 6-10. Note that, between received interrupts, the processor is performing general processing tasks. If, at step 6-12, it was determined that a shut down interrupt (-1 interrupt) was received (i.e., usage of RF circuitry is about to start), then the process proceeds to step 6- 16, where the shut down interrupt service routine is processed.

[45] FIG. 7 (described below) illustrates one embodiment of a shut down interrupt service routine. As is described below with respect to FIG. 7, after step 6-16, the process waits for a restart interrupt (the -2 interrupt). In one example, the -2 interrupt is unmaskable, so the interrupt can not be masked within the interrupt controller. Once a restart interrupt is received, the process proceeds to step 6-20, where the restart interrupt service routine is processed.

[46] FIG. 8 (described below) illustrates one embodiment of a restart interrupt service routine. In an alternative example, rather than using the -2 interrupt, a timer can be used to trigger the restart process, since the length of the RF time-slot should be known. Similarly, a timer could be used to determine when a burst begins, or can be used to trigger the -1 interrupt. After the restart interrupt service routine has completed, the process goes back to step 6-10. Note that the process illustrated in FIG. 6 shows just one example, and that many other examples are possible within the spirit and scope of the present invention.

[47] FIG. 7 illustrates an embodiment of the reception of a shut down interrupt and a shut down interrupt service routine. The main purpose of the shut down interrupt service routine is to quickly transition the processor to a known stable state. In this example, the process begins at step 7-10 where the shut down interrupt was received. At step 7-12, the ability to generate lower priority interrupts (e.g., the normal, legacy interrupts 1 - N shown in FIG. 4) is disabled. The purpose of this step is to prevent additional interrupts from being sent to the processor while the radio is being used, causing the processor restart unexpectedly.

[48] Next, at step 7-14, any unneeded buses may be idled. Similarly, the ability to access memory (e.g., RAM) is limited to the processor. In addition, if desired, any peripheral devices that are not designed to operate during use of the radio can be isolated from the processor. Isolating peripheral devices from the processor can be achieved in response to the shut down interrupt, or can be achieved separately from

the process outlined in FIG. 7. Examples of techniques for idling a bus and isolating peripheral devices are described in U.S. Patent Application Serial No. 10/814,426, filed on March 31, 2004, Attorney Docket No. 5797-00500, titled "COMMUNICATION APPARATUS IMPLEMENTING TIME DOMAIN ISOLATION WITH RESTRICTED BUS ACCESS", which is incorporated by reference herein.

[49] At step 7-16, the processor finishes any processing needed to prepare for radio use. Next, at step 7-18, the processor is instructed to wait for an interrupt. The wait for interrupt instruction is a general instruction that causes the processor to essentially stall until such time as an interrupt is received by the processor. Since no lower priority interrupt can be generated, the normal operations of the processor are stopped. In another example, the processor can be stalled by instructing the processor to wait for a bit to be set. The combination of these tasks results in a quiet processor environment.

[50] Since there are no further operations on the processor, the processor clock can be disabled (step 7-20) without the need to precisely coordinate the clock stop time between the processor, the processor's bus, memories, and peripherals. This essentially decouples the clock stopping and starting from the processor stopping and starting. There are numerous ways of disabling a clock signal. In one example, a clock signal can be disabled by gating a clock source. In another example, a clock signal can be disabled by removing the assertion of a clock enable from state holding

elements. Note that the steps outlined in FIG. 7 may be performed in a different order, as desired. Also, not all of the steps are required, depending on the system, noise requirements, etc. Similarly, other steps could also be included in the process.

[51] In another example, a processor can be disabled by simply stopping the processor clock, and hoping that all is well when the clock signal returns. This method would seem to be less reliable than the method illustrated in FIG. 7. In another example it is possible that the code is in the correct function to stop the processor at precisely the right time to stop the processor in the same manner described, through the use of the wait for interrupt instruction. In this example, the method relies on the tasks executed on the processor to be synchronized to the network timebase. While this method would work, a new software protocol may be desired for implementation. If the use of a simple clock stop is attempted, it would be possible to have synchronization of all clock domains to ensure that no domain received additional clocks relative to the others.

[52] FIG. 8 illustrates an embodiment of the reception of a restart interrupt and a restart interrupt service routine. The main purpose of the restart interrupt service routine is to restart the processor as quickly as possible. Since the processor has been instructed to wait for an interrupt, and the general purpose interrupts are disabled, then the processor will wait until the restart interrupt is received. In this example, the process begins at step 8-10, where the restart interrupt is received. At step 8-12, the processor clock is enabled, and clocking is restored to the internal circuitry of the

processor. At step 8-14, any desired processing is performed prior to returning to normal operation. In addition, the processor may perform data movement operations that may be desired to satisfy peripheral data movement needs. Next, at step 8-16, any disabled buses may be enabled. Also, full memory access may be enabled. The peripheral devices are enabled at this time to use system resources. At step 8-18, the general interrupts are enabled. Finally, at step 8-20, the context is restored to the normal system task that was running immediately prior to the radio use interrupts. Note that the steps outlined in FIG. 8 may be performed in a different order, as desired. Also, not all of the steps are desired, depending on the system, noise requirements, etc. Similarly, other steps could also be included in the process.

