KR20100038938A - Method for hybrid arbitrating a bus - Google Patents

Abstract

PURPOSE: A hybrid bus arbitration method is provided to reduce data transmission time and to improve data transmission efficiency. CONSTITUTION: A fixed priority method is applied between master groups which are grouped into at least one master which execute a same function. A bus arbitration method, through which bus using authority is able to be uniformly distributed, is applied between masters in each group. The bus arbitration method is one among round robin, TDMA, lottery bus methods. The priority of bus utilization is provided from a first mater group(130) to third master group(142) which are included in the master groups.

Description

Method for hybrid arbitrating a bus}

The present invention relates to a bus system, and more particularly, to a hybrid bus arbitration policy that can improve data transmission efficiency.

With the development of the semiconductor manufacturing process, SoC (System on Chip) technology, which integrates various components performing different functions into one chip, has been developed and widely used in circuit design. Such a system on chip (SoC) can reduce chip size and design cost in system design and implementation, and can provide improved performance in terms of low power consumption, real-time processing power, and system flexibility.

The components of the SoC consist of a plurality of masters and arbiters, a plurality of slaves, and a shared bus for data transmission between the master and the slaves. In general, a master refers to a processor (CPU), direct memory access (DMA), and a digital signal processor (DSP), and a slave refers to memory and input / output means such as SDRAM and SRAM.

In the case of an SoC chip in which several functions are integrated in one chip, a bus system in which a master, an arbiter, and a slave are connected to one bus is required. When a plurality of masters want to use a single bus at the same time, a plurality of masters receiving a bus grant signal by the arbiter sequentially transmit data to a slave (memory) or load data from the slave using the bus.

When multiple masters want to use a single bus at the same time, the masters receiving the bus permission signal by the arbiter sequentially transfer data to the slaves (memory) using the buses. To this end, the arbiter determines priorities when multiple masters request the bus to be used simultaneously. The performance of the chip varies according to the arbiter's prioritization method.

1 illustrates a bus general bus system for memory access. In FIG. 1, an advanced micro-controller bus architecture (AMBA) system is used as an example among several conventional bus systems, and only two of the plurality of masters and slaves constituting the bus system are illustrated.

Referring to FIG. 1, a general bus system 1 includes an arbiter that arbitrates the use of one bus 10, a plurality of masters 30, 32, a plurality of slaves 40, 42, and a bus 10. 20).

When the plurality of masters 30 and 32 want to connect to the plurality of slaves 40 and 42, that is, the masters 30 and 32 transmit data to the slaves 40 and 42, or the slaves 40, If the data is to be loaded from 42, the arbiter 20 is requested to use the bus 10. At this time, the arbiter 20 determines the priority of using the bus 10 of each of the masters 30 and 32 according to a predetermined priority determination method, and grants the approval of the use of the bus 10 from the arbiter 20. The obtained master is connected to the slaves 40 and 42 via the bus 10.

For example, when the first master 30 requests a connection to the first slave 40 and at the same time the second master 32 requests a connection to the second slave 42, the arbiter 20 may advance in advance. Priority of bus 10 usage is determined using a predetermined priority determination method, and in general, priority for bus 10 usage is given to a master that performs many operations and requires speed, such as a CPU and an ARM processor. Here, as the priority determination method applied to the arbiter 20, the fixed priority determination method and the round robin priority determination method are most widely used.

2A and 2B are diagrams illustrating a prioritization method of the prior art.

Referring to FIG. 2A, the fixed priority determination method refers to a method of pre-determining a plurality of masters and granting bus usage rights according to the determined rank. Priority is given to each master in order of processor (ARM, CPU), DSP, DMA1, and DMA2 in order, and when multiple masters want to access the bus at the same time, The bus ticket will be given to the higher ranking master.

In addition, referring to FIG. 2B, the round robin determination method is a method of giving priority to bus use so that the bus priority of the master is evenly distributed over time. In other words, it allocates a certain amount of time to the master, and when the allocated time passes, the master temporarily suspends and gives the bus priority to another master, and after that time, gives the bus priority to the next master. .

However, the fixed priority determination method does not consider the characteristics of the data to be processed by the master, and simply determines the priority according to the type of the master, which is not efficient in terms of data processing time.

In addition, in the case of the round robin priority determination method, the priority is determined by referring to time information allocated to the master without a fixed priority, which causes a problem of delaying processing of specific important data to be processed quickly.

