US20110035489A1 - System and method for interfacing with a management system - Google Patents

Abstract

Systems and methods that interface with a management system are provided. In one embodiment, a system and a method may provide a command protocol and format for communication between a network interface card (NIC) and a management device such as, for example, an intelligent management device (IMD). An interface may be adapted to allow the management device to merge its traffic with that of the NIC to provide a fully integrated management solution. The fully integrated management solution may be implemented, for example, without additional network connections.

Description

CROSS REFERENCE TO RELATED APPLICATION
• This application makes reference to, claims priority to and claims benefit from U.S. Provisional Patent Application Ser. No. 60/453,642, entitled “Intelligent Management Device Interface” and filed on Mar. 11, 2003.
INCORPORATION BY REFERENCE
• The above-referenced United States patent application is hereby incorporated herein by reference in its entirety.
FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
• [Not Applicable]
MICROFICHE/COPYRIGHT REFERENCE
• [Not Applicable]
BACKGROUND OF THE INVENTION
• FIG. 1 illustrates a block diagram of asystem10 connected to anetwork20 viaswitches30a,30b. Thesystem10 includes a network interface card (NIC)40 and an intelligent management device (IMD)50. The NIC40 is connected to theswitch30aand to a host (not shown) of thesystem10. The IMD is connected to theswitch30band to the host of thesystem10. The NIC includes a NIC media access control (MAC)60 and aNIC processor70. Thenetwork20 is connected to theswitch30awhich, in turn, is connected to the NIC MAC60. The NIC MAC60 is connected to the NICprocessor70 which, in turn, is connected to the host of thesystem10. The IMD50 includes a MAC80 andmanagement processor90. Thenetwork20 is also connected to theswitch30bwhich, in turn, is connected to theMAC80. The MAC80 is connected to themanagement processor90 which, in turn, is connected to the host of thesystem10.
• The IMD50 provides, for example, monitoring, management capabilities and remote functionality. For example, the IMD50 can provide monitoring and management capabilities for thesystem10 and can provide remote functionality to or from a device (e.g., a remote device) connected to thenetwork20.
• The IMD50 can have one or more of the following disadvantages. For example, as illustrated inFIG. 1, thesystem10 includes an additional dedicated connection to thenetwork20. Besides the additional cost (e.g., theadditional switch30b) of implementing another system port, theIMD50 is susceptible to a failure, for example, of theswitch30b. Thus, ifswitch30bwere to fail, then theIMD50 would no longer be accessible via thenetwork20. Furthermore, because the IMD50 is connected to thenetwork20, the IMD50 may be needlessly processing some packets carried on thenetwork20. On high-speed networks, in particular, the resources of the IMD50 can be substantially consumed by such unnecessary processing, thereby reducing some resources of the IMD50 that could have been allocated for other tasks. For example, some packets (e.g., packets that can be forwarded as received) may be processed by the IMD50, even though these packets need not be processed by the IMD50.
• Further limitations and disadvantages of conventional and traditional approaches will become apparent to one of ordinary skill in the art through comparison of such systems with one or more aspects of the present invention as set forth in the remainder of the present application with reference to the drawings.
BRIEF SUMMARY OF THE INVENTION
• Aspects of the present invention may be found in, for example, systems and methods that interface with a management system such as, for example, an intelligent management device. In one embodiment according to some aspects of the present invention, a communications system may include, for example, a network interface card (NIC) and a management device. The management device may be coupled to the NIC. The NIC may be adapted, for example, to merge communications traffic of the management device with the NIC.
• In another embodiment according to some aspects of the present invention, a communications system may include, for example, a first NIC, a second NIC and a manager. The first NIC and the second NIC may be coupled to a network. The manager may be coupled to the first NIC and the second NIC. The manager may initially be in two-way communications with the network via the first NIC. However, if the first NIC fails, then the manager may switch from the first NIC to the second NIC and be in two-way communications with the network via the second NIC.
• In yet another embodiment according to some aspects of the present invention, a method of communications may include, for example, one or more of the following: providing access to and from a network for a management device via a NIC; configuring one or more filters of the NIC via one or more commands generated by an management device; filtering incoming packets via the one or more filters; and forwarding the filtered packets based upon one or more matches between information carried by the filtered packets and one or more filtering parameters.
