Public bug reported: Yoga Slim 7 14AKP10 (Krackan Point): ~38.6 s s2idle resume stall ending in NVMe IO_PAGE_FAULT, fixed by iommu=pt System Lenovo Yoga Slim 7 14AKP10, AMD Ryzen AI 5 330 (Krackan Point), Radeon 820M, 16 GB NVMe: Micron 2400 MTFDKCD512TGW-1BP1AABLA, firmware 1002VALN (DRAM-less, uses HMB) Kubuntu 26.04.1 LTS, kernel 7.0.0-31-generic BIOS QTCN33WW (2026-04-24), Pluton disabled in BIOS; no newer firmware on LVFS /sys/power/mem_sleep = [s2idle] (no deep/S3 available) Symptom After s2idle suspend (lid or systemctl suspend), affected resumes take ~38 to 40 s before the login screen appears. The delay is almost constant (38.57 s / 38.61 s / 38.61 s over three cycles). Some resumes are immediate, so the issue is intermittent but easily reproducible. What was measured amd_pmc s0ix_stats: S0i3 entry succeeds, ~151 ms to resume from S0i3, so the SoC/SMU side is fine. pm_print_times: endpoint callbacks are fast (amdgpu ~185 ms, USB max ~266 ms, nvme pci_pm_resume 2 µs since it only schedules an async reset). In dmesg the stall sits between ACPI: EC: interrupt unblocked and the following line, which is logged exactly at the end of the ~38.6 s gap: nvme 0000:bf:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x77e54000 flags=0x0000] immediately followed by amdgpu ... SMU is resuming, nvme nvme0: 8/0/0 default/read/poll queues and PM: suspend exit. The firmware also repeatedly logs, at suspend entry and/or exit: ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.LPC0.EC0.OSFG], AE_NOT_FOUND ACPI Error: Aborting method \_SB.PEP._DSM due to previous error (AE_NOT_FOUND) so the LPS0 (Modern Standby) entry/exit notification to the platform aborts. Workaround (verified) Booting with iommu=pt completely removes the stall and the IO_PAGE_FAULT over repeated systemctl suspend cycles (20–30 s sleep). Nothing else was changed. The ACPI OSFG / PEP._DSM error is still logged with the workaround. Parameters deliberately not used: amd_iommu=off, nvme.noacpi=1, iommu=soft, pcie_aspm=off. Interpretation The IO_PAGE_FAULT is the marker of the end of the stall, not its cause: after leaving D3cold the NVMe controller issues DMA with a stale address, the IOMMU rejects it in the default translated mode, and the controller stays in a bad state until a ~38.6 s timeout expires, after which resume completes normally. With identity mapping (iommu=pt) the stale address stays valid and the device answers immediately. This looks like a platform + NVMe (Micron 2400 / HMB / D3cold) resume-ordering issue rather than a userspace, GPU or SMU delay. The aborted PEP._DSM may contribute to the intermittency (the EC may not be told the OS left standby); that part needs a DSDT review. Requests Kernel/AMD: consider a quirk for this platform/SSD combination (e.g. IOMMU passthrough by default for this DMI, or HMB/D3cold handling on resume) so it works out of the box. Lenovo: fix the firmware ACPI tables so that \_SB.PCI0.LPC0.EC0.OSFG, referenced by \_SB.PEP._DSM, exists. Reproduction Boot Kubuntu without iommu=pt; check /sys/power/mem_sleep shows [s2idle]. systemctl suspend, wait 20–30 s, wake. On an affected cycle, observe ~38 s before the login screen; in dmesg the gap is between ACPI: EC: interrupt unblocked and nvme ... AMD-Vi: Event logged [IO_PAGE_FAULT ...]