mesa/src/vulkan/wsi/wsi_common.c

1707 lines
60 KiB
C
Raw Normal View History

/*
* Copyright © 2017 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice (including the next
* paragraph) shall be included in all copies or substantial portions of the
* Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#include "wsi_common_private.h"
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
#include "wsi_common_entrypoints.h"
#include "util/debug.h"
#include "util/macros.h"
#include "util/os_file.h"
#include "util/xmlconfig.h"
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
#include "vk_device.h"
#include "vk_fence.h"
#include "vk_format.h"
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
#include "vk_instance.h"
#include "vk_physical_device.h"
#include "vk_queue.h"
#include "vk_semaphore.h"
#include "vk_sync.h"
#include "vk_sync_dummy.h"
#include "vk_util.h"
#include <time.h>
#include <stdlib.h>
#include <stdio.h>
#ifndef _WIN32
#include <unistd.h>
#endif
uint64_t WSI_DEBUG;
static const struct debug_control debug_control[] = {
{ "buffer", WSI_DEBUG_BUFFER },
{ "sw", WSI_DEBUG_SW },
{ "noshm", WSI_DEBUG_NOSHM },
{ "linear", WSI_DEBUG_LINEAR },
{ NULL, },
};
VkResult
wsi_device_init(struct wsi_device *wsi,
VkPhysicalDevice pdevice,
WSI_FN_GetPhysicalDeviceProcAddr proc_addr,
vulkan: Add KHR_display extension using DRM [v10] This adds support for the KHR_display extension support to the vulkan WSI layer. Driver support will be added separately. v2: * fix double ;; in wsi_common_display.c * Move mode list from wsi_display to wsi_display_connector * Fix scope for wsi_display_mode andwsi_display_connector allocs * Switch all allocations to vk_zalloc instead of vk_alloc. * Fix DRM failure in wsi_display_get_physical_device_display_properties When DRM fails, or when we don't have a master fd (presumably due to application errors), just return 0 properties from this function, which is at least a valid response. * Use vk_outarray for all property queries This is a bit less error-prone than open-coding the same stuff. * Remove VK_COMPOSITE_ALPHA_INHERIT_BIT_KHR from surface caps Until we have multi-plane support, we shouldn't pretend to have any multi-plane semantics, even if undefined. Suggested-by: Jason Ekstrand <jason@jlekstrand.net> * Simplify addition of VK_USE_PLATFORM_DISPLAY_KHR to vulkan_wsi_args Suggested-by: Eric Engestrom <eric.engestrom@imgtec.com> v3: Add separate 'display_fd' and 'render_fd' arguments to wsi_device_init API. This allows drivers to use different FDs for the different aspects of the device. Use largest mode as display size when no preferred mode. If the display doesn't provide a preferred mode, we'll assume that the largest supported mode is the "physical size" of the device and report that. v4: Make wsi_image_state enumeration values uppercase. Follow more common mesa conventions. Remove 'render_fd' from wsi_device_init API. The wsi_common_display code doesn't use this fd at all, so stop passing it in. This avoids any potential confusion over which fd to use when creating display-relative object handles. Remove call to wsi_create_prime_image which would never have been reached as the necessary condition (use_prime_blit) is never set. whitespace cleanups in wsi_common_display.c Suggested-by: Jason Ekstrand <jason@jlekstrand.net> Add depth/bpp info to available surface formats. Instead of hard-coding depth 24 bpp 32 in the drmModeAddFB call, use the requested format to find suitable values. Destroy kernel buffers and FBs when swapchain is destroyed. We were leaking both of these kernel objects across swapchain destruction. Note that wsi_display_wait_for_event waits for anything to happen. wsi_display_wait_for_event is simply a yield so that the caller can then check to see if the desired state change has occurred. Record swapchain failures in chain for later return. If some asynchronous swapchain activity fails, we need to tell the application eventually. Record the failure in the swapchain and report it at the next acquire_next_image or queue_present call. Fix error returns from wsi_display_setup_connector. If a malloc failed, then the result should be VK_ERROR_OUT_OF_HOST_MEMORY. Otherwise, the associated ioctl failed and we're either VT switched away, or our lease has been revoked, in which case we should return VK_ERROR_OUT_OF_DATE_KHR. Make sure both sides of if/else brace use matches Note that we assume drmModeSetCrtc is synchronous. Add a comment explaining why we can idle any previous displayed image as soon as the mode set returns. Note that EACCES from drmModePageFlip means VT inactive. When vt switched away drmModePageFlip returns EACCES. Poll once a second waiting until we get some other return value back. Clean up after alloc failure in wsi_display_surface_create_swapchain. Destroy any created images, free the swapchain. Remove physical_device from wsi_display_init_wsi. We never need this value, so remove it from the API and from the internal wsi_display structure. Use drmModeAddFB2 in wsi_display_image_init. This takes a drm format instead of depth/bpp, which provides more control over the format of the data. v5: Set the 'currentStackIndex' member of the VkDisplayPlanePropertiesKHR record to zero, instead of indexing across all displays. This value is the stack depth of the plane within an individual display, and as the current code supports only a single plane per display, should be set to zero for all elements Discovered-by: David Mao <David.Mao@amd.com> v6: Remove 'platform_display' bits from the build and use the existing 'platform_drm' instead. v7: Ensure VK_ICD_WSI_PLATFORM_MAX is large enough by setting to VK_ICD_WSI_PLATFORM_DISPLAY + 1 v8: Simplify wsi_device_init failure from wsi_display_init_wsi by using the same pattern as the other wsi layers. Adopt Jason Ekstrand's white space and variable declaration suggestions. Declare variables at first use, eliminate extra whitespace between types and names, add list iterator helpers, switch to lower-case list_ macros. Respond to Jason's April 8 review: * Create a function to convert relative to absolute timeouts to catch overflow issues in one place * use VK_NULL_HANDLE to clear prop->currentDisplay * Get rid of available_present_modes array. * return OUT_OF_DATE_KHR when display_queue_next called after display has been released. * Make errors from mode setting fatal in display_queue_next * Remove duplicate pthread_mutex_init call * Add wsi_init_pthread_cond_monotonic helper function to isolate pthread error handling from wsi_display_init_wsi Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> v9: Fix vscan handling by using MAX2(vscan, 1) everywhere. Vscan can be zero anywhere, which is treated the same as 1. Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> v10: Respond to Vulkan CTS failures. 1. Initialize planeReorderPossible in display_properties code 2. Only report connected displays in get_display_plane_supported_displays 3. Return VK_ERROR_OUT_OF_HOST_MEMORY when pthread cond initialization fails. Signed-off-by: Jason Ekstrand <jason.ekstrand@intel.com> 4. Add vkCreateDisplayModeKHR. This doesn't actually create new modes, it only looks to see if the requested parameters matches an existing mode and returns that. Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Signed-off-by: Keith Packard <keithp@keithp.com>
2018-02-07 18:31:44 +00:00
const VkAllocationCallbacks *alloc,
int display_fd,
const struct driOptionCache *dri_options,
bool sw_device)
{
const char *present_mode;
UNUSED VkResult result;
WSI_DEBUG = parse_debug_string(getenv("MESA_VK_WSI_DEBUG"), debug_control);
memset(wsi, 0, sizeof(*wsi));
wsi->instance_alloc = *alloc;
wsi->pdevice = pdevice;
wsi->sw = sw_device || (WSI_DEBUG & WSI_DEBUG_SW);
wsi->wants_linear = (WSI_DEBUG & WSI_DEBUG_LINEAR) != 0;
#define WSI_GET_CB(func) \
PFN_vk##func func = (PFN_vk##func)proc_addr(pdevice, "vk" #func)
WSI_GET_CB(GetPhysicalDeviceExternalSemaphoreProperties);
WSI_GET_CB(GetPhysicalDeviceProperties2);
WSI_GET_CB(GetPhysicalDeviceMemoryProperties);
WSI_GET_CB(GetPhysicalDeviceQueueFamilyProperties);
#undef WSI_GET_CB
wsi->pci_bus_info.sType =
VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PCI_BUS_INFO_PROPERTIES_EXT;
VkPhysicalDeviceProperties2 pdp2 = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2,
.pNext = &wsi->pci_bus_info,
};
GetPhysicalDeviceProperties2(pdevice, &pdp2);
wsi->maxImageDimension2D = pdp2.properties.limits.maxImageDimension2D;
assert(pdp2.properties.limits.optimalBufferCopyRowPitchAlignment <= UINT32_MAX);
wsi->optimalBufferCopyRowPitchAlignment =
pdp2.properties.limits.optimalBufferCopyRowPitchAlignment;
wsi->override_present_mode = VK_PRESENT_MODE_MAX_ENUM_KHR;
GetPhysicalDeviceMemoryProperties(pdevice, &wsi->memory_props);
GetPhysicalDeviceQueueFamilyProperties(pdevice, &wsi->queue_family_count, NULL);
for (VkExternalSemaphoreHandleTypeFlags handle_type = 1;
handle_type <= VK_EXTERNAL_SEMAPHORE_HANDLE_TYPE_SYNC_FD_BIT;
handle_type <<= 1) {
const VkPhysicalDeviceExternalSemaphoreInfo esi = {
.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_EXTERNAL_SEMAPHORE_INFO,
.handleType = handle_type,
};
VkExternalSemaphoreProperties esp = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_SEMAPHORE_PROPERTIES,
};
GetPhysicalDeviceExternalSemaphoreProperties(pdevice, &esi, &esp);
if (esp.externalSemaphoreFeatures &
VK_EXTERNAL_SEMAPHORE_FEATURE_EXPORTABLE_BIT)
wsi->semaphore_export_handle_types |= handle_type;
}
const struct vk_device_extension_table *supported_extensions =
&vk_physical_device_from_handle(pdevice)->supported_extensions;
wsi->has_import_memory_host =
supported_extensions->EXT_external_memory_host;
list_inithead(&wsi->hotplug_fences);
#define WSI_GET_CB(func) \
