feat(draw/sw): add neon support for rgb888 and xrgb8888 #1
Conversation
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
3 issues found across 4 files
Prompt for AI agents (all 3 issues)
Understand the root cause of the following 3 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:116">
This header declares several functions for `AL88` blending (e.g., `lv_draw_sw_blend_neon_al88_to_rgb888`), but these functions are not implemented in the corresponding `lv_draw_sw_blend_neon_to_rgb888.c` file. This creates a broken API contract that will lead to linker errors if other parts of the system attempt to use these declared-but-not-defined functions.</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:587">
The function `lv_draw_sw_blend_neon_rgb565_to_rgb888_with_mask` is advertised as a NEON fast path via macros in `lv_draw_sw_blend_neon_to_rgb888.h`. However, for XRGB8888 targets, it unconditionally returns `LV_RESULT_INVALID` because the implementation is broken. This creates a misleading architectural path where a 'fast path' is called, only to immediately fail and revert to the software fallback, adding unnecessary overhead.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
Mask stride and mask pointer handling bugs; and uninitialized alpha channel writes in RGB565->XRGB8888 paths.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
|
|
||
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
This header declares several functions for AL88 blending (e.g., lv_draw_sw_blend_neon_al88_to_rgb888), but these functions are not implemented in the corresponding lv_draw_sw_blend_neon_to_rgb888.c file. This creates a broken API contract that will lead to linker errors if other parts of the system attempt to use these declared-but-not-defined functions.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 116:
<comment>This header declares several functions for `AL88` blending (e.g., `lv_draw_sw_blend_neon_al88_to_rgb888`), but these functions are not implemented in the corresponding `lv_draw_sw_blend_neon_to_rgb888.c` file. This creates a broken API contract that will lead to linker errors if other parts of the system attempt to use these declared-but-not-defined functions.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
| } | ||
| } | ||
| else if(dest_px_size == 4) { | ||
| /* The algorithm below might look okay but the demo render test fails with it so it's disabled */ |
There was a problem hiding this comment.
The function lv_draw_sw_blend_neon_rgb565_to_rgb888_with_mask is advertised as a NEON fast path via macros in lv_draw_sw_blend_neon_to_rgb888.h. However, for XRGB8888 targets, it unconditionally returns LV_RESULT_INVALID because the implementation is broken. This creates a misleading architectural path where a 'fast path' is called, only to immediately fail and revert to the software fallback, adding unnecessary overhead.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 587:
<comment>The function `lv_draw_sw_blend_neon_rgb565_to_rgb888_with_mask` is advertised as a NEON fast path via macros in `lv_draw_sw_blend_neon_to_rgb888.h`. However, for XRGB8888 targets, it unconditionally returns `LV_RESULT_INVALID` because the implementation is broken. This creates a misleading architectural path where a 'fast path' is called, only to immediately fail and revert to the software fallback, adding unnecessary overhead.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ }
+ }
+ else if(dest_px_size == 4) {
+ /* The algorithm below might look okay but the demo render test fails with it so it's disabled */
+ return LV_RESULT_INVALID;
+#if 0 /* TODO: Figure out the problem with the algorithm below */
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
Mask stride and mask pointer handling bugs; and uninitialized alpha channel writes in RGB565->XRGB8888 paths.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>Mask stride and mask pointer handling bugs; and uninitialized alpha channel writes in RGB565->XRGB8888 paths.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
|
@cubic-dev-local review this |
@sanxroz I've started the AI code review. It'll take a few minutes to complete. |
![cubic-dev-local[bot]](https://avatars.githubusercontent.com/u/189440881?s=60&v=4){kind=small}
There was a problem hiding this comment.
