понедельник

[Bug 2144584] Re: Suspend/Hibernate missing from GNOME menu and causing hard freeze due to Kernel Lockdown (Secure Boot)

** Description changed:

+ I have attached the nvidia-bug-report.log.gz file in comment. It should
+ contain detailed logs regarding the nvidia-settings failures ('Bad
+ handle' errors) and the GPU state during the DisplayPort link loss.
+ Also, the INT3515 (UCSI) IRQ failure is still present, which I believe
+ is the root cause of the power management deadlock
+
+
Hardware & System Info:
    Kernel: Linux towerq 6.19.0-9-generic #9-Ubuntu SMP PREEMPT_DYNAMIC Thu Mar 5 14:49:21 UTC 2026 x86_64 GNU/Linux
    GPU/Driver: 02:00.0 VGA compatible controller [0300]: NVIDIA Corporation AD104 [GeForce RTX 4070 Ti] [10de:2782] (rev a1)
        Subsystem: ASUSTeK Computer Inc. Device [1043:88dd]
        Kernel driver in use: nvidia
        Kernel modules: nvidiafb, nouveau, nova_core, nvidia_drm, nvidia

    Secure Boot: Enabled (Kernel Lockdown active)

Description: Ubuntu Resolute Raccoon (development branch)

Steps to reproduce:
    Boot the system with Secure Boot enabled.

    Attempt to Suspend or Hibernate via the GNOME power menu.
    System hangs instantly.

"Note: The 'Suspend' option is missing or disabled in the GNOME power
menu. Attempting to force suspend/hibernate via other methods leads to
an immediate system lockup. This appears to be a regression in how the
kernel handles power states when Lockdown is active."

yakub@towerq:~/Documents$ systemctl can-suspend && echo "Suspend OK" || echo "Suspend BLOCKED"
Unknown command verb 'can-suspend', did you mean 'suspend'?
Suspend BLOCKED

yakub@towerq:~/Documents$ systemctl can-hibernate && echo "Hibernate OK" || echo "Hibernate BLOCKED"
Unknown command verb 'can-hibernate', did you mean 'hibernate'?
Hibernate BLOCKED

yakub@towerq:~/Documents$ busctl get-property org.freedesktop.login1 /org/freedesktop/login1 org.freedesktop.login1.Manager CanSuspend
Failed to get property CanSuspend on interface org.freedesktop.login1.Manager: Unknown interface org.freedesktop.login1.Manager or property CanSuspend.

Logs analysis:
Journalctl shows a significant gap in logs between the suspend attempt and the hard reset. However, the following critical error was found right before the session start:
Lockdown: systemd-logind: hibernation is restricted; see man kernel_lockdown.7

ProblemType: Bug
DistroRelease: Ubuntu 26.04
Package: linux-image-6.19.0-9-generic 6.19.0-9.9
ProcVersionSignature: Ubuntu 6.19.0-9.9-generic 6.19.5
Uname: Linux 6.19.0-9-generic x86_64
NonfreeKernelModules: zfs
ApportVersion: 2.33.1-0ubuntu3
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: Unity
Date: Mon Mar 16 17:53:01 2026
InstallationDate: Installed on 2026-03-13 (3 days ago)
InstallationMedia: Ubuntu 26.04 LTS "Resolute Raccoon" - Daily amd64 (20260313)
MachineType: ASUS System Product Name
ProcEnviron:
 LANG=en_US.UTF-8
 PATH=(custom, no user)
 TERM=xterm-256color
 XDG_RUNTIME_DIR=<set>
ProcFB: 0 nvidia-drmdrmfb
ProcKernelCmdLine: BOOT_IMAGE=/BOOT/ubuntu_o2oj4n@/vmlinuz-6.19.0-9-generic root=ZFS=rpool/ROOT/ubuntu_o2oj4n 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: 04/25/2025
dmi.bios.release: 30.1
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: 3001
dmi.board.asset.tag: Default string
dmi.board.name: ROG MAXIMUS Z790 HERO
dmi.board.vendor: ASUSTeK COMPUTER INC.
dmi.board.version: Rev 1.xx
dmi.chassis.asset.tag: Default string
dmi.chassis.type: 3
dmi.chassis.vendor: Default string
dmi.chassis.version: Default string
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr3001:bd04/25/2025:br30.1:svnASUS:pnSystemProductName:pvrSystemVersion:rvnASUSTeKCOMPUTERINC.:rnROGMAXIMUSZ790HERO:rvrRev1.xx:cvnDefaultstring:ct3:cvrDefaultstring:skuSKU:
dmi.product.family: To be filled by O.E.M.
dmi.product.name: System Product Name
dmi.product.sku: SKU
dmi.product.version: System Version
dmi.sys.vendor: ASUS

