lxc GLX blit vs. flip

[mismacku]

• Distribution: Ubuntu
• Distribution version: 16.04.6
• Nvidia P5000 and K5000
• Nvidia driver: 418.67
• cuda version: 10.1
• lxc info:

  config:
  core.https_address: '[::]:8443'
  core.trust_password: true
  api_extensions:
  - storage_zfs_remove_snapshots
  - container_host_shutdown_timeout
  - container_stop_priority
  - container_syscall_filtering
  - auth_pki
  - container_last_used_at
  - etag
  - patch
  - usb_devices
  - https_allowed_credentials
  - image_compression_algorithm
  - directory_manipulation
  - container_cpu_time
  - storage_zfs_use_refquota
  - storage_lvm_mount_options
  - network
  - profile_usedby
  - container_push
  - container_exec_recording
  - certificate_update
  - container_exec_signal_handling
  - gpu_devices
  - container_image_properties
  - migration_progress
  - id_map
  - network_firewall_filtering
  - network_routes
  - storage
  - file_delete
  - file_append
  - network_dhcp_expiry
  - storage_lvm_vg_rename
  - storage_lvm_thinpool_rename
  - network_vlan
  - image_create_aliases
  - container_stateless_copy
  - container_only_migration
  - storage_zfs_clone_copy
  - unix_device_rename
  - storage_lvm_use_thinpool
  - storage_rsync_bwlimit
  - network_vxlan_interface
  - storage_btrfs_mount_options
  - entity_description
  - image_force_refresh
  - storage_lvm_lv_resizing
  - id_map_base
  - file_symlinks
  - container_push_target
  - network_vlan_physical
  - storage_images_delete
  - container_edit_metadata
  - container_snapshot_stateful_migration
  - storage_driver_ceph
  - storage_ceph_user_name
  - resource_limits
  - storage_volatile_initial_source
  - storage_ceph_force_osd_reuse
  - storage_block_filesystem_btrfs
  - resources
  - kernel_limits
  - storage_api_volume_rename
  - macaroon_authentication
  - network_sriov
  - console
  - restrict_devlxd
  - migration_pre_copy
  - infiniband
  - maas_network
  - devlxd_events
  - proxy
  - network_dhcp_gateway
  - file_get_symlink
  - network_leases
  - unix_device_hotplug
  - storage_api_local_volume_handling
  - operation_description
  - clustering
  - event_lifecycle
  - storage_api_remote_volume_handling
  - nvidia_runtime
  - container_mount_propagation
  - container_backup
  - devlxd_images
  - container_local_cross_pool_handling
  - proxy_unix
  - proxy_udp
  - clustering_join
  - proxy_tcp_udp_multi_port_handling
  - network_state
  - proxy_unix_dac_properties
  - container_protection_delete
  - unix_priv_drop
  - pprof_http
  - proxy_haproxy_protocol
  - network_hwaddr
  - proxy_nat
  - network_nat_order
  - container_full
  - candid_authentication
  - backup_compression
  - candid_config
  - nvidia_runtime_config
  - storage_api_volume_snapshots
  - storage_unmapped
  - projects
  - candid_config_key
  - network_vxlan_ttl
  - container_incremental_copy
  - usb_optional_vendorid
  - snapshot_scheduling
  - container_copy_project
  - clustering_server_address
  - clustering_image_replication
  - container_protection_shift
  - snapshot_expiry
  - container_backup_override_pool
  - snapshot_expiry_creation
  - network_leases_location
  - resources_cpu_socket
  - resources_gpu
  - resources_numa
  - kernel_features
  - id_map_current
  - event_location
  - storage_api_remote_volume_snapshots
  - network_nat_address
  - container_nic_routes
  - rbac
  - cluster_internal_copy
  - seccomp_notify
  - lxc_features
  - container_nic_ipvlan
  - network_vlan_sriov
  - storage_cephfs
  - container_nic_ipfilter
  - resources_v2
  - container_exec_user_group_cwd
  - container_syscall_intercept
  api_status: stable
  api_version: "1.0"
  auth: trusted
  public: false
  auth_methods:
  - tls
  environment:
  addresses:
  - 10.137.124.131:8443
  - 172.28.0.11:8443
  - 10.206.37.1:8443
  - 172.19.0.1:8443
  - 172.17.0.1:8443
  architectures:
  - x86_64
  - i686
  driver: lxc
  driver_version: 3.1.0 (devel)
  kernel: Linux
  kernel_architecture: x86_64
  kernel_features:
  netnsid_getifaddrs: "false"
  seccomp_listener: "false"
  shiftfs: "false"
  uevent_injection: "false"
  unpriv_fscaps: "true"
  kernel_version: 4.15.0-15-generic
  lxc_features:
  mount_injection_file: "true"
  network_gateway_device_route: "true"
  network_ipvlan: "true"
  network_l2proxy: "true"
  network_phys_macvlan_mtu: "true"
  seccomp_notify: "true"
  project: default
  server: lxd
  server_clustered: false
  server_name: test-rig
  server_pid: 21863
  server_version: "3.15"
  storage: zfs
  storage_version: 0.7.5-1ubuntu13

