Announcement

Collapse
No announcement yet.

Realtek 802.11ax WiFi Driver "rtw89" Queued Ahead Of Linux 5.16

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by direc85 View Post

    I've always had great experience (and luck) with Intel chips. Now that I use an AX201, and its Bluetooth stability is barely adequate, I think that's the only real issue I have ever had.

    Realtek... I should zip my lips. I have not found a stable USB WiFi adapter yet. No matter the OS; Windows and Linux both suffer from the poor quality of the drivers and/or chips. My current ZyXEL w/Realtek is a pain to set up. I have to use an out of the tree driver from GitHub with a pinned version in order for it to work at all, and even then it's once or twice a week I'll have to replug it or reboot the computer. (Model details coming later today.)

    The few times I have used an Atheros adapter it has gone quite well, not much to say about it.

    A few Ralink chips has also passed my hands, they also worked without any considerable fight.

    My second laptop used a Broadcom chip, with b32-legacy driver and fwcutter and all the fun stuff, but it did work like a clock!

    Does anyone know if there are any Intel USB WiFi adapters available? I couldn't find any...
    There were stability issues with (at least) the 9000 and AX series, however those were resolved, I believe in Linux 5.14.

    Regarding Realtek, I went through the same thing with the RTL8211AE back in the day, and have read (and helped out people) with Realtek WiFi and Bluetooth in the past few years, so I can confidently say that Realtek WiFi is less problematic on Linux (even with all the headaches), due to volunteers fixing Realtek's issues (their code quality and coding practices are pretty bad), so over time it gets better (but always underperforming).

    Qualcomm/Atheros is certainly stable, though they usually are behind in specs when it comes to products that are not things like their router and mobile SoCs (and I still don't get why they do not have any products that support 160MHz channels).

    Mediatek/Ralink, I believe should be alright (since Mediatek WiFi is based upon Ralink WiFi, and based upon my experience with the Xiaomi Mi 3 and Mi 3G (V1) routers) over time (Mediatek is a known violator of the GPL, however they at least do not pull the same tricks that Broadcom/Avago and NVidia do, so the community can work with Mediatek netowrking), though not as performant as things from Intel, Qualcomm and Broadcom.

    There are no Intel USB WiFi adapters, since the WLAN part of their solutions do not support any protocol other than PCIe. If you want USB, the best you can find are Broadcom and Qualcomm (both appear to be utterly rare, with it being easier to use an SDIO to USB adapter with their SDIO cards.

    That reminds me, I need to raise some money outside of my usual budget to order some QCA9377 SDIO chips (in several different sizes) to solder to some boards (Smart TV, Tablets, all with either Android or Windows) instead of the Realtek garbage there.
    Last edited by moriel5; 12 December 2022, 12:33 PM. Reason: Accidentally wrote Realtek instead of Atheros, so fixed.

    Comment


    • #22
      Originally posted by moriel5 View Post

      Realtek WiFi on Linux, takes some time to be brought up on Linux (no thanks to Realtek themselves, it is volunteers such as lwfinger that fix Realtek's garbage code)
      This is not borne out by my current experience of using this driver rtw89 on a new Linux installation. There are multiple issues with the driver (low latency, sometimes high power usage, disconnections) and there is no direct bug reporting procedure to the developers. lwfinger has said repeatedly in issues he/she did not write the driver code, can not debug it and is not responsible for it.

      Comment


      • #23
        Intel will always be the best working drivers in linux I believe.

        for qualcom I have ,
        ath11k qca6391 , and wcn6856 for ath11k.
        For ath10k I have qca9984 standard size , qc9984 cus239 , qca9880 , qca6174.

        Currently the ath10k chips are flawless and work equal to intel but that wasn't always the case it did take a while to get up and running proper board files. The candela firmware worked a bit better for most the ath10k chips. A few years back anyways.

        For ath11k the qca6391 wifi is fine but there are some strange things you need to do. Virtualization needs to be turned on and used to have to map out memory for the card to work. Bluetooth never worked on this card. It uses UART6 for bluetooth which could be a issue why I could never get it to work. I have been using this chip for a year or so maybe longer it works fine and is very low power consumption compared to the intel 7260.

        The wcn6856 chip I have bluetooth currently works but so far I have not been able to get a proper board file for this to work. This chip does support 160mhz and is triband. I have not been able to test this chip all that much due to not being able to find a working board file.

        Qualcomm from my understanding is not open source firmware but the driver is open source. Qualcomm is odd in that they have chips qca6390 and qca6391 and wcn6855 and wcn6856 but it is not clear to me what the differences are.