-
-
akub@towerq:~/Documents$ systemd-inhibit --list
WHO UID USER PID COMM WHAT >
ModemManager 0 root 4070 ModemManager sleep >
NetworkManager 0 root 3968 NetworkManager sleep >
Realtime Kit 0 root 4326 rtkit-daemon sleep >
UPower 0 root 4645 upowerd sleep >
Unattended Upgrades Shutdown 0 root 4149 unattended-upgr shutdown >
GNOME Shell 1000 yakub 8474 gnome-shell sleep >
GNOME Shell 1000 yakub 8474 gnome-shell sleep >
yakub 1000 yakub 8603 gsd-power handle-lid-switch >
yakub 1000 yakub 8602 gsd-media-keys handle-power-key:handle-susp>
yakub 1000 yakub 8602 gsd-media-keys sleep >
yakub 1000 yakub 8603 gsd-power sleep
-
-

Critical update on hardware symptoms:

I've confirmed that all USB Type-C ports are non-functional on this
system.

This correlates with the kernel error: Serial bus multi instantiate
pseudo device driver INT3515:00: error -ENXIO: IRQ index 1 not found.

The monitor is connected via DisplayPort.

The failure of the UCSI driver (INT3515) might be causing a cascade
failure in power management, preventing the GPU from re-initializing the
DisplayPort link after the monitor is turned off or the system attempts
to sleep.

-
Important Evidence:
I've confirmed that all hardware (DisplayPort and USB-C ports) works perfectly on Windows on the same machine. This proves the hardware is functional. The issues (Type-C non-functional, INT3515 IRQ error, and Black Screen after monitor power cycle) are strictly related to the Linux kernel/driver implementation on this specific hardware configuration.

** Description changed:

- I have attached the nvidia-bug-report.log.gz file in comment. It should
- contain detailed logs regarding the nvidia-settings failures ('Bad
- handle' errors) and the GPU state during the DisplayPort link loss.
- Also, the INT3515 (UCSI) IRQ failure is still present, which I believe
- is the root cause of the power management deadlock
+ Revised Bug Description (Copy-paste this)

+ Summary: System-wide Power Management failure on ASUS Z790 + NVIDIA RTX
+ 4070 Ti, leading to hard lockups, non-functional USB-C ports, and
+ display recovery failure.

- Hardware & System Info:
-     Kernel: Linux towerq 6.19.0-9-generic #9-Ubuntu SMP PREEMPT_DYNAMIC Thu Mar 5 14:49:21 UTC 2026 x86_64 GNU/Linux
-     GPU/Driver: 02:00.0 VGA compatible controller [0300]: NVIDIA Corporation AD104 [GeForce RTX 4070 Ti] [10de:2782] (rev a1)
-         Subsystem: ASUSTeK Computer Inc. Device [1043:88dd]
-         Kernel driver in use: nvidia
-         Kernel modules: nvidiafb, nouveau, nova_core, nvidia_drm, nvidia
+ Environment:

-     Secure Boot: Enabled (Kernel Lockdown active)
+ OS: Ubuntu 26.04 "Resolute Raccoon" (Development Branch)

- Description: Ubuntu Resolute Raccoon (development branch)
+ Kernel: 6.19.0-9-generic