• In yet another embodiment according to some aspects of the present invention, a method of communications between a NIC and a management device may include, for example, one or more of the following: generating a command in the management device, the command comprising a particular sequence number; storing the command in the management device; sending the command to the NIC; executing the command in the NIC; and generating a response to the command, the response comprising the particular sequence number. In some embodiments according to the present invention, the command may also include, for example, an identifier-type field and a command structure. The response may also include, for example, an identifier-type field and a response structure.
• In yet still another embodiment according to some aspects of the present invention, a method of remote management over a network may include, for example, one or more of the following: accessing the network via a plurality of NICs of a local server system; communicating between a local manager of the local server system and a remote manager over the network through a NIC selected by the local manager, the selected NIC being one of the plurality of NICs; managing the local server system via the local manager; and responding locally to management commands sent over the network from the remote manager.
• These and other features and advantages of the present invention may be appreciated from a review of the following detailed description of the present invention, along with the accompanying figures in which like reference numerals refer to like parts throughout.
BRIEF DESCRIPTION OF THE DRAWING
• FIG. 1 shows a block diagram of a system connected to a network via switches.
• FIG. 2 shows a block diagram of a system coupled to a network according to an embodiment of the present invention.
• FIG. 3 shows a block diagram of a system coupled to a network according to an embodiment of the present invention.
• FIG. 4 shows a flowchart illustrating an embodiment of a process for receiving and forwarding packets according to the present invention.
• FIG. 5 shows a flowchart illustrating an embodiment of a process for handling command packets according to the present invention.
• FIG. 6 shows a flowchart illustrating an embodiment of a process for handling response packets according to the present invention.
DETAILED DESCRIPTION OF THE INVENTION
• Aspects of the present invention may be found, for example, in systems and methods that interface with a management system such as, for example, a management system including an intelligent management device (IMD). In some embodiments according to the present invention, a command protocol and format for communication between an interface card (e.g., a network interface card (NIC), a controller, an adapter, etc.) and a management system may be provided.
• In some embodiments according to the present invention, an interface that may allow a management system to merge its traffic with that NIC (e.g., a standard NIC, a network interface controller, etc.) to provide a fully integrated management solution may be provided. The fully integrated management solution may be implemented, for example, without adding further network connections.
• In some embodiments according to the present invention, a separate Ethernet connection port on a NIC may be provided. The separate Ethernet connection port may be, for example, a universal management port (UMP). Via this interface, a management system may see packets that it would see if it were directly connected to the network as well as many other types of packets (e.g., commands, responses, etc.)
• In some embodiments according to the present invention, a system that passes command packets over an Ethernet interface between a management system and an UMP may be provided.
• FIG. 2 shows an embodiment of asystem100 including an IMD110 according to the present invention. Thesystem100 may be, for example, a server system, a server blade, a desktop system, a computer system, a network system, a set top box, etc.) Thesystem100 may include, for example, the IMD110 and a NIC120. Thesystem100 may be coupled to anetwork130 via theNIC120. TheIMD110 may include, for example, amanagement processor140 and a media access control (MAC)150 (e.g., a 10/100 MAC). TheNIC120 may include, for example, aNIC processor160, a set offilters170, a MAC180 (e.g., a 10/100 MAC or UMP) and amain NIC MAC190. Themanagement processor140 may be coupled to theMAC150 which, in turn, may be coupled to theMAC180. TheMAC180 may be coupled to theNIC processor160 which, in turn, may be coupled to themain NIC MAC190. TheNIC processor160 may also be coupled to the set offilters170 which, in turn, may be coupled to theNIC MAC190. Themain NIC MAC190 may be coupled to thenetwork130 via aswitch200.
• Thesystem100 may also include, for example,system sensors210, system controls220, asystem interconnect230, a central processing unit (CPU)240, anotherCPU250, a system storage device260 (e.g., a system memory) and peripheral devices270 (e.g., disk devices, video devices, etc.) Thesystem sensors210 and the system controls220 may be coupled to theIMD110 and, in particular, may be coupled to themanagement processor140. TheNIC120, theIMD110, theCPUs240,250, thesystem storage device260 and theperipheral devices270 may each be coupled to thesystem interconnect230. In particular, theNIC processor160 and themanagement processor140 may be coupled to thesystem interconnect230 via afirst host connection280 and asecond host connection290. Thehost connections280,290 may be, for example, peripheral component interconnects (PCIs).
