ig4(4): Add an EMAG device type

[gallatin]

Summary

Sponsored by: Ampere Computing LLC, Netflix

[allanjude]

sys/dev/ichiic/ig4_acpi.c

92–100

In the original review, Warner asks: https://reviews.freebsd.org/D28707#inline-178924

Is there some way this can be part of a table of hids to matchs? that makes the device play nicer with devmatch.

[manu]

I agree with Warner here, having a table would be nicer.

[wulf]

In D28746#643683, @manu wrote:

I agree with Warner here, having a table would be nicer.

There is no devmatch info in ig4_acpi.c at all for now. It can be added with placing of following line at the bottom of ig4_acpi.c

ACPI_PNP_INFO(ig4iic_ids);

Along with checking of ACPI_ID_PROBE output instead of extra acpi_MatchHid() call, it will make devmatch happier.

sys/dev/ichiic/ig4_acpi.c

92–100

ACPI_ID_PROBE returns a pointer to matched HID/CID in ig4iic_ids array in fouth parameter. It should be used instead of acpi_MatchHid() to detect "APMC0D0F" device.

[gallatin]

• Tried to address @wulf 's feedback

[imp]

sys/dev/ichiic/ig4_acpi.c

92–100

I'm OK with this change... but if we start to have more, we should move to something table driven.
But the table driven thing doesn't appear to have been implemented yet...