[PATCH v7 08/16] ALSA: virtio: account staged capture reads before RX refill

[Joonho Lee <mynameisjune111@gmail.com>]

The RX callback holds the queue lock across a batch of completions. A
reader that has copied samples can wait for that lock in ACK before its
consumed bytes are removed from capture_ready. A later completion in the
same batch can then reject data using stale ring occupancy.

Account the staged read under the substream lock before acquiring the RX
queue lock. Release that lock before taking the queue and substream locks
in their established order to refill requests.

Once accounting is committed, latch refill errors as XRUN and return
success from ACK. A negative return would make ALSA restore appl_ptr while
the driver retained the consumed-byte accounting. Keep pre-commit errors,
ring-capacity checks, and normal period-work scheduling unchanged.

Source-extracted host and kernel fixtures reproduce the stale-capacity
race and post-commit ALSA pointer rollback. The candidate preserves wrapped
data and pointer accounting. KASAN/lockdep QEMU lifecycle checks passed.
The fixtures substitute selected backend helpers and do not independently
establish physical-device acceptance.

This follows pending staged capture support, whose introducing patch has
no upstream commit ID for a Fixes tag yet.

Link: https://lore.kernel.org/all/rfc-v2-vz-capture-compat-1.b9ab4c1540dc4691@gmail.com/
Assisted-by: LLM
Signed-off-by: Joonho Lee <mynameisjune111@gmail.com>
---
 sound/virtio/virtio_pcm_ops.c | 128 ++++++++++++++++++++--------------
 1 file changed, 77 insertions(+), 51 deletions(-)

diff --git a/sound/virtio/virtio_pcm_ops.c b/sound/virtio/virtio_pcm_ops.c
index 003832d112551..85e7576903d56 100644
--- a/sound/virtio/virtio_pcm_ops.c
+++ b/sound/virtio/virtio_pcm_ops.c
@@ -1229,70 +1229,96 @@ static int virtsnd_pcm_pb_ack(struct snd_pcm_substream *substream)
 	return 0;
 }
 
-static int virtsnd_pcm_cp_ack(struct snd_pcm_substream *substream)
+static int virtsnd_pcm_capture_staged_ack(struct virtio_pcm_substream *vss,
+					  struct snd_pcm_runtime *runtime)
 {
-	struct virtio_pcm_substream *vss = snd_pcm_substream_chip(substream);
 	struct virtio_snd_queue *queue = virtsnd_pcm_queue(vss);
-	struct snd_pcm_runtime *runtime = substream->runtime;
-	struct virtsnd_pcm_transfer transfer = {};
-	int rc;
+	int rc = 0;
+
+	/*
+	 * The RX callback holds the queue lock while draining completions.
+	 * Account ALSA's read under the substream lock alone so completions later
+	 * in that drain observe the newly released ring space.
+	 */
+	scoped_guard(spinlock_irqsave, &vss->lock) {
+		snd_pcm_uframes_t appl_ptr;
+		snd_pcm_sframes_t diff;
+		size_t bytes;
+		u64 next_acked_total;
+
+		if (!vss->runtime_active || !runtime || !runtime->control ||
+		    !runtime->dma_area)
+			return -EBADFD;
+		if (vss->xfer_xrun)
+			return -EPIPE;
+		if (!virtsnd_pcm_capture_counters_valid(vss)) {
+			vss->xfer_xrun = true;
+			schedule_work(&vss->elapsed_period);
+			return -EPIPE;
+		}
 
