Root cause of G7 silence: the ACPI dump added in the previous commit confirmed RTL5682 (headset codec) and MX98360A (speaker amp) are both declared on the ACPI bus (status=15, enabled+functional+visible), and SOF's JSL match table correctly selected them → jsl_rt5682_def platform device registered cleanly. But the sof_rt5682 machine driver probe was silently deferring forever waiting for the RT5682 codec to appear on its I2C bus — and that bus didn't exist, because the Intel LPSS I2C controllers (the ones at 00:15.0-3 etc that were logging "Failed to create debugfs entries") had no I2C master driver to bind to. Turn on CONFIG_I2C_DESIGNWARE_{CORE,PLATFORM,PCI,BAYTRAIL}=y so the LPSS I2C controllers enumerate their buses, ACPI client enumeration finds RTL5682 there, the RT5682 codec driver binds, and the sof_rt5682 machine driver can finally build its card. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>