        Comment


        • #24
          Originally posted by dan77g View Post

          This is not borne out by my current experience of using this driver rtw89 on a new Linux installation. There are multiple issues with the driver (low latency, sometimes high power usage, disconnections) and there is no direct bug reporting procedure to the developers. lwfinger has said repeatedly in issues he/she did not write the driver code, can not debug it and is not responsible for it.
          That is true, lwfinger usually hunts down GPL-compatible drivers released by OEMs, and does not write the code themselves, however this allows others to repair the code.

          Comment


          • #25
            Originally posted by Spike_Spiegel View Post
            Intel will always be the best working drivers in linux I believe.

            for qualcom I have ,
            ath11k qca6391 , and wcn6856 for ath11k.
            For ath10k I have qca9984 standard size , qc9984 cus239 , qca9880 , qca6174.

            Currently the ath10k chips are flawless and work equal to intel but that wasn't always the case it did take a while to get up and running proper board files. The candela firmware worked a bit better for most the ath10k chips. A few years back anyways.

            For ath11k the qca6391 wifi is fine but there are some strange things you need to do. Virtualization needs to be turned on and used to have to map out memory for the card to work. Bluetooth never worked on this card. It uses UART6 for bluetooth which could be a issue why I could never get it to work. I have been using this chip for a year or so maybe longer it works fine and is very low power consumption compared to the intel 7260.

            The wcn6856 chip I have bluetooth currently works but so far I have not been able to get a proper board file for this to work. This chip does support 160mhz and is triband. I have not been able to test this chip all that much due to not being able to find a working board file.

            Qualcomm from my understanding is not open source firmware but the driver is open source. Qualcomm is odd in that they have chips qca6390 and qca6391 and wcn6855 and wcn6856 but it is not clear to me what the differences are.
            Almost no one is open source, including Intel (just the drivers are open source), Qualcomm merely has chipsets endorsed by the FSF because they can run without firmware loaded from the drive (closed source blob). The only difference between Intel and Broadcom and Avago in this regard is that Intel's firmwares are redistributable blobs while Broadcom/Avago's blobs are not redistributable without a proprietary CLA.

            I believe that there may have been a Ralink chipset in the past that people got to run on it open source firmware, however that would have been before I started to read up on these matters, so I haven't gotten around to it yet, and thus do not know for sure,

            Comment


            • #26
              On my new Xiaomi Book Pro 14 Ryzen with Realtek b852 using newest kernel Linux 6.1 Wifi is not recognized at all. Can anyone give me a hint whom I should contact to check if the device ID is missing or something like that?

              Code:
              01:00.0 Network controller: Realtek Semiconductor Co., Ltd. Device b852
                      Subsystem: Device 1d3e:c826
                      Flags: fast devsel, IRQ 255, IOMMU group 11
                      I/O ports at 2000 [disabled] [size=256]
                      Memory at c0900000 (64-bit, non-prefetchable) [disabled] [size=1M]
                      Capabilities: [40] Power Management version 3
                      Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit+
                      Capabilities: [70] Express Endpoint, MSI 00
                      Capabilities: [100] Advanced Error Reporting
                      Capabilities: [148] Device Serial Number 00-e0-4c-ff-fe-88-52-01
                      Capabilities: [158] Latency Tolerance Reporting
                      Capabilities: [160] L1 PM Substates
              ​

              Comment


              • #27
                Originally posted by Junix View Post
                On my new Xiaomi Book Pro 14 Ryzen with Realtek b852 using newest kernel Linux 6.1 Wifi is not recognized at all. Can anyone give me a hint whom I should contact to check if the device ID is missing or something like that?

                Code:
                01:00.0 Network controller: Realtek Semiconductor Co., Ltd. Device b852
                Subsystem: Device 1d3e:c826
                Flags: fast devsel, IRQ 255, IOMMU group 11
                I/O ports at 2000 [disabled] [size=256]
                Memory at c0900000 (64-bit, non-prefetchable) [disabled] [size=1M]
                Capabilities: [40] Power Management version 3
                Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit+
                Capabilities: [70] Express Endpoint, MSI 00
                Capabilities: [100] Advanced Error Reporting
                Capabilities: [148] Device Serial Number 00-e0-4c-ff-fe-88-52-01
                Capabilities: [158] Latency Tolerance Reporting
                Capabilities: [160] L1 PM Substates
                ​
                Driver for Realtek 8852AE, an 802.11ax device. Contribute to lwfinger/rtw89 development by creating an account on GitHub.


                this might help maybe the section about modprobe and the bottom about the firmware.

                Comment

                Working...
                X