mirror of https://gitlab.freedesktop.org/mesa/mesa
75b1f3d39d
I had this as abort() in my original implementation since I was doing drm-shim and my kernel driver in parallel based around using a SW simulator, and I wanted to always update both, but it means that people's new feature detection code can easily end up breaing their drm-shim shader-db runs (such as intel's kernel_has_dynamic_config_support() checking for -ENOENT instead of -EINVAL for a feature, which showed up on my personal runner but not fd.o's for reasons I'm unclear on). Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5994> |
||
|---|---|---|
| .. | ||
| README.md | ||
| device.c | ||
| drm_shim.c | ||
| drm_shim.h | ||
| meson.build | ||
README.md
DRM shim - Fake GEM kernel drivers in userspace for CI
On CI systems where we don't control the kernel, it would be nice to be able to present either no-op GEM devices (for shader-db runs) or simulator-backed GEM devices (for testing against a software simulator or FPGA). This lets us do that by intercepting libc calls and exposing render nodes.
Limitations
- Doesn't know how to handle DRM fds getting passed over the wire from X11 (Could we use kmsro to support the X11 case?).
- libc interception is rather glibc-specific and fragile.
- Can easily break gdb if the libc interceptor code is what's broken. (ulimit -c unlimited and doing gdb on the core after the fact can help)
Using
You choose the backend by setting LD_PRELOAD to the shim you want.
Since this will effectively fake another DRM device to your system,
you may need some work on your userspace to get your test application
to use it if it's not the only DRM device present. Setting
DRM_SHIM_DEBUG=1 in the environment will print out what path the
shim initialized on.
For piglit tests, you can set:
PIGLIT_PLATFORM=gbm
WAFFLE_GBM_DEVICE=<path from DRM_SHIM_DEBUG>
See your drm-shim backend's README for details on how to use it.