mirror of
https://github.com/torvalds/linux.git
synced 2024-12-11 21:52:04 +00:00
4c813d4d75
So it looks like for virtual hw cursors on QXL we need to inform the "hw" device what the cursor hotspot parameters are. This makes sense if you think the host has to draw the cursor and interpret clicks from it. However the current modesetting interface doesn't support passing the hotspot information from userspace. This implements a new cursor ioctl, that takes the hotspot info as well, userspace can try calling the new interface and if it gets -ENOSYS it means its on an older kernel and can just fallback. Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com> |
||
---|---|---|
.. | ||
i2c | ||
ttm | ||
drm_buffer.h | ||
drm_cache.h | ||
drm_core.h | ||
drm_crtc_helper.h | ||
drm_crtc.h | ||
drm_dp_helper.h | ||
drm_edid.h | ||
drm_encoder_slave.h | ||
drm_fb_cma_helper.h | ||
drm_fb_helper.h | ||
drm_fixed.h | ||
drm_gem_cma_helper.h | ||
drm_global.h | ||
drm_hashtab.h | ||
drm_mem_util.h | ||
drm_memory.h | ||
drm_mm.h | ||
drm_os_linux.h | ||
drm_pciids.h | ||
drm_rect.h | ||
drm_sysfs.h | ||
drm_usb.h | ||
drmP.h | ||
exynos_drm.h | ||
gma_drm.h | ||
i915_drm.h | ||
intel-gtt.h |