You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
MacBook Pro 2015-2017 audio broken — need snd_hda_macbookpro driver #11619
MacBook Pro 2015-2017 with Cirrus Logic CS8409 audio codec
Omarchy 4.x / kernel 7.2.x
Problem
The in-tree snd_hda_codec_cs8409 driver doesn't properly configure the CS8409's speaker pins. Speakers don't work, and headphone jack switching is broken.
Fix confirmed working, but it doesn't survive a reboot
Omarchy 4.0.4, MacBookPro14,1, kernel 7.2.5-3-omarchy. The snd-hda-macbookpro DKMS module IS installed (dkms status → snd-hda-macbookpro/0.1, 7.2.5-3-omarchy: installed), but at every boot the stock snd-hda-codec-cs8409 bundled in the initramfs loads first, so the patched Apple path never runs and the internal speakers stay silent.
Symptom at boot: dmesg shows the stock autoconfig and no patch_cs8409 ... trying APPLE line. After sudo modprobe -r snd-hda-codec-cs8409 && sudo modprobe snd-hda-codec-cs8409, the out-of-tree module loads, trying APPLE appears, and speakers work.
Fix on this machine: sudo mkinitcpio -P while the patched module is loaded so the initramfs bundles the patched build.
Suggestion: the install hook (fix-snd-hda-macbookpro.sh in the pending PRs) should regenerate the initramfs after DKMS install, otherwise the fix silently regresses on every reboot.
#11625 gate (106b:3900 || CS8409) correctly skips this board. Suggest scoping title/gate to CS8409 / MacBookPro13,x — 11,5 is 2015 but unaffected. Same family-split as #12286 (CS4208 silent, speaker_outs=0 on MacBook9,1/10,1) vs working CS4208 here.
Another confirmation on MacBookPro14,1 (CS8409, subsystem 0x106b3300), Omarchy 4.0.4-1, official kernel 7.2.5-3-omarchy.
The stock codec driver left the internal speakers silent or briefly audible. After installing the davidjo/snd_hda_macbookpro driver through DKMS and reloading it, CLIamp and browser audio now work continuously through the speakers. The loaded driver reports trying APPLE and exposes CS8409/CS42L83 Analog. Not yet reboot-tested.
Status note plus a control data point, since the fix PR changed underneath this thread: #11625 was closed Sep 21 as a duplicate of #11064, which is now the surviving fix — and it is pkg-add only (no initramfs rebuild, no modprobe ordering). So @blaurans' reboot report currently has no answer in the surviving PR.
Control evidence from MacBookPro11,5 (CS4208, no audio DKMS — cannot reproduce, but can bound the mechanism), kernel 7.2.5-3-omarchy:
The booted UKI contains no HDA drivers at all: only snd core stubs, no snd-hda-*, no codec modules, and no extra/ or updates/ dirs. Sound here binds purely post-switch-root via udev.
DKMS override-by-search-path provably works on this kernel: modinfo -n applesmc resolves to updates/dkms/applesmc.ko.zst, same for sbs. Upstream's dkms.conf names the module snd-hda-codec-cs8409 into extra/, so the same override should hold for it.
modprobe.d IS baked into every UKI via modconf (verified brcmfmac.conf present) — so a blacklist/ordering conf would take effect at early boot if one were shipped.
That frames the open question narrowly: with the override mechanism working, why stock wins at boot on 14,1 needs three data points from an affected machine: (1) is stock snd-hda-codec-cs8409 actually baked into the booted UKI (lsinitcpio — a yes would contradict my 11,5 finding and point at modalias-specific autodetect), (2) modinfo -n snd-hda-codec-cs8409 + dkms status + listing of extra//updates/ for the running kernel, (3) dmesg bind timing for the codec (before vs after switch-root).
@Clowdyffs — your confirmation is the newest data here; a reboot test would settle whether blaurans' report reproduces on a second 14,1.
Checked against current quattro (9480d1b). Claude Opus 5.5 did the diagnosis. Codex at medium effort reviewed it separately as a second opinion and agreed, but it could not reach GitHub, so it only checked part of this. Its independence is not guaranteed. Nothing was reproduced: a VM has no CS8409, so this is a reading of the source and the open pull requests, not a test.
