[REGRESSION] [BISECTED] MT7925 wireless speed drops to zero since kernel 6.14.3 - wifi: mt76: mt7925: integrate *mlo_sta_cmd and *sta_cmd

[Mingyen Hsieh (謝明諺)]

On Fri, 2025-08-08 at 17:37 +0300, Tal Inbar wrote:
> 
> External email : Please do not click links or open attachments until
> you have verified the sender or the content.
> 
> 
> typo, laptop model is:
> Lenovo IdeaPad Slim 5 16AKP10
> 
> On Fri, Aug 8, 2025 at 4:14 PM Tal Inbar <inbartdev at gmail.com> wrote:
> > 
> > Hi,
> > 
> > Since kernel v6.14.3, when using wireless to connect to my home
> > router
> > on my laptop, my wireless connection slows down to unusable speeds.
> > 
> > 
> > More specifically, since kernel 6.14.3, when connecting to the
> > wireless networks of my OpenWRT Router on my Lenovo IdeaPad Slim 15
> > 16AKP10 laptop,
> > either a 2.4ghz or a 5ghz network, the connection speed drops down
> > to
> > 0.1-0.2 Mbps download and 0 Mbps upload when measured using
> > speedtest-cli.
> > My laptop uses an mt7925 chip according to the loaded driver and
> > firmware.
> > 
> > 
> > Detailed Description:
> > 
> > As mentioned above, my wireless connection becomes unusable when
> > using
> > linux 6.14.3 and above, dropping speeds to almost 0 Mbps,
> > even when standing next to my router. Further, pinging
> > archlinux.org
> > results in "Temporary failure in name resolution".
> > Any other wireless device in my house can successfully connect to
> > my
> > router and properly use the internet with good speeds, eg. iphones,
> > ipads, raspberry pi and a windows laptop.
> > When using my Lenovo laptop on a kernel 6.14.3 or higher to connect
> > to
> > other access points, such as my iPhone's hotspot and some TPLink
> > and
> > Zyxel routers - the connection speed is good, and there are no
> > issues,
> > which makes me believe there's something going on with my OpenWRT
> > configuration in conjunction with a commit introduced on kernel
> > 6.14.3
> > for the mt7925e module as detailed below.
> > 
> > I have followed a related issue previously reported on the mailing
> > list regarding a problem with the same wifi chip on kernel 6.14.3,
> > but
> > the merged fix doesn't seem to fix my problem:
> > https://lore.kernel.org/linux-mediatek/EmWnO5b-acRH1TXbGnkx41eJw654vmCR-8_xMBaPMwexCnfkvKCdlU5u19CGbaapJ3KRu-l3B-tSUhf8CCQwL0odjo6Cd5YG5lvNeB-vfdg=@pm.me/
> > 
> > I've tested stable builds of 6.15 as well up to 6.15.9 in the last
> > month, which also do not fix the problem.
> > I've also built and bisected v6.14 on june using guides on the Arch
> > Linux wiki, for the following bad commit, same as the previously
> > mentioned reported issue:
> > 
> > [80007d3f92fd018d0a052a706400e976b36e3c87] wifi: mt76: mt7925:
> > integrate *mlo_sta_cmd and *sta_cmd
> > 
> > Testing further this week, I cloned mainline after 6.16 was
> > released,
> > built and tested it, and the issue still persists.
> > I reverted the following commits on mainline and retested, to
> > successfully see good wireless speeds:
> > 
> > [0aa8496adda570c2005410a30df963a16643a3dc] wifi: mt76: mt7925: fix
> > missing hdr_trans_tlv command for broadcast wtbl
> > [cb1353ef34735ec1e5d9efa1fe966f05ff1dc1e1] wifi: mt76: mt7925:
> > integrate *mlo_sta_cmd and *sta_cmd
> > 
> > Then, reverting *only* 0aa8496adda570c2005410a30df963a16643a3dc
> > causes
> > the issue to reproduce, which confirms the issue is caused by
> > commit
> > cb1353ef34735ec1e5d9efa1fe966f05ff1dc1e1 on mainline.
> > 
> > I've attached the following files to a bugzilla ticket:
> > 
> > - lspci -nnk output:
> > https://urldefense.com/v3/__https://bugzilla.kernel.org/attachment.cgi?id=308466__;!!CTRNKA9wMg0ARbw!gTAaJeIN6s3eF_Kp6U3iwfv0wCCO_BVcRkNVpJYnwR8qeqIIh134mEJ5dxdeNqzuWSPF2ocLAt43LxaBTIY-7w$
> > 
> > - dmesg output:
> > https://urldefense.com/v3/__https://bugzilla.kernel.org/attachment.cgi?id=308465__;!!CTRNKA9wMg0ARbw!gTAaJeIN6s3eF_Kp6U3iwfv0wCCO_BVcRkNVpJYnwR8qeqIIh134mEJ5dxdeNqzuWSPF2ocLAt43LxY1pkjZCA$
> > 
> > - .config for the built mainline kernel:
> > https://urldefense.com/v3/__https://bugzilla.kernel.org/attachment.cgi?id=308467__;!!CTRNKA9wMg0ARbw!gTAaJeIN6s3eF_Kp6U3iwfv0wCCO_BVcRkNVpJYnwR8qeqIIh134mEJ5dxdeNqzuWSPF2ocLAt43LxYcu-Db2g$
> > 
> > 
> > More information:
> > 
> > OS Distribution: Arch Linux
> > 
> > Linux build information from /proc/version:
> > Linux version 6.16.0linux-mainline-11853-g21be711c0235
> > (tal at arch-debug) (gcc (GCC) 15.1.1 20250729, GNU ld (GNU Binutils)
> > 2.45.0) #3 SMP PREEMPT_DYNAMIC
> > 
> > OpenWRT Version on my Router: 24.10.2
> > 
> > Laptop Hardware:
> > - Lenovo IdeaPad Slim 15 16AKP10 laptop (x86_64 Ryzen AI 350 CPU)
> > - Network device as reported by lscpi: 14c3:7925
> > - Network modules and driver in use: mt7925e
> > - mediatek chip firmware as of dmesg:
> >   HW/SW Version: 0x8a108a10, Build Time: 20250526152947a
> >   WM Firmware Version: ____000000, Build Time: 20250526153043
> > 
> > 
> > Referencing regzbot:
> > 
> > #regzbot introduced: 80007d3f92fd018d0a052a706400e976b36e3c87
> > 
> > 
> > Please let me know if any other information is needed, or if there
> > is
> > anything else that I can test on my end.
> > 
> > Thanks,
> > Tal Inbar

Hi,

Can you capture a sniffer log? There should be some configuration
differences that we can identify from the packet frames during the
connection process. So please provide the sniffer logs when connecting
to your OpenWRT, TPLink, and Zyxel.

Or you can check for any configuration differences between OpenWRT and
the other routers, which might also help with debugging.

Thanks~

Yen.