CS35L41 amplifiers in an ASUS Zenbook on linux

cs35l41_spi.md

Asus Zenbook UX3402 speakers on Linux

Important

THIS IS NOW OBSOLETE WITH KERNEL VERSIONS ≥ 6.7.0

A recent announcement in the kernel mail list by Cirrus developers will solve the problem described here. Therefore, the proposed solutions will be soon obsolete. See this comment (thanks, @flukejones, for the tip).

I got the speakers working on my Asus Zenbook 14 OLED UX3402, the one with Intel CPU and the two CS35L41 audio amplifiers connected over SPI (not the UM3402YA, with AMD and I²C). The amplifiers are supported by the snd_hda_scodec_cs35l41 module in recent kernel versions, but they require some model-specific configuration paramaters, that should be provided by the BIOS, via ACPI. Of course, they are missing. It seems that the Windows drivers have those parameters somehow hardcoded, instead of using ACPI (bad idea, but who am I to criticise). Since Asus has been nasty and refused/was not able to update the BIOS with the missing data, I have patched the ACPI tables myself.

Warning

As stated by Cirrus developers, if you try this there is a non-zero (low, I hope) possibility of blowing your speakers. Take into account that this procedure is not just configuring your box, it is hardware-level hacking with very scarce information about the hardware (thank you, ASUS, and thank you, Cirrus, for not publishing the CS35L41 datasheet and making it only available by request). Additionally, You should know how to work with DSDT or SSDT files (I have a brief guide here) and how to recover from a non booting system. You are on your own. You have been warned.

Note

I have used the patch for some months, and everything seems OK. As you can see, the most recent versions do not require electrical parameters to be specified, just providing the functions of several GPIO pins, that are already present in the original BIOS, and the behavior associated to them. I am almost sure that the patch is totally safe.

My BIOS is the most recent one (version 310). I have added two _DSD objects in the SSDT file (ssdt_csc3551.dsl, or its already assembled version, ssdt_csc3551.aml):

• One inside the SPK1 device, providing the missing "cirrus,dev-index" and other properties, just like the I2C models. You can also provide these properties by patching the kernel module (the linux/sound/pci/hda/cs35l41_hda.c file).

• Another one, after the SPK1 device, with the chip select GPIOs, "cs-gpios", required by the SPI bus (the somehow simpler I²C bus does not use chip select pins, but addresses). Without this _DSD object, only one amplifier can be addressed and activated. In this case, the driver refuses to load even for this single working amplifier.

Obsolete: DSDT patch

Instead of applying a SSDT file, you may patch the full DSDT dissassembled DSDT file (dsdt.dsl). This was my original solution (thanks to @thor2002ro for pointing to the simpler, nicer SSDT solution, see the comment, especially if you use grub). The patch is available for reference, just apply the SSDT file.

Obsolete: internal boost

I keep the old, original patch, that produced AMP short error messages when the volume was high enough, for reference. Please try first the simpler patch with external boost, that seems to work without flaws (thanks to @bdandy, see the comment).

Boot log

Output from journalctl -b -g CSC3551 --output short-monotonic after the old patch:

