Public bug reported: Expected Behavior: The discrete NVIDIA GeForce MX250 GPU should successfully initialize and wake from low-power runtime states during system boot and user session initialization. Actual Behavior: During the early system boot phase, the kernel fails to wake the discrete GPU from its power-saving D3cold state to D0. Because the hardware remains unresponsive, it reports a broken 64-bit BAR allocation window mapped above 4GB, and the NVIDIA proprietary driver fails to initialize entirely. This bug is persistent across modern HWE kernel stacks (tested on 6.11.0-29-generic and 6.17.0-35-generic). Steps taken that did NOT resolve the issue: - Verified DKMS modules compile completely cleanly for NVIDIA 535.309.01 against both kernels. - Attempted ACPI OSI overrides via GRUB including 'acpi_osi="Windows 2015"' and 'acpi_osi="Windows 2009"' to force legacy vendor firmware paths; the D3cold -> D0 timeout persisted. - Tested kernel and driver power-state management parameters including 'pcie_port_pm=off', 'pci=realloc', and 'nvidia.NVreg_DynamicPowerManagement=1'. None of these flags successfully intercepted or prevented the hardware from falling into an un-wakeable D3cold state. - System is only stable when completely bypassing the device initialization tree using 'prime-select intel'. Hardware Platform: - Laptop: HP Envy x360 15-dr0003np - GPU: NVIDIA GeForce MX250 / Integrated Intel Graphics - OS: Linux Mint 22.3 (Noble Base) ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Attachment added: "Kernel logs" https://bugs.launchpad.net/bugs/2158075/+attachment/5978734/+files/kernel_power_bug.txt -- You received this bug notification because you are subscribed to linux in Ubuntu. Matching subscriptions: Bgg, Bmail, Nb https://bugs.launchpad.net/bugs/2158075 Title: PCI/PM Regression: NVIDIA MX250 / PCI Bridge unable to change power state from D3cold to D0 on kernels >= 6.8 (HP Envy) Status in linux package in Ubuntu: New Bug description: Expected Behavior: The discrete NVIDIA GeForce MX250 GPU should successfully initialize and wake from low-power runtime states during system boot and user session initialization. Actual Behavior: During the early system boot phase, the kernel fails to wake the discrete GPU from its power-saving D3cold state to D0. Because the hardware remains unresponsive, it reports a broken 64-bit BAR allocation window mapped above 4GB, and the NVIDIA proprietary driver fails to initialize entirely. This bug is persistent across modern HWE kernel stacks (tested on 6.11.0-29-generic and 6.17.0-35-generic). Steps taken that did NOT resolve the issue: - Verified DKMS modules compile completely cleanly for NVIDIA 535.309.01 against both kernels. - Attempted ACPI OSI overrides via GRUB including 'acpi_osi="Windows 2015"' and 'acpi_osi="Windows 2009"' to force legacy vendor firmware paths; the D3cold -> D0 timeout persisted. - Tested kernel and driver power-state management parameters including 'pcie_port_pm=off', 'pci=realloc', and 'nvidia.NVreg_DynamicPowerManagement=1'. None of these flags successfully intercepted or prevented the hardware from falling into an un-wakeable D3cold state. - System is only stable when completely bypassing the device initialization tree using 'prime-select intel'. Hardware Platform: - Laptop: HP Envy x360 15-dr0003np - GPU: NVIDIA GeForce MX250 / Integrated Intel Graphics - OS: Linux Mint 22.3 (Noble Base) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2158075/+subscriptions
Комментариев нет:
Отправить комментарий