Add implementation for a simple sub-allocator in order to save
memory when doing internal driver allocations.
Signed-off-by: Luigi Santivetti <luigi.santivetti@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22940>
Fixes random aborts in CSB handling.
Signed-off-by: Matt Coster <matt.coster@imgtec.com>
Reported-by: James Glanville <james.glanville@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22868>
This code will not be used by the new KMD, so it is being moved out of
this shared code area.
Signed-off-by: Donald Robson <donald.robson@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22826>
Avoids duplicating the struct in pvr_device and fixes uninitialized
accesses when ppEnabledFeatures was NULL.
Noticed when running valgrind over dEQP.
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22331>
rogue_compiler is allocated for each VkPhysicalDevice which means it can
be used from multiple threads when compiling. Allocating the build context
from this will lead to race conditions as ralloc is not thread safe.
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22332>
S8_UINT is not a requirement for Vulkan 1.0.
Revert "pvr: Add initial support for VK_FORMAT_S8_UINT".
This reverts commit 220356e083.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22822>
pvrsrvkm type sync objects can have a pending state where,
the fence is unsignaled but does not have a valid sync file
due to not yet being submitted to kernel.
The wait function therefore needs to handle these types of syncs
through a spin loop.
This was seen as crashes in dEQP-VK.synchronization.timeline_semaphore.*
Signed-off-by: SoroushIMG <soroush.kashani@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22822>
From the Vulkan spec (§11.2.13. Freeing Device Memory):
If a memory object is mapped at the time it is freed, it is implicitly
unmapped.
Signed-off-by: Matt Coster <matt.coster@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22762>
For indirect draw commands the data segment's address was being
calculated with an unaligned code size which could potentially
send an unaligned address to the hw causing problems.
Reported-by: James Glanville <james.glanville@imgtec.com>
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22746>
The code to increment the destination buffer address was wrong. Instead of
incrementing an address on every iteration, the buffer slice size was added
as a static offset to all blits.
Seen in dEQP-VK.draw.renderpass.depth_clamp.d16_unorm, though issues remain.
Signed-off-by: James Glanville <james.glanville@imgtec.com>
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22747>
When pvr_graphics_pipeline_alloc_vertex_special_vars() was
allocating vtxin regs for special vars, it wasn't returning back
the updated amount of allocated regs so when the PDS was filling
them in, it would try to write out of bounds.
Signed-off-by: James Glanville <james.glanville@imgtec.com>
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22750>
Depth/stencil surfaces are not PBE renderable, so a compatible format must
be used instead. The code to calculate this compatible format was not called
when configuring the PBE, and it was missing formats.
Also the code to calculate PBE swizzles was throwing an error in unhandled
cases, rather than using the pre-calculated defaults which was the correct
behaviour.
Fixes blits in dEQP-VK.draw.renderpass.depth_clamp.d32_sfloat
Signed-off-by: James Glanville <james.glanville@imgtec.com>
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22748>
Previously the code was saving the mask as a VkShaderStageFlags
but when allocating shareds it checked against pvr_stage_allocation.
This causes problems as only the vertex bit matches the
VkShaderStageFlagBits so the push constants utilized in fragment
shaders weren't picked up properly.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22731>
The list logic checks for list->next->next (+ some other checks)
to point to the list itself to determine that there is just one
single element.
┌───────────────────────┐
└< { HEAD } >─< { E0 } >┘
When the list_head is copied as was being done previously, the
list element's next pointer still points at the old head so
the `list_is_singular()` check fails.
Fixes pvr_cmd_buffer.c:605:`list_is_singular(&bo_list)` assertion
dEQP-VK.api.image_g.core.clear_color_attachment.cube_layers.b8g8r8a8_unorm
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22730>
We use dwords (32 bit) quite a bit around the code base. Previously
we used '* 4', '<< 2', or '* sizeof(uint32_t)' to go from dwords to
bytes. The conversion isn't always clear when other operations
happen in the same line, which can leave one wondering where the
multiplication came from.
PVR_DW_TO_BYTES() should make the code more obvious as well as
making the conversion more consistent.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22658>
This function was returning the size of a single region header as the stride
when it was supposed to be returning the total size of the region headers for a
single render target. This went unnoticed due to the fact this function had two
variables with basically identical names. To avoid any future confusion, rename
rgn_header_size to single_rgn_header_size throughout the code.
Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22574>
For now always allocate space for the special built-in variables
that need filling in for the vertex shader. Ideally we should get
the compiler to analyze the shader and tell the driver if they are
used so that we don't reserve and fill them needlessly.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21590>
The comment was incorrectly showing which DDMAD param corresponed
to which DDMAD field.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21588>
This also adds pvr_pbe_src_format_ds(), the equivalent to
pvr_pbe_src_format_normal() for depth/stencil formats.
Signed-off-by: Matt Coster <matt.coster@imgtec.com>
Acked-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21550>
This commit fixes the triangle demo on the AM62 by implementing
the paths that were left unimplemented due to the Chromebook not
having the simple internal format v2 feature.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Acked-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21550>
Take the opportunity to tweak the naming of pvr_transfer_ctx_setup_shaders and
pvr_transfer_ctx_fini_shaders to make them fit in with the rest of the naming in
the driver.
Signed-off-by: Frank Binns <frank.binns@imgtec.com>
Acked-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21550>
The robustness buffer is used when no vertex attribute can be loaded and
robustBufferAccess is enabled.
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21579>
As per the Vulkan spec, this is a required feature for graphics implementations.
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21579>
This change also means the texture state words will be based off of the
image view's format instead of the image's format.
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21594>
depth_stencil_attachment has been changed from a pointer to the attachment idx
to just the attachment idx, as this avoids the driver having to check for NULL
when comparing attachments indexes with depth_stencil_attachment.
Anyplace that relies on depth_stencil_attachment being a valid index must
already check that depth_stencil_attachment is not VK_ATTACHMENT_UNUSED, so
this change avoids having to check both the pointer and the index for the same
information.
Noticed when running dEQP-VK.api.smoke.triangle
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21690>
For optimalTilingFeatures we should be advertising
STORAGE_IMAGE_BIT for B10G11R11_UFLOAT_PACK32 too.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21649>
This commit removes the advertising of features that are currently
unsupported by the driver and aren't strictly necessary for
Vulkan 1.0.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21649>
Adds a winsys feature flag to enable threaded submit.
Currently pvrsrvkm can't support threaded submit as pvrsrvkm syncs don't
support VK_SYNC_FEATURE_WAIT_PENDING.
Signed-off-by: Jarred Davies <jarred.davies@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21577>