[5.358062] zenbook kernel: Serial bus multi instantiate pseudo device driver CSC3551:00: Instantiated 2 SPI devices.
[5.643640] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: Cirrus Logic CS35L41 (35a40), Revision: B2
[5.680302] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Cirrus Logic CS35L41 (35a40), Revision: B2
[5.996217] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: Falling back to default firmware.
[5.996608] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: Firmware version: 3
[5.996611] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: cirrus/cs35l41-dsp1-spk-prot.wmfw: Fri 24 Jun 2022 14:55:56 GMT Daylight Time
[6.069598] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: Firmware: 400a4 vendor: 0x2 v0.58.0, 2 algorithms
[6.070173] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: 0: ID cd v29.78.0 XM@94 YM@e
[6.070185] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: 1: ID f20b v0.1.0 XM@17c YM@0
[6.070192] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: spk-prot: e:\workspace\workspace\tibranch_release_playback_6.76_2\ormis\staging\default_tunings\internal\CS35L53\Fixed_Attenuation_Mono_48000_29.78.0\full\Fixed_Attenuation_Mono_48000_29.78.0_full.bin
[6.083485] zenbook kernel: snd_hda_codec_realtek ehdaudio0D0: bound spi1-CSC3551:00-cs35l41-hda.0 (ops cs35l41_hda_comp_ops [snd_hda_scodec_cs35l41])
[6.083899] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Falling back to default firmware.
[6.083962] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: Firmware version: 3
[6.083966] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: cirrus/cs35l41-dsp1-spk-prot.wmfw: Fri 24 Jun 2022 14:55:56 GMT Daylight Time
[6.145224] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: Firmware: 400a4 vendor: 0x2 v0.58.0, 2 algorithms
[6.145817] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: 0: ID cd v29.78.0 XM@94 YM@e
[6.145828] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: 1: ID f20b v0.1.0 XM@17c YM@0
[6.145834] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: spk-prot: e:\workspace\workspace\tibranch_release_playback_6.76_2\ormis\staging\default_tunings\internal\CS35L53\Fixed_Attenuation_Mono_48000_29.78.0\full\Fixed_Attenuation_Mono_48000_29.78.0_full.bin
[6.161484] zenbook kernel: snd_hda_codec_realtek ehdaudio0D0: bound spi1-CSC3551:00-cs35l41-hda.1 (ops cs35l41_hda_comp_ops [snd_hda_scodec_cs35l41])

The "Falling back to default firmware" disappears if spk-id-gpios are provided (the block is commented in the patch). As a result, a different firmware is loaded, resulting in very low volume and strong distortion. If the new patch is used, that uses external boost, the sound is correct, and the log is

[0.004000] zenbook kernel: ACPI: Table Upgrade: install [SSDT-CUSTOM- CSC3551]
[0.004000] zenbook kernel: ACPI: SSDT 0x00000000368FD000 0001A0 (v01 CUSTOM CSC3551  00000001 INTL 20200925)
[5.607253] zenbook kernel: Serial bus multi instantiate pseudo device driver CSC3551:00: Instantiated 2 SPI devices.
[5.906936] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: Cirrus Logic CS35L41 (35a40), Revision: B2
[5.907354] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Reset line busy, assuming shared reset
[5.938136] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Cirrus Logic CS35L41 (35a40), Revision: B2
[6.239406] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: Firmware version: 3
[6.239415] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: cirrus/cs35l41-dsp1-spk-prot-10431e02.wmfw: Fri 27 Aug 2021 14:58:19 W. Europe Daylight Time
[6.313109] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: Firmware: 400a4 vendor: 0x2 v0.43.1, 2 algorithms
[6.313524] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: 0: ID cd v29.63.1 XM@94 YM@e
[6.313530] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: 1: ID f20b v0.1.0 XM@176 YM@0
[6.313535] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: DSP1: spk-prot: C:\Users\tyang\Desktop\Product Setting\SmartAMP\ASUS\ASUS_Zenbook\UX3402\Tuning Release\220304\ASUS_UX3402_L_tuning_IDYC_ReDc_PICL_RTL_0304.bin
[6.337211] zenbook kernel: snd_hda_codec_realtek ehdaudio0D0: bound spi1-CSC3551:00-cs35l41-hda.0 (ops cs35l41_hda_comp_ops [snd_hda_scodec_cs35l41])
[6.337534] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: Firmware version: 3
[6.337540] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: cirrus/cs35l41-dsp1-spk-prot-10431e02.wmfw: Fri 27 Aug 2021 14:58:19 W. Europe Daylight Time
[6.399070] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: Firmware: 400a4 vendor: 0x2 v0.43.1, 2 algorithms
[6.399612] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: 0: ID cd v29.63.1 XM@94 YM@e
[6.399623] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: 1: ID f20b v0.1.0 XM@176 YM@0
[6.399629] zenbook kernel: cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: DSP1: spk-prot: C:\Users\tyang\Desktop\Product Setting\SmartAMP\ASUS\ASUS_Zenbook\UX3402\Tuning Release\220304\ASUS_UX3402_R_tuning_IDYC_ReDc_PICL_RTL_0304.bin
[6.427637] zenbook kernel: snd_hda_codec_realtek ehdaudio0D0: bound spi1-CSC3551:00-cs35l41-hda.1 (ops cs35l41_hda_comp_ops [snd_hda_scodec_cs35l41])

