Re: [PATCH v3 2/8] keyhandler: drop regs parameter from handle_keyregs()

[Julien Grall <julien@xxxxxxx>]

Hi Jan,

On 05/02/2024 13:28, Jan Beulich wrote:
> In preparation for further removal of regs parameters, drop it here. In
> the two places where it's actually needed, retrieve IRQ context if
> available, or else guest context.
>
> Suggested-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> ---
> As an alternative to the new boolean parameter, I wonder if we couldn't
> special-case the idle vCPU case: It's only there where we would not have
> proper context retrievable via guest_cpu_user_regs().
I am trying to understand the implication. Looking at the code, it seems 
in the case where we pass NULL, we would expect to call 
run_in_exception_handler().
If I am not mistaken, at least for Arm, regs would not be the same as 
guest_cpu_user_regs(). So I think your current approach is more correct.
Did I miss anything?

Cheers,

--
Julien Grall