[PATCH v2 1/3] firmware: arm_ffa: Split the response out of ffa_msg_send_direct_req2()

[Jamie Nguyen]

> The kernel document was updated for this, please update the
> rest of two patches as well.

I'll update all three patches to use:

    Assisted-by: LLM

> As you noted in the cover letter, the AML thread will stall here if no
> response, block the ACPI interpreter and the ACPI global lock when
> the field is Lock-flagged, so do we have some warnings if that happened?

There's no warning today. Would a rate-limited warning after a prolonged
wait be useful here? I'm hesitant because that would also affect non-ACPI
callers of this helper, and I don't believe FF-A defines a threshold after
which the wait is abnormal.

> resp_regs[1] is confusing, comment here it links to X1 register
> is better.

I'll clarify that resp_regs[1] is X2; resp_regs[0] is X1.

I'll address these in v3.  Thanks!