Marvell International Limited

Texas Instruments Corporation

Solarflare Communications, Inc.

Marvell Asia Pte., Ltd.

Brett Anthony McClellan

Kuoruey Han

William Lo

Xing Wu

Divya Srinivasan Breed

Leo Fu

Michael Leung

Taehyun Jeon

Seid Alireza Razavi Majomard

George Zimmerman

Zubir Adal

Ragnar Hlynur Jonsson

Ming-Tak Leung

Brian Edem

Leon Yang

A network interface device operates in a normal transmit operating mode in which the network interface device continually receives transmission symbols from a link partner via the communication link. The network interface device determines that receive circuitry of the network interface device is to transition to a low power mode in response to receiving a sleep signal from the link partner. The network interface device then operates according to a quiet/refresh cycle of the low power mode to conserve power. The quiet/refresh cycle corresponds to a time schedule that includes a refresh time window in which receive circuitry of the network interface device is to be powered to receive a refresh signal from the link partner Immediately after transmission of the sleep signal, the network interface device transitions to a quiet time window of the time schedule in which the network interface device ignores transmissions from the link partner.

Energy efficient ethernet (EEE) operation

A physical layer transceiver for a data channel includes receiver circuitry configured to receive signals on the data channel, transmit circuitry configured to transmit signals onto the data channel, and adaptive filter circuitry coupled to the receiver circuitry and the transmit circuitry and configured to filter the data channel by operating on input frequency-domain data samples to output filtered data samples. The adaptive filter circuitry includes error sample generation circuitry configured to generate error samples representing a difference between a target response and the filtered data samples, arithmetic-only circuitry configured to approximate a windowing function to operate on the error samples, and output sample generation circuitry configured to operate on windowed error samples to provide the output filtered data samples. The comparison circuitry may be configured for time-domain operation and may further be configured to transform the error signals into frequency-domain error signals.

Simplified frequency-domain filter adaptation window

Disclosed is aUGMIIsystem to interface multirate devices including 10 gigabit per second data exchange rates. Mode selection is enabled to provide for automatic detection and adaptation to any transmit rate including 10M, 100M, 1G, and 10G. Mode selection comprises the negotiation between the UGMII extension sublayers located at the MAC and PHY to select between one of several operational modes including: XGIVIII communication, GIYIII encapsulation, Clause 22 MDIO register management and Clause 45 MDIO register management. Selection of UGMII and XGMII operating modes are negotiated between the MAC and PHY using ordered sets to announce and acknowledgement a mode change. In one embodiment 802.3 Clause 46 defined ordered sets are utilized.

Method and system for a multi-rate gigabit media independent interface

A method for communication, including, in a Physical Layer (PHY) transceiver, selecting a transmission bitrate from a plurality of transmission bitrates, for transmitting over a media interface bits received from an external device. The received bits are processed by generating, using a framing and encoding scheme that depends on at least the selected transmission bitrate, frames having a common frame length among the framing and encoding schemes. The frames are encoded to produce code words of a predefined Forward Error Correction Code (FEC) code, using a single FEC encoder that accepts a number of bits for encoding equal to the frame length. Sub-units of the code words are mapped into symbols using one of at least two mapping schemes that employ different voltage amplitude levels to define a transmission symbol, the mapping scheme being selected according to the selected transmission bitrate. The symbols are transmitted over the media interface.

Efficient ethernet multi-mode coding and modulation for twisted-pair

The systems and methods for a twisted pair transceiver with correlation detection includes a transceiver system operating on a cable. The transceiver system includes a receiver to obtain one or more data samples related to one or more encoded data symbols. The transceiver system further includes a first correlation filter to generate a first correlation output based on the one or more data samples, and a second correlation filter to generate a second correlation output based on the one or more data samples. The transceiver system further includes a detector. The detector compares the first correlation output with the second correlation output, generates an output data bit based on a comparison result, and sends the output data bit for data decoding.

