Hi Linus
I'm working on gpio for an AMD Family 16h Model 30h system[1]. The SoC is the
same as the GX412-TC used in the PC Engines APU2.
There is an out-of-tree gpio driver (gpio-amd) for this SoC in the meta-amd
yocto layer[2].
Another driver (gpio-sb8xx) was submitted for upstream inclusion, but was
knocked back with the suggestion that pinctrl is the way forward[3].
I would much prefer to use a mainline driver for the system I'm working on, so
I'm looking at the pinctrl-amd driver to see whether it applies to our SoC, or
whether it could be extended, or used as starting point for a new driver.
The out-of-tree drivers apply to the GX412-TC SoC and uses PCI for probing:
gpio-amd registers a platform driver that applies to { PCI_VENDOR_ID_AMD,
PCI_DEVICE_ID_AMD_HUDSON2_SMBUS }
gpio-sb8xx applies to { PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_HUDSON2_SMBUS }
and { PCI_DEVICE(PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS }
These IDs make it easy to cross reference with the datasheet, and keeps the
coupling between ACPI and the driver low.
These drivers do not provide a mechanism for firmware (ACPI or DT) to specify
which gpios are safe to use or how to use them.
In contrast, the pinctrl-amd driver only mentions the newer KERNCZ platform
name and uses ACPI for probing without disclosing any Family or Model numbers.
pinctrl-amd applies to "AMD0030" and "AMDI0030"
The ACPI HID matching makes it difficult to determine what family and model the
driver applies to, or rather, I have not been able to find such a mapping of
HIDs to family and model numbers. It's also impossible to guess an ACPI _HID
that may or may not exist for the Family 16h Model 30h platform and even if I
allocate a new HID for our ACPI implementation, that HID has little hope of
being accepted into the mainline driver.
I would like to extend the generically named, but very specifically implemented
pinctrl-amd driver to also work on Family 16h, Model 30h (specifically the FT3b
package), and I propose to use the PCI_DEVICE_ID_AMD_16H_M30H_NB_F3 symbol to
probe for the device.
Does this seem like a sensible way forward?
Thanks!
Andrew
1. http://support.amd.com/TechDocs/52740_16h_Models_30h-3Fh_BKDG.pdf
2.
http://git.yoctoproject.org/cgit/cgit.cgi/meta-amd/tree/meta-steppeeagle/recipes-kernel/amd-gpio/files/gpio-amd.c
3. https://lkml.org/lkml/2015/6/20/202
