1
0
Fork 0
mutter-performance-source/src/backends/native/meta-kms-impl-device.c

504 lines
13 KiB
C
Raw Normal View History

backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
/*
* Copyright (C) 2019 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-impl-device.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 <errno.h>
#include <xf86drm.h>
#include "backends/native/meta-kms-connector-private.h"
#include "backends/native/meta-kms-connector.h"
#include "backends/native/meta-kms-crtc-private.h"
#include "backends/native/meta-kms-crtc.h"
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
#include "backends/native/meta-kms-impl.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-page-flip-private.h"
#include "backends/native/meta-kms-plane-private.h"
#include "backends/native/meta-kms-plane.h"
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
#include "backends/native/meta-kms-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-update.h"
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
struct _MetaKmsImplDevice
{
GObject parent;
MetaKmsDevice *device;
MetaKmsImpl *impl;
int fd;
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
GSource *fd_source;
GList *crtcs;
GList *connectors;
GList *planes;
MetaKmsDeviceCaps caps;
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
};
G_DEFINE_TYPE (MetaKmsImplDevice, meta_kms_impl_device, G_TYPE_OBJECT)
MetaKmsDevice *
meta_kms_impl_device_get_device (MetaKmsImplDevice *impl_device)
{
return impl_device->device;
}
GList *
meta_kms_impl_device_copy_connectors (MetaKmsImplDevice *impl_device)
{
return g_list_copy (impl_device->connectors);
}
GList *
meta_kms_impl_device_copy_crtcs (MetaKmsImplDevice *impl_device)
{
return g_list_copy (impl_device->crtcs);
}
GList *
meta_kms_impl_device_copy_planes (MetaKmsImplDevice *impl_device)
{
return g_list_copy (impl_device->planes);
}
const MetaKmsDeviceCaps *
meta_kms_impl_device_get_caps (MetaKmsImplDevice *impl_device)
{
return &impl_device->caps;
}
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
page_flip_handler (int fd,
unsigned int sequence,
unsigned int sec,
unsigned int usec,
void *user_data)
{
MetaKmsPageFlipData *page_flip_data = user_data;
MetaKmsImpl *impl;
meta_kms_page_flip_data_set_timings_in_impl (page_flip_data,
sequence, sec, usec);
impl = meta_kms_page_flip_data_get_kms_impl (page_flip_data);
meta_kms_impl_handle_page_flip_callback (impl, page_flip_data);
}
gboolean
meta_kms_impl_device_dispatch (MetaKmsImplDevice *impl_device,
GError **error)
{
drmEventContext drm_event_context;
meta_assert_in_kms_impl (meta_kms_impl_get_kms (impl_device->impl));
drm_event_context = (drmEventContext) { 0 };
drm_event_context.version = 2;
drm_event_context.page_flip_handler = page_flip_handler;
while (TRUE)
{
if (drmHandleEvent (impl_device->fd, &drm_event_context) != 0)
{
struct pollfd pfd;
int ret;
if (errno != EAGAIN)
{
g_set_error_literal (error, G_IO_ERROR,
g_io_error_from_errno (errno),
strerror (errno));
return FALSE;
}
pfd.fd = impl_device->fd;
pfd.events = POLL_IN | POLL_ERR;
do
{
ret = poll (&pfd, 1, -1);
}
while (ret == -1 && errno == EINTR);
}
else
{
break;
}
}
return TRUE;
}
static gpointer
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
kms_event_dispatch_in_impl (MetaKmsImpl *impl,
gpointer user_data,
GError **error)
{
MetaKmsImplDevice *impl_device = user_data;
gboolean ret;
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
ret = meta_kms_impl_device_dispatch (impl_device, error);
return GINT_TO_POINTER (ret);
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
}
drmModePropertyPtr
meta_kms_impl_device_find_property (MetaKmsImplDevice *impl_device,
drmModeObjectProperties *props,
const char *prop_name,
int *out_idx)
{
unsigned int i;
meta_assert_in_kms_impl (meta_kms_impl_get_kms (impl_device->impl));
for (i = 0; i < props->count_props; i++)
{
drmModePropertyPtr prop;
prop = drmModeGetProperty (impl_device->fd, props->props[i]);
if (!prop)
continue;
if (strcmp (prop->name, prop_name) == 0)
{
*out_idx = i;
return prop;
}
drmModeFreeProperty (prop);
}
return NULL;
}
static void
init_caps (MetaKmsImplDevice *impl_device)
{
int fd = impl_device->fd;
uint64_t cursor_width, cursor_height;
if (drmGetCap (fd, DRM_CAP_CURSOR_WIDTH, &cursor_width) == 0 &&
drmGetCap (fd, DRM_CAP_CURSOR_HEIGHT, &cursor_height) == 0)
{
impl_device->caps.has_cursor_size = TRUE;
impl_device->caps.cursor_width = cursor_width;
impl_device->caps.cursor_height = cursor_height;
}
}
static void