Where the fix stands.quattro has no path that installs snd-hda-macbookpro or otherwise brings up CS8409 speakers. install/hardware/apple/fix-t2.sh only covers T2 Macs, which these models predate. Four open pull requests propose the fix, and none has landed:
None of them can land yet, because neither package name is in the Omarchy package repository. The packaging pull requests, omacom/omarchy-pkgs#155 and #249, are still open. As they stand, #8285 and #11064 fail with target not found on exactly these Macs, which stops setup. Choosing an approach and shipping the package is the maintainer's decision. None of these pull requests is linked to this issue as its fix, since none has been seen to fix it on hardware.
The reboot report is still open. None of the four rebuilds the initramfs or orders the codec module. #12689 only documents "reboot once". The source does not settle whether @blaurans's report reproduces. Omarchy's mkinitcpio hooks add nothing that would bundle snd-hda-codec-cs8409: autodetect filters modules, and modconf copies only modprobe configuration. That agrees with @rand0mdud3's finding on an 11,5. But it does not explain what @blaurans saw on a 14,1. Whichever fix lands needs a cold-boot test on a MacBookPro14,1, including the three checks @rand0mdud3 listed: lsinitcpio on the booted image, modinfo -n snd-hda-codec-cs8409 with dkms status, and the codec's bind time in dmesg.
This now waits on the maintainer's packaging decision. After that, the fix that lands needs a reboot test on one of these Macs.
Same CS8409 pre-T2 speaker stack as #11064 / #8285 (driver/firmware path) and #12968 (recover without reboot). Suggest tracking under those PRs — if you can name the exact model and try one of them, that'll help prioritize.
Speaker audio survives reboot on my MacBookPro14,1 with 7.2.5-3-omarchy and local snd-hda-macbookpro-dkms 0.1-0.1. DKMS reports snd_hda_macbookpro/0.1 installed for that kernel, and modinfo resolves the CS8409 module under updates/dkms.
This boot’s kernel log shows Primary patch_cs8409 NOT FOUND trying APPLE, and hdaudioC0D0 is bound to snd_hda_codec_cs8409. Basic music playback through the built-in speakers works normally without a manual reload after boot. I haven’t stress-tested it or checked every audio path.
Thanks @Clowdyffs, that answers part of the reboot question. Re-checked this thread and the open pull requests. Claude Opus 5.5 did the reading, and Codex at medium effort reviewed the conclusion as a second opinion and agreed. Codex could not reach GitHub this time, so it checked only the summary it was given, and its independence is not guaranteed. Nothing was run: a VM has no CS8409, so this is not a test.
Reboot. There are now two reports from a MacBookPro14,1 on 7.2.5-3-omarchy. In @Clowdyffs's, the speakers work after a reboot with no manual reload. modinfo -n resolves the CS8409 module under updates/dkms, and the boot log shows the Apple path (trying APPLE). In @blaurans's, the stock codec ran at every boot until it was reloaded. That makes "it never survives a reboot" unlikely, but it does not explain the failure. @blaurans did not post modinfo -n snd-hda-codec-cs8409, so we cannot tell whether the stock module was shadowing the DKMS build there. @blaurans, if that machine still shows it, the three checks @rand0mdud3 listed would settle it: lsinitcpio on the booted image, modinfo -n snd-hda-codec-cs8409 with dkms status, and when the codec binds in dmesg.
The fix. Nothing has changed since the last update. #8285, #11064, #12689 (draft) and #12968 (draft) are all still open, and so are the packaging pull requests omacom/omarchy-pkgs#155 and #249. A locally built snd-hda-macbookpro-dkms working here supports the driver approach. It does not verify either packaging pull request. #8285 also asks for snd-hda-macbookpro-dkms-git, which neither packaging pull request provides.
This still waits on the maintainer's decision about packaging and which fix to take.
ALSA still listed the analog device and PipeWire routed playback to it, but the speakers were silent.
modinfo -n snd_hda_codec_cs8409 selected the stock module under kernel/sound/hda/codecs/cirrus/.
dkms status had no snd_hda_macbookpro entry. Previous-boot kernel logs showed the custom Apple initialization path.
The evidence points to a manually installed custom module being lost with the old kernel, rather than a demonstrated DKMS audio build failure.
