Skip to content

docs: add displayport audio desync page with the dcn201 spread-spectrum root cause#35

Merged
elektricM merged 1 commit into
mainfrom
docs/audio-dp-desync
Jul 18, 2026
Merged

docs: add displayport audio desync page with the dcn201 spread-spectrum root cause#35
elektricM merged 1 commit into
mainfrom
docs/audio-dp-desync

Conversation

@elektricM

Copy link
Copy Markdown
Owner

What's in this PR

  • New troubleshooting/audio.md: the DisplayPort audio desync/delay/crackle root cause (DCN201 spread-spectrum bug), diagnosis, and the fix.
  • Corrects the "resolved in 6.19.10+" claim: that is downstream/community kernels, not mainline. Mainline 7.1.x still calls dce_clock_read_ss_info() unconditionally, so it is still affected. Verified against the mainline source.
  • Cross-reference from the active-vs-passive adapter section in display.md (the SS desync is a separate, kernel-level issue from the active-adapter re-encode one).
  • Nav entry in mkdocs.yml.

Why this and not more

  • Diagnostic commands are tested live on the board (card0 dev3 QCQ90, ELD valid, EDID via cat, hw_params). The fix (patched amdgpu) is NOT tested on our board yet, so it is flagged as reported, not verified.
  • Left the display.md active-adapter content intact and only added a cross-reference, since the two issues are distinct.
  • Root cause credited to the community (Trov, essdee, mzk10); code path re-confirmed against mainline.

…um root cause

The DP audio desync/delay/crackle only had the active-adapter hand-wave in
display.md. This adds a dedicated troubleshooting/audio.md documenting the real
root cause: dcn201_clk_mgr applies the shared GPU-clock spread spectrum to the DP
reference clock, ignoring the VBIOS dp_ss_control=0. Verified against mainline
source. Corrects the resolved-in-6.19.10 claim: that is downstream only, mainline
7.1.x still calls dce_clock_read_ss_info unconditionally. Diagnostic commands
tested live on the board. Trov's dcn201 patch is shown but flagged as not
verified on our board.

Refs the BC-250 Discord audio thread (Trov, essdee, mzk10).
@bangstk

bangstk commented Jul 11, 2026

Copy link
Copy Markdown
Contributor

Trov here - this version of the kernel patch doesn't work.

I assumed that the VBIOS was setting dp_ss_control to 0, because it's what's in BIOS chip.
However, it turns out that some VBIOS table params are altered at boot time, so dp_ss_control is actually 1 by the time the kernel reads it, therefore the patch has no effect.

@elektricM
elektricM merged commit 23afa3a into main Jul 18, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants