Re: [PATCH drm-next v5 00/14] [RFC] DRM GPUVA Manager & Nouveau VM_BIND UAPI

From: Danilo Krummrich
Date: Tue Jun 20 2023 - 10:21:33 EST


Hi Boris,

On 6/20/23 11:25, Boris Brezillon wrote:
Hi Danilo,

On Tue, 20 Jun 2023 02:42:03 +0200
Danilo Krummrich <dakr@xxxxxxxxxx> wrote:

This patch series provides a new UAPI for the Nouveau driver in order to
support Vulkan features, such as sparse bindings and sparse residency.

Furthermore, with the DRM GPUVA manager it provides a new DRM core feature to
keep track of GPU virtual address (VA) mappings in a more generic way.

The DRM GPUVA manager is indented to help drivers implement userspace-manageable
GPU VA spaces in reference to the Vulkan API. In order to achieve this goal it
serves the following purposes in this context.

1) Provide infrastructure to track GPU VA allocations and mappings,
making use of the maple_tree.

2) Generically connect GPU VA mappings to their backing buffers, in
particular DRM GEM objects.

3) Provide a common implementation to perform more complex mapping
operations on the GPU VA space. In particular splitting and merging
of GPU VA mappings, e.g. for intersecting mapping requests or partial
unmap requests.

The new VM_BIND Nouveau UAPI build on top of the DRM GPUVA manager, itself
providing the following new interfaces.

1) Initialize a GPU VA space via the new DRM_IOCTL_NOUVEAU_VM_INIT ioctl
for UMDs to specify the portion of VA space managed by the kernel and
userspace, respectively.

2) Allocate and free a VA space region as well as bind and unbind memory
to the GPUs VA space via the new DRM_IOCTL_NOUVEAU_VM_BIND ioctl.

3) Execute push buffers with the new DRM_IOCTL_NOUVEAU_EXEC ioctl.

Both, DRM_IOCTL_NOUVEAU_VM_BIND and DRM_IOCTL_NOUVEAU_EXEC, make use of the DRM
scheduler to queue jobs and support asynchronous processing with DRM syncobjs
as synchronization mechanism.

By default DRM_IOCTL_NOUVEAU_VM_BIND does synchronous processing,
DRM_IOCTL_NOUVEAU_EXEC supports asynchronous processing only.

The new VM_BIND UAPI for Nouveau makes also use of drm_exec (execution context
for GEM buffers) by Christian König. Since the patch implementing drm_exec was
not yet merged into drm-next it is part of this series, as well as a small fix
for this patch, which was found while testing this series.

This patch series is also available at [1].

There is a Mesa NVK merge request by Dave Airlie [2] implementing the
corresponding userspace parts for this series.

The Vulkan CTS test suite passes the sparse binding and sparse residency test
cases for the new UAPI together with Dave's Mesa work.

There are also some test cases in the igt-gpu-tools project [3] for the new UAPI
and hence the DRM GPU VA manager. However, most of them are testing the DRM GPU
VA manager's logic through Nouveau's new UAPI and should be considered just as
helper for implementation.

However, I absolutely intend to change those test cases to proper kunit test
cases for the DRM GPUVA manager, once and if we agree on it's usefulness and
design.

[1] https://gitlab.freedesktop.org/nouvelles/kernel/-/tree/new-uapi-drm-next /
https://gitlab.freedesktop.org/nouvelles/kernel/-/merge_requests/1
[2] https://gitlab.freedesktop.org/nouveau/mesa/-/merge_requests/150/
[3] https://gitlab.freedesktop.org/dakr/igt-gpu-tools/-/tree/wip_nouveau_vm_bind

Changes in V2:
==============
Nouveau:
- Reworked the Nouveau VM_BIND UAPI to avoid memory allocations in fence
signalling critical sections. Updates to the VA space are split up in three
separate stages, where only the 2. stage executes in a fence signalling
critical section:

1. update the VA space, allocate new structures and page tables
2. (un-)map the requested memory bindings
3. free structures and page tables

- Separated generic job scheduler code from specific job implementations.
- Separated the EXEC and VM_BIND implementation of the UAPI.
- Reworked the locking parts of the nvkm/vmm RAW interface, such that
(un-)map operations can be executed in fence signalling critical sections.