1 issue found across 4 files
Prompt for AI agents (all 1 issues)
Understand the root cause of the following 1 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:78">
single-agent-filter: This PR introduces NEON optimization paths that are intentionally disabled because they are incomplete, buggy, or perform worse than the non-accelerated C code. Merging known-broken or sub-optimal code, even if disabled by preprocessor directives, adds significant technical debt and creates a maintenance burden. Incomplete features should remain in feature branches, not be merged to main.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Confirmed disabled NEON paths via #if 0 and LV_RESULT_INVALID fallbacks, but they are intentionally gated and do not affect runtime. This is a low-impact maintainability concern, not a functional bug; senior developers often keep experimental code disabled for future work. Not worth reporting at medium sensitivity.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| #define LV_DRAW_SW_RGB888_BLEND_NORMAL_TO_RGB888_MIX_MASK_OPA(dsc, dest_px_size, src_px_size) lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask(dsc, dest_px_size, src_px_size) | ||
| #endif | ||
|
|
||
| #if 0 /* These seem to produce worse results than sw rendering, also RGB888 is not implemented, only XRGB8888. So they are disabled for now*/ |
There was a problem hiding this comment.
single-agent-filter: This PR introduces NEON optimization paths that are intentionally disabled because they are incomplete, buggy, or perform worse than the non-accelerated C code. Merging known-broken or sub-optimal code, even if disabled by preprocessor directives, adds significant technical debt and creates a maintenance burden. Incomplete features should remain in feature branches, not be merged to main.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Confirmed disabled NEON paths via #if 0 and LV_RESULT_INVALID fallbacks, but they are intentionally gated and do not affect runtime. This is a low-impact maintainability concern, not a functional bug; senior developers often keep experimental code disabled for future work. Not worth reporting at medium sensitivity.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 78:
<comment>single-agent-filter: This PR introduces NEON optimization paths that are intentionally disabled because they are incomplete, buggy, or perform worse than the non-accelerated C code. Merging known-broken or sub-optimal code, even if disabled by preprocessor directives, adds significant technical debt and creates a maintenance burden. Incomplete features should remain in feature branches, not be merged to main.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Confirmed disabled NEON paths via #if 0 and LV_RESULT_INVALID fallbacks, but they are intentionally gated and do not affect runtime. This is a low-impact maintainability concern, not a functional bug; senior developers often keep experimental code disabled for future work. Not worth reporting at medium sensitivity.</comment>
<file context>
@@ -0,0 +1,154 @@
+#define LV_DRAW_SW_RGB888_BLEND_NORMAL_TO_RGB888_MIX_MASK_OPA(dsc, dest_px_size, src_px_size) lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask(dsc, dest_px_size, src_px_size)
+#endif
+
+#if 0 /* These seem to produce worse results than sw rendering, also RGB888 is not implemented, only XRGB8888. So they are disabled for now*/
+#ifndef LV_DRAW_SW_ARGB8888_BLEND_NORMAL_TO_RGB888
+#define LV_DRAW_SW_ARGB8888_BLEND_NORMAL_TO_RGB888(dsc, dest_px_size) lv_draw_sw_blend_neon_argb888_to_rgb888(dsc, dest_px_size)
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
|
@cubic-dev-local review this |
@sanxroz I've started the AI code review. It'll take a few minutes to complete. |
There was a problem hiding this comment.
5 issues found across 4 files
Prompt for AI agents (all 5 issues)
Understand the root cause of the following 5 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:30">
single-agent-filter: Utility function `drawbuf_next_row()` duplicates an existing implementation. This generic utility function should be centralized to avoid code duplication across blending modules.
• **Libraries consulted**:</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:48">
two-agent-filter: Scalar RGB888 blending logic `lv_color_24_24_mix_1()` duplicates `lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix()`. This logic is not NEON-specific and should be centralized.
• **Libraries consulted**:</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:68">
two-agent-filter: Scalar pre-multiplied RGB888 blending logic `lv_color_24_24_mix_premult()` duplicates `lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix_premult()`. This logic is not NEON-specific and should be centralized.
• **Libraries consulted**:</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
two-agent-filter: The function `lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask` incorrectly uses `dsc->src_stride` for the mask stride instead of `dsc->mask_stride`. This will cause visual artifacts when blending with a mask that has a different stride than the source image, breaking a core feature of the rendering system for this accelerated path.
• **Libraries consulted**:</violation>
<violation number="5" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1299">
two-agent-filter: Uninitialized 4th channel in lv_rgb565_to_xrgb8888_8 leads to undefined bytes written to X/A component.
