Public bug reported: ## Summary Dell Pro Max 16 Plus laptop running Ubuntu 26.04 with Intel Meteor Lake iGPU (`i915`) and NVIDIA RTX PRO 5000 (hybrid graphics) repeatedly freezes with black screen(s). The system remains running (compute continues, fans spin) but the desktop becomes completely unresponsive. Recovery requires dock power-cycle, suspend/resume, SSH reboot, or hard reset. The kernel logs show repeated `i915` display driver failures in the DPLL/PHY/pipe A path, especially during suspend/resume and display power-state transitions. A Dell Thunderbolt 5 dock with a 5120x2160 external monitor exacerbates the issue. Kernel parameter `i915.enable_psr=0` was applied but did not resolve suspend/resume failures. --- ## Hardware | Component | Details | |---|---| | Laptop | Dell Pro Max 16 Plus MB16250/0K8XPY | | BIOS | 2.6.1 (03/16/2026) | | CPU/GPU (integrated) | Intel Meteor Lake, PCI `8086:7d67`, driver `i915` | | GPU (discrete) | NVIDIA RTX PRO 5000 Blackwell, driver `nvidia 595.71.05` | | RAM | 128 GB | | Internal display | eDP-1 (built-in OLED panel) | | External monitor | Dell U4025QW 40" ultrawide, 5120x2160 | | Dock | Dell Thunderbolt 5 Dock (SD25TB5 / K2 DOCK) | | Dock link | TB5, 120 Gb/s, left-side Thunderbolt port | | Monitor connection | DisplayPort (dock DP port; previously USB-C, same bug) | --- ## Software | Component | Version | |---|---| | OS | Ubuntu 26.04 LTS (resolute) | | Kernel | 7.0.0-27-generic (#27-Ubuntu SMP PREEMPT_DYNAMIC) | | Mesa | 26.0.3-1ubuntu1 | | libdrm | 2.4.131-1 | | linux-firmware-intel-graphics | 20260319.git217ca6e4-0ubuntu2.1 | | Desktop | GNOME Shell 50.1, Wayland | | Kernel cmdline | `i915.enable_psr=0` (PSR disabled, bug persists) | --- ## Symptoms 1. One or both screens go black 2. Keyboard/mouse input stops responding 3. System is NOT fully dead — background processes (e.g. vLLM compute) may continue 4. No clean shutdown — logs stop abruptly until hard reset or recovery action 5. Disconnecting monitor/dock during freeze does not help (compositor blocked in kernel ioctl) 6. Toggling dock power off/on DOES recover the display link --- ## Triggers (confirmed across multiple incidents) | Trigger | Confirmed | |---|---| | Suspend / resume (lid close, idle suspend) | Yes — most reliable trigger | | Overnight idle (display power management) | Yes | | Stopping GPU workload (vLLM exit) | Yes — likely GPU power-state change | | External monitor via dock at 5120x2160 | Contributes — DP link retraining failures | | Monitor hotplug / dock reconnect | Yes | --- ## Kernel error signature (repeated in every incident) ``` i915 0000:00:02.0: [drm] *ERROR* flip_done timed out i915 0000:00:02.0: [drm] *ERROR* [CRTC:150:pipe A] commit wait timed out i915 0000:00:02.0: [drm] DPLL 0: pll hw state mismatch WARNING: drivers/gpu/drm/i915/display/intel_dpll_mgr.c:4945 at verify_single_dpll_state WARNING: drivers/gpu/drm/i915/display/intel_modeset_verify.c:225 at verify_crtc_state i915 0000:00:02.0: [drm] *ERROR* Timeout waiting for DDI BUF A to get active i915 0000:00:02.0: [drm] *ERROR* Timed out waiting for DP idle patterns i915 0000:00:02.0: [drm] *ERROR* PHY A failed to change powerdown state i915 0000:00:02.0: [drm] Port A PLL not unlocked i915 0000:00:02.0: [drm] port_clock mismatch: expected 540000, found 61440 i915 0000:00:02.0: [drm] pixel_rate mismatch: expected 317250, found 36096 ``` During suspend/resume, errors occur in: - `intel_display_driver_resume` → `i915_drm_resume` - `intel_dp_retrain_link` → `intel_encoder_link_check_work_fn` Also seen: ``` pcieport 0000:00:01.0: PCIe Bus Error: ACSViol (Thunderbolt controller) NVRM: RmHandleDNotifierEvent: Failed to handle ACPI D-Notifier event, status=0x11 ``` No `NMI watchdog`, `soft lockup`, or `hung_task` messages — kernel appears alive, display commit path is deadlocked. --- ## Incident timeline ### Incident 1 — 2026-07-10 ~20:23 - Hard reset required - Last log: `flip_done timed out` at 20:23:43 - Reboot at 20:26:48 ### Incident 2 — 