wsi->func = (PFN_vk##func)proc_addr(pdevice, "vk" #func)
WSI_GET_CB(AllocateMemory);
WSI_GET_CB(AllocateCommandBuffers);
WSI_GET_CB(BindBufferMemory);
WSI_GET_CB(BindImageMemory);
WSI_GET_CB(BeginCommandBuffer);
WSI_GET_CB(CmdPipelineBarrier);
WSI_GET_CB(CmdCopyImageToBuffer);
WSI_GET_CB(CreateBuffer);
WSI_GET_CB(CreateCommandPool);
WSI_GET_CB(CreateFence);
WSI_GET_CB(CreateImage);
WSI_GET_CB(CreateSemaphore);
WSI_GET_CB(DestroyBuffer);
WSI_GET_CB(DestroyCommandPool);
WSI_GET_CB(DestroyFence);
WSI_GET_CB(DestroyImage);
WSI_GET_CB(DestroySemaphore);
WSI_GET_CB(EndCommandBuffer);
WSI_GET_CB(FreeMemory);
WSI_GET_CB(FreeCommandBuffers);
WSI_GET_CB(GetBufferMemoryRequirements);
WSI_GET_CB(GetImageDrmFormatModifierPropertiesEXT);
WSI_GET_CB(GetImageMemoryRequirements);
WSI_GET_CB(GetImageSubresourceLayout);
if (!wsi->sw)
WSI_GET_CB(GetMemoryFdKHR);
WSI_GET_CB(GetPhysicalDeviceFormatProperties);
WSI_GET_CB(GetPhysicalDeviceFormatProperties2KHR);
WSI_GET_CB(GetPhysicalDeviceImageFormatProperties2);
WSI_GET_CB(GetSemaphoreFdKHR);
WSI_GET_CB(ResetFences);
WSI_GET_CB(QueueSubmit);
WSI_GET_CB(WaitForFences);
WSI_GET_CB(MapMemory);
WSI_GET_CB(UnmapMemory);
#undef WSI_GET_CB
#ifdef VK_USE_PLATFORM_XCB_KHR
result = wsi_x11_init_wsi(wsi, alloc, dri_options);
if (result != VK_SUCCESS)
goto fail;
#endif
#ifdef VK_USE_PLATFORM_WAYLAND_KHR
result = wsi_wl_init_wsi(wsi, alloc, pdevice);
if (result != VK_SUCCESS)
goto fail;
#endif
#ifdef VK_USE_PLATFORM_WIN32_KHR
result = wsi_win32_init_wsi(wsi, alloc, pdevice);
if (result != VK_SUCCESS)
goto fail;
#endif
vulkan: Add KHR_display extension using DRM [v10] This adds support for the KHR_display extension support to the vulkan WSI layer. Driver support will be added separately. v2: * fix double ;; in wsi_common_display.c * Move mode list from wsi_display to wsi_display_connector * Fix scope for wsi_display_mode andwsi_display_connector allocs * Switch all allocations to vk_zalloc instead of vk_alloc. * Fix DRM failure in wsi_display_get_physical_device_display_properties When DRM fails, or when we don't have a master fd (presumably due to application errors), just return 0 properties from this function, which is at least a valid response. * Use vk_outarray for all property queries This is a bit less error-prone than open-coding the same stuff. * Remove VK_COMPOSITE_ALPHA_INHERIT_BIT_KHR from surface caps Until we have multi-plane support, we shouldn't pretend to have any multi-plane semantics, even if undefined. Suggested-by: Jason Ekstrand <jason@jlekstrand.net> * Simplify addition of VK_USE_PLATFORM_DISPLAY_KHR to vulkan_wsi_args Suggested-by: Eric Engestrom <eric.engestrom@imgtec.com> v3: Add separate 'display_fd' and 'render_fd' arguments to wsi_device_init API. This allows drivers to use different FDs for the different aspects of the device. Use largest mode as display size when no preferred mode. If the display doesn't provide a preferred mode, we'll assume that the largest supported mode is the "physical size" of the device and report that. v4: Make wsi_image_state enumeration values uppercase. Follow more common mesa conventions. Remove 'render_fd' from wsi_device_init API. The wsi_common_display code doesn't use this fd at all, so stop passing it in. This avoids any potential confusion over which fd to use when creating display-relative object handles. Remove call to wsi_create_prime_image which would never have been reached as the necessary condition (use_prime_blit) is never set. whitespace cleanups in wsi_common_display.c Suggested-by: Jason Ekstrand <jason@jlekstrand.net> Add depth/bpp info to available surface formats. Instead of hard-coding depth 24 bpp 32 in the drmModeAddFB call, use the requested format to find suitable values. Destroy kernel buffers and FBs when swapchain is destroyed. We were leaking both of these kernel objects across swapchain destruction. Note that wsi_display_wait_for_event waits for anything to happen. wsi_display_wait_for_event is simply a yield so that the caller can then check to see if the desired state change has occurred. Record swapchain failures in chain for later return. If some asynchronous swapchain activity fails, we need to tell the application eventually. Record the failure in the swapchain and report it at the next acquire_next_image or queue_present call. Fix error returns from wsi_display_setup_connector. If a malloc failed, then the result should be VK_ERROR_OUT_OF_HOST_MEMORY. Otherwise, the associated ioctl failed and we're either VT switched away, or our lease has been revoked, in which case we should return VK_ERROR_OUT_OF_DATE_KHR. Make sure both sides of if/else brace use matches Note that we assume drmModeSetCrtc is synchronous. Add a comment explaining why we can idle any previous displayed image as soon as the mode set returns. Note that EACCES from drmModePageFlip means VT inactive. When vt switched away drmModePageFlip returns EACCES. Poll once a second waiting until we get some other return value back. Clean up after alloc failure in wsi_display_surface_create_swapchain. Destroy any created images, free the swapchain. Remove physical_device from wsi_display_init_wsi. We never need this value, so remove it from the API and from the internal wsi_display structure. Use drmModeAddFB2 in wsi_display_image_init. This takes a drm format instead of depth/bpp, which provides more control over the format of the data. v5: Set the 'currentStackIndex' member of the VkDisplayPlanePropertiesKHR record to zero, instead of indexing across all displays. This value is the stack depth of the plane within an individual display, and as the current code supports only a single plane per display, should be set to zero for all elements Discovered-by: David Mao <David.Mao@amd.com> v6: Remove 'platform_display' bits from the build and use the existing 'platform_drm' instead. v7: Ensure VK_ICD_WSI_PLATFORM_MAX is large enough by setting to VK_ICD_WSI_PLATFORM_DISPLAY + 1 v8: Simplify wsi_device_init failure from wsi_display_init_wsi by using the same pattern as the other wsi layers. Adopt Jason Ekstrand's white space and variable declaration suggestions. Declare variables at first use, eliminate extra whitespace between types and names, add list iterator helpers, switch to lower-case list_ macros. Respond to Jason's April 8 review: * Create a function to convert relative to absolute timeouts to catch overflow issues in one place * use VK_NULL_HANDLE to clear prop->currentDisplay * Get rid of available_present_modes array. * return OUT_OF_DATE_KHR when display_queue_next called after display has been released. * Make errors from mode setting fatal in display_queue_next * Remove duplicate pthread_mutex_init call * Add wsi_init_pthread_cond_monotonic helper function to isolate pthread error handling from wsi_display_init_wsi Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> v9: Fix vscan handling by using MAX2(vscan, 1) everywhere. Vscan can be zero anywhere, which is treated the same as 1. Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> v10: Respond to Vulkan CTS failures. 1. Initialize planeReorderPossible in display_properties code 2. Only report connected displays in get_display_plane_supported_displays 3. Return VK_ERROR_OUT_OF_HOST_MEMORY when pthread cond initialization fails. Signed-off-by: Jason Ekstrand <jason.ekstrand@intel.com> 4. Add vkCreateDisplayModeKHR. This doesn't actually create new modes, it only looks to see if the requested parameters matches an existing mode and returns that. Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Signed-off-by: Keith Packard <keithp@keithp.com>
2018-02-07 18:31:44 +00:00
#ifdef VK_USE_PLATFORM_DISPLAY_KHR
result = wsi_display_init_wsi(wsi, alloc, display_fd);
if (result != VK_SUCCESS)
goto fail;
#endif
present_mode = getenv("MESA_VK_WSI_PRESENT_MODE");
if (present_mode) {
if (!strcmp(present_mode, "fifo")) {
wsi->override_present_mode = VK_PRESENT_MODE_FIFO_KHR;
} else if (!strcmp(present_mode, "relaxed")) {
wsi->override_present_mode = VK_PRESENT_MODE_FIFO_RELAXED_KHR;
} else if (!strcmp(present_mode, "mailbox")) {
wsi->override_present_mode = VK_PRESENT_MODE_MAILBOX_KHR;
} else if (!strcmp(present_mode, "immediate")) {
wsi->override_present_mode = VK_PRESENT_MODE_IMMEDIATE_KHR;
} else {
fprintf(stderr, "Invalid MESA_VK_WSI_PRESENT_MODE value!\n");
}
}
if (dri_options) {
if (driCheckOption(dri_options, "adaptive_sync", DRI_BOOL))
wsi->enable_adaptive_sync = driQueryOptionb(dri_options,
"adaptive_sync");
if (driCheckOption(dri_options, "vk_wsi_force_bgra8_unorm_first", DRI_BOOL)) {
wsi->force_bgra8_unorm_first =
driQueryOptionb(dri_options, "vk_wsi_force_bgra8_unorm_first");
}
}
return VK_SUCCESS;
#if defined(VK_USE_PLATFORM_XCB_KHR) || \
defined(VK_USE_PLATFORM_WAYLAND_KHR) || \
defined(VK_USE_PLATFORM_WIN32_KHR) || \
defined(VK_USE_PLATFORM_DISPLAY_KHR)
fail:
wsi_device_finish(wsi, alloc);
return result;
#endif
}
void
wsi_device_finish(struct wsi_device *wsi,
const VkAllocationCallbacks *alloc)
{