• **Libraries consulted**:</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
two-agent-filter: The function lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask incorrectly uses dsc->src_stride for the mask stride instead of dsc->mask_stride. This will cause visual artifacts when blending with a mask that has a different stride than the source image, breaking a core feature of the rendering system for this accelerated path.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>two-agent-filter: The function `lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask` incorrectly uses `dsc->src_stride` for the mask stride instead of `dsc->mask_stride`. This will cause visual artifacts when blending with a mask that has a different stride than the source image, breaking a core feature of the rendering system for this accelerated path.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
|
|
||
| static inline uint32x4_t lv_color_24_24_mix_4_premul_internal(uint32x4_t src, uint32x4_t dst, uint32x4_t mix); | ||
| static inline uint32x2_t lv_color_24_24_mix_2_premul_internal(uint32x2_t src, uint32x2_t dst, uint32x2_t mix); | ||
| static inline void lv_color_24_24_mix_premult(const uint8_t * src, uint8_t * dest, uint8_t mix); |
There was a problem hiding this comment.
two-agent-filter: Scalar pre-multiplied RGB888 blending logic lv_color_24_24_mix_premult() duplicates lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix_premult(). This logic is not NEON-specific and should be centralized.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 68:
<comment>two-agent-filter: Scalar pre-multiplied RGB888 blending logic `lv_color_24_24_mix_premult()` duplicates `lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix_premult()`. This logic is not NEON-specific and should be centralized.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+
+static inline uint32x4_t lv_color_24_24_mix_4_premul_internal(uint32x4_t src, uint32x4_t dst, uint32x4_t mix);
+static inline uint32x2_t lv_color_24_24_mix_2_premul_internal(uint32x2_t src, uint32x2_t dst, uint32x2_t mix);
+static inline void lv_color_24_24_mix_premult(const uint8_t * src, uint8_t * dest, uint8_t mix);
+
+static inline uint8x16x3_t rgb_mix_8_internal(uint8x16x3_t src, uint8x16x3_t dst, uint8x16_t mix);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
|
|
||
| static inline uint8x8x3_t rgb565_rgb888_mix_8_internal(uint16x8_t src, uint8x8x3_t dst, uint8x8_t mix); | ||
|
|
||
| static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix); |
There was a problem hiding this comment.
two-agent-filter: Scalar RGB888 blending logic lv_color_24_24_mix_1() duplicates lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix(). This logic is not NEON-specific and should be centralized.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 48:
<comment>two-agent-filter: Scalar RGB888 blending logic `lv_color_24_24_mix_1()` duplicates `lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix()`. This logic is not NEON-specific and should be centralized.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+
+static inline uint8x8x3_t rgb565_rgb888_mix_8_internal(uint16x8_t src, uint8x8x3_t dst, uint8x8_t mix);
+
+static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix);
+
+static inline void rgb565_mix_1(const uint16_t * src, uint8_t * dest, uint8_t mix);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
| * STATIC PROTOTYPES | ||
| **********************/ | ||
|
|
||
| static inline void * LV_ATTRIBUTE_FAST_MEM drawbuf_next_row(const void * buf, uint32_t stride); |
There was a problem hiding this comment.
single-agent-filter: Utility function drawbuf_next_row() duplicates an existing implementation. This generic utility function should be centralized to avoid code duplication across blending modules.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 30:
<comment>single-agent-filter: Utility function `drawbuf_next_row()` duplicates an existing implementation. This generic utility function should be centralized to avoid code duplication across blending modules.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ * STATIC PROTOTYPES
+ **********************/
+
+static inline void * LV_ATTRIBUTE_FAST_MEM drawbuf_next_row(const void * buf, uint32_t stride);
+
+static inline uint32x4_t rgb565_xrgb_mix_4_internal(uint32x4_t src, uint32x4_t dst, uint32x4_t mix);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
| rgb.val[0] = b8; | ||
| rgb.val[1] = g8; | ||
| rgb.val[2] = r8; | ||
| return rgb; |
There was a problem hiding this comment.
two-agent-filter: Uninitialized 4th channel in lv_rgb565_to_xrgb8888_8 leads to undefined bytes written to X/A component.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 1299:
<comment>two-agent-filter: Uninitialized 4th channel in lv_rgb565_to_xrgb8888_8 leads to undefined bytes written to X/A component.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ rgb.val[0] = b8;
+ rgb.val[1] = g8;
+ rgb.val[2] = r8;
+ return rgb;
+}
+static inline uint8x8x4_t lv_rgb565_to_xrgb8888_8(const uint16_t * color)
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
|
@cubic-dev-local review this |
@sanxroz I've started the AI code review. It'll take a few minutes to complete. |
There was a problem hiding this comment.
