[plbossart]

Somehow we've added features that get in the way of a top-level override with kernel parameters or DMI quirks.

This will be helpful to solve the PDM1 issue and e.g. use

options snd-sof-pci tplg_filename=sof-hda-generic-2ch-pdm1.tplg

[aiChaoSONG]

if dmic_topology_fixup is called with fixup==false, this means we called a function named ...fixup... but actually it fixup nothing, this looks unfamiliar to me.
I am more comfortable with the on demand calling to dmic_topology_fixup, but we lose dev_info(sdev->dev, "DMICs detected in NHLT tables: %d\n", dmic_num); information if tplg_name desn't need fixup.

I don't have better idea, so approve

[plbossart]

if dmic_topology_fixup is called with fixup==false, this means we called a function named ...fixup... but actually it fixup nothing, this looks unfamiliar to me. I am more comfortable with the on demand calling to dmic_topology_fixup, but we lose dev_info(sdev->dev, "DMICs detected in NHLT tables: %d\n", dmic_num); information if tplg_name desn't need fixup.

I don't have better idea, so approve

Actually no we don't lose the dev_info message. I renamed the function dmic_detect_topology_fixup() with the second part optional and gated by a parameter value. Hope this helps.

[plbossart]

This will need to be rebased and merged after #3730 is in topic/sof-dev. There's a conflict with the dmic handling.