Systems and methods for a twisted pair transceiver with correlation detection

Disclosed is aUGMIIsystem to interface multirate devices including 10 gigabit per second data exchange rates. Mode selection is enabled to provide for automatic detection and adaptation to any transmit rate including 10M, 100M, 1G, and 10G. Mode selection comprises the negotiation between the UGMII extension sublayers located at the MAC and PHY to select between one of several operational modes including: XGMII communication, GMII encapsulation, Clause 22 MDIO register management and Clause 45 MDIO register management. Selection of UGMII and XGMII operating modes are negotiated between the MAC and PHY using ordered sets to announce and acknowledgement a mode change. In one embodiment 802.3 Clause 46 defined ordered sets are utilized.

In a method for coding a packet of information words into a frame for transmission, blocks of input data are received, the input data including control words and the packet of information words, the packet having (a) a start preceded by ones of the control words and (b) an ending followed by others of the control words, the blocks being smaller than the packet. When the block consists exclusively of information words, a one bit block header having a first sense is appended to the block to form the frame. When the block does not consist exclusively of information words: the block is condensed to accommodate a TYPE word, the TYPE word is inserted into the block, and a one bit block header having a second sense, opposite to the first sense, is appended to the block to form the frame.

Communication system and encoding method having low overhead

A method includes generating a pulse, transmitting the pulse on a first cable of a multi-cable communication system, monitoring a second cable of the multi-cable communication system for the pulse or a reflection of the pulse on the second cable, and, if the monitoring detects the pulse or the reflection on the second cable, outputting a notification of a short between the first cable and the second cable.

Cable diagnostics for Base-T systems

A method for converting a data stream from a first rate to a second rate Ntimes greater than the first rate, such that the converted data stream conforms to a coding requirement that symbols be arranged in units of symbol length M, includes receiving, at the first rate, a data packet, and an inter-packet gap containing Yidle symbols, generating Nrepetitions of each data symbol of the data packet, determining whether replicating each symbol of the inter-packet gap Ntimes would violate the coding requirement, and, at least partially in response to determining that replicating each symbol of the inter-packet gap Ntimes would violate the requirement, generating greater than or less than Y*Nidle symbols. The method also includes outputting, at the second rate, the generated Nrepetitions of each data symbol and the generated idle symbols.

Multi-rate media independent interface over a physical coding sublayer

A method for inserting ordered sets into a data stream converted from a first rate to a second rate Ntimes greater than the first rate, while conforming to a requirement that symbols be arranged in units of symbol length M, includes receiving, at the first rate, a data packet containing data symbols, receiving Ysymbols of control information, and identifying an ordered set based on the Ysymbols. The method also includes generating Nrepetitions of each data symbol of the data packet, determining whether Y*N/M has a non-zero remainder, and, at least partially in response to determining that Y*N/M has a non-zero remainder, generating a number of repetitions of the ordered set containing either more or fewer than Y*Nsymbols. The method also includes outputting, at the second rate, the data symbol repetitions and first ordered set repetitions.

A communication system processes blocks of input data that include control words and a packet of information words are received. The packet has a start preceded by ones of the control words and an end followed by others of the control words. When the block consists exclusively of information words, a one bit block header having a first sense is appended to the block to form a frame. When the block does not consist exclusively of information words, the block is condensed to accommodate a TYPE word, the TYPE word is generated and inserted into the block and a one bit block header is appended to the block to form the frame. When the block does not consist exclusively of information words the one bit block header has a second sense, opposite to the first sense. Use of a one bit block header reduces overhead. Forward error correction is also utilized.

An apparatus and method for cable diagnostics is disclosed for deployment as part of an Ethernet communication system to conduct diagnostics. The system transmits one or more pulses which are Ethernet compatible and of finite duration. Reflections are detected and through processing and measurement of the amplitude and round-trip delay of the pulse's reflection, cable faults are located. In one embodiment, this innovation uses an IEEE 802.3 compliant transmit pulse, such as an auto-negotiation signal (AN pulse), to conduct cable diagnostics. The benefits of a standard compliant allow for use with any vendor on the far-end and the signal requires no special hardware or software to produce and, therefore, reduces system complexity and cost. To reduce incorrect measurements, the apparatus measures cable length and termination with multiple AN pulses. It then applies non-linear filters to redundant measurements in such a way that it produces accurate cable diagnostics information.

Growing community of inventors

Laguna Hills, CA, United States of America

Brett Anthony McClellan