Recovery confirmed on this machine
An existing prepared source tree of davidjo/snd_hda_macbookpro, including its local resume patch, was rebuilt against the installed 7.2.15-1-omarchy headers with an explicit KERNELRELEASE=7.2.15-1-omarchy. Installing it as an override and reloading the audio controller/codec while the user audio services were stopped restored playback. The user confirmed audio worked; no reboot was needed for this recovery.
The original sources and separate NVMe suspend configuration were preserved. Existing audio power settings were also unchanged:
A root-owned permanent copy of the already prepared, working source tree, including the generated/patched build/hda sources and local resume patch, was placed at:
/usr/src/snd_hda_macbookpro-0.1.local1
This is not just a fresh repository checkout. The configuration below deliberately skips upstream source preparation and requires that prepared tree to exist.
Build first. Back up and move aside any manually installed override before DKMS installation; restore it if installation fails. Keep the distribution's stock module available for rollback.
Arch's existing /usr/share/libalpm/hooks/70-dkms-install.hook already invokes DKMS after kernel/header changes. The permanent source registration and AUTOINSTALL="yes" let that hook discover this driver, without a separate timer or boot-time reload service. This local setup avoids network downloads during builds.
Verified and still untested
DKMS build and installation succeeded for 7.2.15-1-omarchy.
Module resolution now selects updates/dkms/snd-hda-codec-cs8409.ko.zst.
The custom module had already been loaded and playback confirmed before switching installation management to DKMS.
Local resume patch and suspend configuration were preserved, but suspend/resume was not retested in this session.
A fresh boot and the next kernel upgrade have not been tested. This does not resolve the separate reboot/initramfs question raised above.
Headphones and microphone were not tested.
Prepared HDA sources can become incompatible with later kernels. A successful compilation alone does not guarantee private structure compatibility; matching distribution sources may be needed. This is not a universal replacement for upstream source preparation.
It would help users if installation documentation clearly distinguished manual installation from persistent DKMS registration, explained how to verify the selected module, and covered preservation of local patches.
Related upstream discussion: davidjo/snd_hda_macbookpro#183 . This report does not claim to reproduce its source-download or output-path failures.
Prepared with Codex from commands run on the affected machine and the user's playback confirmation.
My patch davidjo/snd_hda_macbookpro#183 is working fine on iMac 18,3 (2017 27") on Omarchy. The only issue is that sound below 15% volume is effectively 0%.
Correction to my comment above: davidjo/snd_hda_macbookpro#183 is no longer the fix for the DKMS build itself. Its original change (BUILT_MODULE_LOCATION[0]="build/hda/codecs/cirrus", needed for 6.17+/7.x kbuild output) is already in upstream master (davidjo/snd_hda_macbookpro@6c324f6). Packaging upstream as-is (omacom/omarchy-pkgs#155/#249) therefore does not need #183 to build on 7.2.
I've rebased #183 onto upstream and retitled it. What remains is a kernel-source cache for the DKMS PRE_BUILD step (/var/cache/cs8409-src). It reuses any cached tarball from the same minor series instead of downloading from kernel.org on every kernel bump. That covers the failure @carlosa896 described above: a kernel upgrade losing the driver, and their workaround of a pre-prepared tree with no download step. Smoke test: an offline build (no network namespace) for 7.2.15-1-omarchy from a cached linux-7.2.3.tar.xz produced a snd-hda-codec-cs8409.ko with matching vermagic.
Not affected by #183: the initramfs/module-ordering question raised by @blaurans, and the low-volume behaviour I mentioned (below ~15% is effectively silent on my iMac18,3).
Prepared with Claude Opus 5.5 (agent), reviewed by me.
Confirming on MacBookPro13,3 (15", 2016, MAX98706 amps), Omarchy 4.5.0-3, kernel 7.2.15-1-omarchy: davidjo/snd_hda_macbookpro at 89b22ff builds and installs through DKMS (install.cirrus.driver.sh -i). It prints "Kernel version later than implemented version", but the patches apply with offsets/fuzz and both speaker sides play.
One thing an installer would need: the DKMS PRE_BUILD step downloads the kernel source with wget, which Omarchy doesn't ship. Without it the build fails, including the automatic rebuild after a kernel update, so a fix script should omarchy-pkg-add wget dkms (and the machine needs network access when a new kernel is installed).
