mutter/src/backends/native/meta-kms-update.c

730 lines
20 KiB
C
Raw Normal View History

backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
/*
* Copyright (C) 2018 Red Hat
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation; either version 2 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
* 02111-1307, USA.
*/
#include "config.h"
#include "backends/native/meta-kms-update.h"
#include "backends/native/meta-kms-update-private.h"
#include "backends/meta-display-config-shared.h"
#include "backends/native/meta-kms-connector.h"
#include "backends/native/meta-kms-crtc.h"
#include "backends/native/meta-kms-mode-private.h"
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
#include "backends/native/meta-kms-plane.h"
struct _MetaKmsUpdate
{
MetaKmsDevice *device;
gboolean is_locked;
uint64_t sequence_number;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
GList *mode_sets;
GList *plane_assignments;
GList *connector_updates;
GList *crtc_gammas;
MetaKmsCustomPageFlip *custom_page_flip;
GList *page_flip_listeners;
GList *result_listeners;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
};
void
meta_kms_plane_feedback_free (MetaKmsPlaneFeedback *plane_feedback)
{
g_error_free (plane_feedback->error);
g_free (plane_feedback);
}
MetaKmsPlaneFeedback *
meta_kms_plane_feedback_new_take_error (MetaKmsPlane *plane,
MetaKmsCrtc *crtc,
GError *error)
{
MetaKmsPlaneFeedback *plane_feedback;
plane_feedback = g_new0 (MetaKmsPlaneFeedback, 1);
*plane_feedback = (MetaKmsPlaneFeedback) {
.plane = plane,
.crtc = crtc,
.error = error,
};
return plane_feedback;
}
MetaKmsPlaneFeedback *
meta_kms_plane_feedback_new_failed (MetaKmsPlane *plane,
MetaKmsCrtc *crtc,
const char *error_message)
{
GError *error;
error = g_error_new_literal (G_IO_ERROR, G_IO_ERROR_FAILED, error_message);
return meta_kms_plane_feedback_new_take_error (plane, crtc, error);
}
MetaKmsFeedback *
meta_kms_feedback_new_passed (GList *failed_planes)
{
MetaKmsFeedback *feedback;
feedback = g_new0 (MetaKmsFeedback, 1);
*feedback = (MetaKmsFeedback) {
.result = META_KMS_FEEDBACK_PASSED,
.failed_planes = failed_planes,
};
return feedback;
}
MetaKmsFeedback *
meta_kms_feedback_new_failed (GList *failed_planes,
GError *error)
{
MetaKmsFeedback *feedback;
feedback = g_new0 (MetaKmsFeedback, 1);
*feedback = (MetaKmsFeedback) {
.result = META_KMS_FEEDBACK_FAILED,
.error = error,
.failed_planes = failed_planes,
};
return feedback;
}
void
meta_kms_feedback_free (MetaKmsFeedback *feedback)
{
g_list_free_full (feedback->failed_planes,
(GDestroyNotify) meta_kms_plane_feedback_free);
g_clear_error (&feedback->error);
g_free (feedback);
}
MetaKmsFeedbackResult
meta_kms_feedback_get_result (const MetaKmsFeedback *feedback)
{
return feedback->result;
}
GList *
meta_kms_feedback_get_failed_planes (const MetaKmsFeedback *feedback)
{
return feedback->failed_planes;
}
const GError *
meta_kms_feedback_get_error (const MetaKmsFeedback *feedback)
{
return feedback->error;
}
static void
meta_kms_fb_damage_free (MetaKmsFbDamage *fb_damage)
{
g_free (fb_damage->rects);
g_free (fb_damage);
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
static void
meta_kms_plane_assignment_free (MetaKmsPlaneAssignment *plane_assignment)
{
g_clear_pointer (&plane_assignment->fb_damage, meta_kms_fb_damage_free);
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
g_free (plane_assignment);
}
static void
meta_kms_mode_set_free (MetaKmsModeSet *mode_set)
{
g_list_free (mode_set->connectors);
g_free (mode_set);
}
kms/page-flip: Pass ownership of listener user data along with closure In order to reliably manage the reference count of the user data passed to page flip listeners - being the stage view - make the ownership of this data travel through the different objects that take responsibility of the next step. Initially this is the MetaKmsPageFlipListener that belongs to a MetaKmsUpdate. When a page flip is successfully queued, the ownership is transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. In the simple impl device, the MetaKmsPageFlipData is passed to drmModePageFlip(), then returned back via the DRM event. In the future atomic impl device, the MetaKmsPageFlipData is stored in a table, then retrieved when DRM event are handled. When the DRM events are handled, the page flip listener's interface callbacks are invoked, and after that, the user data is freed using the passed GDestroyNotify function, in the main context, the same as where the interface callbacks were called. When a page flip fails, the ownership is also transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. This page flip data will be passed to the main context via a callback, where it will discard the page flip, and free the user data using the provided GDestroyNotify. Note that this adds back a page flip listener type flag for telling the KMS implementation whether to actively discard a page flip via the interface, or just free the user data. Avoiding discarding via the interface is needed for the direct scanout case, where we immediately need to know the result in order to fall back to the composite pipeline if the direct scanout failed. We do in fact also need active discard via the interface paths, e.g. in the simple impl device when we're asynchronously retrying a page flip, so replace the ad-hoc discard paths in meta-renderer-native.c and replace them by not asking for no-discard page flip error handling. Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1488>
2020-12-16 07:41:14 +00:00
static void
meta_kms_page_flip_listener_free (MetaKmsPageFlipListener *listener)
{
g_clear_pointer (&listener->user_data, listener->destroy_notify);
g_free (listener);
}
static gboolean
drop_plane_assignment (MetaKmsUpdate *update,
MetaKmsPlane *plane,
MetaKmsAssignPlaneFlag *out_flags)
{
GList *l;
for (l = update->plane_assignments; l; l = l->next)
{
MetaKmsPlaneAssignment *plane_assignment = l->data;
if (plane_assignment->plane == plane)
{
update->plane_assignments =
g_list_delete_link (update->plane_assignments, l);
if (out_flags)
*out_flags = plane_assignment->flags;
meta_kms_plane_assignment_free (plane_assignment);
return TRUE;
}
}
return FALSE;
}
void
meta_kms_update_drop_plane_assignment (MetaKmsUpdate *update,
MetaKmsPlane *plane)
{
drop_plane_assignment (update, plane, NULL);
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
MetaKmsPlaneAssignment *
meta_kms_update_assign_plane (MetaKmsUpdate *update,
MetaKmsCrtc *crtc,
MetaKmsPlane *plane,
MetaDrmBuffer *buffer,
MetaFixed16Rectangle src_rect,
MetaRectangle dst_rect,
MetaKmsAssignPlaneFlag flags)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
MetaKmsPlaneAssignment *plane_assignment;
MetaKmsAssignPlaneFlag old_flags;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_crtc_get_device (crtc) == update->device);
g_assert (meta_kms_plane_get_device (plane) == update->device);
g_assert (meta_kms_plane_get_plane_type (plane) !=
META_KMS_PLANE_TYPE_PRIMARY ||
!(flags & META_KMS_ASSIGN_PLANE_FLAG_ALLOW_FAIL));
if (drop_plane_assignment (update, plane, &old_flags))
{
if (!(old_flags & META_KMS_ASSIGN_PLANE_FLAG_FB_UNCHANGED))
flags &= ~META_KMS_ASSIGN_PLANE_FLAG_FB_UNCHANGED;
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
plane_assignment = g_new0 (MetaKmsPlaneAssignment, 1);
*plane_assignment = (MetaKmsPlaneAssignment) {
.update = update,
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
.crtc = crtc,
.plane = plane,
.buffer = buffer,
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
.src_rect = src_rect,
.dst_rect = dst_rect,
.flags = flags,
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
};
update->plane_assignments = g_list_prepend (update->plane_assignments,
plane_assignment);
return plane_assignment;
}
MetaKmsPlaneAssignment *
meta_kms_update_unassign_plane (MetaKmsUpdate *update,
MetaKmsCrtc *crtc,
MetaKmsPlane *plane)
{
MetaKmsPlaneAssignment *plane_assignment;
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_crtc_get_device (crtc) == update->device);
g_assert (meta_kms_plane_get_device (plane) == update->device);
plane_assignment = g_new0 (MetaKmsPlaneAssignment, 1);
*plane_assignment = (MetaKmsPlaneAssignment) {
.update = update,
.crtc = crtc,
.plane = plane,
.buffer = NULL,
};
update->plane_assignments = g_list_prepend (update->plane_assignments,
plane_assignment);
return plane_assignment;
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
void
meta_kms_update_mode_set (MetaKmsUpdate *update,
MetaKmsCrtc *crtc,
GList *connectors,
MetaKmsMode *mode)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
MetaKmsModeSet *mode_set;
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_crtc_get_device (crtc) == update->device);
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
mode_set = g_new0 (MetaKmsModeSet, 1);
*mode_set = (MetaKmsModeSet) {