+		appl_ptr = runtime->control->appl_ptr;
+		diff = appl_ptr - vss->pcm_indirect.appl_ptr;
+		if (diff < -(snd_pcm_sframes_t)(runtime->boundary / 2))
+			diff += runtime->boundary;
+		if (diff < 0 ||
+		    (snd_pcm_uframes_t)diff > runtime->buffer_size) {
+			vss->xfer_xrun = true;
+			schedule_work(&vss->elapsed_period);
+			return -EPIPE;
+		}
+
+		/* ACK includes any subperiod tail already copied to the ring. */
+		bytes = frames_to_bytes(runtime, diff);
+		if (bytes > vss->capture_ready ||
+		    check_add_overflow(vss->capture_acked_total, (u64)bytes,
+				       &next_acked_total) ||
+		    next_acked_total > vss->capture_copied_total) {
+			vss->xfer_xrun = true;
+			schedule_work(&vss->elapsed_period);
+			return -EPIPE;
+		}
+
+		vss->capture_acked_total = next_acked_total;
+		vss->capture_ready -= bytes;
+		vss->pcm_indirect.appl_ptr = appl_ptr;
+	}
+
+	/* Preserve the queue -> substream lock order while refilling requests. */
 	scoped_guard(spinlock_irqsave, &queue->lock) {
 		guard(spinlock)(&vss->lock);
-		if (vss->capture_staging) {
-			snd_pcm_uframes_t appl_ptr;
-			snd_pcm_sframes_t diff;
-			size_t bytes;
-			u64 next_acked_total;
-
-			if (!vss->runtime_active || !runtime || !runtime->control ||
-			    !runtime->dma_area)
-				return -EBADFD;
-			if (vss->xfer_xrun)
-				return -EPIPE;
-			if (!virtsnd_pcm_capture_counters_valid(vss)) {
+		if (vss->runtime_active && vss->xfer_enabled &&
+		    !vss->xfer_xrun) {
+			rc = virtsnd_pcm_capture_send(vss);
+			if (rc) {
 				vss->xfer_xrun = true;
 				schedule_work(&vss->elapsed_period);
-				return -EPIPE;
 			}
-			appl_ptr = runtime->control->appl_ptr;
-			diff = appl_ptr - vss->pcm_indirect.appl_ptr;
+		}
+	}
 
-			if (diff < -(snd_pcm_sframes_t)(runtime->boundary / 2))
-				diff += runtime->boundary;
-			if (diff < 0 ||
-			    (snd_pcm_uframes_t)diff > runtime->buffer_size) {
-				vss->xfer_xrun = true;
-				schedule_work(&vss->elapsed_period);
-				return -EPIPE;
-			}
+	if (rc)
+		dev_warn_ratelimited(&vss->snd->vdev->dev,
+				     "SID %u: failed to refill capture queue after ACK: %d\n",
+				     vss->sid, rc);
 
-			/* ACK includes any subperiod tail already copied to the ring. */
-			bytes = frames_to_bytes(runtime, diff);
-			if (bytes > vss->capture_ready ||
-			    check_add_overflow(vss->capture_acked_total, (u64)bytes,
-					       &next_acked_total) ||
-			    next_acked_total > vss->capture_copied_total) {
-				vss->xfer_xrun = true;
-				schedule_work(&vss->elapsed_period);
-				return -EPIPE;
-			}
+	/* The ACK is committed; a refill failure is reported by the next XRUN. */
+	return 0;
+}
 
-			vss->capture_acked_total = next_acked_total;
-			vss->capture_ready -= bytes;
-			vss->pcm_indirect.appl_ptr = appl_ptr;
-			if (vss->xfer_enabled) {
-				rc = virtsnd_pcm_capture_send(vss);
-				if (rc) {
-					vss->xfer_xrun = true;
-					schedule_work(&vss->elapsed_period);
-				}
-				return rc;
-			}
+static int virtsnd_pcm_cp_ack(struct snd_pcm_substream *substream)
+{
+	struct virtio_pcm_substream *vss = snd_pcm_substream_chip(substream);
+	struct virtio_snd_queue *queue = virtsnd_pcm_queue(vss);
+	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct virtsnd_pcm_transfer transfer = {};
+	int rc;
 
-			return 0;
-		}
+	if (vss->capture_staging)
+		return virtsnd_pcm_capture_staged_ack(vss, runtime);
 
+	scoped_guard(spinlock_irqsave, &queue->lock) {
+		guard(spinlock)(&vss->lock);
 		transfer.indirect = vss->pcm_indirect;
 		if (vss->stopped) {
 			rc = -EPIPE;