- Steps to reproduce:
-     Boot the system with Secure Boot enabled.
+ Motherboard: ASUS ROG MAXIMUS Z790 HERO (BIOS 3001)

-     Attempt to Suspend or Hibernate via the GNOME power menu.
-     System hangs instantly.
+ GPU: NVIDIA GeForce RTX 4070 Ti (Driver: 580.126.09)

- "Note: The 'Suspend' option is missing or disabled in the GNOME power
- menu. Attempting to force suspend/hibernate via other methods leads to
- an immediate system lockup. This appears to be a regression in how the
- kernel handles power states when Lockdown is active."
+ Secure Boot: Enabled (Kernel Lockdown active)

- yakub@towerq:~/Documents$ systemctl can-suspend && echo "Suspend OK" || echo "Suspend BLOCKED"
- Unknown command verb 'can-suspend', did you mean 'suspend'?
- Suspend BLOCKED
+ Symptoms:

- yakub@towerq:~/Documents$ systemctl can-hibernate && echo "Hibernate OK" || echo "Hibernate BLOCKED"
- Unknown command verb 'can-hibernate', did you mean 'hibernate'?
- Hibernate BLOCKED
+ Hard Lockup: System freezes instantly when attempting to Suspend or
+ Hibernate. Requires physical hard reset.

- yakub@towerq:~/Documents$ busctl get-property org.freedesktop.login1 /org/freedesktop/login1 org.freedesktop.login1.Manager CanSuspend
- Failed to get property CanSuspend on interface org.freedesktop.login1.Manager: Unknown interface org.freedesktop.login1.Manager or property CanSuspend.
+ Display Recovery Failure: If the monitor is manually turned OFF and
+ back ON (or switched to another input), the screen remains black. SSH
+ access remains functional, confirming a GPU/Driver hang rather than a
+ kernel panic.

- Logs analysis:
- Journalctl shows a significant gap in logs between the suspend attempt and the hard reset. However, the following critical error was found right before the session start:
- Lockdown: systemd-logind: hibernation is restricted; see man kernel_lockdown.7
+ USB-C Failure: All USB Type-C ports are non-functional in Ubuntu.

- ProblemType: Bug
- DistroRelease: Ubuntu 26.04
- Package: linux-image-6.19.0-9-generic 6.19.0-9.9
- ProcVersionSignature: Ubuntu 6.19.0-9.9-generic 6.19.5
- Uname: Linux 6.19.0-9-generic x86_64
- NonfreeKernelModules: zfs
- ApportVersion: 2.33.1-0ubuntu3
- Architecture: amd64
- CasperMD5CheckResult: pass
- CurrentDesktop: Unity
- Date: Mon Mar 16 17:53:01 2026
- InstallationDate: Installed on 2026-03-13 (3 days ago)
- InstallationMedia: Ubuntu 26.04 LTS "Resolute Raccoon" - Daily amd64 (20260313)
- MachineType: ASUS System Product Name
- ProcEnviron:
-  LANG=en_US.UTF-8
-  PATH=(custom, no user)
-  TERM=xterm-256color
-  XDG_RUNTIME_DIR=<set>
- ProcFB: 0 nvidia-drmdrmfb
- ProcKernelCmdLine: BOOT_IMAGE=/BOOT/ubuntu_o2oj4n@/vmlinuz-6.19.0-9-generic root=ZFS=rpool/ROOT/ubuntu_o2oj4n 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: 04/25/2025
- dmi.bios.release: 30.1
- dmi.bios.vendor: American Megatrends Inc.
- dmi.bios.version: 3001
- dmi.board.asset.tag: Default string
- dmi.board.name: ROG MAXIMUS Z790 HERO
- dmi.board.vendor: ASUSTeK COMPUTER INC.
- dmi.board.version: Rev 1.xx
- dmi.chassis.asset.tag: Default string
- dmi.chassis.type: 3
- dmi.chassis.vendor: Default string
- dmi.chassis.version: Default string
- dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr3001:bd04/25/2025:br30.1:svnASUS:pnSystemProductName:pvrSystemVersion:rvnASUSTeKCOMPUTERINC.:rnROGMAXIMUSZ790HERO:rvrRev1.xx:cvnDefaultstring:ct3:cvrDefaultstring:skuSKU:
- dmi.product.family: To be filled by O.E.M.
- dmi.product.name: System Product Name
- dmi.product.sku: SKU
- dmi.product.version: System Version
- dmi.sys.vendor: ASUS
+ UI Regression: The 'Suspend' option is often missing or unresponsive
+ in the GNOME power menu.

