From the API's point of view, border color replacement looks like this:
--------------------
| API Border Color |
--------------------
|
----------- | ---------------- ----------
| API fmt |-------->| User Swizzle |----->| Shader |
----------- ---------------- ----------
From the HW point of view, it looks like this:
-------------------
| HW Border Color |
-------------------
|
---------- ----------- | --------------- ----------
| HW fmt |-----| HW swap |------>| Tex Swizzle |----->| Shader |
---------- ----------- --------------- ----------
When the HW fmt + HW swap isn't enough to represent an API format, we
need to prepend our own swizzle to the the user's swizzle, so the tex
swizzle is a "format swizzle" composed with the user swizzle. However,
we don't want this format swizzle to be applied to border colors, so
there's a workaround in freedreno which is meant to undo the format
swizzle so that the HW border color with the format swizzle applied
equals the API border color, and everything is ok. However, on a6xx at
least it was incorrectly undoing the entire tex swizzle. This broke
border color with a user swizzle, because it was now effectively not
getting applied for the border color. It also made it seem like the user
swizzle is required for the workaround, which would have implications
for VK_EXT_border_color_swizzle with turnip, but it's not.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17177>
virgl: Also drop old pre-trim glxgears trace (cached).
Acked-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Acked-by: Emma Anholt <emma@anholt.net>
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17280>
Early depth test is broken when the linear render target mode is used
and depth is written from the PE stage. It seems RA and PE disagree
about the cache layout in that case, so the RA sees unwritten/invalid
depth cache lines leading to random depth test fails. Early test works
fine if depth is written from the RA stage.
To work around this issue, detect the combination of linear RT, early
test and late write and switch to late test in that case.
Fixes: 53445284a4 ("etnaviv: add linear PE support")
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17215>
On-GPU LRZ direction tracking allows LRZ to support secondary cmdbufs,
reusing LRZ between renderpasses, and in future to support LRZ when
VK_KHR_dynamic_rendering is used.
With on-gpu tracking we have to be careful keeping LRZ state in sync
with underlying depth image, which means we should invalidate LRZ
when underlying image is changed or the view of image is different
from previous renderpass.
All of this resulted in LRZ logic being thinly spread through the code,
making it hard to understand. So most of it was moved to tu_lrz.c.
For more details on past and new LRZ features see comment at the
top of tu_lrz.c.
Note about blob:
- Blob is much more happy to do LRZ_FLUSH, it flushes at the start
of the renderpass, after binning, and at the end of the renderpass.
- Blob seem not to care about changes in depth image done via
vkCmdCopyImage.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6347
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16251>
We are already caching DXIL shaders individually, but that forces us
to retrieve the NIR shader, do the linking and binding translation
steps, to finally query the cache for each DXIL shader. This pipeline
caching is about skipping those steps when we can.
Note that a graphics/compute pipeline cache entry doesn't contain the
DXIL shaders, but hashes to retrieve those shaders from the same cache.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
Binding translation has an impact on the final DXIL shader, and this
binding translation depends on the pipeline layout. Let's extend the
adjust_var_bindings() pass to caculate a hash we can then use in the
DXIL shader hash caculation.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
Basically what vk_pipeline_cache's raw_data_object abstraction provides,
but I'm not sure it makes sense to make it generic so I copied it here.
Might be more appropriate to make raw_data_object public.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
DXIL shaders depend on the vulkan -> d3d12 binding translation done in
adjust_var_bindings(). In order to maximize our chances to re-use those
shaders, we need to hash the binding translation information and take
this hash into account when computing the DXIL shader hash.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
We will need to hash var bindings if we want to cache DXIL shaders.
Let's move this pass to dzn_pipeline.c to prepare this transition.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
We leak the nir_shader object f a failure happens between the NIR
shader creation and the DXIL compilation. Let's drop the local
nir_shader pointer and use the one in the graphics_pipeline object
to avoid that.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
This way we can use d3d12_pipeline_state_stream_new_desc() directly
without doing
if (type == GRAPHICS)
d3d12_gfx_pipeline_state_stream_new_desc()
else
d3d12_compute_pipeline_state_stream_new_desc()
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17140>
In order to get rid of the ntracediff.sh wrapper script, implement
invocation of 'less' internally, if the stdout is determined to
be a tty. Otherwise just print out normally.
Signed-off-by: Matti Hamalainen <ccr@tnsp.org>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17135>
Instead of discarding them at parsing phase, let the difflib
SequenceMatcher always ignore them, and optionally suppress
them from output if -I/--ignore-junk option is given.
Signed-off-by: Matti Hamalainen <ccr@tnsp.org>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17135>
Basically implement the last item on the original feature request list of #4609.
Example: ./pytracediff.py good.xml bad.xml -NM
Or suppress common calls completely via -C, e.g. -NC etc.
Signed-off-by: Matti Hamalainen <ccr@tnsp.org>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17135>
Highlight differing _lines_ in the differing blocks, with somewhat
different ANSI colors.
Signed-off-by: Matti Hamalainen <ccr@tnsp.org>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17135>
The limitations of current approach for Gallium trace diffing via
'tracediff.sh' are becoming all too apparent, as we are first dumping
both trace to text and performing plain line-based sdiff on them.
This obviously loses the context of calls and sometimes results in
confusing diffs when subsequent calls are similar enough. It also
prevents us from formatting the diff output in ways that would
benefit readability.
In attempt to rectify the situation, reimplement the diffing completely
in Python, using difflib and adding the necessary plumbing into the trace
model objects etc.
Signed-off-by: Matti Hamalainen <ccr@tnsp.org>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17135>
Previously tracediff.sh used postprocessing sed-script to remove unwanted
calls from the dump output. Instead of that, add option to parse.py to
ignore a list of calls at parsing phase. Currently this list is hardcoded
in parse.py.
Also clean up the trace model code and pointer tracking a bit to avoid
static state in Pointer class.
Signed-off-by: Matti Hamalainen <ccr@tnsp.org>
Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17135>
This was breaking trace driver serialization, since it relies upon the
transfer box to know what to serialize.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17107>
Ensure the argument name is serialized "state" not, "&tr_ctx->unwrapped_state".
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17107>
Provide a vk_device::check_status() implementation so the code can
call it at key moments to detect when a device-lost event (device-removed
in D3D12) has been received.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17249>
