Re: [PATCH v5 0/2] ALSA: pcm: preserve capture tail during drain

[Takashi Iwai <tiwai@suse.de>]

On Fri, 16 Oct 2026 01:48:03 +0200,
Joonho Lee wrote:
> 
> Capture drain can leave final frames buffered after the device stops.
> This series lets userspace read those frames and refreshes the linked
> streams' hardware pointers before drain-init.
> 
> Patch 1 allows capture reads in DRAINING, preserves the unread count when
> waking a blocked reader, and finishes capture with snd_pcm_drain_done()
> so linked playback can continue draining. DELAY and SYNC_PTR(HWSYNC) use
> the cached final capture pointer after capture stops.
> 
> Patch 2 refreshes each running member under its stream lock while holding
> snd_pcm_link_rwsem for reading. It updates both playback and capture,
> after the existing pause-resume step. The pause and drain-init actions
> remain outside the link rwsem. Pointer errors follow the normal XRUN
> path and return -EPIPE before drain-init.
> 
> Changes since RFC v4:
> - Submit as a non-RFC series following Takashi's response.
> - No changes to either patch body, including the commit message, diff,
>   and existing trailers. This cover and the mail headers are updated.
> 
> Takashi said both patches look OK and that he can try merging a non-RFC
> series:
> https://lore.kernel.org/all/87ld7z6jsk.wl-tiwai@suse.de/
> Previous RFC v4:
> https://lore.kernel.org/all/rfc-v4-core-capture-drain-0.b56779f7976d4657@gmail.com/
> Implementation and test discussion:
> https://lore.kernel.org/all/core-drain-helper-reply.0c1099dbd0ca4a37@gmail.com/
> 
> The five-patch virtio-snd lifetime v3 series is unchanged:
> https://lore.kernel.org/all/cover.1791881787.git.mynameisjune111@gmail.com/
> The base and prerequisite IDs below preserve the tested integration:
> apply the five lifetime patches, then these two core patches. Replace
> both earlier core patches when testing this revision. Core has no
> dependency on the generic control-wait or VZ compatibility RFC. This
> submission does not revise either companion series.
> 
> The following validation is from RFC v4's unchanged source; no new
> runtime tests were run for this mail-only revision.
> 
> Validation used ARM64 QEMU with KASAN and lockdep:
> - 26 linked-stream checks passed across atomic and nonatomic modes,
>   including paused drain, first/later member pointer errors, normal XRUN
>   handling, tail reads, and unlink/prepare/reopen recovery.
> - The three patch-1 failures were reproduced with the helper and the old
>   first patch in both locking modes. All six corresponding fixed cases
>   passed: a blocked reader received all 64 final frames, capture EOF
>   left linked playback in DRAINING, and both position queries succeeded
>   without another pointer callback.
> - Four additional blocked-reader runs compared baseline and fixed modules
>   in both modes with a stronger gate: thread sleep and a pointer callback
>   after START. They confirmed the baseline loss and fixed 64-frame read.
> - The 10 existing lifecycle and timeout/reset scenarios passed, including
>   held STOP/RELEASE, SIGKILL, reset, and subsequent I/O recovery.
> - No KASAN, lockdep, or panic diagnostics appeared in the 54 selected
>   logs, which include the baseline behavioral-failure comparisons.
> 
> The full Image and modules were first built with the new helper and the
> old first patch. After fixing the three tail issues, sound/core was
> rebuilt with W=1 against that same configuration and symbol versions,
> and the corrected PCM module was used for the runs above. The module
> build had no warnings or errors. The full build had 517 existing
> -Woverride-init warnings in arch/arm64 outside sound/. The exported v5
> core source matches the tested source byte for byte.
> 
> These tests use synthetic drivers and modified fault/reset backends.
> 
> LLM assistance was used for implementation, reproducer development,
> analysis, and documentation. I reviewed the changes and test evidence.
> Assisted-by: LLM
> 
> Joonho Lee (2):
>   ALSA: pcm: allow capture reads while draining
>   ALSA: pcm: refresh linked pointers before draining

Applied both to for-next branch now.  Thanks.


Takashi