init_crtcs (MetaKmsImplDevice *impl_device,
drmModeRes *drm_resources)
{
int idx;
for (idx = 0; idx < drm_resources->count_crtcs; idx++)
{
drmModeCrtc *drm_crtc;
MetaKmsCrtc *crtc;
drm_crtc = drmModeGetCrtc (impl_device->fd, drm_resources->crtcs[idx]);
crtc = meta_kms_crtc_new (impl_device, drm_crtc, idx);
drmModeFreeCrtc (drm_crtc);
impl_device->crtcs = g_list_prepend (impl_device->crtcs, crtc);
}
impl_device->crtcs = g_list_reverse (impl_device->crtcs);
}
static MetaKmsConnector *
find_existing_connector (MetaKmsImplDevice *impl_device,
drmModeConnector *drm_connector)
{
GList *l;
for (l = impl_device->connectors; l; l = l->next)
{
MetaKmsConnector *connector = l->data;
if (meta_kms_connector_is_same_as (connector, drm_connector))
return connector;
}
return NULL;
}
static void
update_connectors (MetaKmsImplDevice *impl_device,
drmModeRes *drm_resources)
{
GList *connectors = NULL;
unsigned int i;
for (i = 0; i < drm_resources->count_connectors; i++)
{
drmModeConnector *drm_connector;
MetaKmsConnector *connector;
drm_connector = drmModeGetConnector (impl_device->fd,
drm_resources->connectors[i]);
if (!drm_connector)
continue;
connector = find_existing_connector (impl_device, drm_connector);
if (connector)
connector = g_object_ref (connector);
else
connector = meta_kms_connector_new (impl_device, drm_connector,
drm_resources);
drmModeFreeConnector (drm_connector);
connectors = g_list_prepend (connectors, connector);
}
g_list_free_full (impl_device->connectors, g_object_unref);
impl_device->connectors = g_list_reverse (connectors);
}
static MetaKmsPlaneType
get_plane_type (MetaKmsImplDevice *impl_device,
drmModeObjectProperties *props)
{
drmModePropertyPtr prop;
int idx;
prop = meta_kms_impl_device_find_property (impl_device, props, "type", &idx);
if (!prop)
return FALSE;
drmModeFreeProperty (prop);
switch (props->prop_values[idx])
{
case DRM_PLANE_TYPE_PRIMARY:
return META_KMS_PLANE_TYPE_PRIMARY;
case DRM_PLANE_TYPE_CURSOR:
return META_KMS_PLANE_TYPE_CURSOR;
case DRM_PLANE_TYPE_OVERLAY:
return META_KMS_PLANE_TYPE_OVERLAY;
default:
g_warning ("Unhandled plane type %" G_GUINT64_FORMAT,
props->prop_values[idx]);
return -1;
}
}
MetaKmsPlane *
meta_kms_impl_device_add_fake_plane (MetaKmsImplDevice *impl_device,
MetaKmsPlaneType plane_type,
MetaKmsCrtc *crtc)
{
MetaKmsPlane *plane;
plane = meta_kms_plane_new_fake (plane_type, crtc);
impl_device->planes = g_list_append (impl_device->planes, plane);
return plane;
}
static void
init_planes (MetaKmsImplDevice *impl_device)
{
int fd = impl_device->fd;
drmModePlaneRes *drm_planes;
unsigned int i;
drm_planes = drmModeGetPlaneResources (fd);
if (!drm_planes)
return;
for (i = 0; i < drm_planes->count_planes; i++)
{
drmModePlane *drm_plane;
drmModeObjectProperties *props;
drm_plane = drmModeGetPlane (fd, drm_planes->planes[i]);
if (!drm_plane)
continue;
props = drmModeObjectGetProperties (fd,
drm_plane->plane_id,
DRM_MODE_OBJECT_PLANE);
if (props)
{
MetaKmsPlaneType plane_type;
plane_type = get_plane_type (impl_device, props);
if (plane_type != -1)
{
MetaKmsPlane *plane;
plane = meta_kms_plane_new (plane_type,
impl_device,
drm_plane, props);
impl_device->planes = g_list_prepend (impl_device->planes, plane);
}
}
g_clear_pointer (&props, drmModeFreeObjectProperties);
drmModeFreePlane (drm_plane);
}
impl_device->planes = g_list_reverse (impl_device->planes);
}
void
meta_kms_impl_device_update_states (MetaKmsImplDevice *impl_device)
{
drmModeRes *drm_resources;
meta_assert_in_kms_impl (meta_kms_impl_get_kms (impl_device->impl));
drm_resources = drmModeGetResources (impl_device->fd);
if (!drm_resources)
{
g_list_free_full (impl_device->planes, g_object_unref);
g_list_free_full (impl_device->crtcs, g_object_unref);
g_list_free_full (impl_device->connectors, g_object_unref);
impl_device->planes = NULL;
impl_device->crtcs = NULL;
impl_device->connectors = NULL;
return;
}
update_connectors (impl_device, drm_resources);
g_list_foreach (impl_device->crtcs, (GFunc) meta_kms_crtc_update_state,
NULL);
g_list_foreach (impl_device->connectors, (GFunc) meta_kms_connector_update_state,
drm_resources);
drmModeFreeResources (drm_resources);
}
void
meta_kms_impl_device_predict_states (MetaKmsImplDevice *impl_device,
MetaKmsUpdate *update)
{
g_list_foreach (impl_device->crtcs, (GFunc) meta_kms_crtc_predict_state,
update);
g_list_foreach (impl_device->connectors, (GFunc) meta_kms_connector_predict_state,
update);
}
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
MetaKmsImplDevice *
meta_kms_impl_device_new (MetaKmsDevice *device,