vulkan: Add KHR_display extension using DRM [v10] This adds support for the KHR_display extension support to the vulkan WSI layer. Driver support will be added separately. v2: * fix double ;; in wsi_common_display.c * Move mode list from wsi_display to wsi_display_connector * Fix scope for wsi_display_mode andwsi_display_connector allocs * Switch all allocations to vk_zalloc instead of vk_alloc. * Fix DRM failure in wsi_display_get_physical_device_display_properties When DRM fails, or when we don't have a master fd (presumably due to application errors), just return 0 properties from this function, which is at least a valid response. * Use vk_outarray for all property queries This is a bit less error-prone than open-coding the same stuff. * Remove VK_COMPOSITE_ALPHA_INHERIT_BIT_KHR from surface caps Until we have multi-plane support, we shouldn't pretend to have any multi-plane semantics, even if undefined. Suggested-by: Jason Ekstrand <jason@jlekstrand.net> * Simplify addition of VK_USE_PLATFORM_DISPLAY_KHR to vulkan_wsi_args Suggested-by: Eric Engestrom <eric.engestrom@imgtec.com> v3: Add separate 'display_fd' and 'render_fd' arguments to wsi_device_init API. This allows drivers to use different FDs for the different aspects of the device. Use largest mode as display size when no preferred mode. If the display doesn't provide a preferred mode, we'll assume that the largest supported mode is the "physical size" of the device and report that. v4: Make wsi_image_state enumeration values uppercase. Follow more common mesa conventions. Remove 'render_fd' from wsi_device_init API. The wsi_common_display code doesn't use this fd at all, so stop passing it in. This avoids any potential confusion over which fd to use when creating display-relative object handles. Remove call to wsi_create_prime_image which would never have been reached as the necessary condition (use_prime_blit) is never set. whitespace cleanups in wsi_common_display.c Suggested-by: Jason Ekstrand <jason@jlekstrand.net> Add depth/bpp info to available surface formats. Instead of hard-coding depth 24 bpp 32 in the drmModeAddFB call, use the requested format to find suitable values. Destroy kernel buffers and FBs when swapchain is destroyed. We were leaking both of these kernel objects across swapchain destruction. Note that wsi_display_wait_for_event waits for anything to happen. wsi_display_wait_for_event is simply a yield so that the caller can then check to see if the desired state change has occurred. Record swapchain failures in chain for later return. If some asynchronous swapchain activity fails, we need to tell the application eventually. Record the failure in the swapchain and report it at the next acquire_next_image or queue_present call. Fix error returns from wsi_display_setup_connector. If a malloc failed, then the result should be VK_ERROR_OUT_OF_HOST_MEMORY. Otherwise, the associated ioctl failed and we're either VT switched away, or our lease has been revoked, in which case we should return VK_ERROR_OUT_OF_DATE_KHR. Make sure both sides of if/else brace use matches Note that we assume drmModeSetCrtc is synchronous. Add a comment explaining why we can idle any previous displayed image as soon as the mode set returns. Note that EACCES from drmModePageFlip means VT inactive. When vt switched away drmModePageFlip returns EACCES. Poll once a second waiting until we get some other return value back. Clean up after alloc failure in wsi_display_surface_create_swapchain. Destroy any created images, free the swapchain. Remove physical_device from wsi_display_init_wsi. We never need this value, so remove it from the API and from the internal wsi_display structure. Use drmModeAddFB2 in wsi_display_image_init. This takes a drm format instead of depth/bpp, which provides more control over the format of the data. v5: Set the 'currentStackIndex' member of the VkDisplayPlanePropertiesKHR record to zero, instead of indexing across all displays. This value is the stack depth of the plane within an individual display, and as the current code supports only a single plane per display, should be set to zero for all elements Discovered-by: David Mao <David.Mao@amd.com> v6: Remove 'platform_display' bits from the build and use the existing 'platform_drm' instead. v7: Ensure VK_ICD_WSI_PLATFORM_MAX is large enough by setting to VK_ICD_WSI_PLATFORM_DISPLAY + 1 v8: Simplify wsi_device_init failure from wsi_display_init_wsi by using the same pattern as the other wsi layers. Adopt Jason Ekstrand's white space and variable declaration suggestions. Declare variables at first use, eliminate extra whitespace between types and names, add list iterator helpers, switch to lower-case list_ macros. Respond to Jason's April 8 review: * Create a function to convert relative to absolute timeouts to catch overflow issues in one place * use VK_NULL_HANDLE to clear prop->currentDisplay * Get rid of available_present_modes array. * return OUT_OF_DATE_KHR when display_queue_next called after display has been released. * Make errors from mode setting fatal in display_queue_next * Remove duplicate pthread_mutex_init call * Add wsi_init_pthread_cond_monotonic helper function to isolate pthread error handling from wsi_display_init_wsi Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> v9: Fix vscan handling by using MAX2(vscan, 1) everywhere. Vscan can be zero anywhere, which is treated the same as 1. Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> v10: Respond to Vulkan CTS failures. 1. Initialize planeReorderPossible in display_properties code 2. Only report connected displays in get_display_plane_supported_displays 3. Return VK_ERROR_OUT_OF_HOST_MEMORY when pthread cond initialization fails. Signed-off-by: Jason Ekstrand <jason.ekstrand@intel.com> 4. Add vkCreateDisplayModeKHR. This doesn't actually create new modes, it only looks to see if the requested parameters matches an existing mode and returns that. Suggested-by: Jason Ekstrand <jason.ekstrand@intel.com> Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Signed-off-by: Keith Packard <keithp@keithp.com>
2018-02-07 18:31:44 +00:00
#ifdef VK_USE_PLATFORM_DISPLAY_KHR
wsi_display_finish_wsi(wsi, alloc);
#endif
#ifdef VK_USE_PLATFORM_WAYLAND_KHR
wsi_wl_finish_wsi(wsi, alloc);
#endif
#ifdef VK_USE_PLATFORM_WIN32_KHR
wsi_win32_finish_wsi(wsi, alloc);
#endif
#ifdef VK_USE_PLATFORM_XCB_KHR
wsi_x11_finish_wsi(wsi, alloc);
#endif
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR void VKAPI_CALL
wsi_DestroySurfaceKHR(VkInstance _instance,
VkSurfaceKHR _surface,
const VkAllocationCallbacks *pAllocator)
{
VK_FROM_HANDLE(vk_instance, instance, _instance);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
if (!surface)
return;
vk_free2(&instance->alloc, pAllocator, surface);
}
void
wsi_device_setup_syncobj_fd(struct wsi_device *wsi_device,
int fd)
{
#ifdef VK_USE_PLATFORM_DISPLAY_KHR
wsi_display_setup_syncobj_fd(wsi_device, fd);
#endif
}
VkResult
wsi_swapchain_init(const struct wsi_device *wsi,
struct wsi_swapchain *chain,
VkDevice _device,
const VkSwapchainCreateInfoKHR *pCreateInfo,
const VkAllocationCallbacks *pAllocator,
bool use_buffer_blit)
{
VK_FROM_HANDLE(vk_device, device, _device);
VkResult result;
memset(chain, 0, sizeof(*chain));
vk_object_base_init(device, &chain->base, VK_OBJECT_TYPE_SWAPCHAIN_KHR);
chain->wsi = wsi;
chain->device = _device;
chain->alloc = *pAllocator;
chain->use_buffer_blit = use_buffer_blit || (WSI_DEBUG & WSI_DEBUG_BUFFER);
if (wsi->sw && !wsi->wants_linear)
chain->use_buffer_blit = true;
chain->buffer_blit_queue = VK_NULL_HANDLE;
if (use_buffer_blit && wsi->get_buffer_blit_queue)
chain->buffer_blit_queue = wsi->get_buffer_blit_queue(_device);
int cmd_pools_count = chain->buffer_blit_queue != VK_NULL_HANDLE ? 1 : wsi->queue_family_count;
chain->cmd_pools =
vk_zalloc(pAllocator, sizeof(VkCommandPool) * cmd_pools_count, 8,
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (!chain->cmd_pools)
return VK_ERROR_OUT_OF_HOST_MEMORY;
for (uint32_t i = 0; i < cmd_pools_count; i++) {
int queue_family_index = i;
if (chain->buffer_blit_queue != VK_NULL_HANDLE) {
VK_FROM_HANDLE(vk_queue, queue, chain->buffer_blit_queue);
queue_family_index = queue->queue_family_index;
}
const VkCommandPoolCreateInfo cmd_pool_info = {
.sType = VK_STRUCTURE_TYPE_COMMAND_POOL_CREATE_INFO,
.pNext = NULL,
.flags = 0,
.queueFamilyIndex = queue_family_index,
};
result = wsi->CreateCommandPool(_device, &cmd_pool_info, &chain->alloc,
&chain->cmd_pools[i]);
if (result != VK_SUCCESS)
goto fail;
}
return VK_SUCCESS;
fail:
wsi_swapchain_finish(chain);
return result;
}
static bool
wsi_swapchain_is_present_mode_supported(struct wsi_device *wsi,
const VkSwapchainCreateInfoKHR *pCreateInfo,
VkPresentModeKHR mode)
{
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, pCreateInfo->surface);
struct wsi_interface *iface = wsi->wsi[surface->platform];
VkPresentModeKHR *present_modes;
uint32_t present_mode_count;
bool supported = false;
VkResult result;
result = iface->get_present_modes(surface, &present_mode_count, NULL);
if (result != VK_SUCCESS)
return supported;
present_modes = malloc(present_mode_count * sizeof(*present_modes));
if (!present_modes)
return supported;
result = iface->get_present_modes(surface, &present_mode_count,
present_modes);
if (result != VK_SUCCESS)
goto fail;
for (uint32_t i = 0; i < present_mode_count; i++) {
if (present_modes[i] == mode) {
supported = true;
break;
}
}
fail:
free(present_modes);
return supported;
}
enum VkPresentModeKHR
wsi_swapchain_get_present_mode(struct wsi_device *wsi,
const VkSwapchainCreateInfoKHR *pCreateInfo)
{
if (wsi->override_present_mode == VK_PRESENT_MODE_MAX_ENUM_KHR)
return pCreateInfo->presentMode;
if (!wsi_swapchain_is_present_mode_supported(wsi, pCreateInfo,
wsi->override_present_mode)) {
fprintf(stderr, "Unsupported MESA_VK_WSI_PRESENT_MODE value!\n");
return pCreateInfo->presentMode;
}
return wsi->override_present_mode;
}
void
wsi_swapchain_finish(struct wsi_swapchain *chain)
{
if (chain->fences) {
for (unsigned i = 0; i < chain->image_count; i++)
chain->wsi->DestroyFence(chain->device, chain->fences[i], &chain->alloc);
vk_free(&chain->alloc, chain->fences);
}
if (chain->buffer_blit_semaphores) {
for (unsigned i = 0; i < chain->image_count; i++)
chain->wsi->DestroySemaphore(chain->device, chain->buffer_blit_semaphores[i], &chain->alloc);
vk_free(&chain->alloc, chain->buffer_blit_semaphores);
}
chain->wsi->DestroySemaphore(chain->device, chain->dma_buf_semaphore,
&chain->alloc);
int cmd_pools_count = chain->buffer_blit_queue != VK_NULL_HANDLE ?