.crtc = crtc,
.connectors = connectors,
.mode = mode,
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
};
update->mode_sets = g_list_prepend (update->mode_sets, mode_set);
}
static MetaKmsConnectorUpdate *
ensure_connector_update (MetaKmsUpdate *update,
MetaKmsConnector *connector)
{
GList *l;
MetaKmsConnectorUpdate *connector_update;
for (l = update->connector_updates; l; l = l->next)
{
connector_update = l->data;
if (connector_update->connector == connector)
return connector_update;
}
connector_update = g_new0 (MetaKmsConnectorUpdate, 1);
connector_update->connector = connector;
update->connector_updates = g_list_prepend (update->connector_updates,
connector_update);
return connector_update;
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
void
meta_kms_update_set_underscanning (MetaKmsUpdate *update,
MetaKmsConnector *connector,
uint64_t hborder,
uint64_t vborder)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
MetaKmsConnectorUpdate *connector_update;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_connector_get_device (connector) == update->device);
connector_update = ensure_connector_update (update, connector);
connector_update->underscanning.has_update = TRUE;
connector_update->underscanning.is_active = TRUE;
connector_update->underscanning.hborder = hborder;
connector_update->underscanning.vborder = vborder;
}
void
meta_kms_update_unset_underscanning (MetaKmsUpdate *update,
MetaKmsConnector *connector)
{
MetaKmsConnectorUpdate *connector_update;
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_connector_get_device (connector) == update->device);
connector_update = ensure_connector_update (update, connector);
connector_update->underscanning.has_update = TRUE;
connector_update->underscanning.is_active = FALSE;
}
void
meta_kms_update_set_privacy_screen (MetaKmsUpdate *update,
MetaKmsConnector *connector,
gboolean enabled)
{
MetaKmsConnectorUpdate *connector_update;
g_assert (meta_kms_connector_get_device (connector) == update->device);
connector_update = ensure_connector_update (update, connector);
connector_update->privacy_screen.has_update = TRUE;
connector_update->privacy_screen.is_enabled = enabled;
}
void
meta_kms_update_set_max_bpc (MetaKmsUpdate *update,
MetaKmsConnector *connector,
uint64_t max_bpc)
{
MetaKmsConnectorUpdate *connector_update;
g_assert (meta_kms_connector_get_device (connector) == update->device);
connector_update = ensure_connector_update (update, connector);
connector_update->max_bpc.value = max_bpc;
connector_update->max_bpc.has_update = TRUE;
}
void
meta_kms_crtc_gamma_free (MetaKmsCrtcGamma *gamma)
{
g_free (gamma->red);
g_free (gamma->green);
g_free (gamma->blue);
g_free (gamma);
}
MetaKmsCrtcGamma *
meta_kms_crtc_gamma_new (MetaKmsCrtc *crtc,
int size,
const uint16_t *red,
const uint16_t *green,
const uint16_t *blue)
{
MetaKmsCrtcGamma *gamma;
gamma = g_new0 (MetaKmsCrtcGamma, 1);
*gamma = (MetaKmsCrtcGamma) {
.crtc = crtc,
.size = size,
.red = g_memdup2 (red, size * sizeof (*red)),
.green = g_memdup2 (green, size * sizeof (*green)),
.blue = g_memdup2 (blue, size * sizeof (*blue)),
};
return gamma;
}
void
meta_kms_update_set_crtc_gamma (MetaKmsUpdate *update,
MetaKmsCrtc *crtc,
int size,
const uint16_t *red,
const uint16_t *green,
const uint16_t *blue)
{
MetaKmsCrtcGamma *gamma;
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_crtc_get_device (crtc) == update->device);
gamma = meta_kms_crtc_gamma_new (crtc, size, red, green, blue);
update->crtc_gammas = g_list_prepend (update->crtc_gammas, gamma);
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
void
meta_kms_update_add_page_flip_listener (MetaKmsUpdate *update,
MetaKmsCrtc *crtc,
const MetaKmsPageFlipListenerVtable *vtable,
kms/page-flip: Pass ownership of listener user data along with closure In order to reliably manage the reference count of the user data passed to page flip listeners - being the stage view - make the ownership of this data travel through the different objects that take responsibility of the next step. Initially this is the MetaKmsPageFlipListener that belongs to a MetaKmsUpdate. When a page flip is successfully queued, the ownership is transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. In the simple impl device, the MetaKmsPageFlipData is passed to drmModePageFlip(), then returned back via the DRM event. In the future atomic impl device, the MetaKmsPageFlipData is stored in a table, then