20 issues found across 4 files
Prompt for AI agents (all 20 issues)
Understand the root cause of the following 20 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:30">
two-agent-filter: Utility function `drawbuf_next_row` duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:drawbuf_next_row(). This common helper should be consolidated.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Trivial one-liner; duplication is negligible and not worth reporting.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:30">
single-agent-filter: Utility function `drawbuf_next_row` duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:drawbuf_next_row(). This common helper should be consolidated.
• **Libraries consulted**:</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:48">
two-agent-filter: Color blending logic (24-bit RGB mix) duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix() function. This low-level helper function should be consolidated.
• **Libraries consulted**:</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:48">
single-agent-filter: Color blending logic (24-bit RGB mix) duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix() function. This low-level helper function should be consolidated.
• **Libraries consulted**:</violation>
<violation number="5" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:50">
two-agent-filter: RGB565 to RGB888 conversion and blending logic duplicates the pattern found in src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:rgb565_image_blend(). The logic should be extracted to a common utility or reuse existing blending primitives.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Partial duplication serving scalar tail for NEON; consolidation requires refactor with limited benefit.</violation>
<violation number="6" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:50">
single-agent-filter: RGB565 to RGB888 conversion and blending logic duplicates the pattern found in src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:rgb565_image_blend(). The logic should be extracted to a common utility or reuse existing blending primitives.
• **Libraries consulted**:</violation>
<violation number="7" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:232">
two-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• **Libraries consulted**:</violation>
<violation number="8" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:232">
two-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• **Libraries consulted**:</violation>
<violation number="9" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:232">
single-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• **Libraries consulted**:</violation>
<violation number="10" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:232">
single-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• **Libraries consulted**:</violation>
<violation number="11" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:458">
two-agent-filter: Fallback path for XRGB8888 does not set the 4th byte (alpha/X), leaving it undefined.
• **Libraries consulted**:</violation>
<violation number="12" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:458">
single-agent-filter: Fallback path for XRGB8888 does not set the 4th byte (alpha/X), leaving it undefined.
• **Libraries consulted**:</violation>
<violation number="13" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
two-agent-filter: Mask stride uses src_stride instead of mask_stride, causing incorrect mask row advancement.
• **Libraries consulted**:</violation>
<violation number="14" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
single-agent-filter: Mask stride uses src_stride instead of mask_stride, causing incorrect mask row advancement.
• **Libraries consulted**:</violation>
<violation number="15" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
two-agent-filter: Wrong stride used for mask buffer advancement (uses src_stride instead of mask_stride) causing misaligned masks and possible OOB reads.
• **Libraries consulted**:</violation>
<violation number="16" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
single-agent-filter: Wrong stride used for mask buffer advancement (uses src_stride instead of mask_stride) causing misaligned masks and possible OOB reads.
• **Libraries consulted**:</violation>
<violation number="17" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1101">
two-agent-filter: Scalar fallback ignores global opa; mix should combine alpha, opa, and mask.
• **Libraries consulted**:</violation>
<violation number="18" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1101">
single-agent-filter: Scalar fallback ignores global opa; mix should combine alpha, opa, and mask.
• **Libraries consulted**:</violation>
<violation number="19" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1299">
two-agent-filter: Alpha/X channel is not set in lv_rgb565_to_xrgb8888_8, leaving the 4th lane uninitialized.
• **Libraries consulted**:</violation>
<violation number="20" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1299">
single-agent-filter: Alpha/X channel is not set in lv_rgb565_to_xrgb8888_8, leaving the 4th lane uninitialized.
