Skip to content
GitLab
  • Menu
Projects Groups Snippets
  • Help
    • Help
    • Support
    • Community forum
    • Submit feedback
  • Sign in / Register
  • L linux
  • Project information
    • Project information
    • Activity
    • Labels
    • Members
  • Repository
    • Repository
    • Files
    • Commits
    • Branches
    • Tags
    • Contributors
    • Graph
    • Compare
  • Issues 129
    • Issues 129
    • List
    • Boards
    • Service Desk
    • Milestones
  • Merge requests 23
    • Merge requests 23
  • CI/CD
    • CI/CD
    • Pipelines
    • Jobs
    • Schedules
  • Deployments
    • Deployments
    • Environments
    • Releases
  • Packages & Registries
    • Packages & Registries
    • Package Registry
    • Container Registry
    • Infrastructure Registry
  • Monitor
    • Monitor
    • Incidents
  • Analytics
    • Analytics
    • Value stream
    • CI/CD
    • Repository
  • Wiki
    • Wiki
  • Snippets
    • Snippets
  • Activity
  • Graph
  • Create a new issue
  • Jobs
  • Commits
  • Issue Boards
Collapse sidebar
  • Librem5
  • linux
  • Issues
  • #183
Closed
Open
Created Apr 04, 2020 by Sebastian Krzyszkowiak@sebastian.krzyszkowiakMaintainer

Etnaviv MMU faults, broken state after recovering hung GPU

Sometimes when playing with OpenGL-heavy applications, I get hung GPU due to MMU faults. The driver thinks that the GPU has been recovered, but it's not true - the screen stays frozen until a reboot. When that happens, the system generally works, but some parts of it are unusable - for instance, phoc doesn't blank the screen anymore and trying to restart it just hangs.

[ 8891.411193] etnaviv-gpu 38000000.gpu: MMU fault status 0x00000002
[ 8891.417307] etnaviv-gpu 38000000.gpu: MMU 0 fault addr 0xfe472000
[ 8892.567778] device: 'devcd1': device_add
[ 8892.567851] PM: Adding info for No Bus:devcd1
[ 8892.567952] etnaviv-gpu 38000000.gpu: recover hung GPU!

Seems like the easiest way to trigger it is to run Pure Maps and just play with it for a few minutes.

Right now I'm seeing this on the 5.6 kernel; I don't remember whether it happened on 5.3 as well or not.

Initially mentioned in https://source.puri.sm/Librem5/linux-next/issues/151#note_94840

Edited Apr 04, 2020 by Sebastian Krzyszkowiak
Assignee
Assign to
Time tracking