1 : chain->wsi->queue_family_count;
for (uint32_t i = 0; i < cmd_pools_count; i++) {
chain->wsi->DestroyCommandPool(chain->device, chain->cmd_pools[i],
&chain->alloc);
}
vk_free(&chain->alloc, chain->cmd_pools);
vk_object_base_finish(&chain->base);
}
VkResult
wsi_configure_image(const struct wsi_swapchain *chain,
const VkSwapchainCreateInfoKHR *pCreateInfo,
VkExternalMemoryHandleTypeFlags handle_types,
struct wsi_image_info *info)
{
memset(info, 0, sizeof(*info));
uint32_t queue_family_count = 1;
if (pCreateInfo->imageSharingMode == VK_SHARING_MODE_CONCURRENT)
queue_family_count = pCreateInfo->queueFamilyIndexCount;
/*
* TODO: there should be no reason to allocate this, but
* 15331 shows that games crashed without doing this.
*/
uint32_t *queue_family_indices =
vk_alloc(&chain->alloc,
sizeof(*queue_family_indices) *
queue_family_count,
8, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (!queue_family_indices)
goto err_oom;
if (pCreateInfo->imageSharingMode == VK_SHARING_MODE_CONCURRENT)
for (uint32_t i = 0; i < pCreateInfo->queueFamilyIndexCount; i++)
queue_family_indices[i] = pCreateInfo->pQueueFamilyIndices[i];
info->create = (VkImageCreateInfo) {
.sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO,
.flags = VK_IMAGE_CREATE_ALIAS_BIT,
.imageType = VK_IMAGE_TYPE_2D,
.format = pCreateInfo->imageFormat,
.extent = {
.width = pCreateInfo->imageExtent.width,
.height = pCreateInfo->imageExtent.height,
.depth = 1,
},
.mipLevels = 1,
.arrayLayers = 1,
.samples = VK_SAMPLE_COUNT_1_BIT,
.tiling = VK_IMAGE_TILING_OPTIMAL,
.usage = pCreateInfo->imageUsage,
.sharingMode = pCreateInfo->imageSharingMode,
.queueFamilyIndexCount = queue_family_count,
.pQueueFamilyIndices = queue_family_indices,
.initialLayout = VK_IMAGE_LAYOUT_UNDEFINED,
};
if (handle_types != 0) {
info->ext_mem = (VkExternalMemoryImageCreateInfo) {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_MEMORY_IMAGE_CREATE_INFO,
.handleTypes = handle_types,
};
__vk_append_struct(&info->create, &info->ext_mem);
}
info->wsi = (struct wsi_image_create_info) {
.sType = VK_STRUCTURE_TYPE_WSI_IMAGE_CREATE_INFO_MESA,
};
__vk_append_struct(&info->create, &info->wsi);
if (pCreateInfo->flags & VK_SWAPCHAIN_CREATE_MUTABLE_FORMAT_BIT_KHR) {
info->create.flags |= VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT |
VK_IMAGE_CREATE_EXTENDED_USAGE_BIT;
const VkImageFormatListCreateInfo *format_list_in =
vk_find_struct_const(pCreateInfo->pNext,
IMAGE_FORMAT_LIST_CREATE_INFO);
assume(format_list_in && format_list_in->viewFormatCount > 0);
const uint32_t view_format_count = format_list_in->viewFormatCount;
VkFormat *view_formats =
vk_alloc(&chain->alloc, sizeof(VkFormat) * view_format_count,
8, VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (!view_formats)
goto err_oom;
ASSERTED bool format_found = false;
for (uint32_t i = 0; i < format_list_in->viewFormatCount; i++) {
if (pCreateInfo->imageFormat == format_list_in->pViewFormats[i])
format_found = true;
view_formats[i] = format_list_in->pViewFormats[i];
}
assert(format_found);
info->format_list = (VkImageFormatListCreateInfo) {
.sType = VK_STRUCTURE_TYPE_IMAGE_FORMAT_LIST_CREATE_INFO,
.viewFormatCount = view_format_count,
.pViewFormats = view_formats,
};
__vk_append_struct(&info->create, &info->format_list);
}
return VK_SUCCESS;
err_oom:
wsi_destroy_image_info(chain, info);
return VK_ERROR_OUT_OF_HOST_MEMORY;
}
void
wsi_destroy_image_info(const struct wsi_swapchain *chain,
struct wsi_image_info *info)
{
vk_free(&chain->alloc, (void *)info->create.pQueueFamilyIndices);
vk_free(&chain->alloc, (void *)info->format_list.pViewFormats);
vk_free(&chain->alloc, (void *)info->drm_mod_list.pDrmFormatModifiers);
vk_free(&chain->alloc, info->modifier_props);
}
VkResult
wsi_create_image(const struct wsi_swapchain *chain,
const struct wsi_image_info *info,
struct wsi_image *image)
{
const struct wsi_device *wsi = chain->wsi;
VkResult result;
memset(image, 0, sizeof(*image));
#ifndef _WIN32
image->dma_buf_fd = -1;
#endif
result = wsi->CreateImage(chain->device, &info->create,
&chain->alloc, &image->image);
if (result != VK_SUCCESS)
goto fail;
result = info->create_mem(chain, info, image);
if (result != VK_SUCCESS)
goto fail;
result = wsi->BindImageMemory(chain->device, image->image,
image->memory, 0);
if (result != VK_SUCCESS)
goto fail;
if (info->finish_create) {
result = info->finish_create(chain, info, image);
if (result != VK_SUCCESS)
goto fail;
}
return VK_SUCCESS;
fail:
wsi_destroy_image(chain, image);
return result;
}
void
wsi_destroy_image(const struct wsi_swapchain *chain,
struct wsi_image *image)
{
const struct wsi_device *wsi = chain->wsi;
#ifndef _WIN32
if (image->dma_buf_fd >= 0)
close(image->dma_buf_fd);
#endif
if (image->cpu_map != NULL) {
wsi->UnmapMemory(chain->device, image->buffer.buffer != VK_NULL_HANDLE ?