• Other devices may be coupled to thenetwork130. For example, amanagement console300 may be coupled to thenetwork130 via aswitch310.Other systems320 may be coupled to the network viarespective switches330. Although illustrated as asingle switch330, eachrespective system320 may include its own respective switch or switches330. Some of theother systems320 may be identical to thesystem100 illustrated inFIG. 2 and some of theother systems320 may be identical to thesystem100 illustrated inFIG. 3. However, theother systems320 need not be so limited in scope. In some embodiments, some of theother systems320 and thesystem100 may be remotely controlled (e.g., remotely monitored, remotely activated, remotely managed, remotely accessed, etc.) by themanagement console300.
• FIG. 3 shows an embodiment of the present invention in which thesystem10 is coupled to thenetwork130 via a plurality of NICs120 (i.e., NIC120(1), NIC120(2), . . . , NIC120(n), where n is an integer value). EachNIC120 may include, for example, arespective NIC processor160, a set offilters170, afirst MAC180 and asecond MAC190. EachNIC120 may be coupled to thenetwork130 via arespective switch200. The various components may be coupled together as described above with respect to theNIC120 ofFIG. 2. In addition, eachNIC120 may be coupled to theIMD110. In one embodiment, thesecond MACs190 of therespective NICs120 each may be coupled to theMAC150 of theIMD110.
• In operation and with reference toFIG. 2, themanagement console300 may be in two-way communications with thesystem100 and some of theother systems320 coupled to thenetwork130. Themanagement console300 may provide, for example, management services (e.g., remote management services) for thesystem100 and some of theother systems320. Thus, themanagement console300 may be able to pull up graphical user interfaces (e.g., windows) on its display for each of the systems that themanagement console300 manages. In some embodiments, the activities performed via themanagement console300 may appear seemless even though the systems being managed may be far way.
• TheIMD110 may be in two-way communications with the management console300 (e.g., a remote management console) vianetwork130. Themanagement console300 may be in two-way communications with not only theIMD110 of thesystem100, but also some of theother systems320 coupled to thenetwork130. In some embodiments according to the present invention, theIMD110 may be in two-way communications with themanagement console300 only through theNIC120. In various embodiments according to the present invention, theIMD110 may not have its own direct connection to thenetwork130, but instead may use theNIC120 to access the network. TheIMD110 may provide management services for thesystem100. For example, theIMD110 may monitor the system sensors142 or adjust the system controls144. Thus, for example, theIMD110 may monitor power supply parameters, voltage parameters, current parameters, temperature parameters, status parameters, failure parameters of various components and circuits. The system sensors142 may also provide alerts to theIMD110 such as, for example, that the cover of thesystem100 has not been replaced, that components have been removed or failed, or that the temperature in thesystem100 has exceeded particular thresholds. TheIMD110 may also adjust, activate or set system controls144 in response to monitored parameters or in response to particular commands or requests. For example, theIMD110 may reset power settings or parameters settings, power up thesystem100 or a particular component of thesystem100, or power down the system or a particular component of thesystem100.
• In addition to responding to current conditions, theIMD110 may also respond to requests received from the host of thesystem100 or from thenetwork130. In some embodiments according to the present invention, theIMD110 may receive requests or commands from themanagement console300. In various embodiments according to the present invention, themanagement console300 may provide, for example, a user input device (e.g., a keyboard, a mouse, etc.) and a user output device (e.g., a display, a graphical user interface, a video output, a graphical output, an audio output, etc.) By opening up a window in a display, a user located at themanagement console300 may monitor information that theIMD110 may be receiving from thesystem sensors210 or the system controls220 or other components of thesystem100. The user may then send commands or requests to theIMD110 to which theIMD110 may respond. For example, themanagement console300 may send a request for information to theIMD110. The request may be sent via thenetwork130 to theNIC120. The request may be routed through thefilters170 which may determine whether or not the request is destined for theIMD110. If the request had not been destined for theIMD110, then, for example, it may have been forwarded to theNIC processor160 or to the host of thesystem100 or elsewhere for further processing. If the request is destined for theIMD110, then theNIC120 may route the request to theIMD110. Thus, in some embodiments according to the present invention, theIMD110 may only receive requests or commands or data packets that are destined for theIMD110.
