ath9k: add locking for starting the PCU on RX
authorLuis R. Rodriguez <lrodriguez@atheros.com>
Wed, 20 Oct 2010 23:07:04 +0000 (16:07 -0700)
committerJohn W. Linville <linville@tuxdriver.com>
Wed, 27 Oct 2010 19:41:14 +0000 (15:41 -0400)
There was some locking for starting some parts of
RX but not for starting the PCU. Include this otherwise
we can content against stopping the PCU.

This can potentially lead to races against different
buffers on the PCU which can lead to to the DMA RX
engine writing to buffers which are already freed.

This is part of a series that will help resolve the bug:

https://bugzilla.kernel.org/show_bug.cgi?id=14624

For more details about this issue refer to:

http://marc.info/?l=linux-wireless&m=128629803703756&w=2

Cc: stable@kernel.org
Cc: Ben Greear <greearb@candelatech.com>
Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Tested-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
drivers/net/wireless/ath/ath9k/recv.c

index e581b1f6c40c93d0e066da815e78bd345ec3e125..b3c9baf855c49b86ff314abb2063d11beb850aa0 100644 (file)
@@ -297,11 +297,11 @@ static void ath_edma_start_recv(struct ath_softc *sc)
        ath_rx_addbuffer_edma(sc, ATH9K_RX_QUEUE_LP,
                              sc->rx.rx_edma[ATH9K_RX_QUEUE_LP].rx_fifo_hwsize);
 
-       spin_unlock_bh(&sc->rx.rxbuflock);
-
        ath_opmode_init(sc);
 
        ath9k_hw_startpcureceive(sc->sc_ah, (sc->sc_flags & SC_OP_OFFCHANNEL));
+
+       spin_unlock_bh(&sc->rx.rxbuflock);
 }
 
 static void ath_edma_stop_recv(struct ath_softc *sc)
@@ -504,10 +504,11 @@ int ath_startrecv(struct ath_softc *sc)
        ath9k_hw_rxena(ah);
 
 start_recv:
-       spin_unlock_bh(&sc->rx.rxbuflock);
        ath_opmode_init(sc);
        ath9k_hw_startpcureceive(ah, (sc->sc_flags & SC_OP_OFFCHANNEL));
 
+       spin_unlock_bh(&sc->rx.rxbuflock);
+
        return 0;
 }