Re-checked the new reports against quattro at 419a5a7095d118a218d5f84548dbd83ce90067e6 and the upstream driver sources. GPT-6 in Codex (High) did the reading; Opus 5.5 High and muse-wonder via pi (High thinking) reviewed the same evidence and agreed with the distinctions below. Their independence is not guaranteed. This was source and report review only: no driver or tests were run, and a VM cannot verify these Macs' speakers or boot behavior.
The kernel-update reports describe separate failure modes. @carlosa896 found the stock module selected and no snd_hda_macbookpro DKMS entry after the update. The recovery registered a prepared source tree with DKMS, but it does not demonstrate a failed DKMS rebuild or source download; the next boot and kernel update remain untested. Upstream still invokes wget from its DKMS PRE_BUILD step, which supports @marionuevo's dependency finding. The packaging proposals omacom/omarchy-pkgs#155 and omacom/omarchy-pkgs#249 already declare wget. Upstream davidjo/snd_hda_macbookpro#183 now proposes a source cache for that download step; the report with no DKMS entry does not establish a failure that the cache would fix. The documentation proposal #15776 already covers wget and keeping the cloned source for future rebuilds.
One correction to the earlier status here: #11064 now covers MacBookPro13,1–13,3 and 14,1–14,3 and preserves existing hand-built drivers. It remains open, along with the other proposed integrations. A newer packaging proposal, omacom/omarchy-pkgs#2723, is also open. These local hardware reports do not verify a complete Omarchy pull request at its current head.
The separate stock-module/boot-image report remains unresolved. This still waits on the maintainer's packaging and integration decision, followed by testing the selected head on affected hardware through a reboot and a kernel update.
Activity
blaurans commentedon Sep 18, 2026
Fix confirmed working, but it doesn't survive a reboot
Omarchy 4.0.4, MacBookPro14,1, kernel 7.2.5-3-omarchy. The snd-hda-macbookpro DKMS module IS installed (dkms status → snd-hda-macbookpro/0.1, 7.2.5-3-omarchy: installed), but at every boot the stock snd-hda-codec-cs8409 bundled in the initramfs loads first, so the patched Apple path never runs and the internal speakers stay silent.
Symptom at boot: dmesg shows the stock autoconfig and no patch_cs8409 ... trying APPLE line. After sudo modprobe -r snd-hda-codec-cs8409 && sudo modprobe snd-hda-codec-cs8409, the out-of-tree module loads, trying APPLE appears, and speakers work.
Fix on this machine: sudo mkinitcpio -P while the patched module is loaded so the initramfs bundles the patched build.
Suggestion: the install hook (fix-snd-hda-macbookpro.sh in the pending PRs) should regenerate the initramfs after DKMS install, otherwise the fix silently regresses on every reboot.
Filed by big-pickle via opencode.
rand0mdud3 commentedon Sep 18, 2026
MacBookPro11,5, Omarchy 4.0.4-1, 7.2.5-3-omarchy. Codec CS4208 (0x10134208, subsys 0x106b8000), PCI 8086:8c20 subsys 7270. No 106b:3900, no CS8409 in /proc/asound.
Stock snd_hda_codec_cs420x, no audio DKMS. Speakers live: PipeWire alsa_output.pci-0000_00_1b.0.analog-stereo RUNNING, Speaker 100% [on], Headphone [off], Chrome streaming to CS4208 Analog. Node 0x03 Speaker Playback Volume/Switch, stream=5, D0.
#11625 gate (106b:3900 || CS8409) correctly skips this board. Suggest scoping title/gate to CS8409 / MacBookPro13,x — 11,5 is 2015 but unaffected. Same family-split as #12286 (CS4208 silent, speaker_outs=0 on MacBook9,1/10,1) vs working CS4208 here.
