Patents by Inventor John D. Kenny
John D. Kenny has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 6393506Abstract: A processor system includes an on-chip, split-transaction bus with independent address/control and data buses. Arbitration and bus acquisition protocols are performed on the address/control bus. An arbiter arbitrates I/O requests and regulates concurrent ownership of the split-transaction bus by assigning a virtual channel to each bus request. Data bus access is granted to a virtual channel on a priority basis, s so as to utilize to a maximum extent the available bandwidth of the data bus. In one embodiment, the data bus is preempted by another virtual channel when current virtual channel using the data bus becomes idle due to, for example, latencies in the data stream. Rearbitration, however, is avoided when the interrupted data transfer resumes, owing to state information regarding the data transfer stored in the master and slave modules of each virtual channel.Type: GrantFiled: June 15, 1999Date of Patent: May 21, 2002Assignee: National Semiconductor CorporationInventor: John D. Kenny
-
Patent number: 6281708Abstract: The present invention provides a centralized amplifier-accelerator a tri-state bus. The centralized amplifier-accelerator utilizes the module drivers as pre-drivers to the amplifier-accelerator. The centralized amplifier-accelerator is located physically in the center of the chip. This central amplifier-accelerator consists of a highly sensitive input sense circuit which detects voltage transition at very near the N-channel threshold for rising transitions and at very near the P-channel threshold for falling transitions. Once the sense circuit threshold is met, the output driver is triggered to drive the bus.Type: GrantFiled: June 15, 1999Date of Patent: August 28, 2001Assignee: National Semiconductor CorporationInventor: John D. Kenny
-
Patent number: 5983355Abstract: A method and apparatus is disclosed for controlling the application of a clock stopping signal in a processor to limit power consumption. The system controller receives addresses, signals indicative of primary and secondary system activity, and at least one nap timeout signal. Addresses are compared with fixed software interrupt addresses. Matching non-prefetched addresses trigger a nap mode. Upon nap mode triggering, the clock stopping signal may be throttled until a programmable NAP timer expires. Applying the clock stopping signal with programmable duty cycle during the throttling period ensures that processing necessary for the detection and servicing of primary and secondary activity may occur. A prefetch detect circuit ensures that fixed software interrupt addresses loaded in the middle of a prefetch do not trigger the clock stopping signal. The clock stopping signal is removed or inhibited when primary or secondary activity is detected or when a nap mode is terminated by a nap timer timing out.Type: GrantFiled: May 20, 1996Date of Patent: November 9, 1999Assignee: National Semiconductor CorporationInventors: John D. Kenny, Emilia Lei, Vimi Pandey
-
Patent number: 5954819Abstract: A method and apparatus is disclosed for controlling the application of a clock stopping signal in a processor to limit power consumption. The system controller receives addresses, signals indicative of primary and secondary system activity, and at least one nap timeout signal. Addresses are compared with programmed addresses. Matching addresses trigger a nap mode. Upon nap mode triggering, the clock stopping signal may be applied during a throttling period. Applying the clock stopping signal with programmable duty cycle during the throttling period ensures that processing necessary for the detection and servicing of primary and secondary activity can occur. A prefetch detect circuit ensures that programmed addresses loaded in the middle of a prefetch do not trigger the clock stopping signal. The clock stopping signal is removed or inhibited when primary or secondary activity is detected or when nap mode is terminated by timeout.Type: GrantFiled: May 17, 1996Date of Patent: September 21, 1999Assignee: National Semiconductor CorporationInventors: John D. Kenny, Kenneth Ma, Vimi Pandey
-
Patent number: 5909560Abstract: An apparatus and method for identifying devices on a passive type bus such as an ISA bus where peripheral devices do not identify themselves to the host CPU. The apparatus and method of the present invention has particular application to systems where more than one passive (e.g., ISA type) bus may be implemented and a host CPU has no indication as to which bus a device is coupled to. The data lines of the bus are tied through a pull-up circuit to a logical high level voltage (V.sub.CC) such that, for example in a sixteen-bit data bus, the output data is FF hexadecimal. When a read request is generated on the bus, the bus controller detects whether the data on the bus changes from FF hexadecimal. If a change is detected, then the address device is present on that bus. If the addressed device outputs a data value of FF hexadecimal, that data value is passed through as valid data.Type: GrantFiled: June 6, 1995Date of Patent: June 1, 1999Assignee: National Semiconductor CorporationInventors: John D. Kenny, Steve Wenlung Chang, Emilia Vai-Lun Lei
-
Patent number: 5892930Abstract: An apparatus and method for identifying devices on a passive type bus such as an ISA bus where peripheral devices do not identify themselves to the host CPU. The apparatus and method of the present invention has particular application to systems where more than one passive (e.g., ISA type) bus may be implemented and a host CPU has no indication as to which bus a device is coupled to. The data lines of the bus are tied through a pull-up circuit to a logical high level voltage (V.sub.CC) such that, for example in a sixteen-bit data bus, the output data is FF hexadecimal. When a read request is generated on the bus, the bus controller detects whether the data on the bus changes from FF hexadecimal. If a change is detected, then the address device is present on that bus. If the addressed device outputs a data value of FF hexadecimal, that data value is passed through as valid data.Type: GrantFiled: October 21, 1996Date of Patent: April 6, 1999Assignee: National Semiconductor CorporationInventors: John D. Kenny, Steve Wenlung Chang, Emilia Vai-Lun Lei
