[PATCH] drm/amd/display: Initialize missing TD-LUT programming inputs

[PATCH] drm/amd/display: Initialize missing TD-LUT programming inputs

[Avi Weiss]

dml_core_mode_programming() memsets (clears) its scratch structure
before calculating TD-LUT settings, but does not initialize
tdlut_mpc_width_flag or is_gfx11.

Both inputs therefore remain false regardless of the plane
configuration which can result in incorrect calculations.

Initialize both fields using the same sources as the mode
support path (now both paths are aligned).

Fixes: 70839da63605 ("drm/amd/display: Add new DCN401 sources")
Signed-off-by: Avi Weiss <thnkslprpt@gmail.com>
---
 .../dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c        | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
index 72acf12e9251..0ee385e8ba5f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -11074,6 +11074,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
 		calculate_tdlut_setting_params->cursor_buffer_size = mode_lib->ip.cursor_buffer_size;
 		calculate_tdlut_setting_params->gpuvm_enable = display_cfg->gpuvm_enable;
 		calculate_tdlut_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
+		calculate_tdlut_setting_params->tdlut_mpc_width_flag = display_cfg->plane_descriptors[k].tdlut.tdlut_mpc_width_flag;
+		calculate_tdlut_setting_params->is_gfx11 = dml_get_gfx_version(display_cfg->plane_descriptors[k].surface.tiling);
 
 		// output
 		calculate_tdlut_setting_params->tdlut_pte_bytes_per_frame = &s->tdlut_pte_bytes_per_frame[k];
-- 
2.43.0

Re: [PATCH] drm/amd/display: Initialize missing TD-LUT programming inputs

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] amd/display: Implicit cast of unsigned int to bool in is_gfx11 breaks GFX12
--

commit 6cadd8de043129ed240d5c445ed5c65b848832cc
Author: Avi Weiss <thnkslprpt@gmail.com>

drm/amd/display: Initialize missing TD-LUT programming inputs

This commit aligns the mode programming path with the mode support path by
explicitly initializing tdlut_mpc_width_flag and is_gfx11 inputs in
dml_core_mode_programming() for DCN401 sources.

Fixes: 70839da63605 ("drm/amd/display: Add new DCN401 sources")

> diff --git a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
> --- a/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
> +++ b/drivers/gpu/drm/amd/display/dc/dml2_0/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
> @@ -11052,6 +11052,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
>  		calculate_tdlut_setting_params->cursor_buffer_size = mode_lib->ip.cursor_buffer_size;
>  		calculate_tdlut_setting_params->gpuvm_enable = display_cfg->gpuvm_enable;
>  		calculate_tdlut_setting_params->gpuvm_page_size_kbytes = display_cfg->plane_descriptors[k].overrides.gpuvm_min_page_size_kbytes;
> +		calculate_tdlut_setting_params->tdlut_mpc_width_flag = display_cfg->plane_descriptors[k].tdlut.tdlut_mpc_width_flag;
> +		calculate_tdlut_setting_params->is_gfx11 = dml_get_gfx_version(display_cfg->plane_descriptors[k].surface.tiling);

[Severity: High]
This problem wasn't introduced by this patch, but it appears this copies an
existing bug from dml_core_mode_support() into dml_core_mode_programming().

Looking at dml_get_gfx_version(), it returns an unsigned int 12 for GFX12
swizzle modes. When this result is assigned directly to the boolean field
is_gfx11, it will evaluate to true because any non-zero integer is true in C.

Could this cause GFX12 hardware to incorrectly use GFX11 TD-LUT pitch
alignment (like 256B instead of 128B), leading to display corruption?

Would it be better to use an explicit equality check here and in the mode
support path, such as:
is_gfx11 = (dml_get_gfx_version(...) == 11);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261011074510.2495683-1-thnkslprpt@gmail.com?part=1