Clowdyffs commentedon Sep 26, 2026
Another confirmation on MacBookPro14,1 (CS8409, subsystem
0x106b3300), Omarchy 4.0.4-1, official kernel 7.2.5-3-omarchy.The stock codec driver left the internal speakers silent or briefly audible. After installing the
davidjo/snd_hda_macbookprodriver through DKMS and reloading it, CLIamp and browser audio now work continuously through the speakers. The loaded driver reportstrying APPLEand exposesCS8409/CS42L83 Analog. Not yet reboot-tested.rand0mdud3 commentedon Sep 29, 2026
Status note plus a control data point, since the fix PR changed underneath this thread: #11625 was closed Sep 21 as a duplicate of #11064, which is now the surviving fix — and it is pkg-add only (no initramfs rebuild, no modprobe ordering). So @blaurans' reboot report currently has no answer in the surviving PR.
Control evidence from MacBookPro11,5 (CS4208, no audio DKMS — cannot reproduce, but can bound the mechanism), kernel 7.2.5-3-omarchy:
sndcore stubs, nosnd-hda-*, no codec modules, and noextra/orupdates/dirs. Sound here binds purely post-switch-root via udev.modinfo -n applesmcresolves toupdates/dkms/applesmc.ko.zst, same forsbs. Upstream'sdkms.confnames the modulesnd-hda-codec-cs8409intoextra/, so the same override should hold for it.modprobe.dIS baked into every UKI via modconf (verifiedbrcmfmac.confpresent) — so a blacklist/ordering conf would take effect at early boot if one were shipped.That frames the open question narrowly: with the override mechanism working, why stock wins at boot on 14,1 needs three data points from an affected machine: (1) is stock
snd-hda-codec-cs8409actually baked into the booted UKI (lsinitcpio— a yes would contradict my 11,5 finding and point at modalias-specific autodetect), (2)modinfo -n snd-hda-codec-cs8409+dkms status+ listing ofextra//updates/for the running kernel, (3) dmesg bind timing for the codec (before vs after switch-root).@Clowdyffs — your confirmation is the newest data here; a reboot test would settle whether blaurans' report reproduces on a second 14,1.
omarchybot commentedon Oct 11, 2026
Checked against current
quattro(9480d1b). Claude Opus 5.5 did the diagnosis. Codex at medium effort reviewed it separately as a second opinion and agreed, but it could not reach GitHub, so it only checked part of this. Its independence is not guaranteed. Nothing was reproduced: a VM has no CS8409, so this is a reading of the source and the open pull requests, not a test.Where the fix stands.
quattrohas no path that installssnd-hda-macbookproor otherwise brings up CS8409 speakers.install/hardware/apple/fix-t2.shonly covers T2 Macs, which these models predate. Four open pull requests propose the fix, and none has landed:snd-hda-macbookpro-dkms-git.snd-hda-macbookpro-dkms.None of them can land yet, because neither package name is in the Omarchy package repository. The packaging pull requests, omacom/omarchy-pkgs#155 and #249, are still open. As they stand, #8285 and #11064 fail with
target not foundon exactly these Macs, which stops setup. Choosing an approach and shipping the package is the maintainer's decision. None of these pull requests is linked to this issue as its fix, since none has been seen to fix it on hardware.The reboot report is still open. None of the four rebuilds the initramfs or orders the codec module. #12689 only documents "reboot once". The source does not settle whether @blaurans's report reproduces. Omarchy's mkinitcpio hooks add nothing that would bundle
snd-hda-codec-cs8409:autodetectfilters modules, andmodconfcopies only modprobe configuration. That agrees with @rand0mdud3's finding on an 11,5. But it does not explain what @blaurans saw on a 14,1. Whichever fix lands needs a cold-boot test on a MacBookPro14,1, including the three checks @rand0mdud3 listed:lsinitcpioon the booted image,modinfo -n snd-hda-codec-cs8409withdkms status, and the codec's bind time indmesg.This now waits on the maintainer's packaging decision. After that, the fix that lands needs a reboot test on one of these Macs.
jandrusk commentedon Oct 13, 2026
Same CS8409 pre-T2 speaker stack as #11064 / #8285 (driver/firmware path) and #12968 (recover without reboot). Suggest tracking under those PRs — if you can name the exact model and try one of them, that'll help prioritize.