- akub@towerq:~/Documents$ systemd-inhibit --list
- WHO UID USER PID COMM WHAT >
- ModemManager 0 root 4070 ModemManager sleep >
- NetworkManager 0 root 3968 NetworkManager sleep >
- Realtime Kit 0 root 4326 rtkit-daemon sleep >
- UPower 0 root 4645 upowerd sleep >
- Unattended Upgrades Shutdown 0 root 4149 unattended-upgr shutdown >
- GNOME Shell 1000 yakub 8474 gnome-shell sleep >
- GNOME Shell 1000 yakub 8474 gnome-shell sleep >
- yakub 1000 yakub 8603 gsd-power handle-lid-switch >
- yakub 1000 yakub 8602 gsd-media-keys handle-power-key:handle-susp>
- yakub 1000 yakub 8602 gsd-media-keys sleep >
- yakub 1000 yakub 8603 gsd-power sleep
+ Technical Root Causes Found:

- Critical update on hardware symptoms:
+ UCSI / IRQ Conflict: The kernel log shows a critical failure in the serial bus driver:
+ Serial bus multi instantiate pseudo device driver INT3515:00: error -ENXIO: IRQ index 1 not found.
+ This appears to be the root cause, as the UCSI (USB Type-C Connector System Software Interface) is tied to the system's power state transitions.

- I've confirmed that all USB Type-C ports are non-functional on this
- system.
+ EDID Mismatch: While the monitor is active, cat
+ /sys/class/drm/card1-DP-1/edid returns 0 bytes. However, valid EDID data
+ (384 bytes) was found on card1-DP-3. This suggests a mapping regression
+ between physical ports and logical DRM outputs.

- This correlates with the kernel error: Serial bus multi instantiate
- pseudo device driver INT3515:00: error -ENXIO: IRQ index 1 not found.
+ Lockdown Interference: Kernel Lockdown restricts hibernation, but
+ the system fails to handle this restriction gracefully, leading to a
+ freeze instead of a denied request.

- The monitor is connected via DisplayPort.
+ Evidence of Regression:

- The failure of the UCSI driver (INT3515) might be causing a cascade
- failure in power management, preventing the GPU from re-initializing the
- DisplayPort link after the monitor is turned off or the system attempts
- to sleep.
+ Dual-Boot Verification: All hardware (USB-C ports, Suspend/Resume,
+ DisplayPort link) works perfectly on Windows on the same machine. This
+ confirms the hardware is functional and the issue is strictly a Linux
+ kernel/driver regression.

- Important Evidence:
- I've confirmed that all hardware (DisplayPort and USB-C ports) works perfectly on Windows on the same machine. This proves the hardware is functional. The issues (Type-C non-functional, INT3515 IRQ error, and Black Screen after monitor power cycle) are strictly related to the Linux kernel/driver implementation on this specific hardware configuration.
+ NVIDIA Tooling Failure: nvidia-settings is non-functional under
+ Wayland/Kernel 6.19, returning multiple "Bad handle" errors, preventing
+ standard diagnostic procedures.
+
+ Steps to Reproduce:
+
+ Boot into Ubuntu with Secure Boot active.
+
+ Turn the monitor OFF and then ON via the physical power button
+ (causes Black Screen).
+
+ OR attempt to Suspend via terminal/UI (causes Hard Freeze).
+
+ Attached Logs:
+
+ nvidia-bug-report.log.gz (Detailed GPU/Driver state)
+
+ acpidump.txt (ACPI tables for ASUS Z790)
+
+ /proc/interrupts (Showing the IRQ 1 conflict)

** Description changed:

- Revised Bug Description (Copy-paste this)
-
Summary: System-wide Power Management failure on ASUS Z790 + NVIDIA RTX
4070 Ti, leading to hard lockups, non-functional USB-C ports, and
display recovery failure.