image->buffer.memory : image->memory);
}
if (image->buffer.blit_cmd_buffers) {
int cmd_buffer_count =
chain->buffer_blit_queue != VK_NULL_HANDLE ? 1 : wsi->queue_family_count;
for (uint32_t i = 0; i < cmd_buffer_count; i++) {
wsi->FreeCommandBuffers(chain->device, chain->cmd_pools[i],
1, &image->buffer.blit_cmd_buffers[i]);
}
vk_free(&chain->alloc, image->buffer.blit_cmd_buffers);
}
wsi->FreeMemory(chain->device, image->memory, &chain->alloc);
wsi->DestroyImage(chain->device, image->image, &chain->alloc);
wsi->FreeMemory(chain->device, image->buffer.memory, &chain->alloc);
wsi->DestroyBuffer(chain->device, image->buffer.buffer, &chain->alloc);
}
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfaceSupportKHR(VkPhysicalDevice physicalDevice,
uint32_t queueFamilyIndex,
VkSurfaceKHR _surface,
VkBool32 *pSupported)
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
return iface->get_support(surface, wsi_device,
queueFamilyIndex, pSupported);
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfaceCapabilitiesKHR(
VkPhysicalDevice physicalDevice,
VkSurfaceKHR _surface,
VkSurfaceCapabilitiesKHR *pSurfaceCapabilities)
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
VkSurfaceCapabilities2KHR caps2 = {
.sType = VK_STRUCTURE_TYPE_SURFACE_CAPABILITIES_2_KHR,
};
VkResult result = iface->get_capabilities2(surface, wsi_device, NULL, &caps2);
if (result == VK_SUCCESS)
*pSurfaceCapabilities = caps2.surfaceCapabilities;
return result;
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfaceCapabilities2KHR(
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VkPhysicalDevice physicalDevice,
const VkPhysicalDeviceSurfaceInfo2KHR *pSurfaceInfo,
VkSurfaceCapabilities2KHR *pSurfaceCapabilities)
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, pSurfaceInfo->surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
return iface->get_capabilities2(surface, wsi_device, pSurfaceInfo->pNext,
pSurfaceCapabilities);
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfaceCapabilities2EXT(
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VkPhysicalDevice physicalDevice,
VkSurfaceKHR _surface,
VkSurfaceCapabilities2EXT *pSurfaceCapabilities)
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
assert(pSurfaceCapabilities->sType ==
VK_STRUCTURE_TYPE_SURFACE_CAPABILITIES_2_EXT);
struct wsi_surface_supported_counters counters = {
.sType = VK_STRUCTURE_TYPE_WSI_SURFACE_SUPPORTED_COUNTERS_MESA,
.pNext = pSurfaceCapabilities->pNext,
.supported_surface_counters = 0,
};
VkSurfaceCapabilities2KHR caps2 = {
.sType = VK_STRUCTURE_TYPE_SURFACE_CAPABILITIES_2_KHR,
.pNext = &counters,
};
VkResult result = iface->get_capabilities2(surface, wsi_device, NULL, &caps2);
if (result == VK_SUCCESS) {
VkSurfaceCapabilities2EXT *ext_caps = pSurfaceCapabilities;
VkSurfaceCapabilitiesKHR khr_caps = caps2.surfaceCapabilities;
ext_caps->minImageCount = khr_caps.minImageCount;
ext_caps->maxImageCount = khr_caps.maxImageCount;
ext_caps->currentExtent = khr_caps.currentExtent;
ext_caps->minImageExtent = khr_caps.minImageExtent;
ext_caps->maxImageExtent = khr_caps.maxImageExtent;
ext_caps->maxImageArrayLayers = khr_caps.maxImageArrayLayers;
ext_caps->supportedTransforms = khr_caps.supportedTransforms;
ext_caps->currentTransform = khr_caps.currentTransform;
ext_caps->supportedCompositeAlpha = khr_caps.supportedCompositeAlpha;
ext_caps->supportedUsageFlags = khr_caps.supportedUsageFlags;
ext_caps->supportedSurfaceCounters = counters.supported_surface_counters;
}
return result;
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfaceFormatsKHR(VkPhysicalDevice physicalDevice,
VkSurfaceKHR _surface,
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
uint32_t *pSurfaceFormatCount,
VkSurfaceFormatKHR *pSurfaceFormats)
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
return iface->get_formats(surface, wsi_device,
pSurfaceFormatCount, pSurfaceFormats);
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfaceFormats2KHR(VkPhysicalDevice physicalDevice,
const VkPhysicalDeviceSurfaceInfo2KHR * pSurfaceInfo,
uint32_t *pSurfaceFormatCount,
VkSurfaceFormat2KHR *pSurfaceFormats)
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, pSurfaceInfo->surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
return iface->get_formats2(surface, wsi_device, pSurfaceInfo->pNext,
pSurfaceFormatCount, pSurfaceFormats);
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDeviceSurfacePresentModesKHR(VkPhysicalDevice physicalDevice,
VkSurfaceKHR _surface,
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
uint32_t *pPresentModeCount,
VkPresentModeKHR *pPresentModes)
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
return iface->get_present_modes(surface, pPresentModeCount,
pPresentModes);
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetPhysicalDevicePresentRectanglesKHR(VkPhysicalDevice physicalDevice,
VkSurfaceKHR _surface,
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
uint32_t *pRectCount,
VkRect2D *pRects)
{
VK_FROM_HANDLE(vk_physical_device, device, physicalDevice);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, _surface);
struct wsi_device *wsi_device = device->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
return iface->get_present_rectangles(surface, wsi_device,
pRectCount, pRects);
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
}
VKAPI_ATTR VkResult VKAPI_CALL
wsi_CreateSwapchainKHR(VkDevice _device,
const VkSwapchainCreateInfoKHR *pCreateInfo,
const VkAllocationCallbacks *pAllocator,
VkSwapchainKHR *pSwapchain)
{
VK_FROM_HANDLE(vk_device, device, _device);
ICD_FROM_HANDLE(VkIcdSurfaceBase, surface, pCreateInfo->surface);
struct wsi_device *wsi_device = device->physical->wsi_device;
struct wsi_interface *iface = wsi_device->wsi[surface->platform];
const VkAllocationCallbacks *alloc;
struct wsi_swapchain *swapchain;
if (pAllocator)
alloc = pAllocator;
else
alloc = &device->alloc;
VkResult result = iface->create_swapchain(surface, _device, wsi_device,
pCreateInfo, alloc,
&swapchain);
if (result != VK_SUCCESS)
return result;
swapchain->fences = vk_zalloc(alloc,
sizeof (*swapchain->fences) * swapchain->image_count,
sizeof (*swapchain->fences),
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (!swapchain->fences) {
swapchain->destroy(swapchain, alloc);
return VK_ERROR_OUT_OF_HOST_MEMORY;
}
if (swapchain->buffer_blit_queue != VK_NULL_HANDLE) {
swapchain->buffer_blit_semaphores = vk_zalloc(alloc,
sizeof (*swapchain->buffer_blit_semaphores) * swapchain->image_count,
sizeof (*swapchain->buffer_blit_semaphores),
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (!swapchain->buffer_blit_semaphores) {
swapchain->destroy(swapchain, alloc);
return VK_ERROR_OUT_OF_HOST_MEMORY;
}
}
*pSwapchain = wsi_swapchain_to_handle(swapchain);
return VK_SUCCESS;
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR void VKAPI_CALL
wsi_DestroySwapchainKHR(VkDevice _device,
VkSwapchainKHR _swapchain,
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
const VkAllocationCallbacks *pAllocator)
{
VK_FROM_HANDLE(vk_device, device, _device);
VK_FROM_HANDLE(wsi_swapchain, swapchain, _swapchain);
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
const VkAllocationCallbacks *alloc;
if (!swapchain)
return;
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
if (pAllocator)
alloc = pAllocator;
else
alloc = &device->alloc;
swapchain->destroy(swapchain, alloc);
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
}
VkResult
wsi_common_get_images(VkSwapchainKHR _swapchain,
uint32_t *pSwapchainImageCount,
VkImage *pSwapchainImages)
{
VK_FROM_HANDLE(wsi_swapchain, swapchain, _swapchain);
VK_OUTARRAY_MAKE_TYPED(VkImage, images, pSwapchainImages, pSwapchainImageCount);
for (uint32_t i = 0; i < swapchain->image_count; i++) {
vk_outarray_append_typed(VkImage, &images, image) {
*image = swapchain->get_wsi_image(swapchain, i)->image;
}
}
return vk_outarray_status(&images);
}
VkImage
wsi_common_get_image(VkSwapchainKHR _swapchain, uint32_t index)