• TheIMD110 may then analyze the request and may perform the request. For example, the request may include commands for powering down an overheated component. TheIMD110 may thus adjust one or more of the system controls144 of thesystem100. TheIMD110 may then send a response to themanagement console300 through theNIC120 and thenetwork130. The response may include, for example, graphical information for display on a monitor that the user is viewing. Thus, user seemlessly manages theIMD110 as if the user were located at thesystem100. In some embodiments, themanagement console300 may act as a dummy terminal in which keyboard strokes or other input commands (e.g., mouse interface information) are relayed to theIMD110 which, in turn, then may send graphical information to themanagement console300 which may be displayed on the monitor as alphanumeric symbols (or, for example, cursor movements) corresponding to the keystrokes (or, for example, mouse movements).
• Referring toFIG. 3, theIMD110 may be coupled to a plurality ofNICs120. In some embodiments according to the present invention, only one of the plurality ofNICs120 provides access to and from thenetwork130 for theIMD110. TheIMD110 may select which of theNICs120 to use as a sole connection to and from thenetwork130. In some embodiments according to the present invention, theIMD110 may have the ability to switch between using one NIC or another NIC as the sole connection to and from thenetwork130, for example, at different times. This ability may enhance throughput as well as promote fault tolerance. For example, if the selected NIC is being overused by other resources of thesystem100 or thenetwork130, then theIMD110 may select another NIC (e.g., a NIC that is not being overused by other resources of thesystem100 or the network130). In another example, if the selected NIC has failed or has been removed from thesystem100, then the IMD may select another NIC to provide a connection (e.g., a sole connection) to and from thenetwork130. Furthermore, if themanagement console300 determines that the selected NIC has failed or has been removed from thesystem100, themanagement console300 may wait and try again or themanagement console300 may broadcast the information to the other NICs of thesystem100. TheIMD110 may then be notified via the broadcasted information received from one or more of the other NICs of the system and switch to another NIC. TheIMD110 may then use the newly selected NIC to inform themanagement console300 of the switch in NIC selection.
• In some embodiments according to the present invention, thesystem100 may allow, for example, theIMD110 to merge its traffic with that of theNIC120. Thesystem100 may provide, for example, a fully integrated management solution that does not necessitate additional network connections. TheNIC120 may transmit, for example, non-command packets via themain NIC MAC190. TheNIC120 may receive, for example, packets via themain NIC MAC190 and send the received packets through the set offilters170. The set offilters170 may allow received packets to be passed to theIMD110 if, for example, the received packets meet (e.g., match) at least some of the filter requirements (e.g., programmed filter parameters). TheNIC120 may receive packets from theIMD110 that meet a particular encapsulation format and may process the packets locally. TheNIC120 may receive other types of packets from theIMD110 that may be transmitted exactly as they were received from theIMD110 including, for example, virtual local area network (VLAN) tag information.
• FIG. 4 shows a flowchart illustrating an embodiment of a process for receiving and forwarding packets according to the present invention. Instep340, thesystem100 may receive packets from thenetwork130 via themain NIC MAC190. Inquery350, whether the set offilters170 are operable may be based upon whether the set offilters170 have been configured. The set offilters170 may be configured, for example, via commands generated by theNIC processor160 of theNIC120 or themanagement processor140 of theIMD110. If the set offilters170 have been configured, then, inquery360, it is determined whether the information carried by the packet satisfies (e.g., matches) one or more of the configured filter parameters. If the information carried by the packet does satisfy one or more of the configured filter parameters, then, instep390, the packet may be processed by theNIC processor160 and themanagement processor140 and forwarded to a destination according to the matched filter parameters. In one embodiment, with regard to layer 2 (L2) address values, theNIC120 may forward data that may meet, for example, any of the selected perfect match filters, for example, for the L2 address. In another embodiment, theNIC120 may limit forwarded traffic up to a particular number of VLAN networks. In yet another embodiment, theNIC120 may filter L2 multicast or broadcast traffic and forward such traffic accordingly. In one embodiment, a broadcast packet that is determined to be an address resolution protocol (ARP) packet or other specific types of broadcast packets may be forwarded accordingly.
• If the set of filters are not configured (query350) or if the information carried by the packet does not satisfy any of the configured filter parameters (query360), then, instep370, the packet may not reach theIMD110. Instep380, the packet may be forwarded as received by themain NIC MAC190 to the rest of thesystem100. In one embodiment, if the packets are not filtered, the receive (RX) packets including, for example, VLAN tag information may be forwarded as received via the mainNIC MAC port190.
• In some embodiments, the present invention may provide command packets and/or response packets transmitted and/or received by theNIC120 having a format as set forth below. For example, the commands may be from theIMD110 to theNIC120 and the responses may be from theNIC120 to theIMD110.