Environment:

- OS: Ubuntu 26.04 "Resolute Raccoon" (Development Branch)
+     OS: Ubuntu 26.04 "Resolute Raccoon" (Development Branch)

- Kernel: 6.19.0-9-generic
+     Kernel: 6.19.0-9-generic

- Motherboard: ASUS ROG MAXIMUS Z790 HERO (BIOS 3001)
+     Motherboard: ASUS ROG MAXIMUS Z790 HERO (BIOS 3001)

- GPU: NVIDIA GeForce RTX 4070 Ti (Driver: 580.126.09)
+     GPU: NVIDIA GeForce RTX 4070 Ti (Driver: 580.126.09)

- Secure Boot: Enabled (Kernel Lockdown active)
+     Secure Boot: Enabled (Kernel Lockdown active)

Symptoms:

- Hard Lockup: System freezes instantly when attempting to Suspend or
+     Hard Lockup: System freezes instantly when attempting to Suspend or
Hibernate. Requires physical hard reset.

- Display Recovery Failure: If the monitor is manually turned OFF and
+     Display Recovery Failure: If the monitor is manually turned OFF and
back ON (or switched to another input), the screen remains black. SSH
access remains functional, confirming a GPU/Driver hang rather than a
kernel panic.

- USB-C Failure: All USB Type-C ports are non-functional in Ubuntu.
+     USB-C Failure: All USB Type-C ports are non-functional in Ubuntu.

- UI Regression: The 'Suspend' option is often missing or unresponsive
+     UI Regression: The 'Suspend' option is often missing or unresponsive
in the GNOME power menu.

Technical Root Causes Found:

- UCSI / IRQ Conflict: The kernel log shows a critical failure in the serial bus driver:
- Serial bus multi instantiate pseudo device driver INT3515:00: error -ENXIO: IRQ index 1 not found.
- This appears to be the root cause, as the UCSI (USB Type-C Connector System Software Interface) is tied to the system's power state transitions.
+     UCSI / IRQ Conflict: The kernel log shows a critical failure in the serial bus driver:
+     Serial bus multi instantiate pseudo device driver INT3515:00: error -ENXIO: IRQ index 1 not found.
+     This appears to be the root cause, as the UCSI (USB Type-C Connector System Software Interface) is tied to the system's power state transitions.

- EDID Mismatch: While the monitor is active, cat
+     EDID Mismatch: While the monitor is active, cat
/sys/class/drm/card1-DP-1/edid returns 0 bytes. However, valid EDID data
(384 bytes) was found on card1-DP-3. This suggests a mapping regression
between physical ports and logical DRM outputs.

- Lockdown Interference: Kernel Lockdown restricts hibernation, but
+     Lockdown Interference: Kernel Lockdown restricts hibernation, but
the system fails to handle this restriction gracefully, leading to a
freeze instead of a denied request.

Evidence of Regression:

- Dual-Boot Verification: All hardware (USB-C ports, Suspend/Resume,
+     Dual-Boot Verification: All hardware (USB-C ports, Suspend/Resume,
DisplayPort link) works perfectly on Windows on the same machine. This
confirms the hardware is functional and the issue is strictly a Linux
kernel/driver regression.

- NVIDIA Tooling Failure: nvidia-settings is non-functional under
+     NVIDIA Tooling Failure: nvidia-settings is non-functional under
Wayland/Kernel 6.19, returning multiple "Bad handle" errors, preventing
standard diagnostic procedures.

Steps to Reproduce:

- Boot into Ubuntu with Secure Boot active.
+     Boot into Ubuntu with Secure Boot active.

