Intel i915 reports a GPU hang · Linux Fix Lab

Symptoms & scope

An application freezes and the display may recover after a reset.
Kernel logs identify i915 GPU HANG with an error code.

Relevant environment

Intel GPUs bound to i915. Newer GPUs or configurations using xe have different capture interfaces and should not use i915-specific paths blindly.

Recognizable messages (synthetic examples)

i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dffffb, in render-app [2400]

Preserve the error capture before another incident; this signature deliberately does not classify CPU watchdog events or xe failures.

Possible causes

These are possible explanations, not a confirmed diagnosis. Several independent faults can coexist.

Kernel scheduling or a userspace driver submission bug may stop an engine.
Power-management or firmware interactions and hardware instability remain possible; the error code needs the surrounding capture.

Diagnose safely

Run one command at a time in the relevant session. Read the explanation first. Uppercase placeholders need your own values; tools and privileges vary by distribution. These commands are displayed here and never executed by the website.

Check 1

Read Intel GPU kernel events; journal access may require an administrator.

journalctl -b -k --no-pager --grep='i915|GPU HANG|xe '

Interpret the result: Confirm whether i915 or xe is the active driver and identify the first failing engine. A later reset line is an attempted recovery, not an independent cause.

Check 2

Replace card0 with the i915 card resolved through lspci/sysfs. Read only if this interface exists; access may require administrator rights.

cat /sys/class/drm/card0/error

Interpret the result: A retained error state is valuable for a driver report. No error state collected does not exclude a hang; capture may be disabled or already cleared.

Evidence-guided next steps

Preserve the first error capture

If the system remains accessible, save the first i915 error state and complete kernel messages before rebooting or retriggering. Attach the GPU PCI ID, kernel and Mesa versions to a minimal upstream report.

Precautions: Review captures for process names or paths before publication. Do not clear the error state or manually reset the live display GPU to collect evidence.

Recovery / rollback: Reading capture data changes no GPU configuration; return to a normal supported boot after collecting the report.

Did this solution help you?

Share this solution #

Compare a retained supported stack

If the first hang follows an update, compare an installed supported earlier kernel or Mesa snapshot with the same minimal workload. Remove only optional application overlays first when the issue is limited to that application.

Precautions: Change one layer at a time and retain a working boot entry. Avoid globally disabling power features based only on an unrelated GPU report.

Recovery / rollback: Select the original kernel or restore the package snapshot and application profile if the comparison does not improve reproducibility.