Palutena (Pirika/JSL): no SOF _DSD (speaker-tplg/hp-tplg) in ACPI → Windows SOF driver only exposes Headphones, internal speakers missing

System Details

  • Device: Acer Chromebook 314 CBOA314-1H
  • Board: Palutena (baseboard Pirika rev2, SMBIOS family Google_Dedede), Intel Jasper Lake Celeron N4500
  • Firmware: UEFI Full ROM, MrChromebox-2609.1 (10/01/2026)
  • OS: Windows 11 Pro x64 (build 26300)
  • Audio: Intel SOF (PCI\VEN_8086&DEV_4DC8), codec Realtek RT5650 (ACPI\10EC5650, _SB.PCI0.I2C4.RT58)
  • Windows drivers: CoolStar HD Audio, CoolStar SOF Audio (WDM) 1.0.5.0, Realtek ALC5650 I2S 1.0.1.0

Summary

In Windows, the only internal playback endpoint is “Headphones (CoolStar SOF Audio)”. No Speakers endpoint is ever created. All audio devices report OK in Device Manager. The internal speakers are physically fine (verified in Linux, see below).

Evidence

1. ACPI has no SOF topology properties

Dumped and decompiled DSDT + SSDT from Ubuntu Live (raw and .dsl attached in acpi/): there is no speaker-tplg, hp-tplg or mic-tplg string anywhere. \_SB.PCI0.HDAS (DSDT.dsl, line 1752) only has _ADR, _DDN, _S0W and the NHLT _DSM; there is no _DSD, and the SSDT has no Scope (\_SB.PCI0.HDAS) at all. The only audio-codec node is:

Device (RT58)
{
    Name (_HID, "10EC5650")
    Name (_UID, Zero)
    Name (_DDN, "Realtek RT5650")
    Method (_STA, 0, NotSerialized) { Return (0x0F) }
}

No amplifier node (10EC1015 / RTL1015 / MX98360A) exists, which matches the hardware: the speakers are driven by the RT5650’s internal ClassD amp.
(See salidas/05_* and salidas/07_*.)

2. Hardware path confirmed working in Linux (Ubuntu Live, kernel 7.0.0-30)

  • Card sof-rt5650, machine driver jsl_rt5682_def, topology sof-jsl-rt5650.tplg.
  • PCM 0 “Speakers” → SSP1 → rt5645-aif2. PCM 1 “Headset” → rt5645-aif1.
  • There is no UCM for this card (PipeWire uses stereo-fallback), and DAC1 L/R Mux defaults to IF1 DAC, so the speaker path never closes and DAPM powers nothing during playback (only JD Power / LDO2, both forced). See salidas/04_* and salidas/05_*.
  • After manually setting DAC1 L/R Mux = IF2 DAC and enabling DAC1 MIX → Stereo DAC MIX → SPK MIX → SPKVOL → SPOL/SPOR MIX, DAPM powers AIF2RX → IF2 DAC → DAC L1/R1 → SPK MIX → SPKVOL → SPOL/SPOR → SPK amp → Left/Right Spk. speaker-test -D plughw:0,0 is then clearly audible on both internal speakers. See salidas/06_* and scripts/script6.sh.

3. Windows symptom

Only a Headphones endpoint exists. Plugging or unplugging 3.5 mm headphones does not change the endpoint list. The internal mic works.

Analysis

coreboot added rt5650_sp / rt5650_hp to drivers/sof (commit 7f53e11425) precisely so the Windows SOF drivers can identify RT5650 boards that use the codec for speaker and/or headset output. Palutena’s ACPI has no drivers/sof _DSD at all. Without it, the Windows driver apparently only knows about the headset path, so it never creates the speaker endpoint (SSP1/AIF2).

Suggested fix

Add a drivers/sof chip under the HDA device in the Pirika variant devicetree (or for the Palutena SKU), e.g.:

device ref hda on
    chip drivers/sof
        register "spkr_tplg" = "rt5650_sp"
        register "jack_tplg" = "rt5650_hp"
        register "mic_tplg" = "_2ch_pdm0"    # please verify the DMIC config
        device generic 0 on end
    end
end

If Pirika ships audio SKUs other than ALC5650 (fw_config AUDIO_AMP), this may need a fw_config probe.

I am happy to test a build. I can provide the full ACPI dump and Linux logs.

Attachments

  • salidas/01–07: full console output of each diagnostic step (each file notes whether sound was audible).
  • scripts/script1.sh–script7.sh: the exact commands that produced each output.

acpi/DSDT, acpi/SSDT: raw tables from /sys/firmware/acpi/tables/; acpi/DSDT.dsl, acpi/SSDT.dsl: iasl disassembly.

Logs & ACPI dump (zip):

Contents: full console output of each diagnostic step (salidas/01–07), the exact scripts used (scripts/), raw DSDT/SSDT and their iasl disassembly (acpi/).

This should fix the issue

coreboot_edk2-pirika-mrchromebox_20261118.zip (4.4 MB)

1 Like

Confirmed working, thank you! After flashing the 20261118 build, Windows now shows both Speakers and Headphones endpoints, and audio plays correctly through the internal speakers and the 3.5 mm jack.

1 Like