[coreboot] TPM device has gone missing
Aaron Durbin
adurbin at google.com
Mon Dec 12 23:50:52 CET 2016
On Mon, Dec 12, 2016 at 4:25 PM, Trammell Hudson <hudson at trmm.net> wrote:
> On Mon, Dec 12, 2016 at 03:08:53PM -0600, Aaron Durbin via coreboot wrote:
>> What about the SSDT? With the patch I think the device is in the SSDT
>> -- not DSDT.
>
> Whups, forgot to include it. There is far less change:
So it looks like the acpi TPM device generation didn't run at all.
What I think you need to make this work is a acpi_name() callback that
returns the correct name for the device passed in. See
src/soc/intel/skylake/acpi.c for matching names in ASL with the ACPI
generator. "LPCB" is what you need to return for the LPC device. The
binding of that callback is a little tricky because of the seemingly
independent southbridge and northbridge code. The easiest way might be
to add two acpi_name() callbacks to the device_operations:
1. src/northbridge/intel/sandybridge/northbridge.c that handles the
dev->path.type == DEVICE_PATH_DOMAIN case and return "PCI0"
2. src/southbridge/intel/bd82x6x/lpc.c that returns LPCB.
>
> --- ./no-tpm/SSDT.dsl 2016-12-12 17:23:51.314355365 -0500
> +++ ./yes-tpm/SSDT.dsl 2016-12-12 17:24:08.694610389 -0500
> @@ -5,13 +5,13 @@
> *
> * Disassembling to symbolic ASL+ operators
> *
> - * Disassembly of SSDT, Mon Dec 12 17:23:51 2016
> + * Disassembly of SSDT, Mon Dec 12 17:24:08 2016
> *
> * Original Table Header:
> * Signature "SSDT"
> * Length 0x0000170D (5901)
> * Revision 0x02
> - * Checksum 0xB4
> + * Checksum 0x83
> * OEM ID "CORE "
> * OEM Table ID "COREBOOT"
> * OEM Revision 0x0000002A (42)
> @@ -73,7 +73,7 @@
> Name (_CRS, ResourceTemplate () // _CRS: Current Resource Settings
> {
> Memory32Fixed (ReadOnly,
> - 0xBFEED000, // Address Base
> + 0xBFEF0000, // Address Base
> 0x00008000, // Address Length
> )
> })
>
> --
> Trammell
More information about the coreboot
mailing list