GPUVA Manager:
- made drm_gpuva_regions optional for users of the GPUVA manager
- allow NULL GEMs for drm_gpuva entries
- swichted from drm_mm to maple_tree for track drm_gpuva / drm_gpuva_region
entries
- provide callbacks for users to allocate custom drm_gpuva_op structures to
allow inheritance
- added user bits to drm_gpuva_flags
- added a prefetch operation type in order to support generating prefetch
operations in the same way other operations generated
- hand the responsibility for mutual exclusion for a GEM's
drm_gpuva list to the user; simplified corresponding (un-)link functions

Maple Tree:
- I added two maple tree patches to the series, one to support custom tree
walk macros and one to hand the locking responsibility to the user of the
GPUVA manager without pre-defined lockdep checks.

Changes in V3:
==============
Nouveau:
- Reworked the Nouveau VM_BIND UAPI to do the job cleanup (including page
table cleanup) within a workqueue rather than the job_free() callback of
the scheduler itself. A job_free() callback can stall the execution (run()
callback) of the next job in the queue. Since the page table cleanup
requires to take the same locks as need to be taken for page table
allocation, doing it directly in the job_free() callback would still
violate the fence signalling critical path.
- Separated Nouveau fence allocation and emit, such that we do not violate
the fence signalling critical path in EXEC jobs.
- Implement "regions" (for handling sparse mappings through PDEs and dual
page tables) within Nouveau.
- Drop the requirement for every mapping to be contained within a region.
- Add necassary synchronization of VM_BIND job operation sequences in order
to work around limitations in page table handling. This will be addressed
in a future re-work of Nouveau's page table handling.
- Fixed a couple of race conditions found through more testing. Thanks to
Dave for consitently trying to break it. :-)

GPUVA Manager:
- Implement pre-allocation capabilities for tree modifications within fence
signalling critical sections.
- Implement accessors to to apply tree modification while walking the GPUVA
tree in order to actually support processing of drm_gpuva_ops through
callbacks in fence signalling critical sections rather than through
pre-allocated operation lists.
- Remove merging of GPUVAs; the kernel has limited to none knowlege about
the semantics of mapping sequences. Hence, merging is purely speculative.
It seems that gaining a significant (or at least a measurable) performance
increase through merging is way more likely to happen when userspace is
responsible for merging mappings up to the next larger page size if
possible.
- Since merging was removed, regions pretty much loose their right to exist.
They might still be useful for handling dual page tables or similar
mechanisms, but since Nouveau seems to be the only driver having a need
for this for now, regions were removed from the GPUVA manager.
- Fixed a couple of maple_tree related issues; thanks to Liam for helping me
out.

Changes in V4:
==============
Nouveau:
- Refactored how specific VM_BIND and EXEC jobs are created and how their
arguments are passed to the generic job implementation.
- Fixed a UAF race condition where bind job ops could have been freed
already while still waiting for a job cleanup to finish. This is due to
in certain cases we need to wait for mappings actually being unmapped
before creating sparse regions in the same area.
- Re-based the code onto drm_exec v4 patch.

GPUVA Manager:
- Fixed a maple tree related bug when pre-allocating MA states.
(Boris Brezillion)
- Made struct drm_gpuva_fn_ops a const object in all occurrences.
(Boris Brezillion)

Changes in V5:
==============
Nouveau:
- Link and unlink GPUVAs outside the fence signalling critical path in
nouveau_uvmm_bind_job_submit() holding the dma-resv lock. Mutual exclusion
of BO evicts causing mapping invalidation and regular mapping operations
is ensured with dma-fences.