• **Libraries consulted**:</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
|
|
||
| static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix); | ||
|
|
||
| static inline void rgb565_mix_1(const uint16_t * src, uint8_t * dest, uint8_t mix); |
There was a problem hiding this comment.
two-agent-filter: RGB565 to RGB888 conversion and blending logic duplicates the pattern found in src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:rgb565_image_blend(). The logic should be extracted to a common utility or reuse existing blending primitives.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Partial duplication serving scalar tail for NEON; consolidation requires refactor with limited benefit.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 50:
<comment>two-agent-filter: RGB565 to RGB888 conversion and blending logic duplicates the pattern found in src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:rgb565_image_blend(). The logic should be extracted to a common utility or reuse existing blending primitives.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Partial duplication serving scalar tail for NEON; consolidation requires refactor with limited benefit.</comment>
<file context>
@@ -0,0 +1,1702 @@
+
+static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix);
+
+static inline void rgb565_mix_1(const uint16_t * src, uint8_t * dest, uint8_t mix);
+
+static inline uint32x4_t argb_rgb_mix_4(const uint32_t * src, const uint32_t * dst);
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
| * STATIC PROTOTYPES | ||
| **********************/ | ||
|
|
||
| static inline void * LV_ATTRIBUTE_FAST_MEM drawbuf_next_row(const void * buf, uint32_t stride); |
There was a problem hiding this comment.
two-agent-filter: Utility function drawbuf_next_row duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:drawbuf_next_row(). This common helper should be consolidated.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Trivial one-liner; duplication is negligible and not worth reporting.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 30:
<comment>two-agent-filter: Utility function `drawbuf_next_row` duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:drawbuf_next_row(). This common helper should be consolidated.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Trivial one-liner; duplication is negligible and not worth reporting.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ * STATIC PROTOTYPES
+ **********************/
+
+static inline void * LV_ATTRIBUTE_FAST_MEM drawbuf_next_row(const void * buf, uint32_t stride);
+
+static inline uint32x4_t rgb565_xrgb_mix_4_internal(uint32x4_t src, uint32x4_t dst, uint32x4_t mix);
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
|
|
||
| static inline uint8x8x3_t rgb565_rgb888_mix_8_internal(uint16x8_t src, uint8x8x3_t dst, uint8x8_t mix); | ||
|
|
||
| static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix); |
There was a problem hiding this comment.
two-agent-filter: Color blending logic (24-bit RGB mix) duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix() function. This low-level helper function should be consolidated.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 48:
<comment>two-agent-filter: Color blending logic (24-bit RGB mix) duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:lv_color_24_24_mix() function. This low-level helper function should be consolidated.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+
+static inline uint8x8x3_t rgb565_rgb888_mix_8_internal(uint16x8_t src, uint8x8x3_t dst, uint8x8_t mix);
+
+static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix);
+
+static inline void rgb565_mix_1(const uint16_t * src, uint8_t * dest, uint8_t mix);
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
|
|
||
| static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix); | ||
|
|
||
| static inline void rgb565_mix_1(const uint16_t * src, uint8_t * dest, uint8_t mix); |
There was a problem hiding this comment.
single-agent-filter: RGB565 to RGB888 conversion and blending logic duplicates the pattern found in src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:rgb565_image_blend(). The logic should be extracted to a common utility or reuse existing blending primitives.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 50:
<comment>single-agent-filter: RGB565 to RGB888 conversion and blending logic duplicates the pattern found in src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:rgb565_image_blend(). The logic should be extracted to a common utility or reuse existing blending primitives.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+
+static inline void lv_color_24_24_mix_1(const uint8_t * src, uint8_t * dest, uint8_t mix);
+
+static inline void rgb565_mix_1(const uint16_t * src, uint8_t * dest, uint8_t mix);
+
+static inline uint32x4_t argb_rgb_mix_4(const uint32_t * src, const uint32_t * dst);
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
| * STATIC PROTOTYPES | ||
| **********************/ | ||
|
|
||
| static inline void * LV_ATTRIBUTE_FAST_MEM drawbuf_next_row(const void * buf, uint32_t stride); |
There was a problem hiding this comment.
single-agent-filter: Utility function drawbuf_next_row duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:drawbuf_next_row(). This common helper should be consolidated.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 30:
<comment>single-agent-filter: Utility function `drawbuf_next_row` duplicates src/draw/sw/blend/lv_draw_sw_blend_to_rgb888.c:drawbuf_next_row(). This common helper should be consolidated.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ * STATIC PROTOTYPES
+ **********************/
+
+static inline void * LV_ATTRIBUTE_FAST_MEM drawbuf_next_row(const void * buf, uint32_t stride);
+
+static inline uint32x4_t rgb565_xrgb_mix_4_internal(uint32x4_t src, uint32x4_t dst, uint32x4_t mix);
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: Unmapped Agent in getAgentNameFromViolationSource (undefined)
| uint8_t * dest_buf_8 = dsc->dest_buf; | ||
| for(int32_t y = 0; y < h; y++) { | ||
| uint8_t * row_ptr = dest_buf_8; | ||
| const uint8_t * mask_row_ptr = mask_buf_8; |
There was a problem hiding this comment.
