-
Notifications
You must be signed in to change notification settings - Fork 26
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
drm/virtio: Send flip event based on backend setting #72
base: celadon/u/mr0/master
Are you sure you want to change the base?
drm/virtio: Send flip event based on backend setting #72
Conversation
Improper Commit Message |
virtqueue_enable_cb(vgdev->vblank[output->index].vblank.vq); | ||
do { | ||
virtio_gpu_vblank_poll_arm(vgdev->vblank[output->index].vblank.vq); | ||
} while (!virtqueue_enable_cb(vgdev->vblank[output->index].vblank.vq)); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In which case, virtqueue_enable_cb will return false and need run multiple times, then success?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
robust enhance to avoid dead loop.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should be safe enough to use like this. We can see the same usage in virtio-GPU.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Two commits seems change same logic, should be combined together?
0dc1620
to
0390a3c
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Improper Commit Message |
Introduce per-CRTC flip counter that should be incremented when flip happens in back-end. Atomic commit worker should await change of the new flip sequence rather than vblank sequence. Present fences are also cached and signaled when flip happens. We introduce flip sequence feature for backward compatibility, meaning that we must wait for flip event only when we have flip sequence feature negotiated. This is to avoid breaks old hypervisor/device model. Tracked-On: OAM-128741 Signed-off-by: hangliu1 <[email protected]> Signed-off-by: Xue, Bosheng <[email protected]> Signed-off-by: Weifeng Liu <[email protected]>
78ae568
to
4f02e65
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Cache out fences in atomic commit and signal them when the backend asks to. The backend is supposed to send to frontend the number of fences to signal when the backend presents a frame to users. Only in this way we can preserve the semantics of present fence.
Tracked-On: OAM-128370