Feature Summary - Feature Summary - 1.0 English - PG453

AXI Switch v1.0 LogiCORE IP Product Guide (PG453)

Document ID
PG453
Release Date
2025-11-20
Version
1.0 English
  • Up to 16 Slave Interfaces (SI) and up to 16 Master Interfaces (MI) per instance.
    • Instances of AXI Switch can be cascaded to interconnect a larger number of masters/slaves or for organizing the interconnect topology.
  • AXI Protocol compliant. Each SI and MI of AXI Switch can be configured to be AXI3, AXI4, or AXI4-Lite.
    • Transactions between interfaces of different protocol types are automatically converted by AXI Switch.
    • Burst transactions are automatically split, as required, to remain AXI compliant.
  • Interface Data Widths:
    • AXI4 and AXI3: 32, 64, 128, 256, 512, or 1024 bits.
    • AXI4-Lite: 32 bits.
  • Transactions between interfaces of different data widths are automatically converted by AXI Switch.
  • Supports Multiple Clock Domains:
    • Transactions between interfaces in different clock domains are automatically converted by AXI Switch.
    • Both asynchronous and synchronous (edge-aligned) conversions are supported, per interface.
  • Address Width: Up to 64 bits.
    • AXI Switch decodes up to 64 total address range segments.
    • Supports address ranges beginning and ending on any 4-byte word alignment (non-power-of-2 addressing)
  • User defined signals up to 512 bits wide per channel.
    • User signals on any AXI channel are propagated regardless of internal transaction conversions.
  • Supports read-only and write-only SI and MI.
  • Supports multiple outstanding transactions:
    • Supports connected masters with multiple reordering depth (ID threads).
    • Supports write response reordering, read data reordering, and read data interleaving.
    • Multi-Threaded traffic (masters issuing multiple ID threads) is supported regardless of internal transaction conversions, including data width conversion and transaction splitting.
    • Optionally produces varying ID signals at the MI to allow connected slaves to respond out-of-order (MI multi-threading).
  • Single-Slave per ID method of cyclic dependency (deadlock) avoidance.
    • For each ID thread issued by a connected master, the AXI Switch allows one or more outstanding transactions to only one slave device for writes and one slave device for reads, at a time.
  • Multiple parallel pathways along all AXI channels when connected to multiple masters and multiple slaves:
    • Each AXI channel has independent destination-side arbitration. Transfers from two or more source endpoints to separate destination endpoints can occur concurrently, for any AXI channel.
    • Round-Robin arbitration for each of the AW, AR, R, and B channels. W-channel transfers follow the same order as AW-channel arbitration, per AXI protocol rules.
  • Supports back-to-back transfers (100% duty cycle) on any AXI channel.
    • Single data-beat transactions can traverse the AXI Switch at the same bandwidth as multi-beat bursts.
  • Supports TrustZone security for each connected slave:
    • If configured as a secure address segment, only secure AXI accesses are permitted according to the AXI arprot or awprot signal.
    • Any non-secure accesses are blocked and the AXI Switch core issues a DECERR response at the SI.
  • Internally resynchronized reset:
    • One aresetn input per IP.