Experts,
I will drop a message to the ath9k-devel list for the fix,
but I suspect that it will be long before the fix actually
appears in the compat-wireless (or will not appear at all).
So I wrote an alternative script to do the same thing,
to circumvent my momentum/interest from going.
------
. /lib/ar71xx.sh
board=$(ar71xx_board_name)
ar922x_disable_gpio_jtag()
{
local regidx=0x4054
echo $regidx > /sys/kernel/debug/ieee80211/$1/ath9k/regidx
regval=`cat /sys/kernel/debug/ieee80211/$1/ath9k/regval`
regval=$((regval | 0x20000))
echo regval $regval
echo $regval > /sys/kernel/debug/ieee80211/$1/ath9k/regval
}
if [ "${board}" == "wzr-hp-ag300h" ]; then
ar922x_disable_gpio_jtag phy0
ar922x_disable_gpio_jtag phy1
fi
------
I have tested it by inserting the code at the "done" case of
the set_state() function in /etc/diag.sh, but I believe there
are other more appropriate places and/or mechanism to invoke
this kind of script. It should be executed after the ath9k*.ko
modules are loaded.
I would like to hear ideas and pointers from the openwrt experts.
--- shinoda
On 2012/08/03, at 7:21, LEO Airwarosu Yoichi Shinoda wrote:
> On 2012/08/01, at 15:45, LEO Airwarosu Yoichi Shinoda wrote:
>
>> WZR-HP-AG300H
>> - Current /etc/diag.sh does not support boot/init progress
>> indication.
>>
>> - Wmac leds REQUIRES phy initialization (by means of wifi
>> command execution) to start working. After the first
>> phy initialization, they continue to work fine.
>
>
> Okay, after playing around with files under
> compat-wireless/driver/net/wireless/ath/ath9k
> using numerous printk()s, I tracked down the problem down
> to ath9k_hw_reset(), where
>
> if (AR_SREV_9280_20_OR_LATER(ah))
> REG_SET_BIT(ah, AR_GPIO_INPUT_EN_VAL, AR_GPIO_JTAG_DISABLE);
>
> is executed. After this register write, wmac leds start to work.
>
> By replicating these lines at the end of ath9k_hw_post_init(),
> which is executed during the attach phase, and where it looked
> most appropriate for an ath9k_hw novice, I could successfully
> manipulate all leds on the wmacs without bringing up corresponding
> interfaces.
>
> Being without technical reference manuals for the wmac,
> I wonder why something called INPUT has anything to do
> with outputs, and wonder if I can disable the gpio jtag
> functionality when the wmac is attached.
> But nevertheless, I now have a working image.
>
> So, what should I do next? Report to the ath9k-devel list and
> ask for technically correct patch?
>
> BTW, my current patch is as follows if someone is interested.
>
> ---
> build_dir/linux-ar71xx_generic/compat-wireless-2012-07-16/drivers/net/wireless/ath/ath9k/hw.c-orig
> 2012-07-28 13:12:14.387929391 +0900
> +++
> build_dir/linux-ar71xx_generic/compat-wireless-2012-07-16/drivers/net/wireless/ath/ath9k/hw.c
> 2012-08-03 07:08:04.019333507 +0900
> @@ -578,6 +578,9 @@
> ath9k_hw_ani_init(ah);
> }
>
> + if (AR_SREV_9280_20_OR_LATER(ah))
> + REG_SET_BIT(ah, AR_GPIO_INPUT_EN_VAL, AR_GPIO_JTAG_DISABLE);
> +
> return 0;
> }
>
> --- shinoda
>
>
>
> _______________________________________________
> openwrt-devel mailing list
> [email protected]
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
