crosvm/rutabaga_gfx/src/virgl_renderer.rs

586 lines
20 KiB
Rust
Raw Normal View History

rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
// Copyright 2020 The Chromium OS Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
//! virgl_renderer: Handles 3D virtio-gpu hypercalls using virglrenderer.
//! External code found at https://gitlab.freedesktop.org/virgl/virglrenderer/.
#![cfg(feature = "virgl_renderer")]
use std::cell::RefCell;
use std::ffi::CString;
use std::mem::{size_of, transmute};
use std::os::raw::{c_char, c_void};
use std::ptr::null_mut;
use std::rc::Rc;
use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::Arc;
use base::{
warn, Error as SysError, ExternalMapping, ExternalMappingError, ExternalMappingResult,
FromRawDescriptor, SafeDescriptor,
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
};
use crate::generated::virgl_renderer_bindings::*;
use crate::renderer_utils::*;
use crate::rutabaga_core::{RutabagaComponent, RutabagaContext, RutabagaResource};
use crate::rutabaga_utils::*;
use data_model::VolatileSlice;
use libc::close;
type Query = virgl_renderer_export_query;
/// The virtio-gpu backend state tracker which supports accelerated rendering.
pub struct VirglRenderer {
fence_state: Rc<RefCell<FenceState>>,
}
struct VirglRendererContext {
ctx_id: u32,
}
impl RutabagaContext for VirglRendererContext {
fn submit_cmd(&mut self, commands: &mut [u8]) -> RutabagaResult<()> {
if commands.len() % size_of::<u32>() != 0 {
return Err(RutabagaError::InvalidCommandSize(commands.len()));
}
let dword_count = (commands.len() / size_of::<u32>()) as i32;
// Safe because the context and buffer are valid and virglrenderer will have been
// initialized if there are Context instances.
let ret = unsafe {
virgl_renderer_submit_cmd(
commands.as_mut_ptr() as *mut c_void,
self.ctx_id as i32,
dword_count,
)
};
ret_to_res(ret)
}
fn attach(&mut self, resource: &mut RutabagaResource) {
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
// The context id and resource id must be valid because the respective instances ensure
// their lifetime.
unsafe {
virgl_renderer_ctx_attach_resource(self.ctx_id as i32, resource.resource_id as i32);
}
}
fn detach(&mut self, resource: &RutabagaResource) {
// The context id and resource id must be valid because the respective instances ensure
// their lifetime.
unsafe {
virgl_renderer_ctx_detach_resource(self.ctx_id as i32, resource.resource_id as i32);
}
}
}
impl Drop for VirglRendererContext {
fn drop(&mut self) {
// The context is safe to destroy because nothing else can be referencing it.
unsafe {
virgl_renderer_context_destroy(self.ctx_id);
}
}
}
#[cfg(any(target_arch = "x86", target_arch = "x86_64"))]
extern "C" fn debug_callback(fmt: *const ::std::os::raw::c_char, ap: *mut __va_list_tag) {
let len: u32 = 256;
let mut c_str = CString::new(vec![' ' as u8; len as usize]).unwrap();
unsafe {
let mut varargs = __va_list_tag {
gp_offset: (*ap).gp_offset,
fp_offset: (*ap).fp_offset,
overflow_arg_area: (*ap).overflow_arg_area,
reg_save_area: (*ap).reg_save_area,
};
let raw = c_str.into_raw();
vsnprintf(raw, len.into(), fmt, &mut varargs);
c_str = CString::from_raw(raw);
}
base::debug!("{}", c_str.to_string_lossy());
}
const VIRGL_RENDERER_CALLBACKS: &virgl_renderer_callbacks = &virgl_renderer_callbacks {
version: 1,
write_fence: Some(write_fence),
create_gl_context: None,
destroy_gl_context: None,
make_current: None,
get_drm_fd: None,
};
/// Retrieves metadata suitable for export about this resource. If "export_fd" is true,
/// performs an export of this resource so that it may be imported by other processes.
