Based on the main text of e-trace spec, the synchronization packets are not yielded for privilege change (e.g. xret):
But the Figure 2. Instruction delta trace algorithm implies that synchronization packets are generated for privilege change.
It would be better to explicitly include privilege change as a condition of synchronization in main text.
Besides, the the privilege field in format 3 subformat 0 says:
Regarding the case with instruction xret, ecall, ebreak, it is better to indicate the the privilege field holds the new privilege after execution of the instruction.
Based on the main text of e-trace spec, the synchronization packets are not yielded for privilege change (e.g. xret):
But the Figure 2. Instruction delta trace algorithm implies that synchronization packets are generated for privilege change. It would be better to explicitly include privilege change as a condition of synchronization in main text.
Besides, the the privilege field in format 3 subformat 0 says:
Regarding the case with instruction xret, ecall, ebreak, it is better to indicate the the privilege field holds the new privilege after execution of the instruction.