single-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 232:
<comment>single-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ uint8_t * dest_buf_8 = dsc->dest_buf;
+ for(int32_t y = 0; y < h; y++) {
+ uint8_t * row_ptr = dest_buf_8;
+ const uint8_t * mask_row_ptr = mask_buf_8;
+ int32_t x = 0;
+ for(; x < w - 15; x += 16) {
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: General AI Review Agent
| uint8_t * dest_buf_8 = dsc->dest_buf; | ||
| for(int32_t y = 0; y < h; y++) { | ||
| uint8_t * row_ptr = dest_buf_8; | ||
| const uint8_t * mask_row_ptr = mask_buf_8; |
There was a problem hiding this comment.
single-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 232:
<comment>single-agent-filter: mask_row_ptr is not updated per row, leading to wrong masks on rows after the first.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ uint8_t * dest_buf_8 = dsc->dest_buf;
+ for(int32_t y = 0; y < h; y++) {
+ uint8_t * row_ptr = dest_buf_8;
+ const uint8_t * mask_row_ptr = mask_buf_8;
+ int32_t x = 0;
+ for(; x < w - 15; x += 16) {
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: General AI Review Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
single-agent-filter: Mask stride uses src_stride instead of mask_stride, causing incorrect mask row advancement.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>single-agent-filter: Mask stride uses src_stride instead of mask_stride, causing incorrect mask row advancement.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
two-agent-filter: Wrong stride used for mask buffer advancement (uses src_stride instead of mask_stride) causing misaligned masks and possible OOB reads.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>two-agent-filter: Wrong stride used for mask buffer advancement (uses src_stride instead of mask_stride) causing misaligned masks and possible OOB reads.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
single-agent-filter: Wrong stride used for mask buffer advancement (uses src_stride instead of mask_stride) causing misaligned masks and possible OOB reads.
• Libraries consulted:
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>single-agent-filter: Wrong stride used for mask buffer advancement (uses src_stride instead of mask_stride) causing misaligned masks and possible OOB reads.
• **Libraries consulted**: </comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
|
@cubic-dev-local review this |
@sanxroz I've started the AI code review. It'll take a few minutes to complete. |
There was a problem hiding this comment.
8 issues found across 4 files
Prompt for AI agents (all 8 issues)
Understand the root cause of the following 8 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:232">
mask_row_ptr is initialized outside the row loop and never updated, so all rows use the first mask row instead of the current one.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:458">
Missing write of the 4th byte (X/alpha) in XRGB8888 scalar tail leaves undefined bytes.</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
mask_stride is incorrectly set from src_stride, causing mask rows to be advanced incorrectly and potentially out-of-bounds.</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
Wrong mask stride used (src_stride instead of mask_stride) causing incorrect mask row stepping and potential OOB; and alpha channel is never set in RGB565→XRGB8888 paths (vector and scalar), producing undefined alpha.</violation>
<violation number="5" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1299">
Uninitialized alpha/X channel returned from lv_rgb565_to_xrgb8888_8 leads to undefined 4th byte in XRGB8888 output.</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:116">
Missing AL88→RGB888 macro mappings: AL88 blending to RGB888 won't use the provided NEON implementations and will fall back to LV_RESULT_INVALID instead of NEON paths.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:117">
Inconsistent AL88→RGB888 function prototypes: WITH_OPA/WITH_MASK variants omit dest_px_size unlike other RGB888 targets, risking call-site mismatches.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:135">
Unimplemented ARGB8888→RGB888 prototypes are exposed without corresponding macro wiring, risking unresolved symbols if referenced.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Incorrect: Implementations exist in lv_draw_sw_blend_neon_to_rgb888.c for all ARGB8888→RGB888 functions; macros are disabled via #if 0 and functions safely return LV_RESULT_INVALID for 3-byte dest, so no unresolved symbol risk.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| uint8_t * dest_buf_8 = dsc->dest_buf; | ||
| for(int32_t y = 0; y < h; y++) { | ||
| uint8_t * row_ptr = dest_buf_8; | ||
| const uint8_t * mask_row_ptr = mask_buf_8; |
There was a problem hiding this comment.