{
VK_FROM_HANDLE(wsi_swapchain, swapchain, _swapchain);
assert(index < swapchain->image_count);
return swapchain->get_wsi_image(swapchain, index)->image;
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetSwapchainImagesKHR(VkDevice device,
VkSwapchainKHR swapchain,
uint32_t *pSwapchainImageCount,
VkImage *pSwapchainImages)
{
return wsi_common_get_images(swapchain,
pSwapchainImageCount,
pSwapchainImages);
}
VKAPI_ATTR VkResult VKAPI_CALL
wsi_AcquireNextImageKHR(VkDevice _device,
VkSwapchainKHR swapchain,
uint64_t timeout,
VkSemaphore semaphore,
VkFence fence,
uint32_t *pImageIndex)
{
VK_FROM_HANDLE(vk_device, device, _device);
const VkAcquireNextImageInfoKHR acquire_info = {
.sType = VK_STRUCTURE_TYPE_ACQUIRE_NEXT_IMAGE_INFO_KHR,
.swapchain = swapchain,
.timeout = timeout,
.semaphore = semaphore,
.fence = fence,
.deviceMask = 0,
};
return device->dispatch_table.AcquireNextImage2KHR(_device, &acquire_info,
pImageIndex);
}
static VkResult
wsi_signal_semaphore_for_image(struct vk_device *device,
const struct wsi_swapchain *chain,
const struct wsi_image *image,
VkSemaphore _semaphore)
{
if (device->physical->supported_sync_types == NULL)
return VK_SUCCESS;
VK_FROM_HANDLE(vk_semaphore, semaphore, _semaphore);
vk_semaphore_reset_temporary(device, semaphore);
#ifdef HAVE_LIBDRM
VkResult result = wsi_create_sync_for_dma_buf_wait(chain, image,
VK_SYNC_FEATURE_GPU_WAIT,
&semaphore->temporary);
if (result != VK_ERROR_FEATURE_NOT_PRESENT)
return result;
#endif
if (chain->wsi->signal_semaphore_with_memory) {
return device->create_sync_for_memory(device, image->memory,
false /* signal_memory */,
&semaphore->temporary);
} else {
return vk_sync_create(device, &vk_sync_dummy_type,
0 /* flags */, 0 /* initial_value */,
&semaphore->temporary);
}
}
static VkResult
wsi_signal_fence_for_image(struct vk_device *device,
const struct wsi_swapchain *chain,
const struct wsi_image *image,
VkFence _fence)
{
if (device->physical->supported_sync_types == NULL)
return VK_SUCCESS;
VK_FROM_HANDLE(vk_fence, fence, _fence);
vk_fence_reset_temporary(device, fence);
#ifdef HAVE_LIBDRM
VkResult result = wsi_create_sync_for_dma_buf_wait(chain, image,
VK_SYNC_FEATURE_CPU_WAIT,
&fence->temporary);
if (result != VK_ERROR_FEATURE_NOT_PRESENT)
return result;
#endif
if (chain->wsi->signal_fence_with_memory) {
return device->create_sync_for_memory(device, image->memory,
false /* signal_memory */,
&fence->temporary);
} else {
return vk_sync_create(device, &vk_sync_dummy_type,
0 /* flags */, 0 /* initial_value */,
&fence->temporary);
}
}
VkResult
wsi_common_acquire_next_image2(const struct wsi_device *wsi,
VkDevice _device,
const VkAcquireNextImageInfoKHR *pAcquireInfo,
uint32_t *pImageIndex)
{
VK_FROM_HANDLE(wsi_swapchain, swapchain, pAcquireInfo->swapchain);
VK_FROM_HANDLE(vk_device, device, _device);
VkResult result = swapchain->acquire_next_image(swapchain, pAcquireInfo,
pImageIndex);
if (result != VK_SUCCESS && result != VK_SUBOPTIMAL_KHR)
return result;
struct wsi_image *image =
swapchain->get_wsi_image(swapchain, *pImageIndex);
if (pAcquireInfo->semaphore != VK_NULL_HANDLE) {
VkResult signal_result =
wsi_signal_semaphore_for_image(device, swapchain, image,
pAcquireInfo->semaphore);
if (signal_result != VK_SUCCESS)
return signal_result;
}
if (pAcquireInfo->fence != VK_NULL_HANDLE) {
VkResult signal_result =
wsi_signal_fence_for_image(device, swapchain, image,
pAcquireInfo->fence);
if (signal_result != VK_SUCCESS)
return signal_result;
}
if (wsi->set_memory_ownership)
wsi->set_memory_ownership(swapchain->device, image->memory, true);
return result;
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_AcquireNextImage2KHR(VkDevice _device,
const VkAcquireNextImageInfoKHR *pAcquireInfo,
uint32_t *pImageIndex)
{
VK_FROM_HANDLE(vk_device, device, _device);
return wsi_common_acquire_next_image2(device->physical->wsi_device,
_device, pAcquireInfo, pImageIndex);
}
VkResult
wsi_common_queue_present(const struct wsi_device *wsi,
VkDevice device,
VkQueue queue,
int queue_family_index,
const VkPresentInfoKHR *pPresentInfo)
{
VkResult final_result = VK_SUCCESS;
STACK_ARRAY(VkPipelineStageFlags, stage_flags,
MAX2(1, pPresentInfo->waitSemaphoreCount));
for (uint32_t s = 0; s < MAX2(1, pPresentInfo->waitSemaphoreCount); s++)
stage_flags[s] = VK_PIPELINE_STAGE_ALL_COMMANDS_BIT;
const VkPresentRegionsKHR *regions =
vk_find_struct_const(pPresentInfo->pNext, PRESENT_REGIONS_KHR);
for (uint32_t i = 0; i < pPresentInfo->swapchainCount; i++) {
VK_FROM_HANDLE(wsi_swapchain, swapchain, pPresentInfo->pSwapchains[i]);
uint32_t image_index = pPresentInfo->pImageIndices[i];
VkResult result;
if (swapchain->fences[image_index] == VK_NULL_HANDLE) {
const VkFenceCreateInfo fence_info = {
.sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO,
.pNext = NULL,
.flags = VK_FENCE_CREATE_SIGNALED_BIT,
};
result = wsi->CreateFence(device, &fence_info,
&swapchain->alloc,
&swapchain->fences[image_index]);
if (result != VK_SUCCESS)
goto fail_present;
if (swapchain->use_buffer_blit && swapchain->buffer_blit_queue != VK_NULL_HANDLE) {
const VkSemaphoreCreateInfo sem_info = {
.sType = VK_STRUCTURE_TYPE_SEMAPHORE_CREATE_INFO,
.pNext = NULL,
.flags = 0,
};
result = wsi->CreateSemaphore(device, &sem_info,
&swapchain->alloc,
&swapchain->buffer_blit_semaphores[image_index]);
if (result != VK_SUCCESS)
goto fail_present;
}
} else {
result =
wsi->WaitForFences(device, 1, &swapchain->fences[image_index],
true, ~0ull);
if (result != VK_SUCCESS)
goto fail_present;
}
result = wsi->ResetFences(device, 1, &swapchain->fences[image_index]);
if (result != VK_SUCCESS)
goto fail_present;
VkSubmitInfo submit_info = {
.sType = VK_STRUCTURE_TYPE_SUBMIT_INFO,
};
if (i == 0) {
/* We only need/want to wait on semaphores once. After that, we're
* guaranteed ordering since it all happens on the same queue.
*/
submit_info.waitSemaphoreCount = pPresentInfo->waitSemaphoreCount;
submit_info.pWaitSemaphores = pPresentInfo->pWaitSemaphores;
submit_info.pWaitDstStageMask = stage_flags;
}
struct wsi_image *image =
swapchain->get_wsi_image(swapchain, image_index);
VkQueue submit_queue = queue;
if (swapchain->use_buffer_blit) {
if (swapchain->buffer_blit_queue == VK_NULL_HANDLE) {
submit_info.commandBufferCount = 1;
submit_info.pCommandBuffers =
&image->buffer.blit_cmd_buffers[queue_family_index];
} else {
/* If we are using a blit using the driver's private queue, then
* do an empty submit signalling a semaphore, and then submit the
* blit waiting on that. This ensures proper queue ordering of
* vkQueueSubmit() calls.
*/
submit_info.signalSemaphoreCount = 1;
submit_info.pSignalSemaphores =
&swapchain->buffer_blit_semaphores[image_index];
result = wsi->QueueSubmit(queue, 1, &submit_info, VK_NULL_HANDLE);
if (result != VK_SUCCESS)
goto fail_present;
/* Now prepare the blit submit. It needs to then wait on the
* semaphore we signaled above.
*/
submit_queue = swapchain->buffer_blit_queue;
submit_info.waitSemaphoreCount = 1;
submit_info.pWaitSemaphores = submit_info.pSignalSemaphores;
submit_info.signalSemaphoreCount = 0;
submit_info.pSignalSemaphores = NULL;
submit_info.commandBufferCount = 1;
submit_info.pCommandBuffers = &image->buffer.blit_cmd_buffers[0];
submit_info.pWaitDstStageMask = stage_flags;
}
}
VkFence fence = swapchain->fences[image_index];
bool has_signal_dma_buf = false;
#ifdef HAVE_LIBDRM
result = wsi_prepare_signal_dma_buf_from_semaphore(swapchain, image);
if (result == VK_SUCCESS) {
assert(submit_info.signalSemaphoreCount == 0);
submit_info.signalSemaphoreCount = 1;
submit_info.pSignalSemaphores = &swapchain->dma_buf_semaphore;
has_signal_dma_buf = true;
} else if (result == VK_ERROR_FEATURE_NOT_PRESENT) {
result = VK_SUCCESS;
has_signal_dma_buf = false;
} else {
goto fail_present;
}
#endif
struct wsi_memory_signal_submit_info mem_signal;
if (!has_signal_dma_buf) {
/* If we don't have dma-buf signaling, signal the memory object by
* chaining wsi_memory_signal_submit_info into VkSubmitInfo.
