On Sat, Nov 1, 2008 at 8:33 PM, Iordan Iordanov <iiordanov@xxxxxxxxx> wrote: > Hi, > > I am not a member of the mailing list yet, so please cc me on any > communication regarding this bug report. I wanted to get the word out > as soon as I can. I installed an AR5008 adapter in my laptop tonight. > lspci shows: > > 0c:00.0 Network controller: Atheros Communications Inc. AR5008 > Wireless Network Adapter (rev 01) > > I am running Ubuntu 8.10, but I grabbed the latest source code > (compat-wireless-2008-10-31), built the drivers, installed them, and > loaded the newly built ath9k. > > I connected to a WPA2 protected network without a problem. However, > when I used scp to copy a large file 30 GB) using the wireless > adapter, the kernel started reporting this: > > [18461.865129] ath9k: 0.1 > [18461.865244] ath9k 0000:0c:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 > [18461.865333] ath9k 0000:0c:00.0: setting latency timer to 64 > [18462.007742] phy0: Selected rate control algorithm 'ath9k_rate_control' > [18462.010958] Registered led device: ath9k-phy0:radio > [18462.012169] Registered led device: ath9k-phy0:assoc > [18462.013374] Registered led device: ath9k-phy0:tx > [18462.014500] Registered led device: ath9k-phy0:rx > [18462.016345] phy0: Atheros 5416: mem=0xf90a0000, irq=17 > [18462.026126] udev: renamed network interface wlan0 to wlan3 > [18466.076165] ADDRCONF(NETDEV_UP): wlan3: link is not ready > [18469.168304] wlan3: authenticate with AP 00:14:bf:20:8e:36 > [18469.170528] wlan3: authenticated > [18469.170538] wlan3: associate with AP 00:14:bf:20:8e:36 > [18469.174538] wlan3: RX AssocResp from 00:14:bf:20:8e:36 (capab=0x411 > status=0 aid=2) > [18469.174546] wlan3: associated > [18469.182436] ADDRCONF(NETDEV_CHANGE): wlan3: link becomes ready > [18477.579665] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.581659] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.583162] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.585725] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.587235] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.588736] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.592405] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.593970] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.595999] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.597870] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.599936] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.602325] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.604278] ath_tx_prepare: TX queue: 1 is full, depth: 1004 This is not critical and common, we were supposed to change this to a debug message only so you wouldn't see it. When it occurs we just tell the upper layers to hold its horses. > . > . > . > > [18477.719308] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.720793] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.726773] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.729087] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > [18477.730679] ath_tx_prepare: TX queue: 1 is full, depth: 1004 > etc., etc., etc. > > At this point, I decided to put a nasty hack into the source code to > see if the error disappears. In core.h > > compat-wireless-2008-10-31/drivers/net/wireless/ath9k/core.h > > I changed: > > #define ATH_TXBUF 512 > > to > > #define ATH_TXBUF 2048 You are increasing the size the driver uses with this too, which we want to avoid. > and sure enough, so far I've copied a few GB's and the error has not recurred. I take it you don't see any throughput performance enhancements though right? > On a lighter note, can anybody enlighten me as to why for an AR5008, I > had to load the ath9k and not the ath5k driver?!! This really puzzles > me. 11n Atheros devices begin on AR9xxx, AR5416 and AR5008 (although I think this is just mislabeled and should be AR5416) are the first generation 11n devices and escaped the new nomenclature. Luis -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html