mask_row_ptr is initialized outside the row loop and never updated, so all rows use the first mask row instead of the current one.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 232:
<comment>mask_row_ptr is initialized outside the row loop and never updated, so all rows use the first mask row instead of the current one.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ uint8_t * dest_buf_8 = dsc->dest_buf;
+ for(int32_t y = 0; y < h; y++) {
+ uint8_t * row_ptr = dest_buf_8;
+ const uint8_t * mask_row_ptr = mask_buf_8;
+ int32_t x = 0;
+ for(; x < w - 15; x += 16) {
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: General AI Review Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
mask_stride is incorrectly set from src_stride, causing mask rows to be advanced incorrectly and potentially out-of-bounds.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>mask_stride is incorrectly set from src_stride, causing mask rows to be advanced incorrectly and potentially out-of-bounds.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| const int32_t mask_stride = dsc->src_stride; | |
| const int32_t mask_stride = dsc->mask_stride; |
| const uint8_t b = (((src_row[src_x] & 0x001F)) * 2106) >> 8; | ||
| dest_row[dest_x + 0] = b; | ||
| dest_row[dest_x + 1] = g; | ||
| dest_row[dest_x + 2] = r; |
There was a problem hiding this comment.
Missing write of the 4th byte (X/alpha) in XRGB8888 scalar tail leaves undefined bytes.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 458:
<comment>Missing write of the 4th byte (X/alpha) in XRGB8888 scalar tail leaves undefined bytes.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const uint8_t b = (((src_row[src_x] & 0x001F)) * 2106) >> 8;
+ dest_row[dest_x + 0] = b;
+ dest_row[dest_x + 1] = g;
+ dest_row[dest_x + 2] = r;
+ }
+ dest_buf_u8 = drawbuf_next_row(dest_buf_u8, dest_stride);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| rgb.val[0] = b8; | ||
| rgb.val[1] = g8; | ||
| rgb.val[2] = r8; | ||
| return rgb; |
There was a problem hiding this comment.
Uninitialized alpha/X channel returned from lv_rgb565_to_xrgb8888_8 leads to undefined 4th byte in XRGB8888 output.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 1299:
<comment>Uninitialized alpha/X channel returned from lv_rgb565_to_xrgb8888_8 leads to undefined 4th byte in XRGB8888 output.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ rgb.val[0] = b8;
+ rgb.val[1] = g8;
+ rgb.val[2] = r8;
+ return rgb;
+}
+static inline uint8x8x4_t lv_rgb565_to_xrgb8888_8(const uint16_t * color)
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
Wrong mask stride used (src_stride instead of mask_stride) causing incorrect mask row stepping and potential OOB; and alpha channel is never set in RGB565→XRGB8888 paths (vector and scalar), producing undefined alpha.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>Wrong mask stride used (src_stride instead of mask_stride) causing incorrect mask row stepping and potential OOB; and alpha channel is never set in RGB565→XRGB8888 paths (vector and scalar), producing undefined alpha.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
| uint32_t src_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc, | ||
| uint32_t dest_px_size, uint32_t src_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_argb888_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
Unimplemented ARGB8888→RGB888 prototypes are exposed without corresponding macro wiring, risking unresolved symbols if referenced.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Incorrect: Implementations exist in lv_draw_sw_blend_neon_to_rgb888.c for all ARGB8888→RGB888 functions; macros are disabled via #if 0 and functions safely return LV_RESULT_INVALID for 3-byte dest, so no unresolved symbol risk.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 135:
<comment>Unimplemented ARGB8888→RGB888 prototypes are exposed without corresponding macro wiring, risking unresolved symbols if referenced.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Incorrect: Implementations exist in lv_draw_sw_blend_neon_to_rgb888.c for all ARGB8888→RGB888 functions; macros are disabled via #if 0 and functions safely return LV_RESULT_INVALID for 3-byte dest, so no unresolved symbol risk.</comment>
<file context>
@@ -0,0 +1,154 @@
+ uint32_t src_px_size);
+lv_result_t lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc,
+ uint32_t dest_px_size, uint32_t src_px_size);
+lv_result_t lv_draw_sw_blend_neon_argb888_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_argb888_premultiplied_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc,
+ uint32_t dest_px_size);
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: General AI Review Agent
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
Inconsistent AL88→RGB888 function prototypes: WITH_OPA/WITH_MASK variants omit dest_px_size unlike other RGB888 targets, risking call-site mismatches.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 117:
<comment>Inconsistent AL88→RGB888 function prototypes: WITH_OPA/WITH_MASK variants omit dest_px_size unlike other RGB888 targets, risking call-site mismatches.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: General AI Review Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
Missing AL88→RGB888 macro mappings: AL88 blending to RGB888 won't use the provided NEON implementations and will fall back to LV_RESULT_INVALID instead of NEON paths.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 116:
<comment>Missing AL88→RGB888 macro mappings: AL88 blending to RGB888 won't use the provided NEON implementations and will fall back to LV_RESULT_INVALID instead of NEON paths.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
There was a problem hiding this comment.