*/
result = VK_SUCCESS;
has_signal_dma_buf = false;
mem_signal = (struct wsi_memory_signal_submit_info) {
.sType = VK_STRUCTURE_TYPE_WSI_MEMORY_SIGNAL_SUBMIT_INFO_MESA,
.memory = image->memory,
};
__vk_append_struct(&submit_info, &mem_signal);
}
result = wsi->QueueSubmit(submit_queue, 1, &submit_info, fence);
if (result != VK_SUCCESS)
goto fail_present;
#ifdef HAVE_LIBDRM
if (has_signal_dma_buf) {
result = wsi_signal_dma_buf_from_semaphore(swapchain, image);
if (result != VK_SUCCESS)
goto fail_present;
}
#else
assert(!has_signal_dma_buf);
#endif
if (wsi->sw)
wsi->WaitForFences(device, 1, &swapchain->fences[image_index],
true, ~0ull);
const VkPresentRegionKHR *region = NULL;
if (regions && regions->pRegions)
region = &regions->pRegions[i];
result = swapchain->queue_present(swapchain, image_index, region);
if (result != VK_SUCCESS && result != VK_SUBOPTIMAL_KHR)
goto fail_present;
if (wsi->set_memory_ownership) {
VkDeviceMemory mem = swapchain->get_wsi_image(swapchain, image_index)->memory;
wsi->set_memory_ownership(swapchain->device, mem, false);
}
fail_present:
if (pPresentInfo->pResults != NULL)
pPresentInfo->pResults[i] = result;
/* Let the final result be our first unsuccessful result */
if (final_result == VK_SUCCESS)
final_result = result;
}
STACK_ARRAY_FINISH(stage_flags);
return final_result;
}
vulkan/wsi: Add common wrappers for most entrypoints For a long time, our Vulkan WSI code has acted as something of a layer. The WSI code calls into various Vulkan entrypoints inside the driver to create images, allocate memory, etc. It then implements the API-facing interface almost entirely. The only thing the driver has to provide is little wrappers that wrap around the WSI calls to expose them through the API. However, now that we have a common dispatch framework, we can implement entrypoints directly in the WSI code. As long as the driver uses vk_instance, vk_physical_device, and vk_device, we can provide common wrappers for the vast majority of entrypoints. The only exceptions are vkAcquireNextImage, vkQueuePresent, vkRegisterDeviceEventEXT, and vkRegisterDisplayEventEXT because those may have to manually poke at synchronization primitives. We provide wrappers for vkAcquireNextImage and vkQueuePresent because some drivers can use the default versions. For now, we're intentionally avoiding any link-time dependencies between WSI and the common code. We only use VK_FROM_HANDLE and associated inline helpers and vk_physical_device has a pointer to a wsi_device. Eventually, we may tie the two together closer, but this lets us get 95% of the way there without reworking the universe. Acked-by: Chia-I Wu <olvaffe@gmail.com> Acked-by: Iago Toral Quiroga <itoral@igalia.com> Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13234>
2021-10-06 17:09:12 +01:00
VKAPI_ATTR VkResult VKAPI_CALL
wsi_QueuePresentKHR(VkQueue _queue, const VkPresentInfoKHR *pPresentInfo)
{
VK_FROM_HANDLE(vk_queue, queue, _queue);
return wsi_common_queue_present(queue->base.device->physical->wsi_device,
vk_device_to_handle(queue->base.device),
_queue,
queue->queue_family_index,
pPresentInfo);
}
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetDeviceGroupPresentCapabilitiesKHR(VkDevice device,
VkDeviceGroupPresentCapabilitiesKHR *pCapabilities)
{
memset(pCapabilities->presentMask, 0,
sizeof(pCapabilities->presentMask));
pCapabilities->presentMask[0] = 0x1;
pCapabilities->modes = VK_DEVICE_GROUP_PRESENT_MODE_LOCAL_BIT_KHR;
return VK_SUCCESS;
}
VKAPI_ATTR VkResult VKAPI_CALL
wsi_GetDeviceGroupSurfacePresentModesKHR(VkDevice device,
VkSurfaceKHR surface,
VkDeviceGroupPresentModeFlagsKHR *pModes)
{
*pModes = VK_DEVICE_GROUP_PRESENT_MODE_LOCAL_BIT_KHR;
return VK_SUCCESS;
}
VkResult
wsi_common_create_swapchain_image(const struct wsi_device *wsi,
const VkImageCreateInfo *pCreateInfo,
VkSwapchainKHR _swapchain,
VkImage *pImage)
{
VK_FROM_HANDLE(wsi_swapchain, chain, _swapchain);
#ifndef NDEBUG
const VkImageCreateInfo *swcInfo = &chain->image_info.create;
assert(pCreateInfo->flags == 0);
assert(pCreateInfo->imageType == swcInfo->imageType);
assert(pCreateInfo->format == swcInfo->format);
assert(pCreateInfo->extent.width == swcInfo->extent.width);
assert(pCreateInfo->extent.height == swcInfo->extent.height);
assert(pCreateInfo->extent.depth == swcInfo->extent.depth);
assert(pCreateInfo->mipLevels == swcInfo->mipLevels);
assert(pCreateInfo->arrayLayers == swcInfo->arrayLayers);
assert(pCreateInfo->samples == swcInfo->samples);
assert(pCreateInfo->tiling == VK_IMAGE_TILING_OPTIMAL);
assert(!(pCreateInfo->usage & ~swcInfo->usage));
vk_foreach_struct_const(ext, pCreateInfo->pNext) {
switch (ext->sType) {
case VK_STRUCTURE_TYPE_IMAGE_FORMAT_LIST_CREATE_INFO: {
const VkImageFormatListCreateInfo *iflci =
(const VkImageFormatListCreateInfo *)ext;
const VkImageFormatListCreateInfo *swc_iflci =
&chain->image_info.format_list;
for (uint32_t i = 0; i < iflci->viewFormatCount; i++) {
bool found = false;
for (uint32_t j = 0; j < swc_iflci->viewFormatCount; j++) {
if (iflci->pViewFormats[i] == swc_iflci->pViewFormats[j]) {
found = true;
break;
}
}
assert(found);
}
break;
}
case VK_STRUCTURE_TYPE_IMAGE_SWAPCHAIN_CREATE_INFO_KHR:
break;
default:
assert(!"Unsupported image create extension");
}
}
#endif
return wsi->CreateImage(chain->device, &chain->image_info.create,
&chain->alloc, pImage);
}
VkResult
wsi_common_bind_swapchain_image(const struct wsi_device *wsi,
VkImage vk_image,
VkSwapchainKHR _swapchain,
uint32_t image_idx)
{
VK_FROM_HANDLE(wsi_swapchain, chain, _swapchain);
struct wsi_image *image = chain->get_wsi_image(chain, image_idx);
return wsi->BindImageMemory(chain->device, vk_image, image->memory, 0);
}
uint32_t
wsi_select_memory_type(const struct wsi_device *wsi,
VkMemoryPropertyFlags req_props,
VkMemoryPropertyFlags deny_props,
uint32_t type_bits)
{
assert(type_bits != 0);
VkMemoryPropertyFlags common_props = ~0;
u_foreach_bit(t, type_bits) {
const VkMemoryType type = wsi->memory_props.memoryTypes[t];
common_props &= type.propertyFlags;
if (deny_props & type.propertyFlags)
continue;
if (!(req_props & ~type.propertyFlags))
return t;
}
if ((deny_props & VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT) &&
(common_props & VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT)) {
/* If they asked for non-device-local and all the types are device-local
* (this is commonly true for UMA platforms), try again without denying
* device-local types
*/
deny_props &= ~VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT;
return wsi_select_memory_type(wsi, req_props, deny_props, type_bits);
}
unreachable("No memory type found");
}
uint32_t
wsi_select_device_memory_type(const struct wsi_device *wsi,
uint32_t type_bits)
{
return wsi_select_memory_type(wsi, VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT,
0 /* deny_props */, type_bits);
}
static uint32_t
wsi_select_host_memory_type(const struct wsi_device *wsi,
uint32_t type_bits)
{
return wsi_select_memory_type(wsi, VK_MEMORY_PROPERTY_HOST_COHERENT_BIT,
0 /* deny_props */, type_bits);
}
VkResult
wsi_create_buffer_image_mem(const struct wsi_swapchain *chain,
const struct wsi_image_info *info,
struct wsi_image *image,
VkExternalMemoryHandleTypeFlags handle_types,
bool implicit_sync)
{
const struct wsi_device *wsi = chain->wsi;
VkResult result;
const VkExternalMemoryBufferCreateInfo buffer_external_info = {
.sType = VK_STRUCTURE_TYPE_EXTERNAL_MEMORY_BUFFER_CREATE_INFO,
.pNext = NULL,
.handleTypes = handle_types,
};
const VkBufferCreateInfo buffer_info = {
.sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
.pNext = &buffer_external_info,
.size = info->linear_size,
.usage = VK_BUFFER_USAGE_TRANSFER_DST_BIT,
.sharingMode = VK_SHARING_MODE_EXCLUSIVE,
};
result = wsi->CreateBuffer(chain->device, &buffer_info,
&chain->alloc, &image->buffer.buffer);
if (result != VK_SUCCESS)
return result;
VkMemoryRequirements reqs;
wsi->GetBufferMemoryRequirements(chain->device, image->buffer.buffer, &reqs);
assert(reqs.size <= info->linear_size);
struct wsi_memory_allocate_info memory_wsi_info = {
.sType = VK_STRUCTURE_TYPE_WSI_MEMORY_ALLOCATE_INFO_MESA,
.pNext = NULL,
.implicit_sync = implicit_sync,
};
VkMemoryDedicatedAllocateInfo buf_mem_dedicated_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_ALLOCATE_INFO,
.pNext = &memory_wsi_info,
.image = VK_NULL_HANDLE,
.buffer = image->buffer.buffer,
};
VkMemoryAllocateInfo buf_mem_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.pNext = &buf_mem_dedicated_info,
.allocationSize = info->linear_size,
.memoryTypeIndex =
info->select_buffer_memory_type(wsi, reqs.memoryTypeBits),
};
void *sw_host_ptr = NULL;
if (info->alloc_shm)
sw_host_ptr = info->alloc_shm(image, info->linear_size);
VkExportMemoryAllocateInfo memory_export_info;
VkImportMemoryHostPointerInfoEXT host_ptr_info;
if (sw_host_ptr != NULL) {
host_ptr_info = (VkImportMemoryHostPointerInfoEXT) {
.sType = VK_STRUCTURE_TYPE_IMPORT_MEMORY_HOST_POINTER_INFO_EXT,
