MediaTek MT7925 - USB WiFi 7 EDUP BE1679 BE6500

[ejsc1]

Hello,

How does this dongle work on Linux 7.2?
https://es.aliexpress.com/item/1005012057293862.html
https://www.szedup.com/usb-adaptersBr3/EPBE1679.html?act=specification

Thanks!

[Lucid-Duck]

@ejsc1

Short answer: yes. On 7.2 it should work the moment you plug it in. The chip is a MediaTek MT7925, and Linux has shipped the driver for it since 6.7, so there is nothing to download or build.

The one thing that can spoil that is the id the adapter reports about itself. Linux picks the driver by that id, and this one only recognises four of them. If EDUP put their own id on it rather than MediaTek's, Linux will see the adapter but load nothing for it. If you have one in hand, run lsusb and paste the line here. If the id turns out to be the problem, it is a couple of lines to add rather than a dead end and we can help you out with that here.

One caveat if you are buying it for WiFi 7 in particular. The feature that runs two bands at once (MLO) stops passing traffic after a few minutes on 7.1 and later, while still looking connected. That one is actively being worked: the cause has been narrowed to a single commit and the MediaTek developer behind it is on the thread, so it is moving. Ordinary use on a single band is fine in the meantime.

[Ashcal9669]

@Lucid-Duck I have narrowed down and got 2 links working simultaneously. I'm just cleaning it up for submission. Will post later. Thanks

[morrownr]

Hi @ejsc1

I have been watching that specific adapter for some time. It seems to meet the criteria to be in The Plug and Play List but I have just not found enough evidence to add it to the list. My experience is that Edup generally makes reasonably good adapters... and some are very good. If I you buy it, please return here and have a discussion with us to that we can make a decision on including it in The Plug and Play List.

[morrownr]

@ejsc1

My bad: I should have requested how you are going to use the adapter before answering as this can make a big difference.

What mode will you use? managed (client), AP, monitor?
Is this for a laptop/desktop/RasPi?
Do you have a USB3 extension cable to go with it so that you can position it for best results?

[ejsc1]

Hi @morrownr

What mode will you use? managed (client), AP, monitor?

client

Is this for a laptop/desktop/RasPi?

$ hostnamectl
Static hostname:
Icon name: computer-convertible
Chassis: convertible
Chassis Asset Tag: Chassis Asset Tag
Machine ID: 471a075689204c19b59ab27f1c968cd6
Boot ID: f07d7ae78a934163b5c6c7591e4cfaf1
Operating System:
Kernel: Linux 7.2.0
Architecture: x86-64
Hardware Vendor: HP
Hardware Model: HP Pavilion x360 Convertible 14-cd0xxx
Hardware SKU: 3PX63LA#ABM
Hardware Version: Type1ProductConfigId
Firmware Version: F.41
Firmware Date: Wed 2022-11-16
Firmware Age: 3y 9month 1w 2d

Do you have a USB3 extension cable to go with it so that you can position it for best results?

Yes

[morrownr]

@ejsc1

Let me see: client mode, x86_64, USB3 extension cable and kernel 7.2. The adapter should be plug and play.

If you run into problems. you know where to find us.

[ejsc1]

Hello,

~$ lsusb
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 001 Device 002: ID 04f2:b627 Chicony Electronics Co., Ltd HP Wide Vision HD Camera
Bus 001 Device 003: ID 0bda:b00a Realtek Semiconductor Corp. Realtek Bluetooth 4.2 Adapter
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 002 Device 005: ID 0e8d:7925 MediaTek Inc. Wireless_Device

It's working without compiling anything. As I was told, plug and play. I had some trouble setting up the dual WiFi, but I managed to solve it by turning off the old WiFi 5.

Do you need me to run any tests? Please tell me the commands you need me to run, and I'll send you the results.

Thanks!

[Ashcal9669]

@ejsc1 Thanks for the follow-up and the lsusb output — that confirms the stock MediaTek ID (0e8d:7925), so you're on the fully-supported path.