. Reboot with iommu=pt, repeat 2–3: no stall, no fault. Full report (French, with logs and s0ix_stats) and amd-s2idle output attached 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 CasperMD5CheckResult: unknown CurrentDesktop: KDE Date: Sun Sep 6 21:18:48 2026 InstallationDate: Installed on 2026-09-06 (0 days ago) InstallationMedia: Kubuntu 26.04.1 LTS "Resolute Raccoon" - Release amd64 (20260826.1) MachineType: LENOVO 83JY ProcFB: 0 amdgpudrmfb ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-7.0.0-31-generic root=UUID=764762f1-5e87-46be-869b-a6869d0d88f0 ro quiet splash iommu=pt PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon. SourcePackage: linux UpgradeStatus: No upgrade log present (probably fresh install) dmi.bios.date: 04/24/2026 dmi.bios.release: 1.33 dmi.bios.vendor: LENOVO dmi.bios.version: QTCN33WW dmi.board.asset.tag: NO Asset Tag dmi.board.name: LNVNB161216 dmi.board.vendor: LENOVO dmi.board.version: SDK0T76573 WIN dmi.chassis.asset.tag: NO Asset Tag dmi.chassis.type: 10 dmi.chassis.vendor: LENOVO dmi.chassis.version: Yoga Slim 7 14AKP10 dmi.ec.firmware.release: 1.33 dmi.modalias: dmi:bvnLENOVO:bvrQTCN33WW:bd04/24/2026:br1.33:efr1.33:svnLENOVO:pn83JY:pvrYogaSlim714AKP10:rvnLENOVO:rnLNVNB161216:rvrSDK0T76573WIN:cvnLENOVO:ct10:cvrYogaSlim714AKP10:skuLENOVO_MT_83JY_BU_idea_FM_YogaSlim714AKP10:pfaYogaSlim714AKP10: dmi.product.family: Yoga Slim 7 14AKP10 dmi.product.name: 83JY dmi.product.sku: LENOVO_MT_83JY_BU_idea_FM_Yoga Slim 7 14AKP10 dmi.product.version: Yoga Slim 7 14AKP10 dmi.sys.vendor: LENOVO ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Tags: amd64 apport-bug resolute wayland-session ** Attachment added: "rapport_bug_resume_suspend_lenovo_yoga_slim_7_14AKP10_v2.docx" https://bugs.launchpad.net/bugs/2166622/+attachment/5998010/+files/rapport_bug_resume_suspend_lenovo_yoga_slim_7_14AKP10_v2.docx -- You received this bug notification because you are subscribed to linux in Ubuntu. Matching subscriptions: Bgg, Bmail, Nb https://bugs.launchpad.net/bugs/2166622 Title: Yoga Slim 7 14AKP10: ~38 s s2idle resume stall, NVMe IO_PAGE_FAULT, fixed by iommu=pt Status in linux package in Ubuntu: New Bug description: Yoga Slim 7 14AKP10 (Krackan Point): ~38.6 s s2idle resume stall ending in NVMe IO_PAGE_FAULT, fixed by iommu=pt System Lenovo Yoga Slim 7 14AKP10, AMD Ryzen AI 5 330 (Krackan Point), Radeon 820M, 16 GB NVMe: Micron 2400 MTFDKCD512TGW-1BP1AABLA, firmware 1002VALN (DRAM-less, uses HMB) Kubuntu 26.04.1 LTS, kernel 7.0.0-31-generic BIOS QTCN33WW (2026-04-24), Pluton disabled in BIOS; no newer firmware on LVFS /sys/power/mem_sleep = [s2idle] (no deep/S3 available) Symptom After s2idle suspend (lid or systemctl suspend), affected resumes take ~38 to 40 s before the login screen appears. The delay is almost constant (38.57 s / 38.61 s / 38.61 s over three cycles). Some resumes are immediate, so the issue is intermittent but easily reproducible. What was measured amd_pmc s0ix_stats: S0i3 entry succeeds, ~151 ms to resume from S0i3, so the SoC/SMU side is fine. pm_print_times: endpoint callbacks are fast (amdgpu ~185 ms, USB max ~266 ms, nvme pci_pm_resume 2 µs since it only schedules an async reset). In dmesg the stall sits between ACPI: EC: interrupt unblocked and the following line, which is logged exactly at the end of the ~38.6 s gap: nvme 0000:bf:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0x77e54000 flags=0x0000] immediately followed by amdgpu ... SMU is resuming, nvme nvme0: 8/0/0 default/read/poll queues and PM: suspend exit. The firmware also repeatedly logs, at suspend entry and/or exit: ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.LPC0.EC0.OSFG], AE_NOT_FOUND