.pHostPointer = sw_host_ptr,
.handleType = VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT,
};
__vk_append_struct(&buf_mem_info, &host_ptr_info);
} else if (handle_types != 0) {
memory_export_info = (VkExportMemoryAllocateInfo) {
.sType = VK_STRUCTURE_TYPE_EXPORT_MEMORY_ALLOCATE_INFO,
.handleTypes = handle_types,
};
__vk_append_struct(&buf_mem_info, &memory_export_info);
}
result = wsi->AllocateMemory(chain->device, &buf_mem_info,
&chain->alloc, &image->buffer.memory);
if (result != VK_SUCCESS)
return result;
result = wsi->BindBufferMemory(chain->device, image->buffer.buffer,
image->buffer.memory, 0);
if (result != VK_SUCCESS)
return result;
wsi->GetImageMemoryRequirements(chain->device, image->image, &reqs);
const VkMemoryDedicatedAllocateInfo memory_dedicated_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_ALLOCATE_INFO,
.pNext = NULL,
.image = image->image,
.buffer = VK_NULL_HANDLE,
};
const VkMemoryAllocateInfo memory_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.pNext = &memory_dedicated_info,
.allocationSize = reqs.size,
.memoryTypeIndex =
info->select_image_memory_type(wsi, reqs.memoryTypeBits),
};
result = wsi->AllocateMemory(chain->device, &memory_info,
&chain->alloc, &image->memory);
if (result != VK_SUCCESS)
return result;
image->num_planes = 1;
image->sizes[0] = info->linear_size;
image->row_pitches[0] = info->linear_stride;
image->offsets[0] = 0;
return VK_SUCCESS;
}
VkResult
wsi_finish_create_buffer_image(const struct wsi_swapchain *chain,
const struct wsi_image_info *info,
struct wsi_image *image)
{
const struct wsi_device *wsi = chain->wsi;
VkResult result;
int cmd_buffer_count =
chain->buffer_blit_queue != VK_NULL_HANDLE ? 1 : wsi->queue_family_count;
image->buffer.blit_cmd_buffers =
vk_zalloc(&chain->alloc,
sizeof(VkCommandBuffer) * cmd_buffer_count, 8,
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (!image->buffer.blit_cmd_buffers)
return VK_ERROR_OUT_OF_HOST_MEMORY;
for (uint32_t i = 0; i < cmd_buffer_count; i++) {
const VkCommandBufferAllocateInfo cmd_buffer_info = {
.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO,
.pNext = NULL,
.commandPool = chain->cmd_pools[i],
.level = VK_COMMAND_BUFFER_LEVEL_PRIMARY,
.commandBufferCount = 1,
};
result = wsi->AllocateCommandBuffers(chain->device, &cmd_buffer_info,
&image->buffer.blit_cmd_buffers[i]);
if (result != VK_SUCCESS)
return result;
const VkCommandBufferBeginInfo begin_info = {
.sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_BEGIN_INFO,
};
wsi->BeginCommandBuffer(image->buffer.blit_cmd_buffers[i], &begin_info);
VkImageMemoryBarrier img_mem_barrier = {
.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER,
.pNext = NULL,
.srcAccessMask = 0,
.dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT,
.oldLayout = VK_IMAGE_LAYOUT_PRESENT_SRC_KHR,
.newLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL,
.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED,
.image = image->image,
.subresourceRange = {
.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
.baseMipLevel = 0,
.levelCount = 1,
.baseArrayLayer = 0,
.layerCount = 1,
},
};
wsi->CmdPipelineBarrier(image->buffer.blit_cmd_buffers[i],
VK_PIPELINE_STAGE_TOP_OF_PIPE_BIT,
VK_PIPELINE_STAGE_TRANSFER_BIT,
0,
0, NULL,
0, NULL,
1, &img_mem_barrier);
struct VkBufferImageCopy buffer_image_copy = {
.bufferOffset = 0,
.bufferRowLength = info->linear_stride /
vk_format_get_blocksize(info->create.format),
.bufferImageHeight = 0,
.imageSubresource = {
.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
.mipLevel = 0,
.baseArrayLayer = 0,
.layerCount = 1,
},
.imageOffset = { .x = 0, .y = 0, .z = 0 },
.imageExtent = info->create.extent,
};
wsi->CmdCopyImageToBuffer(image->buffer.blit_cmd_buffers[i],
image->image,
VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL,
image->buffer.buffer,
1, &buffer_image_copy);
img_mem_barrier.srcAccessMask = VK_ACCESS_TRANSFER_READ_BIT;
img_mem_barrier.dstAccessMask = 0;
img_mem_barrier.oldLayout = VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL;
img_mem_barrier.newLayout = VK_IMAGE_LAYOUT_PRESENT_SRC_KHR;
wsi->CmdPipelineBarrier(image->buffer.blit_cmd_buffers[i],
VK_PIPELINE_STAGE_TRANSFER_BIT,
VK_PIPELINE_STAGE_BOTTOM_OF_PIPE_BIT,
0,
0, NULL,
0, NULL,
1, &img_mem_barrier);
result = wsi->EndCommandBuffer(image->buffer.blit_cmd_buffers[i]);
if (result != VK_SUCCESS)
return result;
}
return VK_SUCCESS;
}
VkResult
wsi_configure_buffer_image(UNUSED const struct wsi_swapchain *chain,
const VkSwapchainCreateInfoKHR *pCreateInfo,
uint32_t stride_align, uint32_t size_align,
struct wsi_image_info *info)
{
const struct wsi_device *wsi = chain->wsi;
assert(util_is_power_of_two_nonzero(stride_align));
assert(util_is_power_of_two_nonzero(size_align));
VkResult result = wsi_configure_image(chain, pCreateInfo,
0 /* handle_types */, info);
if (result != VK_SUCCESS)
return result;
info->create.usage |= VK_IMAGE_USAGE_TRANSFER_SRC_BIT;
info->wsi.buffer_blit_src = true;
const uint32_t cpp = vk_format_get_blocksize(pCreateInfo->imageFormat);
info->linear_stride = pCreateInfo->imageExtent.width * cpp;
info->linear_stride = ALIGN_POT(info->linear_stride, stride_align);
/* Since we can pick the stride to be whatever we want, also align to the
* device's optimalBufferCopyRowPitchAlignment so we get efficient copies.
*/
assert(wsi->optimalBufferCopyRowPitchAlignment > 0);
info->linear_stride = ALIGN_POT(info->linear_stride,
wsi->optimalBufferCopyRowPitchAlignment);
info->linear_size = info->linear_stride * pCreateInfo->imageExtent.height;
info->linear_size = ALIGN_POT(info->linear_size, size_align);
info->finish_create = wsi_finish_create_buffer_image;
return VK_SUCCESS;
}
static VkResult
wsi_create_cpu_linear_image_mem(const struct wsi_swapchain *chain,
const struct wsi_image_info *info,
struct wsi_image *image)
{
const struct wsi_device *wsi = chain->wsi;
VkResult result;
VkMemoryRequirements reqs;
wsi->GetImageMemoryRequirements(chain->device, image->image, &reqs);
VkSubresourceLayout layout;
wsi->GetImageSubresourceLayout(chain->device, image->image,
&(VkImageSubresource) {
.aspectMask = VK_IMAGE_ASPECT_COLOR_BIT,
.mipLevel = 0,
.arrayLayer = 0,
}, &layout);
assert(layout.offset == 0);
const VkMemoryDedicatedAllocateInfo memory_dedicated_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_DEDICATED_ALLOCATE_INFO,
.image = image->image,
.buffer = VK_NULL_HANDLE,
};
VkMemoryAllocateInfo memory_info = {
.sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO,
.pNext = &memory_dedicated_info,
.allocationSize = reqs.size,
.memoryTypeIndex =
wsi_select_host_memory_type(wsi, reqs.memoryTypeBits),
};
void *sw_host_ptr = NULL;
if (info->alloc_shm)
sw_host_ptr = info->alloc_shm(image, layout.size);
VkImportMemoryHostPointerInfoEXT host_ptr_info;
if (sw_host_ptr != NULL) {
host_ptr_info = (VkImportMemoryHostPointerInfoEXT) {
.sType = VK_STRUCTURE_TYPE_IMPORT_MEMORY_HOST_POINTER_INFO_EXT,
.pHostPointer = sw_host_ptr,
.handleType = VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT,
};
__vk_append_struct(&memory_info, &host_ptr_info);
}
result = wsi->AllocateMemory(chain->device, &memory_info,
&chain->alloc, &image->memory);
if (result != VK_SUCCESS)
return result;
result = wsi->MapMemory(chain->device, image->memory,
0, VK_WHOLE_SIZE, 0, &image->cpu_map);
if (result != VK_SUCCESS)
return result;
image->num_planes = 1;
image->sizes[0] = reqs.size;
image->row_pitches[0] = layout.rowPitch;
image->offsets[0] = 0;
return VK_SUCCESS;
}
static VkResult
wsi_create_cpu_buffer_image_mem(const struct wsi_swapchain *chain,
const struct wsi_image_info *info,
struct wsi_image *image)
{
VkResult result;
result = wsi_create_buffer_image_mem(chain, info, image, 0,
false /* implicit_sync */);
if (result != VK_SUCCESS)
return result;
result = chain->wsi->MapMemory(chain->device, image->buffer.memory,
0, VK_WHOLE_SIZE, 0, &image->cpu_map);
if (result != VK_SUCCESS)
return result;
return VK_SUCCESS;
}
VkResult
wsi_configure_cpu_image(const struct wsi_swapchain *chain,
const VkSwapchainCreateInfoKHR *pCreateInfo,
uint8_t *(alloc_shm)(struct wsi_image *image,
unsigned size),
struct wsi_image_info *info)
{
const VkExternalMemoryHandleTypeFlags handle_types =
alloc_shm ? VK_EXTERNAL_MEMORY_HANDLE_TYPE_HOST_ALLOCATION_BIT_EXT : 0;
if (chain->use_buffer_blit) {
VkResult result = wsi_configure_buffer_image(chain, pCreateInfo,
1 /* stride_align */,
1 /* size_align */,
info);
if (result != VK_SUCCESS)
return result;
info->select_buffer_memory_type = wsi_select_host_memory_type;
info->select_image_memory_type = wsi_select_device_memory_type;
info->create_mem = wsi_create_cpu_buffer_image_mem;
} else {
VkResult result = wsi_configure_image(chain, pCreateInfo,
handle_types, info);
if (result != VK_SUCCESS)
return result;
/* Force the image to be linear */
info->create.tiling = VK_IMAGE_TILING_LINEAR;
info->create_mem = wsi_create_cpu_linear_image_mem;
}
info->alloc_shm = alloc_shm;
return VK_SUCCESS;
}