fn export_query(resource_id: u32) -> RutabagaResult<Query> {
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
let mut query: Query = Default::default();
query.hdr.stype = VIRGL_RENDERER_STRUCTURE_TYPE_EXPORT_QUERY;
query.hdr.stype_version = 0;
query.hdr.size = size_of::<Query>() as u32;
query.in_resource_id = resource_id;
query.in_export_fds = 0;
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
// Safe because the image parameters are stack variables of the correct type.
let ret =
unsafe { virgl_renderer_execute(&mut query as *mut _ as *mut c_void, query.hdr.size) };
ret_to_res(ret)?;
Ok(query)
}
#[allow(unused_variables)]
fn map_func(resource_id: u32) -> ExternalMappingResult<(u64, usize)> {
#[cfg(feature = "virgl_renderer_next")]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
{
let mut map: *mut c_void = null_mut();
let map_ptr: *mut *mut c_void = &mut map;
let mut size: u64 = 0;
// Safe because virglrenderer wraps and validates use of GL/VK.
let ret = unsafe { virgl_renderer_resource_map(resource_id, map_ptr, &mut size) };
if ret != 0 {
return Err(ExternalMappingError::LibraryError(ret));
}
Ok((map as u64, size as usize))
}
#[cfg(not(feature = "virgl_renderer_next"))]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
Err(ExternalMappingError::Unsupported)
}
#[allow(unused_variables)]
fn unmap_func(resource_id: u32) {
#[cfg(feature = "virgl_renderer_next")]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
{
unsafe {
// Usually, process_gpu_command forces ctx0 when processing a virtio-gpu command.
// During VM shutdown, the KVM thread releases mappings without virtio-gpu being
// involved, so force ctx0 here. It's a no-op when the ctx is already 0, so there's
// little performance loss during normal VM operation.
virgl_renderer_force_ctx_0();
virgl_renderer_resource_unmap(resource_id);
}
}
}
impl VirglRenderer {
pub fn init(
virglrenderer_flags: VirglRendererFlags,
) -> RutabagaResult<Box<dyn RutabagaComponent>> {
if cfg!(debug_assertions) {
let ret = unsafe { libc::dup2(libc::STDOUT_FILENO, libc::STDERR_FILENO) };
if ret == -1 {
warn!("unable to dup2 stdout to stderr: {}", SysError::last());
}
}
// virglrenderer is a global state backed library that uses thread bound OpenGL contexts.
// Initialize it only once and use the non-send/non-sync Renderer struct to keep things tied
// to whichever thread called this function first.
static INIT_ONCE: AtomicBool = AtomicBool::new(false);
if INIT_ONCE
.compare_exchange(false, true, Ordering::Acquire, Ordering::Acquire)
.is_err()
{
rutabaga_gfx: rutabaga_gralloc: a shimmering beacon of hope rutabaga_gralloc is a cross-platform, Rust-based buffer manager. The rationale for this change is: 1) For the {cross-domain, wayland} context type, we need to have a good story for the crucial "wl-dmabuf" feature. As minigbm has been thoroughly tested on ChromeOS and currently powers the "wl-dmabuf" feature, it only makes sense for us to have a path to minigbm for the cross-domain prototype. This will be used by Sommelier. 2) While minigbm allocation works well on Chromebooks, it is not sufficient for cross-platform purposes. For their Virtual Graphics Interface (VGI) initiative, Android graphics virtualization experts have expressed their desire for a Vulkan based allocator. This will to go alongside cros_gralloc in minigbm, which is considered by many to be the ""world's premiere gralloc implementation". 