Visible to Intel only — GUID: suc1462793462915
Ixiasoft
Visible to Intel only — GUID: suc1462793462915
Ixiasoft
A.7.2.1. Handling an FPGA Configuration Failure after Early I/O Release
If an FPGA configuration failure occurs after early I/O release, the HPS shared I/O and hard memory controller I/O enter input tri-state mode. In this situation, any accesses to peripherals or memory interfacing these I/O fails. HPS dedicated I/O are not affected by FPGA configuration failures.
You must implement one of the following solutions in software.
- To prevent FPGA configuration failure, load the FPGA core configuration file into SDRAM and perform an FPGA integrity check before initiating configuration. For example, you can run a custom integrity check, such as a CRC, on the .rbf file to validate that there are no errors in the file. Intel also recommends that you enable ECC capabilities to avoid bit-stream corruption issues.
-
To recover from an FPGA configuration failure, execute the HPS-initiated configuration from on-chip RAM. If the HPS is executing entirely from on-chip RAM, you can recover your application by reconfiguring the peripheral .rbf file to reenable the HPS shared I/O and hard memory controller I/O. After these I/O are recovered, your application can either attempt to configure the same FPGA image that caused the original failure or use a fallback core .rbf file to prevent the same failure from occurring repeatedly.