[53] Referring more specifically to the example of a GSM application, FIG. 9 shows an example of a GSM data frame 70 and the relative timing of the shut down and restart interrupts. Frame 70 includes 8 slots, which may correspond to bursts of data. In one example, for GSM speech, slot 2 represents a receive slot, slot 4 is a monitor slot, and slot 6 constitutes a transmit slot. The remaining slots are unused for RF activity. Of course, the transmit, receive, and monitor slots could also be in different positions. In addition, it is possible to have multiple transmit or receive slots in the same frame. In the GSM frame example, and referring to the discussion of FIG. 2 above, the GSM slots 2, 4, and 6 are RF time-slots, with the remaining slots, in the context of this example, being signal -processing time-slots. For each of the RF time- slots, FIG. 9 also shows the relative timing of the -1 and -2 interrupts (a more detailed view of the relative timing of the -1 and -2 interrupts is shown in FIG. 10). As shown,

the -1 interrupt is generated slightly before the beginning of each RF time-slot, due to the time used to process the shut down interrupt service routine (see FIG. 7). Similarly, the -2 interrupt is generated after the end of each RF time-slot to restart the processor. The precise timing of the interrupts can vary, as desired.

[54] FIG. 10 is a timing diagram illustrating the system operation context for an RF time-slot. Note that the diagram in FIG. 10 is not to scale, and is used to merely to illustrate the relative timing of the various items shown. In FIG. 10, the top line represents the wait for interrupt state, which begins during the shut down interrupt service routine (step 7-18 of FIG. 7) and ends when the -2 restart interrupt is received (step 8-10 of FIG. 8). As shown in FIG. 10, prior to the -1 interrupt, the system processes normal, legacy tasks. At some time between the -1 interrupt, the -2 interrupt, the RF circuitry is in use, and the -1 handler puts the processor in a sleep- like state. The -2 interrupt initiates the -2 handler, which restarts the processor. Part of the function of the -2 handler is to end the -1 handler. Finally, the system again processes normal, legacy tasks.

[55] In examples where a DSP is time domain isolated from other parts of an apparatus (e.g., isolating a DSP from RF circuitry), the present invention may include additional features. For example, during times where it may be desirable to shut down a DSP, an apparatus may be receiving information for the DSP to process. In a mobile telephone apparatus, for example, the apparatus may receive RF signals while the RF circuitry (e.g., transceiver) is operating and the DSP is disabled. Sources that provide

data to the DSP, or receive data from the DSP, may continue to have data needs that may need to be processed while the DSP is shut down. Therefore, a system using the present invention should manage the data, where needed. Following is a detailed description of an embodiment of the present invention applied to a DSP.

[56] As described above, the present invention provides techniques for isolating signal processing circuitry from RF circuitry, which may reside on the same circuit partition (e.g., an IC, etc.). FIG. 11 is a block diagram of an interrupt control system 80. The system 80 includes a digital signal processor (DSP) 82. The DSP 82 is coupled to an interrupt controller 84 by lines 86. Each of the lines 86 is coupled to one of the interrupt inputs INTO, INTl, and INT2 of the DSP 82. The interrupt inputs described are merely examples, and other configurations and notations may also be used. A system timer 88 is coupled to the interrupt controller 84. The system timer 88 acts as a sequencer, or interrupt generator for generating interrupts for the interrupt controller 84. In one example, the system timer 88 is the same system timer that is used with other parts of the apparatus (e.g., the system timer 56 shown in FIG. 3). In another example, the system timer 88 is separate. The interrupt controller 84 may also be coupled to one or more peripheral devices 90, which may also generate interrupts for the interrupt controller 84.

[57] FIG. 11 also shows clock management logic 92 coupled to the DSP 82. The clock management logic 92 provides a clock signal to the DSP 82, which, when disabled, will disable the DSP 82, forcing the DSP 82 into a low noise mode.

Connections between the DSP 82, clock management logic 92, and the interrupt controller 84, allow the DSP 82 to configure each device. An RF input (RF IN) is coupled to the DSP 82 to provide a source of RF input signals (e.g., from RF front end circuitry, not shown).

[58] In the embodiment shown in FIG. 11 , interrupts can be generated by the system timer 88, as well as by other devices (e.g., peripheral devices 90). Interrupts from the system timer 88 are provided to the interrupt controller 84 via lines 100, 102, and 104. In one example, line 100 is used to send the shut down interrupt, line 102 is used to send the restart interrupt, and line 104 is used to send other interrupts (e.g., legacy interrupts) .

