This never appeared until now because it only failed if two records appeared at once.

Debugging with bad information sucks!

  Don't overflow the buffer! Compare the index for overflow not the position.
  Read/write of 0 returns 0 even when non-blocking.

This would break the read (non-posted) and write (posted) ordering that Wishbone depends on.

Currently hardcode SDWB address as hardware does not support it.

Fixed bugs in kernel etherbone:
  record_len was shadowed by redeclation
  wrote maximum of possible bytes instead of minimum
  added non-blocking IO support

Bug: std_match is needed when comparing '-'s

stub wishbone device added for testing

Implement a wishbone streaming interface bridge from the Avalon RX stream

Add driver stub code

Modified PLL for two outputs
Added power on reset
Re-instantiated the PCIe megafunction