MetaKmsImpl *impl,
int fd,
GError **error)
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
{
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
MetaKms *kms = meta_kms_impl_get_kms (impl);
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
MetaKmsImplDevice *impl_device;
int ret;
drmModeRes *drm_resources;
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
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
meta_assert_in_kms_impl (kms);
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
ret = drmSetClientCap (fd, DRM_CLIENT_CAP_UNIVERSAL_PLANES, 1);
if (ret != 0)
{
g_set_error (error, G_IO_ERROR, g_io_error_from_errno (-ret),
"Failed to activate universal planes: %s",
g_strerror (-ret));
return NULL;
}
drm_resources = drmModeGetResources (fd);
if (!drm_resources)
{
g_set_error (error, G_IO_ERROR, g_io_error_from_errno (errno),
"Failed to activate universal planes: %s",
g_strerror (errno));
return NULL;
}
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
impl_device = g_object_new (META_TYPE_KMS_IMPL_DEVICE, NULL);
impl_device->device = device;
impl_device->impl = impl;
impl_device->fd = fd;
init_caps (impl_device);
init_crtcs (impl_device, drm_resources);
init_planes (impl_device);
update_connectors (impl_device, drm_resources);
drmModeFreeResources (drm_resources);
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
impl_device->fd_source =
meta_kms_register_fd_in_impl (kms, fd,
kms_event_dispatch_in_impl,
impl_device);
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
return impl_device;
}
int
meta_kms_impl_device_get_fd (MetaKmsImplDevice *impl_device)
{
meta_assert_in_kms_impl (meta_kms_impl_get_kms (impl_device->impl));
return impl_device->fd;
}
int
meta_kms_impl_device_leak_fd (MetaKmsImplDevice *impl_device)
{
return impl_device->fd;
}
int
meta_kms_impl_device_close (MetaKmsImplDevice *impl_device)
{
int fd;
meta_assert_in_kms_impl (meta_kms_impl_get_kms (impl_device->impl));
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_clear_pointer (&impl_device->fd_source, g_source_destroy);
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
fd = impl_device->fd;
impl_device->fd = -1;
return fd;
}
static void
meta_kms_impl_device_finalize (GObject *object)
{
MetaKmsImplDevice *impl_device = META_KMS_IMPL_DEVICE (object);
g_list_free_full (impl_device->planes, g_object_unref);
g_list_free_full (impl_device->crtcs, g_object_unref);
g_list_free_full (impl_device->connectors, g_object_unref);
G_OBJECT_CLASS (meta_kms_impl_device_parent_class)->finalize (object);
}
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
static void
meta_kms_impl_device_init (MetaKmsImplDevice *device)
{
}
static void
meta_kms_impl_device_class_init (MetaKmsImplDeviceClass *klass)
{
GObjectClass *object_class = G_OBJECT_CLASS (klass);
object_class->finalize = meta_kms_impl_device_finalize;
backends/native: Add basic KMS abstraction building blocks The intention with KMS abstraction is to hide away accessing the drm functions behind an API that allows us to have different kind of KMS implementations, including legacy non-atomic and atomic. The intention is also that the code interacting with the drm device should be able to be run in a different thread than the main thread. This means that we need to make sure that all drm*() API usage must only occur from within tasks that eventually can be run in the dedicated thread. The idea here is that MetaKms provides a outward facing API other places of mutter can use (e.g. MetaGpuKms and friends), while MetaKmsImpl is an internal implementation that only gets interacted with via "tasks" posted via the MetaKms object. These tasks will in the future potentially be run on the dedicated KMS thread. Initially, we don't create any new threads. Likewise, MetaKmsDevice is a outward facing representation of a KMS device, while MetaKmsImplDevice is the corresponding implementation, which only runs from within the MetaKmsImpl tasks. This commit only moves opening and closing the device to this new API, while leaking the fd outside of the impl enclosure, effectively making the isolation for drm*() calls pointless. This, however, is necessary to allow gradual porting of drm interaction, and eventually the file descriptor in MetaGpuKms will be removed. For now, it's harmless, since everything still run in the main thread. https://gitlab.gnome.org/GNOME/mutter/issues/548 https://gitlab.gnome.org/GNOME/mutter/merge_requests/525
2019-01-29 09:24:44 +00:00
}