On 10/10/2019 20:46, Varun Wadekar wrote:

Hello,

First of all, thanks a lot for posting the Hercules-AE patches.

We picked them up and used them internally. Unfortunately, the CPU sees garbage in it's I$ when we enable the I cache for the processor. If we keep I$ disabled, TF-A boots properly. We are using TF-A v1.4 for verification.

We booted Linux kernel v4.14 on the processor and don't see this problem there. So, we suspect something going wrong inside TF-A. Have you seen this problem internally? Any hints or clues to solve it would be helpful.

Thanks.

Hi Varun, As indicated in the commit message of the patch, we have not tested the CPU support internally due to non-availability of FVP for the CPU.

TF-A assumes that I$ will be invalidated when CPU is reset. Perhaps this is not true for your setup. Could you try adding a `IC IALLU` in the CPU reset handler prior to I$ enable and see if that improves anything ?

Best Regards Soby Mathew

---

This email message is for the sole use of the intended recipient(s) and may contain confidential information.  Any unauthorized review, use, disclosure or distribution is prohibited.  If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.

---