  profile

  config:
  environment.DISPLAY: :0
  nvidia.driver.capabilities: compute, display, graphics, utility, video
  nvidia.runtime: "true"
  description: Default LXD profile
  devices:
  X0:
  path: /tmp/.X11-unix/X0
  source: /tmp/.X11-unix/X0
  type: disk
  Xauthority:
  path: /root/.Xauthority
  source: /home/oblong/.Xauthority
  type: disk
  eth0:
  name: eth0
  nictype: macvlan
  parent: enp0s25
  type: nic
  gpu:
  type: gpu
  root:
  path: /
  pool: default
  type: disk
  name: default

  Issue description

Let me first say, thanks for all of the hard work!
I'm revisiting an issue from several years ago because I really want this to work! https://github.com/lxc/lxd/issues/3826
^^^ That issue being poor framerate with GLX apps in lxc containers. I'm getting similar framerates as I did two years ago, but I discovered that on the host the nvidia driver always shows FLIP and in lxc containers (Ubuntu 16.04 and 18.04) it shows BLIT regardless of what nvidia-settings shows (Attribute 'AllowFlipping' (test-rig:0.0): 1). I'm guessing that was the performance issue then and still is now. Any ideas? Any help is greatly appreciated.

Steps to reproduce

1. using nvidia-settings under OpenGL Settings, tick Enable Graphics API Visual Indicator
2. Use the lxc profile above
3. start an Ubuntu 18.04 container
4. apt install glmark2
5. run glmark (glmark2 --fullscreen)
6. notice the visual indicator shows BLIT - VSYNC OFF Note: the VI shows BLIT regardless of the GLX app I'm running (glxgears, glmark2, in-house GLX apps)

   Information to attach

   container log:

   
   lxc info --show-log test1804
   Name: test1804
   Location: none
   Remote: unix://
   Architecture: x86_64
   Created: 2019/07/18 14:54 UTC
   Status: Running
   Type: persistent
   Profiles: default
   Pid: 9388
   Ips:
   eth0: inet    10.137.124.115
   eth0: inet6   fe80::216:3eff:fe53:8224
   lo:   inet    127.0.0.1
   lo:   inet6   ::1
   Resources:
   Processes: 25
   Disk usage:
   root: 147.16MB
   CPU usage:
   CPU usage (in seconds): 10
   Memory usage:
   Memory (current): 76.13MB
   Memory (peak): 124.46MB
   Network usage:
   eth0:
     Bytes received: 325.03kB
     Bytes sent: 2.13kB
     Packets received: 2798
     Packets sent: 22
   lo:
     Bytes received: 332B
     Bytes sent: 332B
     Packets received: 4
     Packets sent: 4

Log:

lxc test1804 20190718161841.826 ERROR start - start.c:proc_pidfd_open:1644 - Function not implemented - Failed to send signal through pidfd

`lxc monitor` shows nothing.

[stgraber]

Hmm, so LXD doesn't interact with the graphics driver in any way. All it does is pass /dev/drm/ and /dev/nvidia* to the container.

I'd recommend maybe opening an issue with https://github.com/NVIDIA/libnvidia-container as that's what's used to pass the various GL libraries. Maybe your issue is caused by a missing library or something similar.

[mismacku]

I'll open an issue there. Thanks Stephane!

[stgraber]

Perfect, thanks!