Skip site navigation (1)Skip section navigation (2) 
Date:      Fri, 10 May 2013 10:06:46 +0000 (UTC)
From:      Adrian Chadd <adrian@FreeBSD.org>
To:        src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org
Subject:   svn commit: r250444 - head/sys/dev/ath
Message-ID:  <201305101006.r4AA6kRh015135@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help 
Author: adrian
Date: Fri May 10 10:06:45 2013
New Revision: 250444
URL: http://svnweb.freebsd.org/changeset/base/250444

Log:
  Make sure the holding descriptor and link pointer are both freed during
  a non-loss reset.
  
  When the drain functions are called, the holding descriptor and link pointers
  are NULLed out.
  
  But when the processq function is called during a non-loss reset, this
  doesn't occur.  So the next time a DMA occurs, it's chained to a descriptor
  that no longer exists and the hardware gets angry.
  
  Tested:
  
  * AR5416, STA mode; use sysctl dev.ath.X.forcebstuck=1 to force a non-loss
    reset.
  
  TODO:
  
  * Further AR9380 testing just to check that the behaviour for the EDMA
    chips is sane.
  
  PR:		kern/178477

Modified:
  head/sys/dev/ath/if_ath.c
  head/sys/dev/ath/if_ath_tx_edma.c

Modified: head/sys/dev/ath/if_ath.c
==============================================================================
--- head/sys/dev/ath/if_ath.c	Fri May 10 09:58:32 2013	(r250443)
+++ head/sys/dev/ath/if_ath.c	Fri May 10 10:06:45 2013	(r250444)
@@ -4668,9 +4668,21 @@ ath_legacy_tx_drain(struct ath_softc *sc
 			if (sc->sc_debug & ATH_DEBUG_RESET)
 				ath_tx_dump(sc, &sc->sc_txq[i]);
 #endif	/* ATH_DEBUG */
-			if (reset_type == ATH_RESET_NOLOSS)
+			if (reset_type == ATH_RESET_NOLOSS) {
 				ath_tx_processq(sc, &sc->sc_txq[i], 0);
-			else
+				ATH_TXQ_LOCK(&sc->sc_txq[i]);
+				/*
+				 * Free the holding buffer; DMA is now
+				 * stopped.
+				 */
+				ath_txq_freeholdingbuf(sc, &sc->sc_txq[i]);
+				/*
+				 * Reset the link pointer to NULL; there's
+				 * no frames to chain DMA to.
+				 */
+				sc->sc_txq[i].axq_link = NULL;
+				ATH_TXQ_UNLOCK(&sc->sc_txq[i]);
+			} else
 				ath_tx_draintxq(sc, &sc->sc_txq[i]);
 		}
 	}

Modified: head/sys/dev/ath/if_ath_tx_edma.c
==============================================================================
--- head/sys/dev/ath/if_ath_tx_edma.c	Fri May 10 09:58:32 2013	(r250443)
+++ head/sys/dev/ath/if_ath_tx_edma.c	Fri May 10 10:06:45 2013	(r250444)
@@ -551,6 +551,22 @@ ath_edma_tx_drain(struct ath_softc *sc, 
 	 */
 	if (reset_type == ATH_RESET_NOLOSS) {
 		ath_edma_tx_processq(sc, 0);
+		for (i = 0; i < HAL_NUM_TX_QUEUES; i++) {
+			if (ATH_TXQ_SETUP(sc, i)) {
+				ATH_TXQ_LOCK(&sc->sc_txq[i]);
+				/*
+				 * Free the holding buffer; DMA is now
+				 * stopped.
+				 */
+				ath_txq_freeholdingbuf(sc, &sc->sc_txq[i]);
+				/*
+				 * Reset the link pointer to NULL; there's
+				 * no frames to chain DMA to.
+				 */
+				sc->sc_txq[i].axq_link = NULL;
+				ATH_TXQ_UNLOCK(&sc->sc_txq[i]);
+			}
+		}
 	} else {
 		for (i = 0; i < HAL_NUM_TX_QUEUES; i++) {
 			if (ATH_TXQ_SETUP(sc, i))

Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201305101006.r4AA6kRh015135>