Hello, I have an older computer, that acts as a kind of media player. It has a GK208B [GeForce GT 710] and until recently I used the proprietary NVIDIA 470 drivers. Currently they are brokwn with linux 6.12, that's why I used the chance and wanted to switch to nouveau. Graphics with this work fine - but the sound over HDMI does not work. I installed older kernel versions to roughly debug the problem and with linux 6.6.10 everything works fine - problems start with linux 6.7. Currently I use the LTS kernel 6.6.63 that also works just fine. I attach here several outputs of the working 6.6.10 version and the non-working 6.7 version: 00 and 03: $ dmesg|grep -iE "nouve|snd" 01 and 04: $ aplay -L 02 and 05: speaker-test -Dhdmi:NVidia -c 2 Maybe anybody can help me here - I can provide more information, if needed. Many greetings Mathias
Created attachment 307320 [details] dmesg, aplay and speaker-test of 6.6.10 (working) and 6.7 (not working)
Please bisect: https://docs.kernel.org/admin-guide/bug-bisect.html
@Artem just a short intermediate update: I started bisecting this issue, unfortunately this is tougher than I thought. I could bisect successfully only with "--first-parent" switch, here the result was: > 7d461b291e65938f15f56fe58da2303b07578a76 is the first bad commit > f56fe58da2303b07578a76 (HEAD) > Merge: 8bc9e6515183 631808095a82 > Author: Linus Torvalds <torvalds@linux-foundation.org> > Date: Wed Nov 1 06:28:35 2023 -1000 > > Merge tag 'drm-next-2023-10-31-1' of git://anongit.freedesktop.org/drm/drm > > Pull drm updates from Dave Airlie: > "Highlights: > - AMD adds some more upcoming HW platforms > - Intel made Meteorlake stable and started adding Lunarlake > - nouveau has a bunch of display rework in prepartion for the NVIDIA > GSP firmware support > - msm adds a7xx support Without using "--first-parent" I stumbled across multiple other problems and was not able to boot. I tried to bisect further between 7d461b291e65938f15f56fe58da2303b07578a76 and git merge-base 7d461b291e65938f15f56fe58da2303b07578a76^1 7d461b291e65938f15f56fe58da2303b07578a76^2 which is 05d3ef8bba77c1b5f98d941d8b2d4aeab8118ef1 And found the following good commits: 5a6a09e97199d6600d31383055f9d43fbbcbe86f 5cbff4b2d9e2a59f4096af8b8f967e2b30f025f2 and the following bad commits: 27442758e9b4e083bef3f164a1739475c01f3202 614351f41e8c557068f1898eef5bacbca3b20911 Unfortunately, when I try to test many other commits, e.g. e5f9d543419c78ac58f3b3557bc5a76b20ff600b I was not able to boot due to a "BUG: unable to handle page fault for address: XXX" error directly after loading my initramfs. So I first bisected this boot problem and found that my machine relies on 3ac873c76d79f055e1712754d9f2e90e41582f04: If I cherry-pick this commit, I can at least boot. But even then e5f9d543419c78ac58f3b3557bc5a76b20ff600b does not load "nouveau" and does not boot into my graphical.target. Trying to "modprove v nouveau" results in: insmod /lib/modules/6.5.0-1-git-v6.5-rc2-764-ge5f9d543419c/kernel/drivers/gpu/drm/nouveau/nouveau.ko.zst modprobe: ERROR: could not insert 'nouveau': Cannot allocate memory I will put some more time into trying to find the commit that leads to this situation and (not being able to load nouveau). I just wanted to keep you updated. Many greetings Mathias
Hello, I could finally bisect the problem to the end. In order to be able to compile intermediate commits, I needed to cherry-pick - 3ac873c76d79f055e1712754d9f2e90e41582f04 (wrong parentheses) - a419beac4a070aff63c520f36ebf7cb8a76a8ae5 (vmalloc usage for zstd decompression) The result of the bisection was: -------------------------------------------------------------------------------------------------- 6c6abab20b99169f5fb11a8619012225ecd02f1e is the first bad commit commit 6c6abab20b99169f5fb11a8619012225ecd02f1e Author: Ben Skeggs <bskeggs@redhat.com> Date: Tue Sep 19 17:56:14 2023 -0400 drm/nouveau/disp: add output hdmi config method - was previously part of acquire() - preparation for GSP-RM Signed-off-by: Ben Skeggs <bskeggs@redhat.com> Reviewed-by: Lyude Paul <lyude@redhat.com> Acked-by: Danilo Krummrich <me@dakr.org> Signed-off-by: Lyude Paul <lyude@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20230919220442.202488-20-lyude@redhat.com -------------------------------------------------------------------------------------------------- The commit: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=6c6abab20b99169f5fb11a8619012225ecd02f1e changes the logic of nv50_hdmi_enable() in drivers/gpu/drm/nouveau/dispnv50/disp.c Before a local variable u8 scdc = 0; was created in the function, filled and later nvif_outp_acquire_tmds() was called (where scdc was passed). The new logic creates also u8 scdc; locally, but does not pass it to nvif_outp_hdmi() - the function that replaced the call to nvif_outp_acquire_tmds(). Instead it passes a bool variable hdmi->scdc.supported These values are later used in uoutp.c in drivers/gpu/drm/nouveau/nvkm/engine/disp/uoutp.c in the function nvkm_uoutp_mthd_hdmi(), respectively nvkm_uoutp_mthd_acquire_tmds() (before the last commit). Here is a debug output of the last good commit (c0f7b729425b8f0764b9e89f19d33a0deba89811) at the beginning of nvkm_uoutp_mthd_acquire_tmds(): outp=000000007ea185c3, head=0, hdmi=1, hdmi_max_ac_packet=7, hdmi_rekey=56, hdmi_scdc=0, hdmi_hda=1 and here the output after the bad commit (6c6abab20b99169f5fb11a8619012225ecd02f1e) of nvkm_uoutp_mthd_hdmi(): args->v0.head=0, args->v0.enable=1, args->v0.max_ac_packet=7, args->v0.rekey=56, args->v0.scdc=1, args->v0.scdc_scrambling=1, args->v0.scdc_low_rates=0, args->v0.khz=297000 You see hdmi_scdc was 0 before and is now replaced with bool args->v0.scdc, which is 1 on the affected machine. I updated disp.c to use the old logic before (create a local _u8 and use this in nvif_outp_hdmi) that fixes the problem for me and attach the patch here in the file fix_v6.12.patch It can be applied to v6.12 (adc218676eef25575469234709c2d87185ca223a) and results in working sound for me - tested with v6.7 and v6.12. Maybe the problem can be generally fixed upsteram? Thank you very much and many greetings Mathias
Created attachment 307360 [details] This patch fixes the sound issue for me
And I forgot one thing: the different value of scdc results in the (for me) broken commit, that in nvkm_uoutp_mthd_hdmi() we take the branch: if (args->v0.scdc && !ior->func->hdmi->scdc) and return -EINVAL; In the former function, nvkm_uoutp_mthd_acquire_tmds(), we did not enter this branch and returned 0.
Note that nouveau driver team uses gitlab.freedesktop.org: https://nouveau.freedesktop.org/Bugs.html . I suggest to move your findings there.
@Jaroslav: Thanks for the hint! Just for completeness: I created a bug report at the gitlab bug tracker here: https://gitlab.freedesktop.org/drm/nouveau/-/issues/402 Many greetings Mathias