• Byte	31	23	15	7
  0	XXXX	XXXX	DA	DA
  4	DA	DA	DA	DA
  8	SA	SA	SA	SA

  12

  SA

  SA

  BRCM Num.

  16	0x5706		Cmd. Seq. Num.
  20	Cmd. Type		Data Length

  24	Data0	Data1	Data2	Data3
  28	Data4	Data5	Data6	Data7
  . . .	. . .	. . .	. . .	. . .

  . . .

  DataN

  Optional Padding to 32 bits

  . . .	2s Complement Checksum
  	Compensation
  . . .	Optional Zero Pad to 64 Byte Legal
  	Ethernet Frame

• In some embodiments, the present invention may provide for fields with definitions as set forth below.
• XXXX Padding for the purposes of this table to make protocol values align on the 32-bit rows. These bytes may not be a part of any packet.
• DA Ethernet Destination Address. Value is configurable, but typically is equal to the address the receiver represents on the main interface.
• SA Ethernet Source Address. Value is configurable, but typically is equal to the address the sender represents on the main interface.
• BRCM Number
  • Well-known IEEE packet type number.
• 0x5706 Packet type number used to define a packet sub-type for the BRCM number.
• Cmd Type Command Type. When bit15 of the type is set, it indicates a response packet sent from the NIC to the IMD.
• Data Length
  • This value is the length in bytes of the data payload of the command packet. This is the number of “Data” bytes in the packet. This count may not include this value or any bytes before this value. This count may not include any Padding or Checksum bytes. The checksum may sum bytes not in this count if this count is not a multiple of four.
• Data . . . . Command payload.
• Optional Padding to 32-bits
  • If the Data portion of the command is not a multiple of four in length (i.e., Bottom two bits of Data Length field are not zero) then from one to three bytes of zero value padding may be added so that the 2s complement checksum may be calculated over full 32-bit values.
• 2s Complement Checksum Compensation
  • When this value is added to the 2's complement sum of all the 32-bit words in the “Data” and “Optional Padding” areas, the resulting value may be 0xffffffff. The checksum checking may optionally be disabled at each receiver. The sender may then take advantage of this by setting this value to 0xffffffff.
• Optional Zero Pad to 64 Byte Legal Ethernet Frame
  • For frames to be compliant with IEEE 820.3, all frames are padded up to a total of 64 bytes. This is typically done automatically by most Ethernet MAC devices.
• In some embodiments, the present invention may provide that command frames may not use additional L2 encapsulation techniques such as, for example, VLAN, SNAP, etc.
• In some embodiments, the present invention may employ a transport mechanism by which reliable reception of control frames may be tracked via sequence numbers. Each command packet may be associated with a respective response packet by each carrying a common sequence number in a particular sequence number space. Accordingly, the same sequence number space may be used for command/response packets in either direction between theNIC120 and theIMD110.
• FIG. 5 shows a flowchart illustrating an embodiment of a process for handling command packets according to the present invention, for example, as executed by theIMD110. Instep400, a sender of a command (e.g., the IMD110) may generate a command packet. The command packet may include, for example, a sequence number with which the command packet is associated. Instep410, the sender of the command may then store the generated command. Instep420, the sequence space counter of the sender may then be incremented. In one embodiment, thestep420 may proceedstep400 andstep410. In another embodiment, the command is sent from theIMD110 to theNIC120. TheNIC120 may respond to the command with a response. Instep422, theNIC120 may receive a packet from theIMD110. Inquery424, a BRCM number may be detected in a particular field (e.g., an identifier-type field). If the BRCM is not detected, then, instep426, the packet may be processed as a management protocol frame for themanagement console300. If the BRCM is detected, then query430 follows. Inquery430, if sender of command receives a response, then, instep450, the command corresponding to the received response may be deleted from storage. In one embodiment, the received response may include, for example, a particular sequence number associated with a particular command, thereby identifying the stored command which can be deleted.
• If a response to a particular command is not received, then, instep440, the particular command, which has been stored, may be retransmitted. In one embodiment, the command sender may be able to retransmit any command until a response with the same sequence number is received and is processed. Accordingly, a retransmission capability is provided in the case a command packet is lost, corrupted or dropped. In another embodiment, if the sender of commands detects a particular sequence number of a response that has already been received, then the sender of commands may resend all the commands that have not had responses received that follow the latest received sequence number. In yet another embodiment, via a periodic timer, a requestor may verify commands to which have been responded. If a command has not been responded to with a response, then a retransmit of the command may be commenced. The requestor may time-out on any request without a response and may start retransmission after that timer has expired. This may be done, for example, by setting the timeout each time a command is transmitted and each time a response is received. When the timer expires, retransmit may be started if all the outstanding commands have not been responded to with corresponding responses.
• FIG. 6 shows a flowchart illustrating an embodiment of a process for handling response packets according to the present invention as executed, for example, by theNIC120. Instep460, a packet may be received. In one embodiment, a command packet may be received by a responder to commands (e.g., the NIC120) from the sender of commands (e.g., the IMD110). Inquery462, a BRCM number may be detected in a particular field (e.g., an identifier-type field). If the BRCM number is not detected, then, instep464, the packet may be forwarded to a main transmission port. If the BRCM number is detected, then, instep470, the sequence number carried by the command packet may be determined. For example, theNIC processor160 may parse the command packet to determine the sequence number carried by the command packet. Inquery480, it may be determined whether the determined sequence number is the expected sequence number. If the determined sequence number is the expected sequence number, then, instep490, the command may be executed and a response packet may be generated that may include, for example, the expected sequence number. In one embodiment, theNIC processor160 may configure (e.g., program), for example, one or more filters in the set offilters170 in light of the received command packet. Instep500, the generated response packet may be stored and returned to the sender of the command packet. In one embodiment, the responder to commands may store only the last response packet sent to the sender of commands. Thus, the latest response packet may be written over the previous response packet or the previous response packet may be deleted or invalidated. In another embodiment, the responder to commands may store one or more response packets sent to the sender of commands.
