Public bug reported: System
- ASUS Vivobook S16 S3607AA, BIOS S3607AA.302 (BIOS date shown in kernel log: 01/07/2026). - Ubuntu 26.04.1 LTS, x86_64. - Kernel: 7.0.0-31-generic, Ubuntu package 7.0.0-31.31. The kernel version string also reports upstream 7.0.14. - linux-firmware: 20260319.git217ca6e4.1ubuntu. - Wi-Fi: MediaTek MT7920, PCI ID 14c3:7920, subsystem 1a3b:6470, PCI address 0000:01:00.0, driver mt7921e. - Upstream PCIe port: 0000:00:1c.4. - Wi-Fi firmware log: ASIC revision 79200010; HW/SW 0x8a108a10; build 20260224110857a; WM firmware build 20260224110939. - GPU: Intel Panther Lake 8086:b090, driver xe. Mesa 26.0.8-1ubuntu0.3. - /sys/power/mem_sleep: [s2idle] deep. All reported sleep tests used s2idle. ## Reproduction 1. Boot normally with mt7921e loaded. Wi-Fi and the desktop work. 2. Select Suspend from the desktop menu. 3. Wake the laptop. 4. Check Wi-Fi, launch Chrome, and attempt a normal Power Off. Expected: Wi-Fi reconnects, applications remain responsive, and Power Off completes. Actual: the desktop resumes, but the Wi-Fi device fails to resume. Some applications can hang, and Power Off after suspend can fail to turn off the machine. Power Off from an ordinary session without suspend works. The failure was reproduced multiple times, but not exhaustively tested under every power-source or timing combination. ## Kernel evidence On 2026-09-15, at both 15:35 and 17:13 (UTC+03:00), resume logged: ```text pcieport 0000:00:1c.4: Data Link Layer Link Active not set in 100 msec mt7921e 0000:01:00.0: driver own failed mt7921e 0000:01:00.0: PM: dpm_run_callback(): pci_pm_resume returns -5 mt7921e 0000:01:00.0: PM: failed to resume async: error -5 ``` At 17:13:18, a warning followed in kthread_park, from the mt76 workqueue, mt7921_mac_reset_work -> mt7921e_mac_reset. An earlier incident on 2026-09-09, on the same 7.0.0-31 kernel, includes blocked-task stacks lasting over 122 seconds: - NetworkManager waits for a mutex in mt7921_start. - The kernel identifies the likely mutex owner as a worker running mt7921_mac_reset_work, stuck in napi_disable_locked via mt7921e_mac_reset. - Chrome_ChildIOT waits in rtnl_lock -> ip_mc_drop_socket. The kernel identifies NetworkManager as the likely mutex owner. This suggests a Wi-Fi recovery failure can propagate into kernel networking lock waits affecting desktop applications. I do not claim that the initial PCIe failure and every subsequent hang have been traced to a single root cause. ## Tests already performed 1. Turning Wi-Fi off in the desktop UI before suspend did not solve it. At 18:41:25, the same PCIe/resume error occurred, followed by the reset-work warning. Shutdown reached systemd-shutdown before logging ended; its exact final blocking point was not captured. 2. Unloading mt7921e before suspend allowed a sleep/resume/power-off test to complete normally. The PCIe Link Active warning remained even without the driver loaded. 3. Reloading with `modprobe mt7921e disable_aspm=1` did not solve it. The kernel confirmed `disabling ASPM L0s L1` at 19:02:21, but the same resume failure and reset warning occurred at 19:04:05-06. This was a module reload test, not a boot-command-line ASPM test. 4. A manual unload-before-sleep/reload-after-wake test was reported successful once. An automated version was then tested twice (19:24 and 19:25), and Wi-Fi failed to return both times: `probe with driver mt7921e failed with error -5`. modprobe itself returned exit status 0, despite the device probe failure. Therefore this is NOT a reliable workaround. 