Debugging

To adjust the logging verbosity level, configure the following environment variables:

SRM_DEBUG=N (default = 0)

SRM_EGL_DEBUG=N (default = 0)

Where N can be 0 = Disabled, 1 = Fatal, 2 = Error, 3 = Warning or 4 = Debug.

DRM API

SRM defaults to using the Atomic DRM API for all devices (when avaliable). To enforce the use of the legacy API set the following variable to 1.

SRM_FORCE_LEGACY_API=[0,1] (default = 0)

Writeback connectors are specialized virtual connectors designed to access pixel data from offscreen plane compositions. They are particularly useful for capturing and processing the final output of the display pipeline.

SRM_ENABLE_WRITEBACK_CONNECTORS=[0,1] (default = 0)

Cursor Planes

SRM_FORCE_LEGACY_CURSOR=[0,1] (default = 1)
The legacy cursor IOCTLs are used by default, even when the atomic API is enabled, as they allow the cursor to be updated asynchronously, providing a much smoother experience.
SRM_NVIDIA_CURSOR=[0,1] (default = 0)
Cursor planes are disabled by default for proprietary NVIDIA drivers, as updating the cursor can sometimes cause screen stutter. Nouveau drivers are not affected by this option.
SRM_DISABLE_CURSOR=[0,1] (default = 0)
This setting can be used to disable cursor planes for all drivers.

Device Blacklisting

To disable specific DRM devices, list the devices separated by ":", for example:

SRM_DEVICES_BLACKLIST=/dev/dri/card0:/dev/dri/card1

Buffer Allocation

By default, SRM automatically selects the integrated GPU for buffer allocation. To override the default allocator, use:

SRM_ALLOCATOR_DEVICE=/dev/dri/card[N]

SRM uses GBM for buffer allocation from main memory, resorting to OpenGL when the former fails. To enforce the use of OpenGL employ:

SRM_FORCE_GL_ALLOCATION=1 (default = 0)

Render Buffering

All connectors use double buffering by default. You can customize the number of framebuffers used for each rendering mode using the following environment variables:

SRM_RENDER_MODE_ITSELF_FB_COUNT=N

SRM_RENDER_MODE_PRIME_FB_COUNT=N

SRM_RENDER_MODE_DUMB_FB_COUNT=N

SRM_RENDER_MODE_CPU_FB_COUNT=N

Where N can be 2 = Double or 3 = Triple buffering.

Triple buffering can provide a smoother experience by allowing a new frame to be rendered while a page flip is pending. However, it consumes more memory and power and can introduce some latency.

Direct Scanout

Scanning out custom buffers is allowed by default (see srmConnectorSetCustomScanoutBuffer()). This allows, for example, compositors to directly present fullscreen application windows without needing to render them using OpenGL.

To disable it, set:

SRM_DISABLE_CUSTOM_SCANOUT=1 (default = 0)