• If the determined sequence number is not the expected sequence number (query480), then, instep510, the stored response packet (e.g., the previously stored response packet) may be resent. In one embodiment, if the responder receives a command with a sequence number other than the one expected (e.g., the last one plus one), then theNIC120 may not execute that command, but instead may send the response for the previously executed command. In another embodiment, the responder to commands may re-execute a command without any adverse effects (e.g., side effects).
• In one embodiment, no more than approximately 2¹⁵−1 commands may be outstanding at any time, for example, due to protocol limitations. Memory limitations in the sender may become evident long before the protocol limit is reached. The responder may only keep the last response that was sent so that it might not have any retransmit memory limitations for outstanding packets. However, other limitations may become factors for consideration.
• Since the requestor may be the only station saving the outstanding commands, the responder may be able to execute all commands more than once without any adverse effects. Accordingly, the retransmission of a command that exhibits modal effects may cause problems. These modal effects can be avoided by the design of the command packets such that repeated execution has no side effects.
• The following are some examples of commands and responses according to an embodiment of the present invention.
• Hello—0x0001. The Hello command may solicit a presence response from theNIC120. TheNIC120 may respond as long as the message is received without error. The command does not make a commitment to work further. The Hello command may be intended for use as a flush command, if needed, possibly during initial negotiation. Error Codes may include, for example, OK.
• ID Request—0x0002. The ID Request command may indicate the IMD type and version as a string up to, for example, forty characters long. The response to the ID Request command may return the NIC type and version as a string up to, for example, forty characters long. Error Codes may include, for example, OK; UNAVAIL; BAD_ID; or FATAL
• Reset To Default—0x0003. The Reset To Default command may request that defaults be set for settings relating to, for example, filters, flow controls, etc. This may be an equivalent state to a reset NIC. Error Codes may include, for example, OK or FATAL.
• Set NIC<->IMD Flow Control Method—0x0101. The Set NIC<->IMD Flow Control Method command may set, for example, the flow control method used between, for example, theIMD110 and theNIC120. In some embodiments, only symmetrical flow control may be supported in which both ends use the same method. An OK response may indicate that theNIC120 has set itself to the same flow control method. Settings may include, for example, NONE or PAUSE. Error Codes may include, for example, OK; NOT SUPPORTED; or FATAL.
• Set NIC Drop Policy—0x0102. The Set NIC Drop Policy command may set the drop policy used by theNIC120 when passing frames to theIMD110. Settings may include, for example, DROP or BACKPRESSURE. Responses may include, for example, DROP or BACKPRESSURE—setting now in; or DRIVER_CTL—bit set when driver is running and BACKPRESSURE mode could not be selected. Error Codes may include, for example, OK or FATAL.
• Set NIC Port—0x0103. The Set NIC Port command may request, for example, that theNIC120 set the main port to a particular speed. If the main port speed is already at the specified speed, then no action may be taken and the link might not be dropped. The speed may be limited, for example, by power management limitations. Setting the requested speed may jeopardize exceeding the power supply for a particular design. This may not be returned on a card or LAN-on-motherboard (LOM) that may provide enough power in all modes. The speed change may be denied due to the OS driver being loaded. The speed change may be denied because of, for example, interface limits such as Fiber or because an external “in-line” device connected that may not be controlled for some reason. Settings may include, for example, GET, 10, 100, 1000—GET requests that current state be returned with no effect; HD_FLAG—half duplex flag, invalid with 1000 setting; PAUSE—Pause Flow Control Enable Flag; or AUTO—flag indicating that a setting is a “maximum setting” to advertise for auto-negotiation, when set, the link may be dropped as the link is re-negotiated. Responses may include, for example, 10, 100, 1000—speed of link; HD_FLAG—set if link is half duplex; PAUSE—set if pause flow control is enabled; LINK UP—flag set if link was attained in new mode; FIBER—flag set if fiber connection; PWR_LIMIT—flag set if requested speed was denied due to particular card power limitations; DRV_LIMIT—flag set if request speed was denied due to driver being loaded and forcing link type; INTF_LIMIT—flag is set if requested speed was denied due to limitations in the PHY device (e.g., Fiber that cannot support 10/100); AUTO_NEG—flag is set if link was attained using auto-negotiation; AUTO_NEG_PAR—flag is set if link was attained using auto-negotiation and parallel detection was used to detect link (AUTO_NEG may also be set if this flag is set); CAP10, CAP100, CAP1000—capability advertised by link partner (valid only if AUTO_NEG=1 && AUTO_NEG_PAR=0); CAP_HD_FLAG—capability advertised by link partner (valid only if AUTO_NEG=1 && AUTO_NEG_PAR==0); or CAP PAUSE—capability advertised by link partner (valid only if AUTO_NEG=1&& AUTO NE_PAR=0). Error Codes may include, for example, OK or FATAL.
• Drop NIC Port Link—0x0104. The Drop NIC Port Link command may drop a link on the main port if it has not already been dropped. The link may be re-established, for example, by using the Set NIC Port command. Error Codes may include, for example, OK or FATAL.
• Filter All Packets—0x0201.