5. The custom automation has been disabled. Wi-Fi works again in the current normal boot. No permanent ASPM setting or kernel replacement was installed during these tests. The custom service used ExecStart=modprobe -r mt7921e before systemd- suspend.service, and ExecStop=modprobe mt7921e after it completed, with Type=oneshot, RemainAfterExit=yes and StopWhenUnneeded=yes. The attached logs show the timing. Bluetooth was blocked and bluetooth.service masked before the automated tests; earlier failures predate this change. ## Other observations / scope No xe GPU hang/reset was identified in the inspected failing resume sequences. Separate Mesa INTEL_FORCE_PROBE errors occur in Snap Store even without suspend; those may be unrelated. Boot logs also show ACPI AE_ALREADY_EXISTS and AE_NOT_FOUND errors. Their relationship to this problem is unknown. The first affected kernel version is not established. Although 7.0.0-30 was previously installed, the attached evidence is from 7.0.0-31; this is not a confirmed -30 to -31 regression. Possibly related upstream discussion (not a confirmed duplicate): https://lists.infradead.org/pipermail/linux-mediatek/2026-January/102545.html Please advise whether this should be investigated as a PCIe/platform power restoration issue, a mt7921e recovery issue, or both, and which supported kernel/firmware update or additional diagnostics would be most useful. ## Attachment `suspend-diagnostic-logs.txt` contains selected historical kernel and service logs, with the hostname and MAC addresses redacted. Timestamps are UTC+03:00. No new suspend or failure reproduction was triggered to collect this report. ProblemType: Bug DistroRelease: Ubuntu 26.04 Package: linux-image-7.0.0-31-generic 7.0.0-31.31 ProcVersionSignature: Ubuntu 7.0.0-31.31-generic 7.0.14 Uname: Linux 7.0.0-31-generic x86_64 ApportVersion: 2.34.1-0ubuntu0.1 Architecture: amd64 AudioDevicesInUse: USER PID ACCESS COMMAND /dev/snd/controlC0: slava 1803 F.... pipewire slava 1832 F.... wireplumber /dev/snd/seq: slava 1803 F.... pipewire CasperMD5CheckResult: pass CurrentDesktop: ubuntu:GNOME Date: Tue Sep 15 20:35:07 2026 InstallationDate: Installed on 2026-08-22 (24 days ago) InstallationMedia: Ubuntu 26.04 "Resolute Raccoon" - Release amd64 (20260423.1) MachineType: ASUS Vivobook S16 S3607AA ProcEnviron: LANG=en_US.UTF-8 PATH=(custom, no user) SHELL=/bin/bash TERM=xterm-256color XDG_RUNTIME_DIR=<set> ProcFB: 0 xedrmfb ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-7.0.0-31-generic root=UUID=79ac42df-6a6a-4d98-8ca9-887ed711332d ro quiet splash crashkernel=2G-4G:320M,4G-32G:512M,32G-64G:1024M,64G-128G:2048M,128G-:4096M SourcePackage: linux UpgradeStatus: No upgrade log present (probably fresh install) dmi.bios.date: 01/07/2026 dmi.bios.release: 5.39 dmi.bios.vendor: American Megatrends International, LLC. dmi.bios.version: S3607AA.302 dmi.board.asset.tag: ATN12345678901234567 dmi.board.name: S3607AA dmi.board.vendor: ASUSTeK COMPUTER INC. dmi.board.version: 1.0 dmi.chassis.asset.tag: No Asset Tag dmi.chassis.type: 10 dmi.chassis.vendor: ASUSTeK COMPUTER INC. dmi.chassis.version: 1.0 dmi.ec.firmware.release: 3.1 dmi.modalias: dmi:bvnAmericanMegatrendsInternational,LLC.:bvrS3607AA.302:bd01/07/2026:br5.39:efr3.1:svnASUS:pnVivobookS16S3607AA:pvr1.0:rvnASUSTeKCOMPUTERINC.:rnS3607AA:rvr1.0:cvnASUSTeKCOMPUTERINC.:ct10:cvr1.0:sku:pfaVivobookS16: dmi.product.family: Vivobook S16 dmi.product.name: Vivobook S16 S3607AA dmi.product.version: 1.0 dmi.sys.vendor: ASUS ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Tags: amd64 apport-bug resolute wayland-session ** Attachment added: "suspend-diagnostic-logs.txt" https://bugs.launchpad.net/bugs/2167384/+attachment/6000439/+files/suspend-diagnostic-logs.txt -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2167384 Title: mt7921e / MT7920 [14c3:7920]: Wi-Fi fails after s2idle on ASUS S3607AA, applications and shutdown hang To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2167384/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