retrieved when DRM event are handled. When the DRM events are handled, the page flip listener's interface callbacks are invoked, and after that, the user data is freed using the passed GDestroyNotify function, in the main context, the same as where the interface callbacks were called. When a page flip fails, the ownership is also transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. This page flip data will be passed to the main context via a callback, where it will discard the page flip, and free the user data using the provided GDestroyNotify. Note that this adds back a page flip listener type flag for telling the KMS implementation whether to actively discard a page flip via the interface, or just free the user data. Avoiding discarding via the interface is needed for the direct scanout case, where we immediately need to know the result in order to fall back to the composite pipeline if the direct scanout failed. We do in fact also need active discard via the interface paths, e.g. in the simple impl device when we're asynchronously retrying a page flip, so replace the ad-hoc discard paths in meta-renderer-native.c and replace them by not asking for no-discard page flip error handling. Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1488>
2020-12-16 07:41:14 +00:00
MetaKmsPageFlipListenerFlag flags,
gpointer user_data,
GDestroyNotify destroy_notify)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
MetaKmsPageFlipListener *listener;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
g_assert (!meta_kms_update_is_locked (update));
g_assert (meta_kms_crtc_get_device (crtc) == update->device);
listener = g_new0 (MetaKmsPageFlipListener, 1);
*listener = (MetaKmsPageFlipListener) {
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
.crtc = crtc,
.vtable = vtable,
kms/page-flip: Pass ownership of listener user data along with closure In order to reliably manage the reference count of the user data passed to page flip listeners - being the stage view - make the ownership of this data travel through the different objects that take responsibility of the next step. Initially this is the MetaKmsPageFlipListener that belongs to a MetaKmsUpdate. When a page flip is successfully queued, the ownership is transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. In the simple impl device, the MetaKmsPageFlipData is passed to drmModePageFlip(), then returned back via the DRM event. In the future atomic impl device, the MetaKmsPageFlipData is stored in a table, then retrieved when DRM event are handled. When the DRM events are handled, the page flip listener's interface callbacks are invoked, and after that, the user data is freed using the passed GDestroyNotify function, in the main context, the same as where the interface callbacks were called. When a page flip fails, the ownership is also transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. This page flip data will be passed to the main context via a callback, where it will discard the page flip, and free the user data using the provided GDestroyNotify. Note that this adds back a page flip listener type flag for telling the KMS implementation whether to actively discard a page flip via the interface, or just free the user data. Avoiding discarding via the interface is needed for the direct scanout case, where we immediately need to know the result in order to fall back to the composite pipeline if the direct scanout failed. We do in fact also need active discard via the interface paths, e.g. in the simple impl device when we're asynchronously retrying a page flip, so replace the ad-hoc discard paths in meta-renderer-native.c and replace them by not asking for no-discard page flip error handling. Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1488>
2020-12-16 07:41:14 +00:00
.flags = flags,
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
.user_data = user_data,