GPUVA Manager:
- Removed the separate GEMs GPUVA list lock. Link and unlink as well as
iterating the GEM's GPUVA list should be protected with the GEM's dma-resv
lock instead.
- Renamed DRM_GPUVA_EVICTED flag to DRM_GPUVA_INVALIDATED. Mappings do not
get eviced, they might get invalidated due to eviction.
- Maple tree uses the 'unsinged long' type for node entries. While this
works for GPU VA spaces larger than 32-bit on 64-bit kernel, the GPU VA
space is limited to 32-bit on 32-bit kernels as well.
As long as we do not have a 64-bit capable maple tree for 32-bit kernels,
the GPU VA manager contains checks to throw warnings when GPU VA entries
exceed the maple tree's storage capabilities.
- Extended the Documentation and added example code as requested by Donald
Robson.

Christian König (1):
drm: execution context for GEM buffers v4

Danilo Krummrich (13):
maple_tree: split up MA_STATE() macro
drm: manager to keep track of GPUs VA mappings
drm: debugfs: provide infrastructure to dump a DRM GPU VA space

Core drm patches are

Reviewed-by: Boris Brezillon <boris.brezillon@xxxxxxxxxxxxx>

The only thing I'm worried about is the 'sync mapping requests have to
go through the async path and wait for all previous async requests to
be processed' problem I mentioned in one of your previous submission,
but I'm happy leave that for later.

Yes, I'm aware of this limitation.

Let me quickly try to explain where this limitation comes from and how I intend to address it.

In order to be able to allocate the required page tables for a mapping request and in order to free corresponding page tables once the (async) job finished I need to know the corresponding sequence of operations (drm_gpuva_ops) to fulfill the mapping request.

This requires me to update the GPUVA space in the ioctl() rather than in the async stage, because otherwise I would need to wait for previous jobs to finish before being able to submit subsequent jobs to the job queue, since I need an up to date view of the GPUVA space in order to calculate the sequence of operations to fulfill a mapping request.

As a consequence all jobs need to be processed in the order they were submitted, including synchronous jobs.

@Matt: I think you will have the same limitation with synchronous jobs as your implementation in XE should be similar?

In order to address it I want to switch to using callbacks rather than 'pre-allocated' drm_gpuva_ops and update the GPUVA space within the asynchronous stage. This would allow me to 'fit' synchronous jobs between jobs waiting in the async job queue. However, to do this I have to re-work how the page table handling in Nouveau is implemented, since this would require me to be able to manage the page tables without knowing the exact sequence of operations to fulfill a mapping request.

- Danilo


drm/nouveau: new VM_BIND uapi interfaces
drm/nouveau: get vmm via nouveau_cli_vmm()
drm/nouveau: bo: initialize GEM GPU VA interface
drm/nouveau: move usercopy helpers to nouveau_drv.h
drm/nouveau: fence: separate fence alloc and emit
drm/nouveau: fence: fail to emit when fence context is killed
drm/nouveau: chan: provide nouveau_channel_kill()
drm/nouveau: nvkm/vmm: implement raw ops to manage uvmm
drm/nouveau: implement new VM_BIND uAPI
drm/nouveau: debugfs: implement DRM GPU VA debugfs

