Changes between Version 3 and Version 4 of nic_device_api


Ignore:
Timestamp:
Jan 26, 2020, 10:55:01 PM (5 years ago)
Author:
alain
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • nic_device_api

    v3 v4  
    66
    77
    8 This device provide access to an external generic Gigabit Ethernet network controller. It assume that the NIC hardware peripheral has a DMA capability, and can access two packets queues in kernel memory for sent (TX) and received (RX) packets. Packets are (Ethernet/IPV4).
     8This device allows the kernel to access an external generic Gigabit Ethernet network controller. It assume that the NIC hardware peripheral has a DMA capability, and can access two packets queues in kernel memory for sent (TX) and received (RX) packets. Packets are Ethernet/IPV4.
    99
    10 The NIC device is handling two (infinite) streams of packets to or from the network. It is the driver responsibility to move the RX packets from the NIC to the RX queue, and the TX packets from the TX queue to the NIC.
     10The NIC device is handling two (infinite) streams of packets to or from the network. It is the driver responsibility to move the RX packets from the hardware NIC to the software RX queue, and to move the TX packets from the software TX queue to the hardware NIC.
    1111
    12 AS the RX and TX queues are independant, there is one NIC-RX device descriptor to handle RX packets, and another NIC-TX device descriptor to handle TX packets. In order to improve throughput, the hardware NIC controller can (optionnally) implements multiple (N) channels:
     12AS the RX and TX queues are independant, there is one NIC-RX chdev descriptor to handle RX packets, and another NIC-TX chdev descriptor to handle TX packets. In order to improve throughput, the hardware NIC controller can (optionnally) implements multiple (N) channels. To share the load between channels, the hardware is supposed to use an hash key :
    1313 * The RX channels are indexed by an hash key derived from the source IP address.
    1414 * The TX channels are indexed by an hash key derived from the destination IP address.
    15 These 2*N chdev descriptors, and 2*N associated server threads, are distributed in  clusters.
    16 The  2*N server threads implement the protocols stack. The RX server threads block and deschedule when the RX queue is empty. The TX server threads block and deschedule when the TX queue is full.
     15These 2*N chdev descriptors, and 2*N associated server threads, are distributed in clusters.
    1716
    18 It is the driver responsibily to re-activate a blocked server thread when the queue state is modified: not full for TX, or not empty for RX.
     17The  2*N server threads implement the protocols stack.
    1918
    20 The WTI mailboxes used by the driver ro receive events from the NIC hardware (available RX packet, or free TX slot, for a given channel), must be statically allocated during the kernel initialisation phase, and must be routed to the cluster containing the associated device descriptor and server thread.
     19The RX server threads block and deschedule when the RX queue is empty. The TX server threads block and deschedule when the TX queue is full. It is the driver responsibily to re-activate a blocked server thread when the queue state is modified: not full for TX, or not empty for RX.
     20
     21The WTI mailboxes used to receive the NIC_TX_IRQ[N] and NIC_RX_IRQ[N] (one IRQ per channel and per direction) to signal available RX packet, or free TX slot, for a given channel, must be statically allocated during the kernel initialisation phase. They are routed to the cluster containing the associated chdev descriptor and The associated server thread.
    2122
    2223The generic NIC device "kernel" API defines two functions:
    … …  
    2526This "kernel" API is detailed in section C below.
    2627
    27 All RX or TX paquets are sent or received in standard 2 Kbytes kernel buffers, that are dynamically allocated by the server threads. The structure ''pkd_t'' defining a packet descriptor contains the buffer pointer and the actual Ethernet packet length in bytes.
     28All RX or TX paquets are sent or received in standard 2 Kbytes kernel buffers, that are dynamically allocated by the server threads. The structure ''pkd_t'' defining a packet descriptor contains the buffer pointer and the actual Ethernet packet length (in bytes).
    2829
    29 The actual TX an RX queues structures depends on the hardware NIC implementation, and are defined by the driver code.
     30The actual TX an RX queues structures depends on the hardware NIC implementation, and are only accessed by the driver functions.
    3031
    3132To access the drivers, the NIC device defines a lower-level "driver" API, containing four command types, that are detailed in section D below.