3) Android graphics virtualization experts have expressed their desire for vkMapMemory(..) to be used when crosvm is in multi-process mode. Currently, only dma-buf mmap() is supported for zero-copy blobs in multi-process mode. dma-buf mmap() is not guaranteed to work on Nvidia (a "must have" for Cuttlefish) or any other driver for that matter (we *make* it work for ChromeOS). Possibly only solution: vkMapMemory ;-) With these goals in mind, here's a summary of the revelant changes: * Renamed the {gpu_allocator.rs, GpuMemoryAllocator trait} to be {gralloc.rs, Gralloc trait}. * Moved all GPU allocation out of the resources crate and into the rutabaga_gfx crate. This will allow the resources crate to be focused on managing resources for virtual machines. * Moved the gpu_buffer crate into the gralloc module in the rutabaga_gfx crate. The same functionality is now under "minigbm.rs", "minigbm_bindings.rs" and "rendernode.rs" * Added an optional dependency on vulkano.rs. vulkano.rs is a safe Rust wrapper around the Vulkan api [a]. It's emphasis on type safety makes a good fit for crosvm, though there are other high quality crates out there (gfx-rs, ash.rs). Though development has slowed down, it should satisfy goals (2) and (3) quite easily. * Added a system_gralloc implementation based on memfd. This can be used when minigbm or Vulkano features are not used, to replicate the highly useful "wl-shm" feature in Sommelier. Astute observers will note this can also enable seamless Wayland windowing without GPU features for Android too. Some minor changes to the base crate were needed. * Cut down on the amount of DrmFormats to the subset needed by Sommelier and cros_gralloc. * Moved checked arithmetic into it's own file. * Internally renamed to "wl-dmabuf" feature to be the "minigbm" feature. This is because "wl-dmabuf" has a dependency on minigbm. * Small rutabaga_gfx cleanups [a] https://github.com/vulkano-rs/vulkano/blob/master/DESIGN.md BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=run sommelier with "wl-dmabuf" and "wl-shm" Change-Id: I693a39cef64cd98e56d843d3c60caa7983d4d6e1 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2626487 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Zach Reizner <zachr@chromium.org> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org>
2020-12-09 18:44:13 +00:00
return Err(RutabagaError::AlreadyInUse);
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
}
// Cookie is intentionally never freed because virglrenderer never gets uninitialized.
// Otherwise, Resource and Context would become invalid because their lifetime is not tied
// to the Renderer instance. Doing so greatly simplifies the ownership for users of this
// library.
let fence_state = Rc::new(RefCell::new(FenceState { latest_fence: 0 }));
let cookie: *mut VirglCookie = Box::into_raw(Box::new(VirglCookie {
fence_state: Rc::clone(&fence_state),
}));
#[cfg(any(target_arch = "x86", target_arch = "x86_64"))]
unsafe {
virgl_set_debug_callback(Some(debug_callback))
};
// Safe because a valid cookie and set of callbacks is used and the result is checked for
// error.
let ret = unsafe {
virgl_renderer_init(
cookie as *mut c_void,
virglrenderer_flags.into(),
transmute(VIRGL_RENDERER_CALLBACKS),
)
};
ret_to_res(ret)?;
Ok(Box::new(VirglRenderer { fence_state }))
}
#[allow(unused_variables)]
fn map_info(&self, resource_id: u32) -> RutabagaResult<u32> {
#[cfg(feature = "virgl_renderer_next")]
{
let mut map_info = 0;
let ret =
unsafe { virgl_renderer_resource_get_map_info(resource_id as u32, &mut map_info) };
ret_to_res(ret)?;
Ok(map_info)
}
#[cfg(not(feature = "virgl_renderer_next"))]
Err(RutabagaError::Unsupported)
}
fn query(&self, resource_id: u32) -> RutabagaResult<Resource3DInfo> {
let query = export_query(resource_id)?;
if query.out_num_fds == 0 {
return Err(RutabagaError::Unsupported);
}
// virglrenderer unfortunately doesn't return the width or height, so map to zero.
