This gets rid of our (fairly sketchy) heap resizing via stall-and-copy
and replaces it with VM_BIND. We couldn't do this on the old nouveau
API but now that we can assume VM_BIND, it makes everything simpler.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27205>
The queue now owns the nv_push and just invokes the per-engine functions
to fill it with context state init data. This also splits out 3D and
compute into separate helpers and pulls M2MF off into its own thing.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27205>
Also, rework the fail gotos so they make a little more sense. Since we
were already calling dealloc the first time a subchan fails, we may as
well just always use dealloc. Worst case the ioctl fails.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27205>
GL ARB_sparse_buffer allows unbound regions in buffers.
VK sparseBinding insists all regions must be bound before first use.
This means we need to use sparseResidencyBuffer to back GL
sparse buffers to get the same semantics.
Fixes GL and piglit sparse buffer tests on zink/nvk.
Fixes: c90246b682 ("zink: implement sparse buffer creation/mapping")
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27404>
In clear_depth_stencil() stencil_surf is defined but not initiaized.
Then in the same function if stencil_mask is calculated and if != 0
stencil_surf is initialized.
But blorp_clear_stencil_as_rgba() access stencil_surf before checking
stencil_mask, what could cause a read of a uninitialized valued.
clear_depth_stencil()
struct blorp_surf stencil_surf;
...
uint8_t stencil_mask = clear_stencil && stencil_res ? 0xff : 0;
if (stencil_mask) {
...
iris_blorp_surf_for_resource(&stencil_surf);
}
...
blorp_clear_depth_stencil(stencil_mask, stencil_surf)
blorp_clear_stencil_as_rgba(stencil_mask, stencil)
if (surf->surf->format ...)
....
Just inverting the order and checking stencil_mask first in
blorp_clear_stencil_as_rgba() fixes the issue.
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Dylan Baker <dylan.c.baker@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27390>
This is required to allow gallium drivers to display instead of
forcing software display. On llvmpipe/softpipe for sw display
gdi_sw_winsys will be used. While "hardware" drivers will be
able to call pfnPresentCb to present without CPU copy.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jose Fonseca <jose.fonseca@broadcom.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27333>
In d3d10umd it is required to pass D3DKMT_PRESENT in
context_private of flush_frontbuffer, but wgl passes HDC in it.
To accomidate this when gdi_sw_winsys is created functions that
acquire and release hdc from context_private are passed.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Feng Jiang <jiangfeng@kylinos.cn>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27333>
If a resource is created with sharing mode CONCURRENT when multiple
queues are supported, we can't support the compression since we can't do
FULL_RESOLVE/PARTIAL_RESOLVE to construct the main surface data without
barriers.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27349>
The offset_B member of levels[0] should always be 0. In this case,
we're copying the level from the previous image but we've already taken
the offset into account via the offset_B_out parameter. Leaving the
offset non-zero risks it getting added in twice if nil_image_for_level
gets called again. In particular, NVK can hit this with 3D compressed
textures with VK_IMAGE_CREATE_BLOCK_TEXEL_VIEW_COMPATIBLE_BIT set.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26990>
