You wouldn't think the TileWalk mode matters when TiledSurface is
false. However, it has to be TILEWALK_XMAJOR. Make it so.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
With all the previous commits in place, we can now drop in support for
multiple platforms. First up is gen7 (Ivybridge).
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
This is a more logical place for it, between geometry front end state
and pixel backend state.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
This commit moves all occurances of gen8 specific state into a gen8
substruct. This clearly identifies the state as gen8 specific and
prepares for adding gen7 state structs. In the process we also rename
the field names to exactly match the command or state packet name,
without the 3DSTATE prefix, eg:
3DSTATE_VF -> gen8.vf
3DSTATE_WM_DEPTH_STENCIL -> gen8.wm_depth_stencil
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
The clear pipeline didn't have a vertex shader and relied on the clear
shader being hardcoded by the compiler to accept one attribute. This
necessitated a few special cases in the 3DSTATE_VS setup. Instead,
always provide a vertex shader, even if we disable VS dispatch.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Many of of these fields aren't used for buffer surfaces, so leave them
out for brevity.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
This adds VALIGN_2 and VALIGN_4 defines for IVB and HSW
RENDER_SURFACE_STATE.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
I'd prefer to move anv_CreateAttachmentView() as well, but it's a little
too much generic code to just duplicate for each gen. For now, we'll
add a anv_color_attachment_view_init() to dispatch to the gen specific
implementation, which we then call from anv_CreateAttachmentView().
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
We'll probably want to move some code back into a shared init function,
but this gets one GEN8 surface state initialization out of anv_image.c.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
We need this for generating surface state on the fly for dynamic buffer
views.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
We're going to have to do this differently for earlier gens, so lets do
it in place only.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
Since the extra dword in MI_BATCH_BUFFER_START added in gen8 is at the
end of the struct, we can emit the gen8 packet on all gens as long as we
set the instruction length correctly.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
We used to use a manual GEN8_MI_BATCH_BUFFER_START_pack() call, but this
refactors the code to use anv_batch_emit();
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
We'll organize gen specific code in three files per gen: pipeline,
cmd_buffer and state, eg:
gen8_cmd_buffer.c
gen8_pipeline.c
gen8_state.c
where gen8_cmd_buffer.c holds all vkCmd* entry points, gne8_pipeline.c
all gen specific code related to pipeline building and remaining state
code (sampler, surface state, dynamic state) in gen8_state.c.
Signed-off-by: Kristian Høgsberg Kristensen <kristian.h.kristensen@intel.com>
During vkCreateRenderPass, count the number of clear ops and store them
in new members of anv_render_pass:
uint32_t num_color_clear_attachments
bool has_depth_clear_attachment
bool has_stencil_clear_attachment
Cacheing these 8 bytes (including padding) reduces the number of times
that anv_cmd_buffer_clear_attachments needs to loop over the pass's
attachments.
In struct anv_saved_state, each member's type was a pointer to an Anvil
struct and each member's name was prefixed with "old" except cb_state,
which was a Vulkan handle whose name lacked "old".
The source image's format was incorrectly used for both the source view
and destination view. For vkCmdCopyImage to correctly translate formats,
the destination view's format must be that of the destination image's.
Change type of anv_render_pass_attachment::format from VkFormat to const
struct anv_format*. This elimiates the repetitive lookups into the
VkFormat -> anv_format table when looping over attachments during
anv_cmd_buffer_clear_attachments().
Stop creating a temporary VkImageCreateInfo with overriden
format=VK_FORMAT_S8_UINT. Instead, just pass the format override
directly to anv_image_make_surface().
Stencil formats are often a special case. To reduce the number of lookups
into the VkFormat-to-anv_format translation table when working with
stencil, expose the table's entry for VK_FORMAT_S8_UINT as global
variable anv_format_s8_uint.
Change type of anv_surface_view::format from VkFormat to const struct
anv_format*. This reduces the number of lookups in the VkFormat ->
anv_format table.
This moves the translation of VkFormat to anv_format from
anv_fill_buffer_surface_state() to its caller.
A prep commit to reduce more VkFormat -> anv_format translations.
Store the original VkFormat as anv_format::vk_format. This will be used
to reduce format indirection, such as lookups into the VkFormat ->
anv_format translation table.
Kristian Høgsberg Kristensen