Ok(Resource3DInfo {
width: 0,
height: 0,
drm_fourcc: query.out_fourcc,
strides: query.out_strides,
offsets: query.out_offsets,
modifier: query.out_modifier,
})
}
#[allow(unused_variables)]
fn export_blob(&self, resource_id: u32) -> RutabagaResult<Arc<RutabagaHandle>> {
#[cfg(feature = "virgl_renderer_next")]
{
let mut fd_type = 0;
let mut fd = 0;
let ret = unsafe {
virgl_renderer_resource_export_blob(resource_id as u32, &mut fd_type, &mut fd)
};
ret_to_res(ret)?;
/* Only support dma-bufs until someone wants opaque fds too. */
if fd_type != VIRGL_RENDERER_BLOB_FD_TYPE_DMABUF {
// Safe because the FD was just returned by a successful virglrenderer
// call so it must be valid and owned by us.
unsafe { close(fd) };
return Err(RutabagaError::Unsupported);
}
let dmabuf = unsafe { SafeDescriptor::from_raw_descriptor(fd) };
Ok(Arc::new(RutabagaHandle {
os_handle: dmabuf,
handle_type: RUTABAGA_MEM_HANDLE_TYPE_DMABUF,
}))
}
#[cfg(not(feature = "virgl_renderer_next"))]
Err(RutabagaError::Unsupported)
}
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
}
impl RutabagaComponent for VirglRenderer {
fn get_capset_info(&self, capset_id: u32) -> (u32, u32) {
let mut version = 0;
let mut size = 0;
// Safe because virglrenderer is initialized by now and properly size stack variables are
// used for the pointers.
unsafe {
virgl_renderer_get_cap_set(capset_id, &mut version, &mut size);
}
(version, size)
}
fn get_capset(&self, capset_id: u32, version: u32) -> Vec<u8> {
let (_, max_size) = self.get_capset_info(capset_id);
let mut buf = vec![0u8; max_size as usize];
// Safe because virglrenderer is initialized by now and the given buffer is sized properly
// for the given cap id/version.
unsafe {
virgl_renderer_fill_caps(capset_id, version, buf.as_mut_ptr() as *mut c_void);
}
buf
}
fn force_ctx_0(&self) {
unsafe { virgl_renderer_force_ctx_0() };
}
fn create_fence(&mut self, fence_data: RutabagaFenceData) -> RutabagaResult<()> {
let ret =
unsafe { virgl_renderer_create_fence(fence_data.fence_id as i32, fence_data.ctx_id) };
ret_to_res(ret)
}
fn poll(&self) -> u32 {
unsafe { virgl_renderer_poll() };
self.fence_state.borrow().latest_fence
}
fn create_3d(
&self,
resource_id: u32,
resource_create_3d: ResourceCreate3D,
) -> RutabagaResult<RutabagaResource> {
let mut args = virgl_renderer_resource_create_args {
handle: resource_id,
target: resource_create_3d.target,
format: resource_create_3d.format,
bind: resource_create_3d.bind,
width: resource_create_3d.width,
height: resource_create_3d.height,
depth: resource_create_3d.depth,
array_size: resource_create_3d.array_size,
last_level: resource_create_3d.last_level,
nr_samples: resource_create_3d.nr_samples,
flags: resource_create_3d.flags,
};
// Safe because virglrenderer is initialized by now, and the return value is checked before
// returning a new resource. The backing buffers are not supplied with this call.
let ret = unsafe { virgl_renderer_resource_create(&mut args, null_mut(), 0) };
ret_to_res(ret)?;
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
Ok(RutabagaResource {
resource_id,
handle: self.export_blob(resource_id).ok(),
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
blob: false,
blob_mem: 0,
blob_flags: 0,
map_info: None,
info_2d: None,
info_3d: self.query(resource_id).ok(),
vulkan_info: None,
backing_iovecs: None,
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
})
}
fn attach_backing(
&self,
resource_id: u32,
vecs: &mut Vec<RutabagaIovec>,
) -> RutabagaResult<()> {
// Safe because the backing is into guest memory that we store a reference count for.
let ret = unsafe {
virgl_renderer_resource_attach_iov(
resource_id as i32,
vecs.as_mut_ptr() as *mut iovec,
vecs.len() as i32,
)
};
ret_to_res(ret)
}
fn detach_backing(&self, resource_id: u32) {
// Safe as we don't need the old backing iovecs returned and the reference to the guest
// memory can be dropped as it will no longer be needed for this resource.
