Let iwx(4) use per-Tx-queue interface timers to ensure that the interface

watchdog will trigger a device timeout if a particular Tx queue gets stuck
while other Tx queues keep working.

The Linux driver is using a similar workaround for "stuck queues".

Tested by myself and jmc@

diff --git a/sys/dev/pci/if_iwx.c b/sys/dev/pci/if_iwx.c
index 9d396fe..b7072c1 100644 (file)
--- a/sys/dev/pci/if_iwx.c
+++ b/sys/dev/pci/if_iwx.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: if_iwx.c,v 1.122 2021/11/22 10:47:55 stsp Exp $       */
+/*     $OpenBSD: if_iwx.c,v 1.123 2021/11/22 10:54:36 stsp Exp $       */
 
 /*
  * Copyright (c) 2014, 2016 genua gmbh <info@genua.de>
@@ -4552,8 +4552,6 @@ iwx_rx_tx_cmd(struct iwx_softc *sc, struct iwx_rx_packet *pkt,
        bus_dmamap_sync(sc->sc_dmat, data->map, 0, IWX_RBUF_SIZE,
            BUS_DMASYNC_POSTREAD);
 
-       sc->sc_tx_timer = 0;
-
        /* Sanity checks. */
        if (sizeof(*tx_resp) > len)
                return;
@@ -4563,6 +4561,8 @@ iwx_rx_tx_cmd(struct iwx_softc *sc, struct iwx_rx_packet *pkt,
            tx_resp->frame_count * sizeof(tx_resp->status) > len)
                return;
 
+       sc->sc_tx_timer[qid] = 0;
+
        if (tx_resp->frame_count > 1) /* A-MPDU */
                return;
 
@@ -4658,7 +4658,7 @@ iwx_rx_compressed_ba(struct iwx_softc *sc, struct iwx_rx_packet *pkt,
                idx = le16toh(ba_tfd->tfd_index);
                if (idx >= IWX_TX_RING_COUNT)
                        continue;
-               sc->sc_tx_timer = 0;
+               sc->sc_tx_timer[qid] = 0;
                iwx_txq_advance(sc, ring, idx);
                iwx_clear_oactive(sc, ring);
        }
@@ -5433,6 +5433,9 @@ iwx_tx(struct iwx_softc *sc, struct mbuf *m, struct ieee80211_node *ni)
                sc->qfullmsk |= 1 << ring->qid;
        }
 
+       if (ic->ic_if.if_flags & IFF_UP)
+               sc->sc_tx_timer[ring->qid] = 15;
+
        return 0;
 }
 
@@ -7973,10 +7976,8 @@ iwx_start(struct ifnet *ifp)
                        continue;
                }
 
-               if (ifp->if_flags & IFF_UP) {
-                       sc->sc_tx_timer = 15;
+               if (ifp->if_flags & IFF_UP)
                        ifp->if_timer = 1;
-               }
        }
 
        return;
@@ -8046,7 +8047,8 @@ iwx_stop(struct ifnet *ifp)
                struct iwx_rxba_data *rxba = &sc->sc_rxba_data[i];
                iwx_clear_reorder_buffer(sc, rxba);
        }
-       ifp->if_timer = sc->sc_tx_timer = 0;
+       memset(sc->sc_tx_timer, 0, sizeof(sc->sc_tx_timer));
+       ifp->if_timer = 0;
 
        splx(s);
 }
@@ -8055,21 +8057,30 @@ void
 iwx_watchdog(struct ifnet *ifp)
 {
        struct iwx_softc *sc = ifp->if_softc;
+       int i;
 
        ifp->if_timer = 0;
-       if (sc->sc_tx_timer > 0) {
-               if (--sc->sc_tx_timer == 0) {
-                       printf("%s: device timeout\n", DEVNAME(sc));
-                       if (ifp->if_flags & IFF_DEBUG) {
-                               iwx_nic_error(sc);
-                               iwx_dump_driver_status(sc);
+
+       /*
+        * We maintain a separate timer for each Tx queue because
+        * Tx aggregation queues can get "stuck" while other queues
+        * keep working. The Linux driver uses a similar workaround.
+        */
+       for (i = 0; i < nitems(sc->sc_tx_timer); i++) {
+               if (sc->sc_tx_timer[i] > 0) {
+                       if (--sc->sc_tx_timer[i] == 0) {
+                               printf("%s: device timeout\n", DEVNAME(sc));
+                               if (ifp->if_flags & IFF_DEBUG) {
+                                       iwx_nic_error(sc);
+                                       iwx_dump_driver_status(sc);
+                               }
+                               if ((sc->sc_flags & IWX_FLAG_SHUTDOWN) == 0)
+                                       task_add(systq, &sc->init_task);
+                               ifp->if_oerrors++;
+                               return;
                        }
-                       if ((sc->sc_flags & IWX_FLAG_SHUTDOWN) == 0)
-                               task_add(systq, &sc->init_task);
-                       ifp->if_oerrors++;
-                       return;
+                       ifp->if_timer = 1;
                }
-               ifp->if_timer = 1;
        }
 
        ieee80211_watchdog(ifp);

diff --git a/sys/dev/pci/if_iwxvar.h b/sys/dev/pci/if_iwxvar.h
index 69eebcb..02bdd3a 100644 (file)
--- a/sys/dev/pci/if_iwxvar.h
+++ b/sys/dev/pci/if_iwxvar.h
@@ -1,4 +1,4 @@
-/*     $OpenBSD: if_iwxvar.h,v 1.26 2021/11/22 10:31:58 stsp Exp $     */
+/*     $OpenBSD: if_iwxvar.h,v 1.27 2021/11/22 10:54:36 stsp Exp $     */
 
 /*
  * Copyright (c) 2014 genua mbh <info@genua.de>
@@ -563,7 +563,7 @@ struct iwx_softc {
        struct iwx_nvm_data sc_nvm;
        struct iwx_bf_data sc_bf;
 
-       int sc_tx_timer;
+       int sc_tx_timer[IWX_NUM_TX_QUEUES];
        int sc_rx_ba_sessions;
 
        int sc_scan_last_antenna;