Remove dhd_doflow. iperf result without flow control is unacceptable.
Signed-off-by: Grant Grundler <[email protected]>
---
tested on ARM tegra2_seaboard.
Broadcom sent me a patch with "dhd_doflow = true" and I just don't
see a use case for "= false". ergo remove it.
diff --git a/drivers/staging/brcm80211/brcmfmac/dhd_sdio.c
b/drivers/staging/brcm80211/brcmfmac/dhd_sdio.c
index aaaf593..8f9d895 100644
--- a/drivers/staging/brcm80211/brcmfmac/dhd_sdio.c
+++ b/drivers/staging/brcm80211/brcmfmac/dhd_sdio.c
@@ -333,7 +333,6 @@ uint dhd_txminmax;
#define DONGLE_MIN_MEMSIZE (128 * 1024)
int dhd_dongle_memsize;
-static bool dhd_doflow;
static bool dhd_alignctl;
static bool sd1idle;
@@ -1119,7 +1118,7 @@ int dhd_bus_txdata(struct dhd_bus *bus, struct sk_buff
*pkt)
}
dhd_os_sdunlock_txq(bus->dhd);
- if ((pktq_len(&bus->txq) >= TXHI) && dhd_doflow)
+ if (pktq_len(&bus->txq) >= TXHI)
dhd_txflowcontrol(bus->dhd, 0, ON);
#ifdef DHD_DEBUG
@@ -1217,7 +1216,7 @@ static uint dhdsdio_sendfromq(dhd_bus_t *bus, uint
maxframes)
}
/* Deflow-control stack if needed */
- if (dhd_doflow && dhd->up && (dhd->busstate == DHD_BUS_DATA) &&
+ if (dhd->up && (dhd->busstate == DHD_BUS_DATA) &&
dhd->txoff && (pktq_len(&bus->txq) < TXLOW))
dhd_txflowcontrol(dhd, 0, OFF);
@@ -5074,7 +5073,6 @@ static void *dhdsdio_probe(u16 venid, u16 devid, u16
bus_no,
sd1idle = true;
dhd_readahead = true;
retrydata = false;
- dhd_doflow = false;
dhd_dongle_memsize = 0;
dhd_txminmax = DHD_TXMINMAX;
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
