Re: [PATCH net v3] net: airoha: fix BQL underflow in shared QDMA TX ring

On Sat, Jun 20, 2026 at 05:04:51PM +0200, Lorenzo Bianconi wrote:

When multiple netdevs share a QDMA TX ring and one device is stopped,
netdev_tx_reset_subqueue() zeroes that device's BQL counters while its
pending skbs remain in the shared HW TX ring. When NAPI later completes
those skbs via netdev_tx_completed_queue(), the already-zeroed
dql->num_queued counter underflows.

Fix the issue:
- Remove netdev_tx_reset_subqueue() from airoha_dev_stop() so pending
  skbs are completed naturally by NAPI with proper BQL accounting.
- Rework airoha_qdma_tx_cleanup() to disable TX DMA, flush BQL
  counters, DMA-unmap and free all pending skbs while skb->dev
  references are still valid. Use a per-queue flushing flag checked
  under q->lock in airoha_dev_xmit() to prevent races between teardown
  and transmit. Call airoha_qdma_stop_napi() before
  airoha_qdma_tx_cleanup() at the call sites.
- Move DMA engine start into probe. Split DMA teardown so TX DMA is
  disabled in airoha_qdma_tx_cleanup() and RX DMA in
  airoha_qdma_cleanup().
- Remove qdma->users counter since DMA lifetime is now tied to
  probe/cleanup rather than per-netdev open/stop.

Fixes: a9c2ca61fec7 ("net: airoha: Support multiple net_devices for a single FE GDM port")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
Changes in v3:
- Remove airoha_qdma users counter.
- Drop DEV_STATE_FLUSH bit and add per-queue flushing bool to avoid any
  race between airoha_qdma_tx_flush() and airoha_dev_xmit().
- Refactor airoha_qdma_cleanup_tx_queue().
- Rename airoha_qdma_cleanup_tx_queue() in airoha_qdma_tx_cleanup().
- Link to v2: https://lore.kernel.org/r/20260619-airoha-bql-fixes-v2-1-4351d6a24484@kernel.org (local)

Thanks for the updates.

Reviewed-by: Simon Horman <horms@kernel.org>