kms/page-flip: Pass ownership of listener user data along with closure In order to reliably manage the reference count of the user data passed to page flip listeners - being the stage view - make the ownership of this data travel through the different objects that take responsibility of the next step. Initially this is the MetaKmsPageFlipListener that belongs to a MetaKmsUpdate. When a page flip is successfully queued, the ownership is transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. In the simple impl device, the MetaKmsPageFlipData is passed to drmModePageFlip(), then returned back via the DRM event. In the future atomic impl device, the MetaKmsPageFlipData is stored in a table, then retrieved when DRM event are handled. When the DRM events are handled, the page flip listener's interface callbacks are invoked, and after that, the user data is freed using the passed GDestroyNotify function, in the main context, the same as where the interface callbacks were called. When a page flip fails, the ownership is also transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. This page flip data will be passed to the main context via a callback, where it will discard the page flip, and free the user data using the provided GDestroyNotify. Note that this adds back a page flip listener type flag for telling the KMS implementation whether to actively discard a page flip via the interface, or just free the user data. Avoiding discarding via the interface is needed for the direct scanout case, where we immediately need to know the result in order to fall back to the composite pipeline if the direct scanout failed. We do in fact also need active discard via the interface paths, e.g. in the simple impl device when we're asynchronously retrying a page flip, so replace the ad-hoc discard paths in meta-renderer-native.c and replace them by not asking for no-discard page flip error handling. Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1488>
2020-12-16 07:41:14 +00:00
.destroy_notify = destroy_notify,
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
};
update->page_flip_listeners = g_list_prepend (update->page_flip_listeners,
listener);
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
}
void
meta_kms_update_drop_defunct_page_flip_listeners (MetaKmsUpdate *update)
{
GList *l;
l = update->page_flip_listeners;
while (l)
{
MetaKmsPageFlipListener *listener = l->data;
GList *l_next = l->next;
if (listener->flags & META_KMS_PAGE_FLIP_LISTENER_FLAG_DROP_ON_ERROR)
{
meta_kms_page_flip_listener_free (listener);
update->page_flip_listeners =
g_list_delete_link (update->page_flip_listeners, l);
}
l = l_next;
}
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
void
meta_kms_update_set_custom_page_flip (MetaKmsUpdate *update,
MetaKmsCustomPageFlipFunc func,
gpointer user_data)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
MetaKmsCustomPageFlip *custom_page_flip;
g_assert (!meta_kms_update_is_locked (update));
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
custom_page_flip = g_new0 (MetaKmsCustomPageFlip, 1);
custom_page_flip->func = func;
custom_page_flip->user_data = user_data;
update->custom_page_flip = custom_page_flip;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
}
void
meta_kms_plane_assignment_set_fb_damage (MetaKmsPlaneAssignment *plane_assignment,
const int *rectangles,
int n_rectangles)
{
MetaKmsFbDamage *fb_damage;
struct drm_mode_rect *mode_rects;
int i;
mode_rects = g_new0 (struct drm_mode_rect, n_rectangles);
for (i = 0; i < n_rectangles; ++i)
{
mode_rects[i].x1 = rectangles[i * 4];
mode_rects[i].y1 = rectangles[i * 4 + 1];
mode_rects[i].x2 = mode_rects[i].x1 + rectangles[i * 4 + 2];
mode_rects[i].y2 = mode_rects[i].y1 + rectangles[i * 4 + 3];
}
fb_damage = g_new0 (MetaKmsFbDamage, 1);
*fb_damage = (MetaKmsFbDamage) {
.rects = mode_rects,
.n_rects = n_rectangles,
};
plane_assignment->fb_damage = fb_damage;
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
void
meta_kms_plane_assignment_set_rotation (MetaKmsPlaneAssignment *plane_assignment,
MetaKmsPlaneRotation rotation)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
g_assert (!meta_kms_update_is_locked (plane_assignment->update));
g_warn_if_fail (rotation);
plane_assignment->rotation = rotation;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
}
void
meta_kms_plane_assignment_set_cursor_hotspot (MetaKmsPlaneAssignment *plane_assignment,
int x,
int y)
{
plane_assignment->cursor_hotspot.is_valid = TRUE;
plane_assignment->cursor_hotspot.x = x;
plane_assignment->cursor_hotspot.y = y;
}
void
meta_kms_update_add_result_listener (MetaKmsUpdate *update,
MetaKmsResultListenerFunc func,
gpointer user_data)
{
MetaKmsResultListener *listener;
listener = g_new0 (MetaKmsResultListener, 1);
*listener = (MetaKmsResultListener) {
.func = func,
.user_data = user_data,
};
update->result_listeners = g_list_append (update->result_listeners,