Documentation/gpu/driver-uapi.rst | 11 +
Documentation/gpu/drm-mm.rst | 54 +
drivers/gpu/drm/Kconfig | 6 +
drivers/gpu/drm/Makefile | 3 +
drivers/gpu/drm/drm_debugfs.c | 41 +
drivers/gpu/drm/drm_exec.c | 278 +++
drivers/gpu/drm/drm_gem.c | 3 +
drivers/gpu/drm/drm_gpuva_mgr.c | 1971 ++++++++++++++++
drivers/gpu/drm/nouveau/Kbuild | 3 +
drivers/gpu/drm/nouveau/Kconfig | 2 +
drivers/gpu/drm/nouveau/dispnv04/crtc.c | 9 +-
drivers/gpu/drm/nouveau/include/nvif/if000c.h | 26 +-
drivers/gpu/drm/nouveau/include/nvif/vmm.h | 19 +-
.../gpu/drm/nouveau/include/nvkm/subdev/mmu.h | 20 +-
drivers/gpu/drm/nouveau/nouveau_abi16.c | 24 +
drivers/gpu/drm/nouveau/nouveau_abi16.h | 1 +
drivers/gpu/drm/nouveau/nouveau_bo.c | 204 +-
drivers/gpu/drm/nouveau/nouveau_bo.h | 2 +-
drivers/gpu/drm/nouveau/nouveau_chan.c | 22 +-
drivers/gpu/drm/nouveau/nouveau_chan.h | 1 +
drivers/gpu/drm/nouveau/nouveau_debugfs.c | 39 +
drivers/gpu/drm/nouveau/nouveau_dmem.c | 9 +-
drivers/gpu/drm/nouveau/nouveau_drm.c | 27 +-
drivers/gpu/drm/nouveau/nouveau_drv.h | 94 +-
drivers/gpu/drm/nouveau/nouveau_exec.c | 418 ++++
drivers/gpu/drm/nouveau/nouveau_exec.h | 54 +
drivers/gpu/drm/nouveau/nouveau_fence.c | 23 +-
drivers/gpu/drm/nouveau/nouveau_fence.h | 5 +-
drivers/gpu/drm/nouveau/nouveau_gem.c | 62 +-
drivers/gpu/drm/nouveau/nouveau_mem.h | 5 +
drivers/gpu/drm/nouveau/nouveau_prime.c | 2 +-
drivers/gpu/drm/nouveau/nouveau_sched.c | 461 ++++
drivers/gpu/drm/nouveau/nouveau_sched.h | 123 +
drivers/gpu/drm/nouveau/nouveau_svm.c | 2 +-
drivers/gpu/drm/nouveau/nouveau_uvmm.c | 1979 +++++++++++++++++
drivers/gpu/drm/nouveau/nouveau_uvmm.h | 107 +
drivers/gpu/drm/nouveau/nouveau_vmm.c | 4 +-
drivers/gpu/drm/nouveau/nvif/vmm.c | 100 +-
.../gpu/drm/nouveau/nvkm/subdev/mmu/uvmm.c | 213 +-
drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.c | 197 +-
drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h | 25 +
.../drm/nouveau/nvkm/subdev/mmu/vmmgf100.c | 16 +-
.../drm/nouveau/nvkm/subdev/mmu/vmmgp100.c | 16 +-
.../gpu/drm/nouveau/nvkm/subdev/mmu/vmmnv50.c | 27 +-
include/drm/drm_debugfs.h | 25 +
include/drm/drm_drv.h | 6 +
include/drm/drm_exec.h | 119 +
include/drm/drm_gem.h | 52 +
include/drm/drm_gpuva_mgr.h | 682 ++++++
include/linux/maple_tree.h | 7 +-
include/uapi/drm/nouveau_drm.h | 209 ++
51 files changed, 7566 insertions(+), 242 deletions(-)
create mode 100644 drivers/gpu/drm/drm_exec.c
create mode 100644 drivers/gpu/drm/drm_gpuva_mgr.c
create mode 100644 drivers/gpu/drm/nouveau/nouveau_exec.c
create mode 100644 drivers/gpu/drm/nouveau/nouveau_exec.h
create mode 100644 drivers/gpu/drm/nouveau/nouveau_sched.c
create mode 100644 drivers/gpu/drm/nouveau/nouveau_sched.h
create mode 100644 drivers/gpu/drm/nouveau/nouveau_uvmm.c
create mode 100644 drivers/gpu/drm/nouveau/nouveau_uvmm.h
create mode 100644 include/drm/drm_exec.h
create mode 100644 include/drm/drm_gpuva_mgr.h


base-commit: 2222dcb0775d36de28992f56455ab3967b30d380