Clowdyffs commentedon Oct 15, 2026
Speaker audio survives reboot on my MacBookPro14,1 with
7.2.5-3-omarchyand localsnd-hda-macbookpro-dkms 0.1-0.1. DKMS reportssnd_hda_macbookpro/0.1installed for that kernel, andmodinforesolves the CS8409 module underupdates/dkms.This boot’s kernel log shows
Primary patch_cs8409 NOT FOUND trying APPLE, andhdaudioC0D0is bound tosnd_hda_codec_cs8409. Basic music playback through the built-in speakers works normally without a manual reload after boot. I haven’t stress-tested it or checked every audio path.omarchybot commentedon Oct 15, 2026
Thanks @Clowdyffs, that answers part of the reboot question. Re-checked this thread and the open pull requests. Claude Opus 5.5 did the reading, and Codex at medium effort reviewed the conclusion as a second opinion and agreed. Codex could not reach GitHub this time, so it checked only the summary it was given, and its independence is not guaranteed. Nothing was run: a VM has no CS8409, so this is not a test.
Reboot. There are now two reports from a MacBookPro14,1 on
7.2.5-3-omarchy. In @Clowdyffs's, the speakers work after a reboot with no manual reload.modinfo -nresolves the CS8409 module underupdates/dkms, and the boot log shows the Apple path (trying APPLE). In @blaurans's, the stock codec ran at every boot until it was reloaded. That makes "it never survives a reboot" unlikely, but it does not explain the failure. @blaurans did not postmodinfo -n snd-hda-codec-cs8409, so we cannot tell whether the stock module was shadowing the DKMS build there. @blaurans, if that machine still shows it, the three checks @rand0mdud3 listed would settle it:lsinitcpioon the booted image,modinfo -n snd-hda-codec-cs8409withdkms status, and when the codec binds indmesg.The fix. Nothing has changed since the last update. #8285, #11064, #12689 (draft) and #12968 (draft) are all still open, and so are the packaging pull requests omacom/omarchy-pkgs#155 and #249. A locally built
snd-hda-macbookpro-dkmsworking here supports the driver approach. It does not verify either packaging pull request. #8285 also asks forsnd-hda-macbookpro-dkms-git, which neither packaging pull request provides.This still waits on the maintainer's decision about packaging and which fix to take.
carlosa896 commentedon Nov 2, 2026
Tested recovery after a kernel update: MacBookPro14,1 / Omarchy
Sharing a working recovery and DKMS setup from real hardware. This is a local workaround, not a universal fix for every Mac or future kernel.
Diagnosis
modinfo -n snd_hda_codec_cs8409selected the stock module underkernel/sound/hda/codecs/cirrus/.dkms statushad nosnd_hda_macbookproentry. Previous-boot kernel logs showed the custom Apple initialization path.The evidence points to a manually installed custom module being lost with the old kernel, rather than a demonstrated DKMS audio build failure.
Recovery confirmed on this machine
An existing prepared source tree of davidjo/snd_hda_macbookpro, including its local resume patch, was rebuilt against the installed 7.2.15-1-omarchy headers with an explicit
KERNELRELEASE=7.2.15-1-omarchy. Installing it as an override and reloading the audio controller/codec while the user audio services were stopped restored playback. The user confirmed audio worked; no reboot was needed for this recovery.The original sources and separate NVMe suspend configuration were preserved. Existing audio power settings were also unchanged:
options snd_hda_intel power_save=0 power_save_controller=NPersistent DKMS registration
A root-owned permanent copy of the already prepared, working source tree, including the generated/patched
build/hdasources and local resume patch, was placed at:/usr/src/snd_hda_macbookpro-0.1.local1This is not just a fresh repository checkout. The configuration below deliberately skips upstream source preparation and requires that prepared tree to exist.
Then:
Build first. Back up and move aside any manually installed override before DKMS installation; restore it if installation fails. Keep the distribution's stock module available for rollback.
Arch's existing
/usr/share/libalpm/hooks/70-dkms-install.hookalready invokes DKMS after kernel/header changes. The permanent source registration andAUTOINSTALL="yes"let that hook discover this driver, without a separate timer or boot-time reload service. This local setup avoids network downloads during builds.Verified and still untested
snd_hda_macbookpro/0.1.local1, 7.2.15-1-omarchy, x86_64: installed.updates/dkms/snd-hda-codec-cs8409.ko.zst.It would help users if installation documentation clearly distinguished manual installation from persistent DKMS registration, explained how to verify the selected module, and covered preservation of local patches.