2026-07-11 ~10:14 - Overnight computation, spontaneous freeze (no user interaction) - Both screens black, unresponsive - Hard reset, reboot at 10:31:55 ### Incident 3 — 2026-07-11 ~17:51 - After login, exiting vLLM - Recovered via dock power-cycle (no hard reset) ### Incident 4 — 2026-07-12 ~10:44 - Suspend/resume (lid close) - Same i915 errors during `i915_drm_resume` - Recovered without hard reset (~1 minute) --- ## What was tried | Mitigation | Result | |---|---| | `i915.enable_psr=0` kernel parameter | Active in cmdline; bug persists on suspend/resume | | Monitor USB-C → DisplayPort on dock | Display path changed (DP-7 → DP-10); suspend/resume still fails | | Dock on TB5 port (left side, 120 Gb/s) | Correct port confirmed; bug persists | | Dock power-cycle | Recovers display (workaround, not fix) | --- ## Display topology ``` Laptop (Intel i915, card1) ├── eDP-1: internal panel (always connected) └── DP-10: external Dell U4025QW via TB5 dock DisplayPort NVIDIA (card2): all outputs disconnected (compute only, vLLM) ``` External display is driven by **Intel i915**, not NVIDIA, even with vLLM using the discrete GPU. --- ## Steps to reproduce 1. Dell Pro Max 16 Plus with BIOS 2.6.1 2. Connect Dell TB5 dock to left-side Thunderbolt 5 port 3. Connect Dell U4025QW (or similar 5K display) via DisplayPort on dock 4. Boot Ubuntu 26.04 with kernel 7.0.0-27-generic 5. Use dual-monitor setup (internal + external) 6. Suspend the system (close lid or `systemctl suspend`) 7. Resume **Expected:** Normal desktop **Actual:** Black screen(s), input frozen; kernel logs show i915 DPLL/flip_done errors Alternative trigger: leave system idle overnight with external monitor connected. ** Affects: linux (Ubuntu) Importance: Undecided Status: New ** Attachment added: "logs" https://bugs.launchpad.net/bugs/2160453/+attachment/5981380/+files/dell.zip -- You received this bug notification because you are subscribed to linux in Ubuntu. Matching subscriptions: Bgg, Bmail, Nb https://bugs.launchpad.net/bugs/2160453 Title: i915: DPLL/PHY hang on Meteor Lake (Dell Pro Max 16 Plus) with TB5 dock after suspend/resume - flip_done timeout, system freeze Status in linux package in Ubuntu: New Bug description: ## Summary Dell Pro Max 16 Plus laptop running Ubuntu 26.04 with Intel Meteor Lake iGPU (`i915`) and NVIDIA RTX PRO 5000 (hybrid graphics) repeatedly freezes with black screen(s). The system remains running (compute continues, fans spin) but the desktop becomes completely unresponsive. Recovery requires dock power-cycle, suspend/resume, SSH reboot, or hard reset. The kernel logs show repeated `i915` display driver failures in the DPLL/PHY/pipe A path, especially during suspend/resume and display power-state transitions. A Dell Thunderbolt 5 dock with a 5120x2160 external monitor exacerbates the issue. Kernel parameter `i915.enable_psr=0` was applied but did not resolve suspend/resume failures. --- ## Hardware | Component | Details | |---|---| | Laptop | Dell Pro Max 16 Plus MB16250/0K8XPY | | BIOS | 2.6.1 (03/16/2026) | | CPU/GPU (integrated) | Intel Meteor Lake, PCI `8086:7d67`, driver `i915` | | GPU (discrete) | NVIDIA RTX PRO 5000 Blackwell, driver `nvidia 595.71.05` | | RAM | 128 GB | | Internal display | eDP-1 (built-in OLED panel) | | External monitor | Dell U4025QW 40" ultrawide, 5120x2160 | | Dock | Dell Thunderbolt 5 Dock (SD25TB5 / K2 DOCK) | | Dock link | TB5, 120 Gb/s, left-side Thunderbolt port | | Monitor connection | DisplayPort (dock DP port; previously USB-C, same bug) | --- ## Software | Component | Version | |---|---| | OS | Ubuntu 26.04 LTS (resolute) | | Kernel | 7.0.0-27-generic (#27-Ubuntu SMP PREEMPT_DYNAMIC) | | Mesa | 26.0.3-1ubuntu1 | | libdrm | 2.4.131-1 | | linux-firmware-intel-graphics | 20260319.git217ca6e4-0ubuntu2.1 | | Desktop | GNOME Shell 50.1, Wayland | | Kernel cmdline | `i915.enable_psr=0` (PSR disabled, bug persists) | --- ## Symptoms 1. One or both screens go black 2. Keyboard/mouse input stops responding 3. System