dsdt.dsl-old.patch

--- dsdt.dsl.old	2023-03-26 01:06:48.845795666 +0100
+++ dsdt.dsl	2023-03-26 12:00:45.643851841 +0200
@@ -18,7 +18,7 @@
  *     Compiler ID      "INTL"
  *     Compiler Version 0x20200717 (538969879)
  */
-DefinitionBlock ("", "DSDT", 2, "_ASUS_", "Notebook", 0x01072009)
+DefinitionBlock ("", "DSDT", 2, "_ASUS_", "Notebook", 0x0107200A)
 {
     /*
      * iASL Warning: There were 233 external control methods found during
@@ -90642,7 +90642,48 @@
             Method (_DIS, 0, NotSerialized)  // _DIS: Disable Device
             {
             }
+
+            Name (_DSD, Package ()   // _DSD: Device-Specific Data
+            {
+                ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), 
+                Package ()
+                {
+                    Package () { "cirrus,dev-index", Package () { Zero, One }},
+                    Package () { "reset-gpios", Package () {
+                       SPK1, One, Zero, Zero,
+                       SPK1, One, Zero, Zero,
+                    } },
+                    // Package () { "spk-id-gpios", Package () {
+                    //    SPK1, 0x02, Zero, Zero,
+                    //    SPK1, 0x02, Zero, Zero,
+                    // } },
+                    Package () { "cirrus,speaker-position",     Package () { Zero, One } },
+                    Package () { "cirrus,gpio1-func",           Package () { Zero, Zero } },
+                    Package () { "cirrus,gpio2-func",           Package () { 0x02, 0x02 } },
+                    // boost-type: 0 internal, 1 external
+                    Package () { "cirrus,boost-type",           Package () { Zero, Zero } },
+                    // boost-peak-milliamp: 1600 to 4500, increments of 50 mA
+                    Package () { "cirrus,boost-peak-milliamp",  Package () { 4500, 4500 } },
+                    // boost-ind-nanohenry: 1000 | 1200 | 1500 | 2200 nH
+                    Package () { "cirrus,boost-ind-nanohenry",  Package () { 1000, 1000 } },
+                    // boost-cap-microfarad: rounded to the nearest integer
+                    Package () { "cirrus,boost-cap-microfarad", Package () { 24, 24 } },
+                },
+            })
         }
+
+        Name (_DSD, Package ()
+        {
+            ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), 
+            Package ()
+            {
+                Package () { "cs-gpios", Package () { 
+                    Zero,                    // Native CS
+                    SPK1, Zero, Zero, Zero   // GPIO CS
+                } }
+            }
+        })
+
     }
 
     Scope (_SB.PC00.LPCB)

dsdt.dsl.patch

--- dsdt.dsl.old	2023-03-26 01:06:48.845795666 +0100
+++ dsdt.dsl	2023-03-26 12:00:45.643851841 +0200
@@ -18,7 +18,7 @@
  *     Compiler ID      "INTL"
  *     Compiler Version 0x20200717 (538969879)
  */
-DefinitionBlock ("", "DSDT", 2, "_ASUS_", "Notebook", 0x01072009)
+DefinitionBlock ("", "DSDT", 2, "_ASUS_", "Notebook", 0x0107200A)
 {
     /*
      * iASL Warning: There were 233 external control methods found during
@@ -90642,7 +90642,43 @@
             Method (_DIS, 0, NotSerialized)  // _DIS: Disable Device
             {
             }
+
+            Name (_DSD, Package ()   // _DSD: Device-Specific Data
+            {
+                ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), 
+                Package ()
+                {
+                    Package () { "cirrus,dev-index", Package () { Zero, One }},
+                    Package () { "reset-gpios", Package () {
+                       SPK1, One, Zero, Zero,
+                       SPK1, One, Zero, Zero,
+                    } },
+                    Package () { "spk-id-gpios", Package () {
+                       SPK1, 0x02, Zero, Zero,
+                       SPK1, 0x02, Zero, Zero,
+                    } },
+                    Package () { "cirrus,speaker-position",     Package () { Zero, One } },
+                    // gpioX-func: 0 not used, 1 VPSK_SWITCH, 2: INTERRUPT, 3: SYNC
+                    Package () { "cirrus,gpio1-func",           Package () { One, One } },
+                    Package () { "cirrus,gpio2-func",           Package () { 0x02, 0x02 } },
+                    // boost-type: 0 internal, 1 external
+                    Package () { "cirrus,boost-type",           Package () { One, One } },
+                },
+            })
         }
+
+        Name (_DSD, Package ()
+        {
+            ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), 
+            Package ()
+            {
+                Package () { "cs-gpios", Package () { 
+                    Zero,                    // Native CS
+                    SPK1, Zero, Zero, Zero   // GPIO CS
+                } }
+            }
+        })
+
     }
 
