Same machine here (Nitro AN16S-61, board EV3_SKF, Ryzen AI 7 350) and the same failure, so adding evidence that may narrow it. Two things differ from the original report and are worth recording:
1. It is not fixed by newer firmware. This report is on BIOS V1.03 (2025-04-30). I am on V1.53 (2026-03-06) and it still happens, on kernel 7.2.8, mem_sleep=s2idle. 2. It follows the M.2 slot, not the drive. I swapped the two NVMe drives between slots and re-tested: the drive behind root port 0000:00:02.1 dies every time, whichever SSD is installed there; the drive behind 0000:00:03.2 is never affected. So the slot that kills the drive is 0000:00:02.1 (the one whose LnkCap advertises ASPM L1, and which is the ACPI wake entry GPP3 / power resource LNXPOWER:04; the other port reports "ASPM not supported" and is not a wake entry). On resume the link simply never comes back: pcieport 0000:00:02.1: broken device, retraining non-functional downstream link at 2.5GT/s pcieport 0000:00:02.1: retraining failed pcieport 0000:00:02.1: Data Link Layer Link Active not set in 100 msec nvme nvme0: Disabling device after reset failure: -19 What I've been able to rule out: - ASPM: cleared LnkCtl[1:0] on both ends of that link (controller + upstream port), verified afterwards with lspci (ASPM Disabled), then suspended - still fails. amd_pmc reports 'Last S0i3 Status: Success' during that same sleep, so the deepest state is reached without link L1 (ASPM is not needed for sleep depth here, and 'pcie_aspm=off' alone never disabled it either). - D3cold / shared _PR3: 'd3cold_allowed=0' on the drive and its upstream bridge - still fails. - Driver/PM callbacks: 'pm_test' ladder (devices, platform) is clean; only a real S0ix entry/exit breaks it. Workaround that works: Put the OS whose filesystem must stay writable on 0000:00:03.2 and accept that the drive on 0000:00:02.1 dies per suspend (recoverable only by reboot, the firmware exposes no ACPI hotplug slot for either port, and remove + rescan on the dead controller hangs the machine). Note that a root filesystem on the bad port is unrecoverable, which is exactly your report's symptom. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2129564 Title: Acer nitro 16s AI makes filesystem emergency ro after suspend To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2129564/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
