Hi Takashi, This cumulative resend keeps the twelve-patch series on the same base, a4311e9447bbb219044db38897f95be112f46a8a. It covers PCM stop-threshold sampling, locked partial-progress reporting, the optional reset operation, VirtIO ACK error propagation and recovery, PCM and control DMA ownership, signal-safe teardown, staged capture lifecycle and work prioritization, and capture parameter documentation.
Relative to v19, patch 6 moves the virtqueue initialization summary into the commit body and removes the unsupported event-message ownership claim. Its source diff is unchanged. Patch 7 also rejects ordinary indirect capture appl_ptr advances that are not aligned to runtime->min_align before converting the frame delta to bytes. Validation: - The final ARM64 W=1 build uses KASAN, PROVE_LOCKING, and LOCKDEP. The cumulative stage build checks patches 7–12 incrementally. - A synthetic exact-validator case rejects an odd frame delta for 4-bit samples before indirect accounting changes. An aligned sub-byte delta and a byte-aligned frame delta pass. QEMU does not advertise IMA_ADPCM, so the sub-byte case uses synthetic runtime geometry rather than a real stream. - An actual S16_LE capture accepts two half-period ACKs for one completed period and then recovers a full-period read. - Ordinary and staged capture passed RW and MMAP reset/reopen runs. The 14-segment oracle matched the QEMU sample sequence. - Inherited v19 raw SYNC_PTR and paused-lifecycle evidence covers the existing accounting state machine. The transport timeout and PCI evidence covers unchanged control and teardown paths; these references are listed in the manifest. No physical-device validation or KCSAN run was performed. Previous version: https://lore.kernel.org/r/[email protected]/ AI assistance was used for analysis, implementation, and test preparation. Assisted-by: LLM Thanks, Joonho base-commit: a4311e9447bbb219044db38897f95be112f46a8a
