[coreboot-gerrit] [S] Change in coreboot[main]: soc/intel/cmn/pmc: Add support for early power off

[Angel Pons (Code Review)]

Attention is currently required from: Intel coreboot Reviewers, Julius Werner, Karthik Ramasubramanian, Subrata Banik.

Angel Pons has posted comments on this change by Subrata Banik. ( https://review.coreboot.org/c/coreboot/+/86336?usp=email )

Change subject: soc/intel/cmn/pmc: Add support for early power off
......................................................................

Patch Set 3:

(1 comment)

File src/soc/intel/common/block/pmc/pmclib.c:

https://review.coreboot.org/c/coreboot/+/86336/comment/58cfaed8_f38cb763?usp... :
PS3, Line 635: __weak void platform_do_early_poweroff(void)
> > > I'm unable to follow what is **inappropriate** here. […]

How about guarding `platform_do_early_poweroff()` behind a generic `PLATFORM_HAS_EARLY_POWEROFF` Kconfig option? The idea is that whoever selects this Kconfig has to provide a definition of `platform_do_early_poweroff()`, which could be mainboard code, EC code or whatever.

```
void poweroff(void)
{
    if (!ENV_ROMSTAGE_OR_BEFORE) {
    	pmc_control_poweroff();
    } else if (CONFIG(PLATFORM_HAS_EARLY_POWEROFF)) {
    	platform_do_early_poweroff();
    } else {
    	printk(BIOS_EMERG, "This platform doesn't know how to power off before ramstage, hanging!\n");
    	halt();
    }
}
```
As for making the chipset behave, I know I was able to power off Haswell during romstage (before raminit), but I had to perform a few extra steps for it to work. I don't remember where I put that code but I can look for it tomorrow (it's part of an interactive command-line I use for testing, I never published it).