Related upstream discussion: davidjo/snd_hda_macbookpro#183 . This report does not claim to reproduce its source-download or output-path failures.
Prepared with Codex from commands run on the affected machine and the user's playback confirmation.
edrobertsrayne commentedon Nov 6, 2026
My patch davidjo/snd_hda_macbookpro#183 is working fine on iMac 18,3 (2017 27") on Omarchy. The only issue is that sound below 15% volume is effectively 0%.
edrobertsrayne commentedon Nov 6, 2026
Correction to my comment above: davidjo/snd_hda_macbookpro#183 is no longer the fix for the DKMS build itself. Its original change (
BUILT_MODULE_LOCATION[0]="build/hda/codecs/cirrus", needed for 6.17+/7.x kbuild output) is already in upstreammaster(davidjo/snd_hda_macbookpro@6c324f6). Packaging upstream as-is (omacom/omarchy-pkgs#155/#249) therefore does not need #183 to build on 7.2.I've rebased #183 onto upstream and retitled it. What remains is a kernel-source cache for the DKMS
PRE_BUILDstep (/var/cache/cs8409-src). It reuses any cached tarball from the same minor series instead of downloading from kernel.org on every kernel bump. That covers the failure @carlosa896 described above: a kernel upgrade losing the driver, and their workaround of a pre-prepared tree with no download step. Smoke test: an offline build (no network namespace) for7.2.15-1-omarchyfrom a cachedlinux-7.2.3.tar.xzproduced asnd-hda-codec-cs8409.kowith matching vermagic.Not affected by #183: the initramfs/module-ordering question raised by @blaurans, and the low-volume behaviour I mentioned (below ~15% is effectively silent on my iMac18,3).
Prepared with Claude Opus 5.5 (agent), reviewed by me.
marionuevo commentedon Nov 9, 2026
Confirming on MacBookPro13,3 (15", 2016, MAX98706 amps), Omarchy 4.5.0-3, kernel
7.2.15-1-omarchy: davidjo/snd_hda_macbookpro at89b22ffbuilds and installs through DKMS (install.cirrus.driver.sh -i). It prints "Kernel version later than implemented version", but the patches apply with offsets/fuzz and both speaker sides play.One thing an installer would need: the DKMS
PRE_BUILDstep downloads the kernel source withwget, which Omarchy doesn't ship. Without it the build fails, including the automatic rebuild after a kernel update, so a fix script shouldomarchy-pkg-add wget dkms(and the machine needs network access when a new kernel is installed).omarchybot commentedon Nov 18, 2026
Re-checked the new reports against
quattroat419a5a7095d118a218d5f84548dbd83ce90067e6and the upstream driver sources. GPT-6 in Codex (High) did the reading; Opus 5.5 High and muse-wonder via pi (High thinking) reviewed the same evidence and agreed with the distinctions below. Their independence is not guaranteed. This was source and report review only: no driver or tests were run, and a VM cannot verify these Macs' speakers or boot behavior.The kernel-update reports describe separate failure modes. @carlosa896 found the stock module selected and no
snd_hda_macbookproDKMS entry after the update. The recovery registered a prepared source tree with DKMS, but it does not demonstrate a failed DKMS rebuild or source download; the next boot and kernel update remain untested. Upstream still invokeswgetfrom its DKMSPRE_BUILDstep, which supports @marionuevo's dependency finding. The packaging proposals omacom/omarchy-pkgs#155 and omacom/omarchy-pkgs#249 already declarewget. Upstream davidjo/snd_hda_macbookpro#183 now proposes a source cache for that download step; the report with no DKMS entry does not establish a failure that the cache would fix. The documentation proposal #15776 already coverswgetand keeping the cloned source for future rebuilds.One correction to the earlier status here: #11064 now covers MacBookPro13,1–13,3 and 14,1–14,3 and preserves existing hand-built drivers. It remains open, along with the other proposed integrations. A newer packaging proposal, omacom/omarchy-pkgs#2723, is also open. These local hardware reports do not verify a complete Omarchy pull request at its current head.
The separate stock-module/boot-image report remains unresolved. This still waits on the maintainer's packaging and integration decision, followed by testing the selected head on affected hardware through a reboot and a kernel update.