unsafe {
virgl_renderer_resource_detach_iov(resource_id as i32, null_mut(), null_mut());
}
}
fn unref_resource(&self, resource_id: u32) {
// The resource is safe to unreference destroy because no user of these bindings can still
// be holding a reference.
unsafe {
virgl_renderer_resource_unref(resource_id);
}
}
fn transfer_write(
&self,
ctx_id: u32,
resource: &mut RutabagaResource,
transfer: Transfer3D,
) -> RutabagaResult<()> {
if transfer.is_empty() {
return Ok(());
}
let mut transfer_box = VirglBox {
x: transfer.x,
y: transfer.y,
z: transfer.z,
w: transfer.w,
h: transfer.h,
d: transfer.d,
};
// Safe because only stack variables of the appropriate type are used.
let ret = unsafe {
virgl_renderer_transfer_write_iov(
resource.resource_id,
ctx_id,
transfer.level as i32,
transfer.stride,
transfer.layer_stride,
&mut transfer_box as *mut VirglBox as *mut virgl_box,
transfer.offset,
null_mut(),
0,
)
};
ret_to_res(ret)
}
fn transfer_read(
&self,
ctx_id: u32,
resource: &mut RutabagaResource,
transfer: Transfer3D,
buf: Option<VolatileSlice>,
) -> RutabagaResult<()> {
if transfer.is_empty() {
return Ok(());
}
let mut transfer_box = VirglBox {
x: transfer.x,
y: transfer.y,
z: transfer.z,
w: transfer.w,
h: transfer.h,
d: transfer.d,
};
let mut iov = RutabagaIovec {
base: null_mut(),
len: 0,
};
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
let (iovecs, num_iovecs) = match buf {
Some(buf) => {
iov.base = buf.as_ptr() as *mut c_void;
iov.len = buf.size() as usize;
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
(&mut iov as *mut RutabagaIovec as *mut iovec, 1)
}
None => (null_mut(), 0),
};
// Safe because only stack variables of the appropriate type are used.
let ret = unsafe {
virgl_renderer_transfer_read_iov(
resource.resource_id,
ctx_id,
transfer.level,
transfer.stride,
transfer.layer_stride,
&mut transfer_box as *mut VirglBox as *mut virgl_box,
transfer.offset,
iovecs,
num_iovecs,
)
};
ret_to_res(ret)
}
#[allow(unused_variables)]
fn create_blob(
&mut self,
ctx_id: u32,
resource_id: u32,
resource_create_blob: ResourceCreateBlob,
mut iovecs: Vec<RutabagaIovec>,
) -> RutabagaResult<RutabagaResource> {
#[cfg(feature = "virgl_renderer_next")]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
{
let resource_create_args = virgl_renderer_resource_create_blob_args {
res_handle: resource_id,
ctx_id,
blob_mem: resource_create_blob.blob_mem,
blob_flags: resource_create_blob.blob_flags,
blob_id: resource_create_blob.blob_id,
size: resource_create_blob.size,
iovecs: iovecs.as_mut_ptr() as *const iovec,
num_iovs: iovecs.len() as u32,
};
let ret = unsafe { virgl_renderer_resource_create_blob(&resource_create_args) };
ret_to_res(ret)?;
let iovec_opt = match resource_create_blob.blob_mem {
RUTABAGA_BLOB_MEM_GUEST => Some(iovecs),
_ => None,
};
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
Ok(RutabagaResource {
resource_id,
handle: self.export_blob(resource_id).ok(),
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
blob: true,
blob_mem: resource_create_blob.blob_mem,
blob_flags: resource_create_blob.blob_flags,
map_info: self.map_info(resource_id).ok(),
info_2d: None,
info_3d: self.query(resource_id).ok(),
vulkan_info: None,
backing_iovecs: iovec_opt,
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
})
}
#[cfg(not(feature = "virgl_renderer_next"))]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
Err(RutabagaError::Unsupported)
}
fn map(&self, resource_id: u32) -> RutabagaResult<ExternalMapping> {
let map_result = unsafe { ExternalMapping::new(resource_id, map_func, unmap_func) };
match map_result {
Ok(mapping) => Ok(mapping),
Err(e) => Err(RutabagaError::MappingFailed(e)),
}
}
#[allow(unused_variables)]
fn export_fence(&self, fence_id: u32) -> RutabagaResult<RutabagaHandle> {
#[cfg(feature = "virgl_renderer_next")]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
{
// Safe because the parameters are stack variables of the correct type.