ACPI Error: Aborting method \_SB.PEP._DSM due to previous error (AE_NOT_FOUND) so the LPS0 (Modern Standby) entry/exit notification to the platform aborts. Workaround (verified) Booting with iommu=pt completely removes the stall and the IO_PAGE_FAULT over repeated systemctl suspend cycles (20–30 s sleep). Nothing else was changed. The ACPI OSFG / PEP._DSM error is still logged with the workaround. Parameters deliberately not used: amd_iommu=off, nvme.noacpi=1, iommu=soft, pcie_aspm=off. Interpretation The IO_PAGE_FAULT is the marker of the end of the stall, not its cause: after leaving D3cold the NVMe controller issues DMA with a stale address, the IOMMU rejects it in the default translated mode, and the controller stays in a bad state until a ~38.6 s timeout expires, after which resume completes normally. With identity mapping (iommu=pt) the stale address stays valid and the device answers immediately. This looks like a platform + NVMe (Micron 2400 / HMB / D3cold) resume-ordering issue rather than a userspace, GPU or SMU delay. The aborted PEP._DSM may contribute to the intermittency (the EC may not be told the OS left standby); that part needs a DSDT review. Requests Kernel/AMD: consider a quirk for this platform/SSD combination (e.g. IOMMU passthrough by default for this DMI, or HMB/D3cold handling on resume) so it works out of the box. Lenovo: fix the firmware ACPI tables so that \_SB.PCI0.LPC0.EC0.OSFG, referenced by \_SB.PEP._DSM, exists. Reproduction Boot Kubuntu without iommu=pt; check /sys/power/mem_sleep shows [s2idle]. systemctl suspend, wait 20–30 s, wake. On an affected cycle, observe ~38 s before the login screen; in dmesg the gap is between ACPI: EC: interrupt unblocked and nvme ... AMD-Vi: Event logged [IO_PAGE_FAULT ...]. Reboot with iommu=pt, repeat 2–3: no stall, no fault. Full report (French, with logs and s0ix_stats) and amd-s2idle output attached 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 CasperMD5CheckResult: unknown CurrentDesktop: KDE Date: Sun Sep 6 21:18:48 2026 InstallationDate: Installed on 2026-09-06 (0 days ago) InstallationMedia: Kubuntu 26.04.1 LTS "Resolute Raccoon" - Release amd64 (20260826.1) MachineType: LENOVO 83JY ProcFB: 0 amdgpudrmfb ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-7.0.0-31-generic root=UUID=764762f1-5e87-46be-869b-a6869d0d88f0 ro quiet splash iommu=pt PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon. SourcePackage: linux UpgradeStatus: No upgrade log present (probably fresh install) dmi.bios.date: 04/24/2026 dmi.bios.release: 1.33 dmi.bios.vendor: LENOVO dmi.bios.version: QTCN33WW dmi.board.asset.tag: NO Asset Tag dmi.board.name: LNVNB161216 dmi.board.vendor: LENOVO dmi.board.version: SDK0T76573 WIN dmi.chassis.asset.tag: NO Asset Tag dmi.chassis.type: 10 dmi.chassis.vendor: LENOVO dmi.chassis.version: Yoga Slim 7 14AKP10 dmi.ec.firmware.release: 1.33 dmi.modalias: dmi:bvnLENOVO:bvrQTCN33WW:bd04/24/2026:br1.33:efr1.33:svnLENOVO:pn83JY:pvrYogaSlim714AKP10:rvnLENOVO:rnLNVNB161216:rvrSDK0T76573WIN:cvnLENOVO:ct10:cvrYogaSlim714AKP10:skuLENOVO_MT_83JY_BU_idea_FM_YogaSlim714AKP10:pfaYogaSlim714AKP10: dmi.product.family: Yoga Slim 7 14AKP10 dmi.product.name: 83JY dmi.product.sku: LENOVO_MT_83JY_BU_idea_FM_Yoga Slim 7 14AKP10 dmi.product.version: Yoga Slim 7 14AKP10 dmi.sys.vendor: LENOVO To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2166622/+subscriptions
Комментариев нет:
Отправить комментарий