Electrical computers and digital data processing systems: input/ – Intrasystem connection – Bus interface architecture
Reexamination Certificate
2001-06-01
2004-08-31
Auve, Glenn A. (Department: 2181)
Electrical computers and digital data processing systems: input/
Intrasystem connection
Bus interface architecture
C370S392000, C710S311000
Reexamination Certificate
active
06785758
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention is generally directed to system-on-a-chip (SOC) devices and other large integrated circuits (ICs) and, in particular, to a high-throughput bus architecture for use in a SOC device or other large integrated circuit (IC).
BACKGROUND OF THE INVENTION
In recent years, there have been great advancements in the speed, power, and complexity of integrated circuits, such as application specific integrated circuit (ASIC) chips, random access memory (RAM) chips, microprocessor (uP) chips, and the like. These advancements have made possible the development of system-on-a-chip (SOC) devices. A SOC device integrates into a single chip all (or nearly all) of the components of a complex electronic system, such as a wireless receiver (i.e., cell phone, a television receiver, and the like). SOC devices greatly reduce the size, cost, and power consumption of the system.
However, SOC designs are pushing the limits of existing interconnect topologies and diagnostic capabilities. Many SOC devices, including microprocessors, use a variety of shared tri-state buses (e.g., XBus, fast XBus, PCI, and fast PCI). Currently there are no standard bus topologies and no easy way to mix and match designs for quick integration. In addition, with no consistent bus model, there are no consistent debugging, power management, or validation standards. The existing bus topologies are not scalable and do not support the demanding needs for higher bandwidth, isochronous data, and scalable peripherals.
These problems stem, in part from the lack of a standard interconnect for high-performance devices, such as the central processing unit (CPU) or processor core, 2D/3D graphics blocks, MPEG decoding blocks, 1394 bus, and the like. As device requirements exceed existing bus capabilities, either new derivative buses are created or non-Universal Memory Architecture (non-UMA) solutions are used. These ad-hoc non-standard interfaces preclude the reuse of technology improvements between products.
Another weakness in current bus topologies is the lack of a generalized UMA interface. Allowing multiple devices to use the same unified memory reduces system cost. However, the UMA devices must not adversely effect the processor access latency. Another limitation in many data processing devices is the chip-to-chip peripheral connection interface (PCI) bus. Using a chip-to-chip PCI bus limits bandwidth and the possibility of implementing chip-is-to-chip UMA devices.
Existing bus architectures do not support technology reuse as memory bandwidth increases with new memory speeds and technologies (e.g., SDRAM-166). A new bus standard must support bandwidth matching between older, lower bandwidth devices and newer, higher bandwidth devices. In addition to bandwidth matching, clock matching must be addressed when mixing bus architectures.
New input/output (I/O) standards, such as 1394 and USB, create real-time isochronous data streams which need guaranteed bandwidths and latencies. Most bus topologies do not adequately support these isochronous requirements. Mixing isochronous data, low latency access, and high-bandwidth UMA peripherals requires a new full-featured bus topology.
Peer-to-peer communication is optimal for data streams such as VIP, 1394 and MPEG transport layer. Using peer-to-peer, memory and CPU interactions can be avoided. In addition, data traffic between the CPU and a graphics rendering block requires high bandwidth peer-to-peer communication. A new interconnect bus topology must provide common test strategies, power management, diagnostic and clocking interfaces to address design reuse. Also, a new bus topology must address reuse of legacy bus technologies. It is unreasonable to expect device manufacturers to re-code existing devices to conform to a new standard. Existing PCI and XBus blocks must be able to fit in the new topology without modification.
Still another weakness in current bus topologies is the manner in which machine specific register (MSR) requests are handled by bus interface (IF) units in a bus architecture. In order to transfer a MSR Read request or MSR Write request from a master bus device to a slave bus device, each bus IF unit between the master and slave bus devices requires complex address translation circuitry to correctly route the MSR request. Typically, address maps are used to identify the location of the slave bus device. This requires that each bus IF unit be configured by an external device, such as the CPU core or the master bus device, according to the bus topology in which the bus IF unit is established.
Therefore, there is a need in the art for an improved bus architecture for system-on-a-chip (SOC) devices and other large scale integrated circuits. In particular, there is a need for a bus architecture that implements a simplified MSR processing capability. More particularly, there is a need for a bus architecture in which bus interface units do not need to be configured by external devices in order to handle MSR requests.
SUMMARY OF THE INVENTION
To address the above-discussed deficiencies of the prior art, it is a primary object of the present invention to provide a bus interface unit for transferring machine specific register (MSR) requests between a plurality of bus devices. According to an advantageous embodiment of the present invention, the bus interface unit comprises: 1) a plurality of input ports capable of receiving incoming MSR requests from the plurality of bus devices; 2) a plurality of output ports capable of transmitting data to the plurality of bus devices; and 3) a controller capable of reading N routing associated with a first received MSR request and comparing a first identification (ID) value in a predetermined M-bit field in the N routing bits to a first designated value, wherein the controller, in response to a determination that the first ID value does not equal the first designated value: 1) realigns the N routing bits such that remaining ones of the N-M bits outside the predetermined M-bit field are moved into the predetermined M-bit field and 2) transmits the realigned N routing bits via a first one of the plurality of output ports identified by the first ID value.
According to one embodiment of the present invention, the predetermined M-bit field comprises the M most significant bits of the N routing bits.
According to another embodiment of the present invention, the controller realigns the N routing bits by left-shifting M places the remaining N-M bits.
According to still another embodiment of the present invention, the predetermined M-bit field comprises the M least significant bits of the N routing bits.
According to yet another embodiment of the present invention, the controller realigns the N routing bits by right-shifting M places the remaining N-M bits.
According to a further embodiment of the present invention, the controller, in response to a determination that the first ID value equals the first designated value, transfers the MSR request to an internal machine specific register disposed in the bus interface unit.
According to a still further embodiment of the present invention, the internal machine specific register is identified by an MSR address associated with the N routing bits.
The foregoing has outlined rather broadly the features and technical advantages of the present invention so that those skilled in the art may better understand the detailed description of the invention that follows. Additional features and advantages of the invention will be described hereinafter that form the subject of the claims of the invention. Those skilled in the art should appreciate that they may readily use the conception and the specific embodiment disclosed as a basis for modifying or designing other structures for carrying out the same purposes of the present invention. Those skilled in the art should also realize that such equivalent constructions do not depart from the spirit and scope of the invention in its broadest form.
Before undertaking the DETAILED DESCRIPTION OF THE INVENTION below, it may be a
Kotlowski Kenneth James
Tischler Brett A.
Advanced Micro Devices , Inc.
Auve Glenn A.
LandOfFree
System and method for machine specific register addressing... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for machine specific register addressing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for machine specific register addressing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3334995