- Turn the monitor OFF and then ON via the physical power button
+     Turn the monitor OFF and then ON via the physical power button
(causes Black Screen).

- OR attempt to Suspend via terminal/UI (causes Hard Freeze).
+     OR attempt to Suspend via terminal/UI (causes Hard Freeze).

Attached Logs:

- nvidia-bug-report.log.gz (Detailed GPU/Driver state)
+     nvidia-bug-report.log.gz (Detailed GPU/Driver state)

- acpidump.txt (ACPI tables for ASUS Z790)
+     acpidump.txt (ACPI tables for ASUS Z790)

- /proc/interrupts (Showing the IRQ 1 conflict)
+     /proc/interrupts (Showing the IRQ 1 conflict)

--
You received this bug notification because you are subscribed to linux
in Ubuntu.
Matching subscriptions: Bgg, Bmail, Nb
https://bugs.launchpad.net/bugs/2144584

Title:
Multiple PM Regressions on ASUS Z790/RTX 4070 Ti: Suspend freeze, UCSI
(INT3515) IRQ failure, and Display recovery hang

Status in linux package in Ubuntu:
New

Bug description:
Summary: System-wide Power Management failure on ASUS Z790 + NVIDIA
RTX 4070 Ti, leading to hard lockups, non-functional USB-C ports, and
display recovery failure.

Environment:

    OS: Ubuntu 26.04 "Resolute Raccoon" (Development Branch)

    Kernel: 6.19.0-9-generic

    Motherboard: ASUS ROG MAXIMUS Z790 HERO (BIOS 3001)

    GPU: NVIDIA GeForce RTX 4070 Ti (Driver: 580.126.09)

    Secure Boot: Enabled (Kernel Lockdown active)

Symptoms:

    Hard Lockup: System freezes instantly when attempting to Suspend
or Hibernate. Requires physical hard reset.

    Display Recovery Failure: If the monitor is manually turned OFF
and back ON (or switched to another input), the screen remains black.
SSH access remains functional, confirming a GPU/Driver hang rather
than a kernel panic.

    USB-C Failure: All USB Type-C ports are non-functional in Ubuntu.

    UI Regression: The 'Suspend' option is often missing or
unresponsive in the GNOME power menu.

Technical Root Causes Found:

    UCSI / IRQ Conflict: The kernel log shows a critical failure in the serial bus driver:
    Serial bus multi instantiate pseudo device driver INT3515:00: error -ENXIO: IRQ index 1 not found.
    This appears to be the root cause, as the UCSI (USB Type-C Connector System Software Interface) is tied to the system's power state transitions.

    EDID Mismatch: While the monitor is active, cat
/sys/class/drm/card1-DP-1/edid returns 0 bytes. However, valid EDID
data (384 bytes) was found on card1-DP-3. This suggests a mapping
regression between physical ports and logical DRM outputs.

    Lockdown Interference: Kernel Lockdown restricts hibernation, but
the system fails to handle this restriction gracefully, leading to a
freeze instead of a denied request.

Evidence of Regression:

    Dual-Boot Verification: All hardware (USB-C ports, Suspend/Resume,
DisplayPort link) works perfectly on Windows on the same machine. This
confirms the hardware is functional and the issue is strictly a Linux
kernel/driver regression.

    NVIDIA Tooling Failure: nvidia-settings is non-functional under
Wayland/Kernel 6.19, returning multiple "Bad handle" errors,
preventing standard diagnostic procedures.

Steps to Reproduce:

    Boot into Ubuntu with Secure Boot active.

    Turn the monitor OFF and then ON via the physical power button
(causes Black Screen).

    OR attempt to Suspend via terminal/UI (causes Hard Freeze).

Attached Logs:

    nvidia-bug-report.log.gz (Detailed GPU/Driver state)

    acpidump.txt (ACPI tables for ASUS Z790)

    /proc/interrupts (Showing the IRQ 1 conflict)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2144584/+subscriptions

Комментариев нет:

Отправить комментарий