• Byte	31	23	15	7

  16	0x5706		Cmd. Seq. Num.
  20	0x0201		4

  24	0xFDFEFFFB
  28-63	0

• The Filter All Packets command may provide that packets received on the main port may be filtered and may not be delivered to theIMD110. The command may be used, for example, as a power-up setting and may be used with any filters (e.g., time filters) when reset or re-loaded.
• Filter All Packets Response—0x8201.

• Byte	31	23	15	7

  16	0x5706		Res. Seq. Num.
  20	0x8201		4

  24	0x7DFEFFFB
  28-63	0

• Filter based on settings—0x0202. Packets may be filtered according to particular filter settings (e.g., current filter settings).
• Set Perfect Match Filter—0x0203. It may set the filters defined by Fnum to the value specified in DA. Legal values for Fnum are 0 and 1. If the “OFF” bit is set, then the specified perfect match filter may be disabled.

• Byte	31	23	15	7

  16	0x5706		Cmd. Seq. Num.
  20	0x0203		8

  24

  DA

  28

  DA

  OFF

  FNum

  28	Checksum
  32-63	0

• Set Perfect Match Filter Response—0x8204.
• Set Broadcast Filter—0x0205. The Set Broadcast Filter command may be set if broadcast frames should be forwarded to theIMD110. Settings may include, for example, ON. Error Code may include, for example, OK or FATAL.
• Set ARP Filter—0x0206. The Set ARP Filter command may be set if ARP frames should be forwarded to theIMD110. Settings may include, for example, ON. Error Code may include, for example, OK or FATAL.
• Get Statistics—0x0301. The Get Statistics command may request a response with the requested statistics in it. Results may include, for example:

• u64_t IfHCInOctets;

  	/* Collected from emac_rx_stat_ifhcinoctets. */
  	/* This is the number of octets received on the interface, including framing

  characters. */

  u64_t IfHCOutOctets;

  	/* Collected from emac_tx_stat_ifhcoutoctets. */
  	/* This is the number of octets that have been transmitted on the interface. */

  u64_t IfHCInUcastPkts;

  	/* Collected from emac_rx_stat_ifhcinucastpkts. */
  	/* This is the number of frames received on the wire that were not dropped due

  to errors that have Unicast Ethernet destination addresses. */

  u64_t IfHCInMulticastPkts;

  	/* Collected from emac_rx_stat_ifhcinmulticastpkts. */
  	/* This is the number of frames received on the wire that were not dropped due to

  errors that have multicast Ethernet destination addresses. */

  u64_t IfHCInBroadcastPkts;

  	/* Collected from emac_rx_stat_ifhcinbroadcastpkts. */
  	/* This is the number of frames received on the wire that were not dropped due to

  errors that have the broadcast Ethernet destination addresses. */

  u64_t IfHCOutUcastPkts;

  	/* Collected from emac_tx_stat_ifhcoutucastpkts. */
  	/* This is the number of packets transmitted that have unicast destination addresses. */

  u64_t IfHCOutMulticastPkts;

  	/* Collected from emac_tx_stat_ifhcoutmulticastpkts. */
  	/* This is the number of packets transmitted that have multicast destination addresses. */

  u64_t IfHCOutBroadcastPkts;

  	/* Collected from emac_tx_stat_ifhcoutbroadcastpkts. */
  	/* This is the number of packets transmitted that have the broadcast destination

  address. */

  u32_t Dot3StatsCarrierSenseErrors;

  	/* Collected from emac_rx_stat_dot3statscarriersenseerrors. */
  	/* This is the number of times a false carrier has been detected on the internal

  	PHY device. This is indicated from the PHY by asserting RXER while
  	RXDV is low when the RXD pins are at a state of 0x0e. */

  u32_t Dot3StatsFCSErrors;

  	/* Collected from emac_rx_stat_dot3statsfcserrors. */
  	/* A IfInErrors value is the sum of this value and <b>Dot3StatsFCSErrors</b> and

  <b>Dot3StatsAlignmentErrors</b>. */

  u32_t Dot3StatsAlignmentErrors;

  	/* Collected from emac_rx_stat_dot3statsalignmenterrors. */
  	/* This is the number of frames received on the wire that have an odd number of

  nibbles and fail FCS check and are of legal length. */

  u32_t Dot3StatsSingleCollisionFrames;

  	/* Collected from emac_tx_stat_dot3statssinglecollisionframes. */
  	/* This is the number of collisions that were followed by successful packet

  	transmits. This is the same as the number of packets that were transmitted
  	with only one collision. */

  u32_t Dot3StatsMultipleCollisionFrames;

  	/* Collected from emac_tx_stat_dot3statsmultiplecollisionframes. */
  	/* This is the number of packets that have transmitted with more that one collision. */

  u32_t Dot3StatsDeferredTransmissions;

  	/* Collected from emac_tx_stat_dot3statsdeferredtransmissionsl. */
  	/* This is the number of packets that were delayed in transmission because they

  had to wait for a RX packet to complete. */

  u32_t Dot3StatsExcessiveCollisions;

  	/* Collected from emac_tx_stat_dot3statsexcessivecollisions. */
  	/* This is the number of packets that have been dropped due to having 16 collisions

  in a row. */

  u32_t Dot3StatsLateCollisions;

  	/* Collected from emac_tx_stat_dot3statslatecollisions. */
  	/* This is the number of packets that have been dropped due to having a

  collision received after the 64-byte collision window. */

  u32_t EtherStatsCollisions;

  	/* Collected from emac_tx_stat_etherstatscollisions. */
  	/* This is the number of collisions that have been detected on the interface. */

  u32_t EtherStatsFragments;

  	/* Collected from emac_rx_stat_etherstatsfragments. */
  	/* This is the count of frames less than 64 bytes with bad FCS. */

  u32_t EtherStatsJabbers;

  	/* Collected from emac_rx_stat_etherstatsjabbers. */
  	/* This is the number of frames received that exceed the programmed MTU

  size and have bad FCS. */

  u32_t EtherStatsUndersizePkts;

  	/* Collected from emac_rx_stat_etherstatsundersizepkts. */
  	/* This is the number of frames received that are less than 64 bytes in length. */

  u32_t EtherStatsOverrsizePkts;