In order to compensate for the above-mentioned disadvantages of the fixed priority determination method and the round robin priority determination method, a TDM arbitration method and a rotary bus arbitration method have been proposed.

The TDM arbitration method has the advantage of preventing starvation because the masters are evenly given priority to the use of buses, but there is a problem in that the master needs to process important data and has a long waiting time.

The rotary bus method has a problem of prolonging the waiting time in the case of a master that needs to process important data in a manner that a number of masters are prioritized to use the bus probabilistically.

Due to the problems described above, there is a demand for a bus arbitration method that can efficiently mediate bus usage among a plurality of masters to improve data transmission efficiency.

SUMMARY OF THE INVENTION The present invention has been made in an effort to provide a hybrid bus arbitration method capable of improving data transmission efficiency by improving a bus arbitration method in a multiple bus system.

Another object of the present invention is to provide a hybrid bus arbitration method capable of improving data transmission efficiency by reducing data transmission time in a multi-bus system including a plurality of masters and slaves.

The present invention provides a hybrid bus arbitration method that can improve data transmission efficiency by grouping masters performing the same function and applying a plurality of bus arbitration methods to bus group priority determination of grouped masters. That is another technical problem.

Hybrid bus arbitration method according to an embodiment of the present invention for achieving the above object in the hybrid bus arbitration method for mediating the bus use of a plurality of masters through a plurality of bus arbitration method, at least one performing the same function The fixed priority method is applied between the master groups grouped into masters, and the bus arbitration method can be equally distributed among the masters in each group.

In the hybrid bus arbitration method according to an embodiment of the present invention, the bus arbitration method in which the bus usage rights are equally distributed may be one of a round robin, a TDMA, and a rotary bus method.

In the hybrid bus arbitration method according to an embodiment of the present invention, the master groups include a first master group including a processor (ARM) master, a second master group including an Ethernet master, and a third including a DMA master And a master group, wherein the master groups are given priority to use a bus in order from the first master group to the third master group.

The master group of the hybrid bus arbitration method according to an embodiment of the present invention is characterized by consisting of masters having the same or similar bus request frequency.

When the bus use request frequency of the masters constituting the master group of the hybrid bus arbitration method according to an embodiment of the present invention is changed, reconfiguring the masters constituting the master group according to the bus use request frequency do.

As described above, the hybrid bus arbitration method according to an embodiment of the present invention can improve data transmission efficiency by reducing data transmission time in a multi-bus system including a plurality of masters and slaves.

As described above, the hybrid bus arbitration method according to an embodiment of the present invention groups masters according to characteristics of a plurality of masters, applies a fixed priority method among master groups, and round robin method among masters in the same group. By applying this, the data transmission efficiency can be improved.

The present invention relates to a bus system, and more particularly, to a hybrid bus arbitration method that can improve data transmission efficiency by applying different arbitration methods according to characteristics of masters.

Hereinafter, exemplary embodiments of the present invention will be described with reference to the accompanying drawings.

4 is a diagram illustrating a multiple bus system, and FIG. 5 is a diagram illustrating a hybrid bus arbitration method according to an exemplary embodiment of the present invention.

Hybrid bus arbitration method according to an embodiment of the present invention is applicable to various bus systems, hereinafter, the description will be described by using the multiple bus system shown in FIG. 4 as an example.

Referring to FIG. 4, the multi-bus system 100 includes a plurality of buses 110 and 112, a plurality of masters 130, 140, 142, 150, 152 and 154 and a plurality of slaves 160, 162, 164 and 166. ), The arbiter 120, 122, and the bridge 170.

The plurality of masters 130, 140, 142, 150, 152, and 154 refer to processors such as ARM and CUP, Ethernet, DMA, and DSP, and the plurality of slaves 160, 162, 164, and 166 to SDRAM. It refers to memory such as SRAM, and input / output means such as universal serial bus (USB) and universal asynchronous receiver transmitter (UART).

The plurality of masters 130, 140, 142, 150, 152, and 154 are connected to the slaves 160, 162, 164, and 166 through the buses 110 and 112 to which they are connected, and the plurality of masters simultaneously slave. If the user wants to access (160, 162, 164, 166), the master, which has obtained the permission signal for using the bus (110, 112) from the arbiter (120, 122), uses the bus in order to use the slave (160, 162, Data is transmitted to 164 and 166 or data is loaded from the slaves 160, 162, 164 and 166.