let mut fd: i32 = 0;
let ret = unsafe { virgl_renderer_export_fence(fence_id, &mut fd) };
ret_to_res(ret)?;
// Safe because the FD was just returned by a successful virglrenderer call so it must
// be valid and owned by us.
let fence = unsafe { SafeDescriptor::from_raw_descriptor(fd) };
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
Ok(RutabagaHandle {
os_handle: fence,
handle_type: RUTABAGA_FENCE_HANDLE_TYPE_SYNC_FD,
})
}
#[cfg(not(feature = "virgl_renderer_next"))]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
Err(RutabagaError::Unsupported)
}
rutabaga_gfx: rutabaga_gralloc: a shimmering beacon of hope rutabaga_gralloc is a cross-platform, Rust-based buffer manager. The rationale for this change is: 1) For the {cross-domain, wayland} context type, we need to have a good story for the crucial "wl-dmabuf" feature. As minigbm has been thoroughly tested on ChromeOS and currently powers the "wl-dmabuf" feature, it only makes sense for us to have a path to minigbm for the cross-domain prototype. This will be used by Sommelier. 2) While minigbm allocation works well on Chromebooks, it is not sufficient for cross-platform purposes. For their Virtual Graphics Interface (VGI) initiative, Android graphics virtualization experts have expressed their desire for a Vulkan based allocator. This will to go alongside cros_gralloc in minigbm, which is considered by many to be the ""world's premiere gralloc implementation". 3) Android graphics virtualization experts have expressed their desire for vkMapMemory(..) to be used when crosvm is in multi-process mode. Currently, only dma-buf mmap() is supported for zero-copy blobs in multi-process mode. dma-buf mmap() is not guaranteed to work on Nvidia (a "must have" for Cuttlefish) or any other driver for that matter (we *make* it work for ChromeOS). Possibly only solution: vkMapMemory ;-) With these goals in mind, here's a summary of the revelant changes: * Renamed the {gpu_allocator.rs, GpuMemoryAllocator trait} to be {gralloc.rs, Gralloc trait}. * Moved all GPU allocation out of the resources crate and into the rutabaga_gfx crate. This will allow the resources crate to be focused on managing resources for virtual machines. * Moved the gpu_buffer crate into the gralloc module in the rutabaga_gfx crate. The same functionality is now under "minigbm.rs", "minigbm_bindings.rs" and "rendernode.rs" * Added an optional dependency on vulkano.rs. vulkano.rs is a safe Rust wrapper around the Vulkan api [a]. It's emphasis on type safety makes a good fit for crosvm, though there are other high quality crates out there (gfx-rs, ash.rs). Though development has slowed down, it should satisfy goals (2) and (3) quite easily. * Added a system_gralloc implementation based on memfd. This can be used when minigbm or Vulkano features are not used, to replicate the highly useful "wl-shm" feature in Sommelier. Astute observers will note this can also enable seamless Wayland windowing without GPU features for Android too. Some minor changes to the base crate were needed. * Cut down on the amount of DrmFormats to the subset needed by Sommelier and cros_gralloc. * Moved checked arithmetic into it's own file. * Internally renamed to "wl-dmabuf" feature to be the "minigbm" feature. This is because "wl-dmabuf" has a dependency on minigbm. * Small rutabaga_gfx cleanups [a] https://github.com/vulkano-rs/vulkano/blob/master/DESIGN.md BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=run sommelier with "wl-dmabuf" and "wl-shm" Change-Id: I693a39cef64cd98e56d843d3c60caa7983d4d6e1 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2626487 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Zach Reizner <zachr@chromium.org> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org>
2020-12-09 18:44:13 +00:00
#[allow(unused_variables)]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
fn create_context(
&self,
ctx_id: u32,
context_init: u32,
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
) -> RutabagaResult<Box<dyn RutabagaContext>> {
const CONTEXT_NAME: &[u8] = b"gpu_renderer";
// Safe because virglrenderer is initialized by now and the context name is statically
// allocated. The return value is checked before returning a new context.