listener);
}
void
meta_kms_update_remove_result_listeners (MetaKmsUpdate *update,
MetaKmsResultListenerFunc func,
gpointer user_data)
{
GList *l;
for (l = update->result_listeners; l;)
{
MetaKmsResultListener *listener = l->data;
GList *next = l->next;
if (listener->user_data == func && listener->user_data == user_data)
{
update->result_listeners =
g_list_delete_link (update->result_listeners, l);
}
l = next;
}
}
GList *
meta_kms_update_take_result_listeners (MetaKmsUpdate *update)
{
return g_steal_pointer (&update->result_listeners);
}
void
meta_kms_result_listener_notify (MetaKmsResultListener *listener,
const MetaKmsFeedback *feedback)
{
listener->func (feedback, listener->user_data);
}
void
meta_kms_result_listener_free (MetaKmsResultListener *listener)
{
g_free (listener);
}
static MetaKmsPlaneAssignment *
get_first_plane_assignment (MetaKmsUpdate *update,
MetaKmsCrtc *crtc,
MetaKmsPlaneType plane_type)
{
GList *l;
for (l = meta_kms_update_get_plane_assignments (update); l; l = l->next)
{
MetaKmsPlaneAssignment *plane_assignment = l->data;
if (meta_kms_plane_get_plane_type (plane_assignment->plane) !=
plane_type)
continue;
if (plane_assignment->crtc != crtc)
continue;
return plane_assignment;
}
return NULL;
}
MetaKmsPlaneAssignment *
meta_kms_update_get_primary_plane_assignment (MetaKmsUpdate *update,
MetaKmsCrtc *crtc)
{
return get_first_plane_assignment (update, crtc, META_KMS_PLANE_TYPE_PRIMARY);
}
MetaKmsPlaneAssignment *
meta_kms_update_get_cursor_plane_assignment (MetaKmsUpdate *update,
MetaKmsCrtc *crtc)
{
return get_first_plane_assignment (update, crtc, META_KMS_PLANE_TYPE_CURSOR);
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
GList *
meta_kms_update_get_plane_assignments (MetaKmsUpdate *update)
{
return update->plane_assignments;
}
GList *
meta_kms_update_get_mode_sets (MetaKmsUpdate *update)
{
return update->mode_sets;
}
GList *
meta_kms_update_get_page_flip_listeners (MetaKmsUpdate *update)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
return update->page_flip_listeners;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
}
GList *
meta_kms_update_get_connector_updates (MetaKmsUpdate *update)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
return update->connector_updates;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
}
GList *
meta_kms_update_get_crtc_gammas (MetaKmsUpdate *update)
{
return update->crtc_gammas;
}
void
meta_kms_update_lock (MetaKmsUpdate *update)
{
update->is_locked = TRUE;
}
void
meta_kms_update_unlock (MetaKmsUpdate *update)
{
update->is_locked = FALSE;
}
gboolean
meta_kms_update_is_locked (MetaKmsUpdate *update)
{
return update->is_locked;
}
MetaKmsDevice *
meta_kms_update_get_device (MetaKmsUpdate *update)
{
return update->device;
}
MetaKmsCustomPageFlip *
meta_kms_update_take_custom_page_flip_func (MetaKmsUpdate *update)
{
return g_steal_pointer (&update->custom_page_flip);
}
void
meta_kms_custom_page_flip_free (MetaKmsCustomPageFlip *custom_page_flip)
{
g_free (custom_page_flip);
}
uint64_t
meta_kms_update_get_sequence_number (MetaKmsUpdate *update)
{
return update->sequence_number;
}
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
MetaKmsUpdate *
meta_kms_update_new (MetaKmsDevice *device)
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
{
MetaKmsUpdate *update;
static uint64_t sequence_number = 0;
update = g_new0 (MetaKmsUpdate, 1);
update->device = device;
update->sequence_number = sequence_number++;
return update;
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
}
void
meta_kms_update_free (MetaKmsUpdate *update)
{
g_list_free_full (update->result_listeners,
(GDestroyNotify) meta_kms_result_listener_free);
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
g_list_free_full (update->plane_assignments,
(GDestroyNotify) meta_kms_plane_assignment_free);
g_list_free_full (update->mode_sets,
(GDestroyNotify) meta_kms_mode_set_free);