Each of the plurality of arbitrators 120 and 122 includes a determination unit that determines the priority of using the buses 110 and 112 to which the plurality of arbitrators 120 and 122 are connected. The plurality of masters 130, 140, 142, 150, 152 and 154 When the slaves 160, 162, 164, and 166 are to be connected to each other, as shown in FIG. 5, the priority of using the buses 110 and 112 is determined according to a predetermined priority determination method.

According to an embodiment of the present invention, each of the arbitrators includes a plurality of determination units for determining the bus use priority of the masters. The hybrid bus arbitration method of the present invention includes a fixed priority determination method and a round robin priority. A round robin priority method, a TDMA, and a Lottery Bus method can be used.

The buses 110 and 112 allow each of a plurality of masters to be connected as slaves, that is, to connect a plurality of masters and slaves to perform a read / write function. At this time, the master that has obtained permission to use the buses 110 and 112 from the arbiters 120 and 122 may perform data transmission.

As an example, the bus illustrated in FIG. 4 applies an AHB (Advanced High-performance Bus) capable of high-speed data transmission. In another embodiment of the present invention, an ASB (Advanced System Bus) and an APB (Advanced Peripheral Bus) are used in addition to the AHB. Can be applied.

As shown in FIG. 4, when the bus system is configured, the master 1 130, the master 2 140, and the master 3 142 operate the first bus 110 by arbitration of the first arbiter 120. The master 4 150, the master 5 152, and the master 6 154 use the second bus 112 by arbitration of the second arbiter 122.

The bridge 170 provides a connection interface for transmitting and receiving data between the first bus 110 and the second bus 112, and each of the plurality of masters is connected to a bus other than the bus to which the master is connected. If the slave is to be connected, the slave 170 may be connected to a slave connected to another bus via the bridge 170.

Next, a hybrid bus arbitration method according to an embodiment of the present invention will be described with reference to FIG. 5.

In the hybrid bus arbitration method of the present invention, when a plurality of masters want to connect as slaves, the bus usage of the plurality of masters can be mediated through a plurality of bus arbitration methods, where the plurality of bus arbitration methods have a fixed priority ( Fixed Priority determination method, Round Robin Priority determination method, TDMA, Lottery Bus method can all be used. FIG. 5 illustrates a method of arbitrating a bus of a master by applying a fixed priority determination method and a round robin priority method.

The hybrid bus arbitration method according to an embodiment of the present invention creates a plurality of master groups by grouping at least one master that performs the same function, and applies a fixed priority method among the grouped master groups. Then, the bus usage is arbitrated by applying a bus arbitration method in which bus usage rights can be evenly distributed among the masters in each group. Here, as a bus arbitration method in which the bus usage rights can be equally distributed, one of round robin, TDMA, and rotary bus methods may be used.

Specifically, the master groups are groups in which masters performing the same function as described above are formed, forming processors such as CPU and ARM into one master group, and Ethernet masters into another master group. And DMA masters into another group.

The determination of the priority between the master groups uses a fixed priority method, and a first priority group having the highest priority is given to the first master group including the processor masters 130 such as ARM and CPU, and Ethernet ( Ethernet) A second master group including the masters 140 and 142 is given a second rank after the first rank given to the first master group, and includes the DMA masters 150, 152 and 154. A third rank following the second rank is given to the third master group.

In addition, a bus arbitration method in which bus usage rights are equally distributed among the masters in the first to third master groups is applied. In FIG. 5, the round robin method is applied.

As described above, in mediating the bus use of a plurality of masters, the masters performing the same function are grouped, and the buses are mediated by applying a fixed priority method based on the frequency of access to the slaves between the groups. In addition, by arranging the bus using the round robin method among the masters in the master group performing the same function, data transmission efficiency of the entire bus system can be improved.

In the above description, the masters constituting the master group have been described as being masters performing the same function, but this has been described as an example. In the hybrid bus arbitration method according to an embodiment of the present invention, the master group has a bus usage request frequency. May consist of the same or similar masters.

In addition, when the bus use request frequency of the masters constituting the master group is changed, the masters constituting the master group may be reconfigured according to the bus use request frequency.

6 is a diagram illustrating data transmission efficiency of the present invention on the basis of a transition cycle.

When the hybrid bus arbitration method according to an embodiment of the present invention is applied, it can be confirmed through simulation that data transmission efficiency of the bus system is improved, which is illustrated in FIG. 6.

The simulation result shown in FIG. 6 is a result obtained by assuming that the multi-bus system (IEEE 802.11 network) shown in FIG. 4 is applied and six masters request a connection (using a bus) to a slave at a predetermined time period. to be.

