Closed nearly-big-endian closed 1 year ago
Any update on this?
Hi @nearly-big-endian,
Thanks for providing the logs. There is known issue when using Boot Mode 3 (Factory Secure Boot) with applications targeting eNVM + L2-scratchpad. We are currently working on this. We'll provide a fix as soon as we complete our testing.
Hi @nearly-big-endian ,
This issue was fixed on 2023.02 release. Could you please update your icicle kit with the latest HSS?
Thanks
Hello @vfalanis,
I am happy to report that HSS v0.99.35 now boots as expected in Secure Boot mode 3 on an IcicleKit where all versions of HSS / Yocto BSP / Ref. Design are aligned to the 2023.02 release.
Successfully tested with the HSS programmed through SoftConsole+fpgenprog, and through a LiberoSoC v2022.3 project (in a Boot Mode 3 eNVM Client).
So, thanks and congrats for the good work 👍
Just another question : is the fix relatively easy to back-port to HSS 2022.09 (matter of merging a couple of commits only) or is it too deeply nested with other changes to be a reasonable idea ?
Thanks !
Hi @nearly-big-endian
There are couple of updates (HAL version and some other commits related to cache flushing) required that are split into more than one commit, normally HAL updates involving several files are not that straightforward to back-port. We usually recommend to update to the latest release.
Is there any reason why you need 2022.09 specifically? We strongly recommend to update to latest release as there have been new features and improvements since then.
Thanks
Hey @nearly-big-endian, I'm closing this issue since it has been fixed in 2023.02 release, feel free to open it back up if you need any further clarification :)
Hi,
We are currently using latest HSS (0.99.33, release 2022.10) together with SoftConsole 2022.2 on an IcicleKit. When programmed in Non Secure Boot Mode 1 (via SoftConsole), HSS boots successfully and proceeds with the load of u-boot then our Linux image.
When programming the same HSS binary in Factory Secure Boot Mode 3 (still through SoftConsole), HSS hangs at the Memory testing stage (as logs show below).
The root cause does not seem to be the DDR testing itself though, as disabling CONFIG_MEMTEST just leads the HSS to hang at a later stage. It's more looking like a memory mapping issue.
Any idea what is going on ?
Some additional notes:
HSS Boot log in Non-Secure Boot Mode 1 (succeeds):
HSS Boot log in Factory-Secure Boot Mode 3 (fails):
mpfsBootmodeProgrammer boot log (for boot mode 3):
Thanks for your inputs.