On the dual-band trouble: that lines up with an MLO issue I've been chasing since my Aug 25 update. Real progress since then — root-caused and fixed on real hardware (a related MT7927E chip, same driver code path):

• A firmware band-context collision that broke 5GHz+6GHz specifically
• An MLO remain-on-channel lifecycle bug that silently blocked link transitions after the first one
• A late-added link not receiving its encryption keys (both unicast and broadcast/multicast), which is likely what you hit — it'd look exactly like "dual WiFi doesn't work" even though the connection looks fine

These are committed on my fork but not upstream-clean yet (still has test scaffolding mixed in) — cleaning that up now.

To help nail down whether your case is the same bug: next time you hit the dual-WiFi issue (before disabling WiFi 5), could you grab sudo dmesg | tail -100 right at the point it breaks, and iw dev <iface> link to show which links are active? That'll tell me if it's the same key-reconciliation issue or something adapter-specific.

@morrownr hope you're doing alright — appreciate everything you've put into this project and list.

[ejsc1]

@Ashcal9669

sudo dmesg tail 100 .txt

Is this log useful? I couldn't connect; I had to unplug and plug the dongle back in for it to work properly.

[Ashcal9669]

@ejsc1 Thanks — but this one's actually a different issue than the dual-WiFi problem, and useful in its own right.

What this log shows: the adapter's Bluetooth side (hci0) is failing to initialize (Opcode 0x0c03 failed: -110, repeated ~5 times), and each retry triggers a full USB reset of the whole device — including the WiFi side. It only recovers once you physically unplug/replug and it re-enumerates as a fresh USB device. No MLO or dual-link activity shows up here at all — this is a single-band association at the very end, after the reset loop finally settles.

That's a BT/WiFi co-enumeration issue on this adapter, separate from the dual-band bug I mentioned. Not something I can fix in the wifi driver — it's on the Bluetooth init path. Worth its own report if it keeps happening (does it happen every boot, or only sometimes?).

For the actual dual-WiFi issue: could you reproduce that specifically (get both bands connected, then wait for it to break) and grab sudo dmesg | tail -100 at that point, plus iw dev <iface> link while it's still associated but not passing traffic? That's the one I need to compare against.

[morrownr]

@ejsc1

I see from the log you sent that your firmware, wifi and bt, are dated 20260106. That is actually a few versions old. One thing this repo does during installation is that it updates all of your mt76 firmware to the latest release. In this case the latest release was about a week ago so it is hot off the stove.

The instructions for installation are near the top of the README. It should be an easy installation and the uninstallation is easy. Installing the driver in this repo would allow us to see the state of the situation based on the very latest firmware and driver.

It is interesting that this adapter has bt support. Adapters that include bt support are not something that I normally add to The Plug and Play List as it can be problematic depending on how you intend to use the adapter. If you plan on using bt for streaming and also use the 2.4 GHz band for wifi, there will be sharing which will slow things down and then there is the issue of using USB3 ports which can cause emi on the 2.4 GHz band. I wish I had known this adapter had bt support as I would have warned you. However, it you plan to use wifi only on the 5 and 6 GHz bands, it could be okay for you.

Please give us additional reports after installing the driver. Thanks.

@Ashcal9669

hope you're doing alright — appreciate everything you've put into this project and list.

Thank you for the kind words. I hate that my health is slowing me down but that is life. At least I was able to engage with this issue today. You and Devin and doing a great job.

[Lucid-Duck]

@ejsc1 @Ashcal9669 Two things from my bench that might save some time.

The all-zero version line isn't a symptom. btmtk reads that out of the firmware file, not the chip, so it tells you which firmware image loaded and nothing about the hardware. Another MediaTek controller here prints the same zeros, sets up in under half a second and works fine. The Enhanced Setup warning turns up on my healthy ones too.

The timing is the interesting part. Setup takes 3.7 seconds on yours, where three controllers here are all under half a second. And the first thing that actually fails is the iso setting command, with everything after it timing out too. Then the whole device gets reset, the wifi tears down and re-probes with it, and it goes round again about every ten seconds. That loop is why the link won't hold.

Thanks @ejsc1 for pulling the log, and @Ashcal9669 for spotting the Bluetooth side.