     Scope (_SB.PC00.LPCB)

ssdt_csc3551.aml

ssdt_csc3551.dsl

DefinitionBlock ("", "SSDT", 1, "CUSTOM", "CSC3551", 0x00000001)
{
    External (_SB_.PC00.SPI0, DeviceObj)
    External (_SB_.PC00.SPI0.SPK1, DeviceObj)

    Scope (_SB.PC00.SPI0.SPK1)
    {
        Name (_DSD, Package ()
        {
            ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
            Package ()
            {
                Package () { "cirrus,dev-index", Package () { Zero, One }},
                Package () { "reset-gpios", Package () {
                    SPK1, One, Zero, Zero,
                    SPK1, One, Zero, Zero
                } },
                Package () { "spk-id-gpios", Package () {
                    SPK1, 0x02, Zero, Zero,
                    SPK1, 0x02, Zero, Zero
                } },
                Package () { "cirrus,speaker-position", Package () { Zero, One } },
                // gpioX-func: 0 not used, 1 VPSK_SWITCH, 2: INTERRUPT, 3: SYNC
                Package () { "cirrus,gpio1-func", Package () { One, One } },
                Package () { "cirrus,gpio2-func", Package () { 0x02, 0x02 } },
                // boost-type: 0 internal, 1 external
                Package () { "cirrus,boost-type", Package () { One, One } }
            }
        })
    }

    Scope (_SB.PC00.SPI0)
    {
        Name (_DSD, Package ()
        {
            ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
            Package ()
            {
                Package () { "cs-gpios", Package () { 
                    Zero,                    // Native CS
                    SPK1, Zero, Zero, Zero   // GPIO CS
                } }
            }
        })
    }
}

ASUS Zenbook UM3402YA — CS35L41 right speaker silent after suspend (kernel 6.17.9)

Hardware: ASUS Zenbook UM3402YA, Pop!_OS (Ubuntu-based), kernel 6.17.9, SSID 10431E12
Amplifiers: 2x CS35L41 via I2C (i2c-CSC3551:00-cs35l41-hda.0 = Left, hda.1 = Right)

Problem

After suspend/resume, the right speaker (hda.1) goes completely silent. The left speaker continues working normally. ALSA mixer shows balanced levels — the issue is at the kernel driver level. Both amps boot fine (FW EN: 1), but after resume hda.1 gets stuck:

cs35l41-hda i2c-CSC3551:00-cs35l41-hda.1: Failed to read algorithm info: -16
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.1: Cannot Initialize Firmware. Error: -16
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.1: CS35L41 Bound - SSID: 10431E12, BST: 1, VSPK: 1, CH: R, FW EN: 0

Note FW EN: 0 on the right amp — it binds but has no firmware loaded.

Root cause

The ASUS BIOS is missing the GPIO reset pin in ACPI _DSD. The driver's built-in quirk table for 10431E12 provides some properties ("Using extra _DSD properties, bypassing _DSD in ACPI") but can't provide GPIO reset. Without hardware reset, the driver has no way to recover a stuck I2C bus after suspend.