Here, the bus usage request frequency of each of the six masters is set as shown in Table 1 as an example.

Referring to Table 1, the bus utilization request frequency of the master 1 is the highest, the fraudulent master 1, the bus usage request frequency of the master 2 and master 3 is the highest, and the bus usage request frequency of the master 4, master 5, master 6 It can be seen that is the lowest.

As shown in Table 1, when the bus usage request frequency of the masters is set, the hybrid bus arbitration method according to the embodiment of the present invention groups six masters by the master performing the same function as shown in Table 2 below, A plurality of priority determination methods are applied between groups.

Referring to Table 2, in the hybrid bus arbitration method according to an embodiment of the present invention, the master 1 having the highest bus utilization request is set as the first master group, and then the second master and the third master are set as the second master. The master group is set, and master 4, master 5 and master 6 are set as the third master group. The priority of bus usage is determined by applying a fixed priority method among the master groups, and the priority of bus usage is determined by applying a round robin method between masters of each master group.

First, referring to FIG. 6, when the bus arbitration method according to an exemplary embodiment of the present invention is defined, data transmission efficiency is based on a transition section cycle. When only the conventional round robin method is applied, the data transmission amount of six masters is increased. Since the same, the bus system environment as shown in Table 1 is not satisfied. In addition, TDMA, Lottery, and Fixed Priority (FP) methods reflect the master environment of Table 1, but the amount of data transfer between masters performing the same function, that is, masters having the same or similar bus utilization request frequency. Since there is a difference, the bus system environment of Table 1 may not be fully reflected.

On the other hand, in the hybrid bus arbitration method according to the embodiment of the present invention, it can be confirmed that data transmission amounts are increased in order of masters having a high frequency of using a bus, and masters performing the same function, that is, frequency of using a bus using the same or It can be seen that the data transfer rates of similar masters are the same or similar. Hybrid bus arbitration method according to an embodiment of the present invention can be seen that can fully reflect the bus system environment of Table 1.

As described above, the hybrid bus arbitration method according to an embodiment of the present invention can improve data transmission efficiency by reducing data transmission time in a multi-bus system including a plurality of masters and slaves.

In addition, the hybrid bus arbitration method according to an embodiment of the present invention groups the masters according to the characteristics of a plurality of masters, applies a fixed priority method among master groups, and evenly distributes bus usage rights among masters in the same group. A bus arbitration method (eg, round robin method) may be applied to improve data transmission efficiency.

Configuration and operation as described above is not limited to the claims of the present invention as an embodiment, and various changes and modifications are possible within the scope of not changing the technical spirit of the present invention in the field to which the present invention belongs. It is obvious to those who are engaged.

It is apparent to those skilled in the art that the present invention can be embodied in other specific forms without departing from the spirit and essential features of the present invention. Accordingly, the above detailed description should not be interpreted as limiting in all aspects and should be considered as illustrative. The scope of the invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the invention are included in the scope of the invention.

1 illustrates a bus general bus system for memory access.

2A and 2B are diagrams illustrating a prioritization method of the prior art.

3A illustrates a prior art TDM arbitration method.

FIG. 3B illustrates a prior art rotary bus arbitration method. FIG.

4 illustrates a multiple bus system.

5 illustrates a hybrid bus arbitration method according to an embodiment of the present invention.

6 is a diagram showing the data transmission efficiency of the present invention on the basis of a transition cycle.

<Explanation of symbols for the main parts of the drawings>

100: multiple bus system 110, 112: bus

120, 122: Arbitrators 130, 140, 150: Master

160, 162: Slave 170: Bridge

Claims (5)

In the hybrid bus arbitration method for mediating the bus use of a plurality of masters through a plurality of bus arbitration method, A fixed priority method is applied between master groups grouped into at least one master that performs the same function, and a bus arbitration method can be applied in which bus usage rights can be evenly distributed among masters in each group. Hybrid bus arbitration method. The method of claim 1, The bus arbitration method in which the bus usage rights can be evenly distributed is one of a round robin, a TDMA, and a rotary bus method. The method of claim 1, The master groups include a first master group including a processor (ARM) master, a second master group including an Ethernet master, a third master group including a DMA master, And wherein the master groups are given priority to use a bus in order from the first master group to the third master group. The method of claim 1, And wherein the master groups are made up of masters having the same or similar bus request frequency. The method of claim 1, And reconfiguring the masters constituting the master group according to the bus use request frequency when the bus use request frequencies of the masters constituting the master group change.

Images