is NOT fully dead — background processes (e.g. vLLM compute) may continue 4. No clean shutdown — logs stop abruptly until hard reset or recovery action 5. Disconnecting monitor/dock during freeze does not help (compositor blocked in kernel ioctl) 6. Toggling dock power off/on DOES recover the display link --- ## Triggers (confirmed across multiple incidents) | Trigger | Confirmed | |---|---| | Suspend / resume (lid close, idle suspend) | Yes — most reliable trigger | | Overnight idle (display power management) | Yes | | Stopping GPU workload (vLLM exit) | Yes — likely GPU power-state change | | External monitor via dock at 5120x2160 | Contributes — DP link retraining failures | | Monitor hotplug / dock reconnect | Yes | --- ## Kernel error signature (repeated in every incident) ``` i915 0000:00:02.0: [drm] *ERROR* flip_done timed out i915 0000:00:02.0: [drm] *ERROR* [CRTC:150:pipe A] commit wait timed out i915 0000:00:02.0: [drm] DPLL 0: pll hw state mismatch WARNING: drivers/gpu/drm/i915/display/intel_dpll_mgr.c:4945 at verify_single_dpll_state WARNING: drivers/gpu/drm/i915/display/intel_modeset_verify.c:225 at verify_crtc_state i915 0000:00:02.0: [drm] *ERROR* Timeout waiting for DDI BUF A to get active i915 0000:00:02.0: [drm] *ERROR* Timed out waiting for DP idle patterns i915 0000:00:02.0: [drm] *ERROR* PHY A failed to change powerdown state i915 0000:00:02.0: [drm] Port A PLL not unlocked i915 0000:00:02.0: [drm] port_clock mismatch: expected 540000, found 61440 i915 0000:00:02.0: [drm] pixel_rate mismatch: expected 317250, found 36096 ``` During suspend/resume, errors occur in: - `intel_display_driver_resume` → `i915_drm_resume` - `intel_dp_retrain_link` → `intel_encoder_link_check_work_fn` Also seen: ``` pcieport 0000:00:01.0: PCIe Bus Error: ACSViol (Thunderbolt controller) NVRM: RmHandleDNotifierEvent: Failed to handle ACPI D-Notifier event, status=0x11 ``` No `NMI watchdog`, `soft lockup`, or `hung_task` messages — kernel appears alive, display commit path is deadlocked. --- ## Incident timeline ### Incident 1 — 2026-07-10 ~20:23 - Hard reset required - Last log: `flip_done timed out` at 20:23:43 - Reboot at 20:26:48 ### Incident 2 — 2026-07-11 ~10:14 - Overnight computation, spontaneous freeze (no user interaction) - Both screens black, unresponsive - Hard reset, reboot at 10:31:55 ### Incident 3 — 2026-07-11 ~17:51 - After login, exiting vLLM - Recovered via dock power-cycle (no hard reset) ### Incident 4 — 2026-07-12 ~10:44 - Suspend/resume (lid close) - Same i915 errors during `i915_drm_resume` - Recovered without hard reset (~1 minute) --- ## What was tried | Mitigation | Result | |---|---| | `i915.enable_psr=0` kernel parameter | Active in cmdline; bug persists on suspend/resume | | Monitor USB-C → DisplayPort on dock | Display path changed (DP-7 → DP-10); suspend/resume still fails | | Dock on TB5 port (left side, 120 Gb/s) | Correct port confirmed; bug persists | | Dock power-cycle | Recovers display (workaround, not fix) | --- ## Display topology ``` Laptop (Intel i915, card1) ├── eDP-1: internal panel (always connected) └── DP-10: external Dell U4025QW via TB5 dock DisplayPort NVIDIA (card2): all outputs disconnected (compute only, vLLM) ``` External display is driven by **Intel i915**, not NVIDIA, even with vLLM using the discrete GPU. --- ## Steps to reproduce 1. Dell Pro Max 16 Plus with BIOS 2.6.1 2. Connect Dell TB5 dock to left-side Thunderbolt 5 port 3. Connect Dell U4025QW (or similar 5K display) via DisplayPort on dock 4. Boot Ubuntu 26.04 with kernel 7.0.0-27-generic 5. Use dual-monitor setup (internal + external) 6. Suspend the system (close lid or `systemctl suspend`) 7. Resume **Expected:** Normal desktop **Actual:** Black screen(s), input frozen; kernel logs show i915 DPLL/flip_done errors Alternative trigger: leave system idle overnight with external monitor connected. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2160453/+subscriptions
Комментариев нет:
Отправить комментарий