We confirmed this by attempting a module reload while the issue was active — the re-probe fails on both amps with:

cs35l41-hda i2c-CSC3551:00-cs35l41-hda.0: error -ENOENT: Failed to get reset GPIO
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.0: probe with driver cs35l41-hda failed with error -2

Warning: Do NOT attempt modprobe -r + modprobe while the issue is active — it kills both speakers and only a reboot recovers.

Working fix: systemd system-sleep hook

The key insight is that the modules must be unloaded before suspend (while the I2C bus is still clean). After resume, modprobe re-probes fresh — equivalent to a clean boot.

Create /usr/lib/systemd/system-sleep/cs35l41-fix.sh:

#!/bin/sh
case $1 in
  pre)
    modprobe -r snd_hda_scodec_cs35l41_i2c snd_hda_scodec_cs35l41_spi 2>/dev/null
    modprobe -r snd_hda_scodec_cs35l41 2>/dev/null
    ;;
  post)
    modprobe snd_hda_scodec_cs35l41_i2c 2>/dev/null
    ;;
esac

Then chmod +x it. No PipeWire restart needed — the CS35L41 modules have 0 direct users even while audio is playing (PipeWire uses the HDA codec, not the amp modules directly).

Verified result — clean post-resume kernel log, both amps load firmware successfully:

cs35l41-hda i2c-CSC3551:00-cs35l41-hda.0: Calibration applied: R0=10662
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.0: Firmware Loaded - Type: spk-prot, Gain: 17
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.1: Calibration applied: R0=10920
cs35l41-hda i2c-CSC3551:00-cs35l41-hda.1: Firmware Loaded - Type: spk-prot, Gain: 17

Hope this helps others with the same hardware!

Thanks a lot @pvojnisek! That looked interesting for the ASUS 3405MA as it has the same problem. It has a different SSID: 10431A63 and uses SPI for the audio chip. I replaced the post command above with snd_hda_scodec_cs35l41_spi but no luck (and also used your code as is first: same). I completely loose audio upon restoring from hibernation.

This is the dmesg output that seems relevant:

PM: hibernation: hibernation exit
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: Using extra _DSD properties, bypassing _DSD in ACPI
gpio gpiochip0: (INTC1083:00): gpiochip_lock_as_irq: tried to flag a GPIO set as output for IRQ
gpio gpiochip0: (INTC1083:00): unable to lock HW IRQ 207 for IRQ
genirq: Failed to request resources for cs35l41 IRQ1 Controller (irq 184) on irqchip intel-gpio
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: Failed to request IRQ 184 for cs35l41 IRQ1 Controller: -5
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: IRQ Config Failed. Amp errors may not be recoverable without reboot.
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.0: Cirrus Logic CS35L41 (35a40), Revision: B2
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Using extra _DSD properties, bypassing _DSD in ACPI
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Reset line busy, assuming shared reset
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: Failed waiting for OTP_BOOT_DONE
cs35l41-hda spi1-CSC3551:00-cs35l41-hda.1: probe with driver cs35l41-hda failed with error -110

I'm struggling to debug this and to make sense of it all. I'm stuck now and am giving up for now.

It also ended up screwing up with even cold boot volume issues on one speaker. It took me a while to find but removing the /var/lib/alsa/asound.state file fixed bad states where for example I had the left speaker firmware not loaded (and very quiet)

    control.2 {
        iface CARD
        name 'L0 DSP1 Firmware Load'
        value false
        comment {
            access 'read write'
            type BOOLEAN
            count 1
        }
    }

But the right one was OK (value true):

    control.5 {
        iface CARD
        name 'R0 DSP1 Firmware Load'
        value true
        comment {
            access 'read write'
            type BOOLEAN
            count 1
        }
    }

Alsa kept reloading the bad state file I believe. One has to systemctl stop alsa-restore first before deleting the state file, and reboot.
I hope that bit helps other people who had screwed up their system too.

Turns out for the UM3405MA the speakers share a reset pin and the driver tries to reset both amps at the same time when resuming, which fails.
So unfortunately looks like there are more differences with the UM3402YA :/

I filed a bug about it there with a patch as a workaround: https://bugzilla.kernel.org/show_bug.cgi?id=221161 . I can resume from hibernation with both speakers at full volume now. Hopefully that's the right error diagnosis and they'll know what to do.