openwrt / openwrt/mt76

mt7915: wifi unresponsive - backlog fills - no tx activity

Open
#1,009 17 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C
Stars
888
Forks
436
PR merge metrics
No merged PRs in 30d

Description

It seems, that the same behavior as described by @ptpt52 in https://github.com/openwrt/mt76/issues/518#issuecomment-817173322 does not only affect mt7615 but also mt7915e.
Some packets are buffered in the txq backlog and the txq is blocked.

iw phy phy1 get txq shows that there are often a lot of packages in Backlog.
Sometimes, the backlog is cleared after a few minutes, but in busy environments this takes far too long and gets the system into an unresponsive state.
Pings are not responend, trafffic is stalled.

I did not see anything in the queues though.

root@ffac:~# cat /sys/kernel/debug/ieee80211/phy0/mt76/xmit-queues
     queue | hw-queued |      head |      tail |
      MAIN |         1 |        13 |        12 |
     MCUWM |         0 |        20 |        20 |
     MCUWA |         0 |         9 |         9 |
   MCUFWDL |         0 |        51 |        51 |
root@ffac:~# cat /sys/kernel/debug/ieee80211/phy0/mt76/rx-queues
     queue | hw-queued |      head |      tail |
         0 |      1535 |       626 |       627 |
         1 |       511 |        19 |        20 |
         2 |       511 |       218 |       219 |
         3 |      1535 |      1107 |      1108 |
         4 |       511 |       511 |         0 |
         5 |       511 |        71 |        72 |

I experienced these issues on a Zyxel NWA 55 AXE as well as on an Acer Vero W6M.
This happens on 5GHz as well as on 2.4GHz ifaces, depending on where the higher load is.
Other devices with mt7915 like COVR-X1860 and DAP-X1860 are not affected on openwrt-24.10

Workarounds:

  • It helps to restart the wifi iface using wifi
  • It does help to run echo 7 > /sys/kernel/debug/ieee80211/phy0/mt76/sys_recovery

What does not help:

  • other recovery modes (1-6) did not help
  • turning off either 2.4GHz or 5GHz did not help
  • turning off the additional mesh ifaces did not help

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the stall on the affected mt7915 devices and inspect iw phy phy1 get txq alongside the debugfs xmit-queues and rx-queues paths shown in the report. Compare the listed sys_recovery modes, including echo 7, while monitoring traffic. Done means the txq backlog no longer leaves the interface unresponsive under load, with recovery behavior verified.

Written by the indexing model from the issue text.

Assessment

Tech stack
c
Domain
networking, operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.