Conversation

Jarkko Sakkinen

Edited 1 year ago

I nailed the Keystone SATP change bug on CVA6 after only 1.5 months of constant debugging:

#riscv #keystone

1
0
1
On positive side, this was the best possible RISC-V assembly tutorial...
1
0
1

Jarkko Sakkinen

Edited 1 year ago

75% time went to exclusion, i.e. finding that the bug is in the enclave itself, not in kernel, SDK or the security monitor. Last two weeks was peeking and poking with loader.S. I even wrote a new driver for Keystone in the process so that I could exclude the driver (OTT driver is, well, I don’t want to say anything impolite…).

1
0
1

Jarkko Sakkinen

Edited 1 year ago
Well, if I count total time that went to SGX, it was 8 years (I started on it in 2013). So not as bad as SGX :-) SGX took long partly because of required hardware changes (flexible launch control).
0
0
0