DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org DD79F2086E
Subject: Re: Problems and suggestions to the kernel module r8723bs
To:     Simon Beginn <simonmicro@simonmicro.de>,
        Bastien Nocera <hadess@hadess.net>,
        Hans de Goede <hdegoede@redhat.com>
References: <83c86a2a-0d2a-8119-2831-e9cfab7a43d7@simonmicro.de>
 <f0b67da4-747a-08e3-9744-fce2c74aaddc@lwfinger.net>
 <6e27de67-739f-ee17-eb5f-efb079d1f37f@simonmicro.de>
From:   Larry Finger <Larry.Finger@lwfinger.net>
Cc:     linux-wireless <linux-wireless@vger.kernel.org>
Message-ID: <7a47d05c-2f3b-e4d5-ccf1-5c2106d3c93d@lwfinger.net>
Date:   Wed, 26 Sep 2018 10:28:04 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.0
MIME-Version: 1.0
In-Reply-To: <6e27de67-739f-ee17-eb5f-efb079d1f37f@simonmicro.de>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On 9/26/18 3:54 AM, Simon Beginn wrote:
>>
>> The convention has long been to use rnnnnxx for the Realtek drivers in 
>> staging, and rtlnnnnxx when they are moved to the regular wireless tree. If 
>> you want it renamed, then you need to do the work to get the driver into the 
>> condition necessary for it to be in the regular tree.
> I'll see what I can do...
> 
>> Module parameter ant_sel for rtl8723be and rtl8723de was created because some 
>> vendors saved a few pennies by only supplying one antenna. The problem is that 
>> they encoded the WRONG value into the EFUSE, thus code that uses that setting 
>> (Linux) essentially has no antenna attached. The simple measure of opening the 
>> case and moving the antenna to the other connector was not viable for warranty 
>> conditions. In addition, it broke the Windows driver. Do you have a case where 
>> there is a single antenna for the rtl8723bs, and the PROM shows the wrong 
>> value? If not, then any antenna number option is not needed. 
> Hmm. I cant answer this for sure. The symptoms on my tablet are that the wifi 
> chip works OK using Windows (10) and is slow as **** under Linux. Under Linux 
> the system also didn't connect directly to my router (or any other of its 
> guest-networks) and prefers a far away adapter - which are all spread over my 
> house. And I think the chip does select the wrong antenna, because my router 
> (which the tablet doesn't like) discard all connection attempts if they are too 
> weak. The, as mentioned before, other adapters doesn't have such a feature AND 
> the wifi works only well if I stand directly beside the "i-take-everything" 
> adapters. To make this analysis stand on something: How can I read the PROM 
> values? The tablets BIOS is not the key - it only provides the EFI selection.

This driver is not reading any PROM data. It sets the antenna to 1 unless it is 
overridden by the rtw_ant_num option. We need to use signal strengths to decide 
if using ant_num 2 would be better. Use the following command to get the 
baseline values:

sudo iw dev wlan0 scan | egrep "SSID|signal"

Change the wlan0 to the actual device name on your system.

Next, do the command sequence

sudo modprobe -rv r8723bs
sudo modprobe -v r8723bs rtw_ant_num=2

Then repeat the iw command.

If the second set of signals is something like 30 dBm greater, then prepare an 
options file to force antenna #2. Post all the scan data for the AP you want to use.

Larry