let ret = unsafe {
#[cfg(feature = "virgl_renderer_next")]
match context_init {
0 => virgl_renderer_context_create(
ctx_id,
CONTEXT_NAME.len() as u32,
CONTEXT_NAME.as_ptr() as *const c_char,
),
_ => virgl_renderer_context_create_with_flags(
ctx_id,
context_init,
CONTEXT_NAME.len() as u32,
CONTEXT_NAME.as_ptr() as *const c_char,
),
}
#[cfg(not(feature = "virgl_renderer_next"))]
rutabaga_gfx: an electrifying possibility rutabaga_gfx is a cross platform, Rust-based, Wayland and Vulkan-centric Virtual Graphics Interface (VGI). Apologies for the mega-change, but it was hard to do this piece by piece. The rationale for this change is: 1) Android graphics virtualization experts have been proposing for a VGI for many months (years?). Their goal is to boot Android anywhere, everywhere. 2) For the {wayland, cross-domain} context type prototype, it's desirable to create a {wayland, camera} connection at the appropriate time. Details can be found in the code, though the RutabagaChannels have yet to be hooked up. There's a high chance neither effort will work. As such, rutabaga is just a prototype. However, even (1) and (2) don't end up working, this refactor/cleanup by itself makes a ton of sense. Here's a summary of revelant changes: * Removed auto-generated {p_defines, p_format, virgl_protocol}. These files were added for tests when bringing up crosvm-gpu, and AFAICT these tests are not run. There's actually now a commit queue for virglrenderer changes and container boot tests that provides excellent coverage. * Removed command_buffer.rs. Used only for the previously mentioned tests. It's quite nice, but couldn't determine the right place to put it. Maybe data_model? But removed it in the interim. * Removed {write_from_guest_memory, read_to_volatile}. The same basic functionality has been moved into {transfer_write, transfer_read} in Rutabaga. * Removed VirtioResource, Virtio3DResource, Virtio2DResource, and VirtioGfxStreamResource in favor of VirtioGpuResource and RutabagaResource. This leads to less duplication and clearer separation between external library functions and VMM functions. * Moved display and hypervisor memory management functions to virtio_gpu.rs. This is because external components do not interface with this functionality, and there was a lot of duplication (for example map/unmap blob). * Added context management between gfxstream and virglrenderer. * Added separate gfxstream and virglrenderer flags. * Clearer naming. * Added simple implementations for context init and multiple timelines. These changes have no effect since all Google kernels don't pass the revelant flags, but are useful for theoretical {wayland, cross-domain} prototype. * Unify RESOURCE_CREATE_3D and RESOURCE_CREATE_2D handling. * Better error handling. BUG=b:146066070, b:173630595, b:150239451 TEST=launch virtual machine with 2D mode TEST=launch virtual machine with 3D mode TEST=boot ARCVM Change-Id: I240b0c134a3b562cbc65981837a41f6db7767c92 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/crosvm/+/2522452 Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: kokoro <noreply+kokoro@google.com> Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Zach Reizner <zachr@chromium.org>
2020-10-20 01:31:13 +00:00
virgl_renderer_context_create(
ctx_id,
CONTEXT_NAME.len() as u32,
CONTEXT_NAME.as_ptr() as *const c_char,
)
};
ret_to_res(ret)?;
Ok(Box::new(VirglRendererContext { ctx_id }))
}
}