When one or more interrupts are received by the interrupt controller 84, interrupts are passed on to the DSP 82 via lines 86. In one example, the interrupt input INTO is used for the shut down and restart interrupts. The interrupt inputs INTl and INT2 are used for other, lower priority interrupts. The DSP 82 is configured such that interrupts received at input INTO have the highest priority, interrupts received at input INTl have the next highest priority, and interrupts received at input INT2 have the lowest priority. On advantage of utilizing multi-level interrupt scheme is that a less complex interrupt controller can be used. In another example, the DSP 82 uses a single interrupt input to receive all interrupts from the interrupt controller 84.

[59] As described above in more detail, the present invention can efficiently stop ■ and start a DSP by (1) providing a way to trigger a transition; and (2) providing

procedures to efficiently stop and start the DSP. In one example, a transition is triggered by generating and servicing an interrupt of relatively high priority. Upon receiving the interrupt of relatively high priority, the DSP is stopped or started such that the processor can operate in a normal, processing mode, and in a low noise, low power mode.

[60] In one example, the present invention configures a first interrupt (the shut down or "-1 " interrupt) of relatively high priority that will be generated when the DSP is to be stopped (i.e., immediately before radio use). A second interrupt (the restart or "-2" interrupt) of relatively high priority is configured that will be generated when the DSP is to be restarted (i.e., immediately after radio use). In one example, interrupts can be prioritized as follows: the shut down and restart interrupts have the highest priority; legacy system timer interrupts have the next highest priority; and other peripheral interrupts have the lowest priority.

[61] The -1 and -2 interrupts described above for use in disabling the MCU 42, and the -1 and -2 interrupts for use in disabling the DSP 82 are generated separately in one example, since the timing requirements (e.g., the desired preprocessing, etc.) may be different for the MCU 42 and the DSP 82. hi another example, the -1 and -2 interrupts can be the same, as desired.

[62] FIG. 12 is a flowchart illustrating one example of the operation of the present invention illustrated in FIG. 11. The process begins with step 12-10, where one or

-_O-

more interrupts are generated and received by the interrupt controller 84. Generally, as interrupts are received, the interrupt controller will make a decision as to the next processing step that the DSP will take. If multiple interrupts are received, the interrupt with the highest priority will be serviced first. Also, if a shut down or restart interrupt is received, it is sent to the DSP 82 via input INTO, in the example shown in FIG. 11. At step 12-12, the process determines whether a shut down interrupt (e.g., the -1 interrupt described above) was received from the system timer 88. If not (i.e., a normal, legacy interrupt was received), the process proceeds to step 12-14, where the new interrupt is processed. In the case where a previously received higher priority interrupt has not yet been completed, this processing is continued. Note that while processing interrupts at step 12-14, the interrupt processing can be preempted by the reception of a higher interrupt. The process then goes back to step 12-10. Note that, between received interrupts, the DSP is performing general processing tasks.

[63] If, at step 12-12, it was determined that a shut down interrupt (-1 interrupt) was received (i.e., usage of the RF circuitry is about to start), then the shut down interrupt service routine begins. At step 12-16, the ability to generate lower priority interrupts is disabled. The purpose of this step is to prevent additional interrupts from being sent to the DSP while the radio is being used, causing the DSP restart unexpectedly. Next, at step 12-18, the DSP 82 finishes any desired processing needed to prepare for radio use. The DSP 82 may perform preprocessing to the DSP and the DSP peripherals. For example, the DSP 82 may flush the audio buffer and set up the system for managing RF signals while the DSP 82 is disabled.

[64] Next, at step 12-20, the DSP 82 configures the restart interrupt as the source to re-activate the DSP clock. At step 12-22, the DSP clock is disabled via the clock management logic 92. Next, at step 12-24, the system waits for the restart interrupt (the -2 interrupt). During the time that the DSP 82 is disabled, any other interrupts that might be generated (e.g., from peripheral devices 90) are masked, to prevent the interrupts from restarting the DSP clock. The masking of interrupts is configured to let the restart interrupt through. During radio use, when the DSP 82 is disabled, the system will manage any continuing RF input signals. After radio use, the DSP will process the signals. As a result, data will not be lost as a result of the DSP 82 being disabled. Since the DSP 82 is disabled part of the time, the DSP 82 can be configured to process signals at a faster rate than it would need to if it were never disabled. In one example, the DSP 82 is configured to process signals at a data rate which is greater than the rate at which the data is generated.

[65] Once a restart interrupt is received, the restart interrupt service routine is processed. When the restart interrupt is generated by the system timer 88, the interrupt is received by the clock management logic 92, which will then enable the DSP clock (step 12-26). Finally, at step 12-28, the DSP performs post-processing to the DSP and DSP peripherals.

[66] From the preceding description it is apparent that the present invention can be implemented in many ways to provide various advantages. One novel advantage of the present invention is that the invention provides a processor able to minimize noise

created by digital circuitry. Another advantage of the present invention is the provision of the ability to quickly transition from high performance processing to low noise modes and back, with a minimum amount of overhead. In addition, the low noise modes cannot be accidentally exited, although it could be exited earlier than originally expected.

[67] In the preceding detailed description, the invention is described with reference to specific exemplary embodiments thereof. Various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the claims. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.