So, driver sometimes hangs,... we think that this is reset related,
maybe...
I today did some research on that.
First of all it happens very rarely, so hard to reproduce,
but as old saying goes, this always happens, when you don't want it
to happen.
I managed to find out that there are 2 unrelated hangs in the device.
- one shows up as a gain calibration error.
when this happens, device is still accessible.
this is typical part of dmesg:
> [ 1473.024582] ath5k 0000:03:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
> [ 1473.024628] ath5k 0000:03:00.0: setting latency timer to 64
> [ 1473.027490] ath5k 0000:03:00.0: registered as 'phy0'
> [ 1473.096642] ath5k phy0: (ath5k_attach:714): devid 0x1c
> [ 1473.096663] ath5k phy0: (ath5k_desc_alloc:1297): DMA map: f483c000 (7744)
> -> 3483c000
> [ 1473.100216] wmaster0 (ath5k): not using net_device_ops yet
> [ 1473.102813] phy0: Selected rate control algorithm 'minstrel'
> [ 1473.102944] wlan0 (ath5k): not using net_device_ops yet
> [ 1473.108080] Registered led device: ath5k-phy0::rx
> [ 1473.108145] Registered led device: ath5k-phy0::tx
> [ 1473.108156] ath5k phy0: Atheros AR2425 chip found (MAC: 0xe2, PHY: 0x70)
> [ 1477.169957] ath5k phy0: (ath5k_init:2257): mode 2
> [ 1477.169973] ath5k phy0: (ath5k_stop_locked:2307): invalid 0
> [ 1477.170128] ath5k phy0: (ath5k_txq_cleanup:1511): beacon queue 0
> [ 1477.170271] ath5k phy0: (ath5k_txq_cleanup:1520): txq [0] 0, link (null)
> [ 1477.170322] rx queue 2c782340, link (null)
> [ 1477.170338] ath5k phy0: (ath5k_reset:2731): resetting
> [ 1477.784426] ath5k phy0: gain calibration timeout (2412MHz)
> [ 1478.410893] ath5k phy0: noise floor calibration timeout (2412MHz)
> [ 1478.410980] ath5k phy0: (ath5k_rx_start:1563): cachelsz 64 rxbufsize 2560
> [ 1478.414568] ath5k phy0: (ath5k_chan_set:1033): (2412 MHz) -> (2412 MHz)
> [ 1478.415009] ADDRCONF(NETDEV_UP): wlan0: link is not ready
> [ 1478.416861] ath5k phy0: (ath5k_reset:2731): resetting
> [ 1478.417012] ath5k phy0: (ath5k_txq_cleanup:1511): beacon queue 0
> [ 1478.417153] ath5k phy0: (ath5k_txq_cleanup:1520): txq [0] 0, link (null)
> [ 1478.417193] rx queue 3483c000, link f483c4e0
a suspend/resume cycle fixes the above error.
- full device hang, it happens even more rarely, but when happens,
suspend/resume cycle doesn't help.
device then returns all ff's and when I tried lspci, then device was off the
bus, it just reported ff to everyting,
after a suspend/resume cycle it showed up in lspci, but still device regs all
returned ff's
Best regards,
Maxim Levitsky
_______________________________________________
ath5k-devel mailing list
ath5k-devel@lists.ath5k.org
https://lists.ath5k.org/mailman/listinfo/ath5k-devel