  	/* Collected from emac_rx_stat_dot3statsframestoolong. */
  	/* This is the number of frames received that exceed the programmed MTU size. */

  u32_t EtherStatsPktsRx64Octets;

  	/* Collected form emac_rx_stat_etherstatspkts64octets. */
  	/* This is the number of good frames received of 64 bytes in size. */

  u32_t EtherStatsPktsRx65Octetsto127Octets;

  	/* Collected from emac_rx_stat_etherstatspkts65octetsto127octets. */
  	/* This is the number of good frames received of 65 bytes to 127 bytes in size. */

  u32_t EtherStatsPktsRx128Octetsto255Octets;

  	/* Collected from emac_rx_stat_etherstatspkts128octetsto255octets. */
  	/* This is the number of good frames received of 128 bytes to 255 bytes in size. */

  u32_t EtherStatsPktsRx256Octetsto511Octets;

  	/* Collected from emac_rx_stat_etherstatspkts256octetsto511octets. */
  	/* This is the number of good frames received of 256 bytes to 511 bytes in size. */

  u32_t EtherStatsPktsRx512Octetsto1023Octets;

  	/* Collected from emac_rx_stat_etherstatspkts512octetsto1023octets. */
  	/* This is the number of good frames received of 512 bytes to 1023 bytes in size. */

  u32_t EtherStatsPktsRx1024Octetsto1522Octets;

  	/* Collected from emac_rx_stat_etherstatspkts1024octetsto1522octets. */
  	/* This is the number of good frames received of 1024 bytes to 1522 bytes in size. */

  u32_t EtherStatsPktsRx1523Octetsto9022Octets;

  	/* Collected from emac_rx_stat_etherstatspkts1523octetsto9022octets. */
  	/* This is the number of good frames received of 1523 bytes to 9022 bytes in size. */

  u32_t EtherStatsPktsTx64Octets;

  	/* Collected form emac_tx_stat_etherstatspkts64octets. */
  	/* This is the number of good frames transmitted of 64 bytes in size. */

  u32_t EtherStatsPktsTx65Octetsto127Octets;

  	/* Collected from emac_tx_stat_etherstatspkts65octetsto127octets. */
  	/* This is the number of good frames transmitted of 65 bytes to 127 bytes in size. */

  u32_t EtherStatsPktsTx128Octetsto255Octets;

  	/* Collected from emac_tx_stat_etherstatspkts128octetsto255octets. */
  	/* This is the number of good frames transmitted of 128 bytes to 255 bytes in size. */

  u32_t EtherStatsPktsTx256Octetsto511Octets;

  	/* Collected from emac_tx_stat_etherstatspkts256octetsto511octets. */
  	/* This is the number of good frames transmitted of 256 bytes to 511 bytes in size. */

  u32_t EtherStatsPktsTx512Octetsto1023Octets;

  	/* Collected from emac_tx_stat_etherstatspkts512octetsto1023octets. */
  	/* This is the number of good frames transmitted of 512 bytes to 1023 bytes in size. */

  u32_t EtherStatsPktsTx1024Octetsto1522Octets;

  	/* Collected from emac_tx_stat_etherstatspkts1024octetsto1522octets. */
  	/* This is the number of good frames transmitted of 1024 bytes to 1522 bytes in size. */

  u32_t EtherStatsPktsTx1523Octetsto9022Octets;

  	/* Collected from emac_tx_stat_etherstatspkts1523octetsto9022octets. */
  	/* This is the number of good frames transmitted of 1523 bytes to 9022 bytes in size. */

  u32_t XonPauseFramesReceived;

  	/* Collected from emac_rx_stat_xonpauseframesreceived. */
  	/* This is the number of good MAC control frames received of pause type with

  a back-off value of zero. This register increments regardless of flow control state. */

  u32_t XoffPauseFramesReceived;

  	/* Collected from emac_rx_stat_xoffpauseframesreceived. */
  	/* This is the number of good MAC control frames received of pause type with

  	a back-off value other than zero. This register increments regardless of
  	flow control state. */

  u32_t OutXonSent;

  	/* Collected from emac_tx_stat_outxonsent. */
  	/* This is the number of MAC Control pause packets of value 0xffff that

  have be transmitted. */

  u32_t OutXoffSent;

  	/* Collected from emac_tx_stat_outxoffsent. */
  	/* This is the number of MAC Control pause packets of value 0 that have

  be transmitted. */

  u32_t MacControlFramesReceived;

  	/* Collected from emac_rx_stat_maccontrolframesreceived. */
  	/* This is the number of good MAC control frames received that are not
  	of pause type. */

  
  Error Codes may include, for example, OK or FATAL.
• While the present invention has been described with reference to certain embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the present invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the present invention without departing from its scope. Therefore, it is intended that the present invention not be limited to the particular embodiments disclosed, but that the present invention will include all embodiments falling within the scope of the appended claims.

Claims (2)

1. A communications system, comprising:

a network interface card (NIC); and

a management device coupled to the NIC,

wherein the NIC is adapted to merge communications traffic of the management device with the NIC.

2-56. (canceled)

Images