kms/page-flip: Pass ownership of listener user data along with closure In order to reliably manage the reference count of the user data passed to page flip listeners - being the stage view - make the ownership of this data travel through the different objects that take responsibility of the next step. Initially this is the MetaKmsPageFlipListener that belongs to a MetaKmsUpdate. When a page flip is successfully queued, the ownership is transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. In the simple impl device, the MetaKmsPageFlipData is passed to drmModePageFlip(), then returned back via the DRM event. In the future atomic impl device, the MetaKmsPageFlipData is stored in a table, then retrieved when DRM event are handled. When the DRM events are handled, the page flip listener's interface callbacks are invoked, and after that, the user data is freed using the passed GDestroyNotify function, in the main context, the same as where the interface callbacks were called. When a page flip fails, the ownership is also transferred to a MetaKmsPageFlipClosure that is part of a MetaKmsPageFlipData. This page flip data will be passed to the main context via a callback, where it will discard the page flip, and free the user data using the provided GDestroyNotify. Note that this adds back a page flip listener type flag for telling the KMS implementation whether to actively discard a page flip via the interface, or just free the user data. Avoiding discarding via the interface is needed for the direct scanout case, where we immediately need to know the result in order to fall back to the composite pipeline if the direct scanout failed. We do in fact also need active discard via the interface paths, e.g. in the simple impl device when we're asynchronously retrying a page flip, so replace the ad-hoc discard paths in meta-renderer-native.c and replace them by not asking for no-discard page flip error handling. Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1488>
2020-12-16 07:41:14 +00:00
g_list_free_full (update->page_flip_listeners,
(GDestroyNotify) meta_kms_page_flip_listener_free);
g_list_free_full (update->connector_updates, g_free);
g_list_free_full (update->crtc_gammas, (GDestroyNotify) meta_kms_crtc_gamma_free);
g_clear_pointer (&update->custom_page_flip, meta_kms_custom_page_flip_free);
backend/native: Add and use transactional KMS API This commit introduces, and makes use of, a transactional API used for setting up KMS state, later to be applied, potentially atomically. From an API point of view, so is always the case, but in the current implementation, it still uses legacy drmMode* API to apply the state non-atomically. The API consists of various buliding blocks: * MetaKmsUpdate - a set of configuration changes, the higher level handle for handing over configuration to the impl backend. It's used to set mode, assign framebuffers to planes, queue page flips and set connector properties. * MetaKmsPlaneAssignment - the assignment of a framebuffer to a plane. Currently used to map a framebuffer to the primary plane of a CRTC. In the legacy KMS implementation, the plane assignment is used to derive the framebuffer used for mode setting and page flipping. This also means various high level changes: State, excluding configuring the cursor plane and creating/destroying DRM framebuffer handles, are applied in the end of a clutter frame, in one go. From an API point of view, this is done atomically, but as mentioned, only the non-atomic implementation exists so far. From MetaRendererNative's point of view, a page flip now initially always succeeds; the handling of EBUSY errors are done asynchronously in the MetaKmsImpl backend (still by retrying at refresh rate, but postponing flip callbacks instead of manipulating the frame clock). Handling of falling back to mode setting instead of page flipping is notified after the fact by a more precise page flip feedback API. EGLStream based page flipping relies on the impl backend not being atomic, as the page flipping is done in the EGLStream backend (e.g. nvidia driver). It uses a 'custom' page flip queueing method, keeping the EGLStream logic inside meta-renderer-native.c. Page flip handling is moved to meta-kms-impl-device.c from meta-gpu-kms.c. It goes via an extra idle callback before reaching meta-renderer-native.c to make sure callbacks are invoked outside of the impl context. While dummy power save page flipping is kept in meta-renderer-native.c, the EBUSY handling is moved to meta-kms-impl-simple.c. Instead of freezing the frame clock, actual page flip callbacks are postponed until all EBUSY retries have either succeeded or failed due to some other error than EBUSY. This effectively inhibits new frames to be drawn, meaning we won't stall waiting on the file descriptor for pending page flips. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-04-04 20:36:41 +00:00
g_free (update);
}