Re: [PATCH 2/2] media: rc: meson-ir: allow the receiver to wake the system

[Christain Stefan Köver-Draxl]

On 11/2/26 00:01, Martin Blumenstingl wrote:
> Can you help me out with better understanding: is the firmware
> involved in suspend or is this CPU core idle state suspend (meaning:
> firmware is not involved)?

Hi Martin,

yes the commit message is a bit misleading.

For suspend to RAM ("mem") the firmware is involved: Linux calls PSCI 
SYSTEM_SUSPEND, BL31 puts the DDR into self-refresh and turns vdd_cpu 
off, and the always-on firmware runs its own IR decoder. It wakes the 
board only on the remote's power key code and reports the wake reason on 
the serial console. That wake does not need this patch at all.

The patch is for s2idle ("freeze"), where the firmware is not involved 
and only Linux interrupts can wake the CPUs. There the meson-ir 
interrupt has to be a wake interrupt and the decoder has to stay 
powered, which is what the patch does.

I tested it with suspend to RAM, which proved nothing for this patch. I 
will test s2idle with and without the patch and, if it holds up, send a 
v2 that says it is about s2idle; otherwise I will drop it.

Kind regards,
Christian K.-D.