Larry Finger <Larry.Finger@xxxxxxxxxxxx> writes:
> On 02/29/2016 06:28 AM, Jes Sorensen wrote:
>> That one I have never seen before - could you try and insert some debug
>> prints to see where the RF initialization fails?
>
> The call to usb_control_msg() is returning -EPROTO (-71), but
> sometimes the system works. I added the following patch:
>
> diff --git a/drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c
> b/drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c
> index 38aa3c2..8ab77a0 100644
> --- a/drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c
> +++ b/drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c
> @@ -1979,9 +1979,12 @@ static int rtl8xxxu_write_rfreg(struct rtl8xxxu_priv *priv,
>
> /* Use XB for path B */
> ret = rtl8xxxu_write32(priv, rtl8xxxu_rfregs[path].lssiparm, dataaddr);
> - if (ret != sizeof(dataaddr))
> + if (ret != sizeof(dataaddr)) {
> + pr_info("**** rtl8xxxu_write32() error %d: path %d,
> reg: 0x%x, data 0x%x, dataaddr 0x%x, lssiparm 0x%x\n",
> + ret, path, reg, data, dataaddr,
> rtl8xxxu_rfregs[path].lssiparm);
> + dump_stack();
> retval = -EIO;
> - else
> + } else
> retval = 0;
>
> udelay(1);
>
> When the error happens, my log contains
Odd, I wonder if my code writes the reigsters out to the device too
fast. You could try adding a call to read back the RF register before
exiting the function - preferably after the udelay().
Cheers,
Jes
--
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