-
Patent number: 5734850Abstract: A bridge and a method for interfacing a plurality of buses with the bridge provides electrical isolation between the buses but is transparent so that the plurality of buses is viewed by software as a single logical bridge. Transaction cycles initiated on one bus are reflected on the other bus. A speculative start of a transaction cycle on a secondary bus immediately after the transaction cycle has been started on the first bus provides a significant savings in time to complete transactions in which the target of the transaction is on the secondary bus.Type: GrantFiled: July 5, 1995Date of Patent: March 31, 1998Assignee: National Semiconductor CorporationInventors: John D. Kenny, Pranay D. Shah
-
Patent number: 5673400Abstract: An apparatus and method for controlling devices in a multiple bus system such as a system having two or more ISA type buses. Separate ISA bus controllers may be provided for each ISA bus, linked on a common bus system such as a PCI bus. One ISA controller may be designated as a primary ISA controller whereas other ISA controllers in the system may be designated as secondary ISA controllers. Each ISA controller in the system is provided with IRQ (interrupt request) enable bits to enable different interrupts for the corresponding ISA bus. Each secondary ISA controller outputs a signal IRQSER as a PCI sideband signal to the primary ISA controller to indicate which IRQs have been asserted on the respective ISA buses. The primary controller receives the IRQSER signal as well as the IRQ signals asserted on its own bus and converts these interrupt requests to PCI bus cycles.Type: GrantFiled: June 7, 1995Date of Patent: September 30, 1997Assignee: National Semiconductor CorporationInventor: John D. Kenny
-
Patent number: 5546037Abstract: A NAPNOP circuit for decreasing energy consumption of all or a portion of a microprocessor based system which includes a delay circuit for inhibiting or slowing the output of the system clock pulses for a variable length of time equal to a multiple of N clock pulses where N is a positive integer. The NAPNOP circuit has an input element for inputting a STARTNAP signal which begins a nap period during which the system clock pulses are inhibited or slowed, a clock input device for providing a plurality of selectable clock pulses as inputs to the delay circuit for controlling the operation of the computer system, and a gate element for terminating the nap period.Type: GrantFiled: November 15, 1993Date of Patent: August 13, 1996Assignee: Cirrus Logic, Inc.Inventors: John D. Kenny, Min S. Ma
-
Patent number: 5414863Abstract: A power control circuit for a device such as a personal computer, including a laptop or notebook computer, which can conserve battery use, prevent power surges to promote longer battery charges and longer battery life, and can assure that circuitry is correctly biased. The power control circuitry of the present invention achieves these objectives by appropriately staggering the powering on of circuit components of the computer. A circuit for achieving these objectives may feature at least one memory for storing power control state data and a multiplexer for receiving the power control state data stored in the at least one memory. Further, a plurality of serially connected power control output circuits connected to the multiplexer output power control signals based on the power control state data stored in the at least one memory.Type: GrantFiled: April 2, 1993Date of Patent: May 9, 1995Assignee: Cirrus Logic, Inc.Inventors: Robert H. J. Lee, John D. Kenny
-
Patent number: 5313108Abstract: The time a microprocessor CPU must wait for memory access is controlled to be one of two values by stretching the CPU clock signal either a first time duration or a second time duration, depending on the expected delay caused by the memory access. The clock stretching is in increments of one quarter of the CPU clock period and is done with both the leading and trailing edges of the clock pulse.Type: GrantFiled: April 17, 1992Date of Patent: May 17, 1994Assignee: Picopower Technology, Inc.Inventors: Robert H. J. Lee, John D. Kenny
-
Patent number: 5287292Abstract: The temperature of a circuit is monitored and controlled by accumulating an estimate of heat generated in the circuit, and decreasing heat generation in the circuit when necessary. A periodic sampling of the operating mode of the circuit, as determined by clock speed and bus cycle activity, is used to determine heat accumulation in the circuit. An up/down counter increments when the sampling shows an operating mode indicating heating of the circuit and decrements when the sampled mode indicates cooling of the circuit. The circuit is forced to cool if a count on the up/down counter reaches a programmable threshold. Cooling is accomplished by slowing the clock speed of the circuit.Type: GrantFiled: October 16, 1992Date of Patent: February 15, 1994Assignee: Picopower Technology, Inc.Inventors: John D. Kenny, Emilia V. Lei
-
Patent number: 5254888Abstract: Power dissipation of a CMOS circuit such as a microprocessor is reduced by dynamically slowing down the microprocessor clock during selected system operations such as hold, wait, or AT peripheral bus access cycles. The microprocessor clock is slowed to its minimum allowable frequency with precise synchronous control to maintain the accuracy of high frequency clock edges and to prevent glitches or substandard pulse widths.Type: GrantFiled: March 27, 1992Date of Patent: October 19, 1993Assignee: Picopower Technology Inc.Inventors: Robert H. J. Lee, John D. Kenny