Our sparse implementation will require us to issue partial unbinds,
but partial unbinds are not supported in synchronous vm_bind ioctls,
requiring us to to have our VM be marked with the ASYNC flag. This is
not properly documented and is subject to change in the next
iterations of the API.
Error handling with async binds is also not documented anywhere and is
being actively discussed in the mailing lists, so whatever we decide
to implement here is likely to end up changing in a few weeks. Also, I
haven't seen these errors happening in the real world, so for now
they're a very corner case. So for now just foward errors to
user-space and hope things work.
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24681>
For Sparse Resources we need to be able to specify the address, size
and offsets and we also want to be able to issue multiple binds at the
same time. Extend xe_vm_bind_op() to handle those cases and add
the new vfunc.
v2:
- use STACK_ARRAY() (Lionel)
- no more need to work around xe.ko bug that was fixed (José)
Reviewed-by: José Roberto de Souza <jose.souza@intel.com> (v1)
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24681>
The only driver that has a vm_bind ioctl is xe.ko, and its vm_bind
ioctl is not called GEM vm_bind, it's just DRM_IOCTL_XE_VM_BIND
(without GEM anywhere). Back when i915.ko was going to have a vm_bind
ioctl it had GEM on its name, so I guess that's how "gem" appeared in
the naming here, but now nothing does, so let's get rid of it.
Also, these vfuncs we have are specifically made to bind and unbind
whole BOs, so rename them to vm_bind_bo() and vm_unbind_bo() in order
to try to clarify what they mean. The goal is to add a more generic
vm_bind() later that can do anything.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24681>
While we don't need to emit all of the unused mesh/task states when mesh
is disabled, if we don't have them we fail some assertions in the
difference checks due to the corresponding state being empty.
This may happen when going from a mesh pipeline to a non-mesh one, or
one that uses task shaders to one that doesn't.
It may be possible to avoid having to do this, but I'd rather start from
a working state and optimize it later.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25109>
Flushing and invalidating caches isn't necessary for workgroup scope
fences. In fact, the DP_FLUSH_TYPE docs (BSpec 54041) say:
"If the fence scope is Local or Threadgroup, HW ignores the flush
type and operates as if it was set to None(no flush)"
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24842>
Don't rely on the HW to set values correctly so just emit
STATE_COMPUTE_MODE with default values set to zero.
Also, this change includes workaround changes:-
- 14015808183 (Parent HSD 14015782607) - Need to emit pipe control
with HDC flush and untyped cache flush set to 1 when CCS has
non-pipelined state update with STATE_COMPUTE_MODE.
- 14014427904 (Parent HSD 22013045878) - We need additional
invalidate/flush when emitting non-pipelined state commands with
multiple CCS enabled.
v2: (Tapani)
- Use lineage HSD numbers for check
- Don't use poisoned WA directly
- Use intel_needs_workaround helper
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24508>
When applying barriers for image transitions, we're currently
considering all possible usages of an image. But when running on a
compute only queue for example, the usage of an image will never be
one of those :
- VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT
- VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT
- VK_IMAGE_USAGE_TRANSIENT_ATTACHMENT_BIT
- VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT
- VK_IMAGE_USAGE_FRAGMENT_SHADING_RATE_ATTACHMENT_BIT_KHR
Removing unused usages for the compute queue allows us to reduce the
scope of the VK_IMAGE_LAYOUT_GENERAL for example. This a bunch of
transition operation that are completely useless when dealing with
barriers on the compute queue.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25092>
When we have MSAA copy/clear operation on the compute queue, use the
companion RCS command buffer to carry out copy/clear operations.
v2: (Sagar)
- Flush cache according to command buffer
- Invalidate AUX when we create new companion RCS command buffer if
platform support AUX TT.
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23661>
We need to synchronize main (CCS/BCS) and companion rcs batch, so let's
create an empty batch and make both the batches (CCS/BCS) and companion
RCS batch wait on empty sync batch and signal the fence.
Reason to execute the empty batch is we need to make sure the companion
RCS batch finish as soon as the CCS/BCS batch finish. Preemption could
prevent the companion RCS batch execution and we might end up destroying
the CCS/BCS batch before companion RCS finishes.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23661>
In future patches, we will be creating a separate companion RCS engine
and each engine is created with it's own address space, and we really
don't want. CCS and RCS engine writes should be visible to each other in
order to get the wait/signal mechanism working.
v2:
- Move drm_i915_gem_context_create_ext_setparam out of if block (Lionel)
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23661>
This helper takes the main command buffer as input and then create a
companion RCS command buffer.
v2:
- Rename anv_get_render_queue_index helper to
anv_get_first_render_queue_index (Jose)
- Rename RCS command buffer to companion RCS command buffer (Lionel)
- Add early return in anv_get_first_render_queue_index (Lionel)
- Add lock around the function (Jose)
- Move companion rcs command pool creation in device create (Sagar)
- Reset companion RCS cmd buffer (Sagar)
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23661>
A single Vulkan state can map to multiple fields in different GPU
instructions. This change introduces the bottom half of a simplified
emission mechanism where we do the following :
Vulkan runtime state
|
V
Intermediate driver state
|
V
Instruction programming
This way we can detect that the intermediate state didn't change and
avoid HW instruction emission.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24536>
The goal of this change it to move away from a single batch buffer
containing all kind of pipeline instructions to a list of instructions
we can emit separately.
We will later implement pipeline diffing and finer state tracking that
will allow fewer instructions to be emitted.
This changes the following things :
* instead of having a batch & partially packed instructions, move
everything into the batch
* add a set of pointer in the batch that allows us to point to each
instruction (almost... we group some like URB instructions,
etc...).
At pipeline emission time, we just go through all of those pointers
and emit the instruction into the batch. No additional packing is
involved.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24536>
Paulo Zanoni