i82801gx: SMI from "GPE0_STS: USB1"

[Константин Аладышев]

Hello!

I have some troubles with southbridge i82801gx.

I can boot Windows on my board, but S3 from OS doesn't work properly if i resume with USB keyboard.

In this case, i have Wake signal and coreboot starts to boot, but it stucks with endless messages about SMI# "GPE0_STS: USB1":

..

SMI# #0
GPE0_STS: USB1

SMI# #0
GPE0_STS: USB1

SMI# #0
GPE0_STS: USB1

...

These messages come from file "\southbridge\intel\i82801gx\smihandler.c", function "dump_gpe0_status".

According to i82801gx datasheet:
___

USB1_STS - R/WC. Software clears this bit by writing a 1 to it.
0 = USB UHCI controller 1 does Not need to cause a wake.

1 = Set by hardware when USB UHCI controller 1 needs to cause a wake. Wake event

will be generated if the corresponding USB1_EN bit is set.

___

Clearing of GPE0_STS bits properly done in function reset_gpe0_status

/**
 * @brief read and clear GPE0_STS
 * @return GPE0_STS register
 */
static u32 reset_gpe0_status(void)
{
 u32 reg32;

reg32 = inl(pmbase + GPE0_STS);
 /* set status bits are cleared by writing 1 to them */
 outl(reg32, pmbase + GPE0_STS);

return reg32;
}

So... i can't really understand, why does this SMI# appear again and again?

Thanks,
Aladyshev Konstantin.

[Aaron Durbin]

On Tue, Jun 11, 2013 at 9:33 AM, Константин Аладышев
 wrote:
> Hello!
>
> I have some troubles with southbridge i82801gx.
>
> I can boot Windows on my board, but S3 from OS doesn't work properly if i
> resume with USB keyboard.
>
> In this case, i have Wake signal and coreboot starts to boot, but it stucks
> with endless messages about SMI# "GPE0_STS: USB1":
>
> ...
>
> SMI# #0
> GPE0_STS: USB1
>
> SMI# #0
> GPE0_STS: USB1
>
> SMI# #0
> GPE0_STS: USB1
>
> ....
>
> These messages come from file "\southbridge\intel\i82801gx\smihandler.c",
> function "dump_gpe0_status".
>
> According to i82801gx datasheet:
> ___
>
> USB1_STS -- R/WC. Software clears this bit by writing a 1 to it.
>
> 0 = USB UHCI controller 1 does Not need to cause a wake.
>
> 1 = Set by hardware when USB UHCI controller 1 needs to cause a wake. Wake
> event
>
> will be generated if the corresponding USB1_EN bit is set.
>
> Clearing of GPE0_STS bits properly done in function reset_gpe0_status
>
> /**
> @brief read and clear GPE0_STS
> @return GPE0_STS register
>
> */
> static u32 reset_gpe0_status(void)
> {
>  u32 reg32;
>
> reg32 = inl(pmbase + GPE0_STS);
>  /* set status bits are cleared by writing 1 to them */
>  outl(reg32, pmbase + GPE0_STS);
>
> return reg32;
> }
>
> So... i can't really understand, why does this SMI# appear again and again?

I assume the USB controller is still asserting its event. You'll need
to clear USB1_EN in GPE0_EN. That should stop it. From what I can tell
in the code GPE0_EN is only messed with when entering S5. So you may
want to ensure that is cleared when setting up SMIs in coreboot.

-Aaron