8 issues found across 4 files
Prompt for AI agents (all 8 issues)
Understand the root cause of the following 8 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:458">
Scalar xrgb8888 path omits writing the 4th byte; the X/alpha channel remains uninitialized, leading to undefined pixel data.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
The `mask_stride` is incorrectly initialized from `dsc->src_stride` instead of `dsc->mask_stride`. This will cause the mask buffer to be read at an incorrect offset for each new row, leading to rendering errors.</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1101">
Scalar fallback uses only alpha and mask, ignoring dsc->opa; blending result diverges from vector path and expected behavior.</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1298">
lv_rgb565_to_xrgb8888_8 returns a 4-channel vector but leaves the 4th (X/alpha) lane uninitialized, which vst4_u8 will store as garbage.</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:117">
AL88 with_opa prototype is missing dest_px_size, inconsistent with other blend variants and likely hindering correct dest stride handling.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:118">
AL88 with_mask prototype is missing dest_px_size; align with other blend variants to ensure proper destination pixel stride handling.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:119">
AL88 with_opa_mask prototype is missing dest_px_size; inconsistent with other blend variants and may break stride/offset calculations.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:135">
Use consistent ARGB8888 naming across macros and prototypes to avoid confusion and potential mismatches.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Stylistic naming only (ARGB8888 in macros vs argb888 in function names). Common pattern elsewhere; no functional impact.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
The mask_stride is incorrectly initialized from dsc->src_stride instead of dsc->mask_stride. This will cause the mask buffer to be read at an incorrect offset for each new row, leading to rendering errors.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>The `mask_stride` is incorrectly initialized from `dsc->src_stride` instead of `dsc->mask_stride`. This will cause the mask buffer to be read at an incorrect offset for each new row, leading to rendering errors.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
| const int32_t mask_stride = dsc->src_stride; | |
| const int32_t mask_stride = dsc->mask_stride; |
| for(; x < w; x++) { | ||
| uint8_t alpha = ((uint8_t *)src_buf_u32)[x * sizeof(uint32_t) + 3]; | ||
| lv_color_24_24_mix_1((uint8_t *)&src_buf_u32[x], (uint8_t *)&dest_buf_u32[x], | ||
| LV_OPA_MIX2(alpha, mask_row_ptr[x])); |
There was a problem hiding this comment.
Scalar fallback uses only alpha and mask, ignoring dsc->opa; blending result diverges from vector path and expected behavior.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 1101:
<comment>Scalar fallback uses only alpha and mask, ignoring dsc->opa; blending result diverges from vector path and expected behavior.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ for(; x < w; x++) {
+ uint8_t alpha = ((uint8_t *)src_buf_u32)[x * sizeof(uint32_t) + 3];
+ lv_color_24_24_mix_1((uint8_t *)&src_buf_u32[x], (uint8_t *)&dest_buf_u32[x],
+ LV_OPA_MIX2(alpha, mask_row_ptr[x]));
+ }
+ src_buf_u32 = drawbuf_next_row(src_buf_u32, src_stride);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| uint8x8x3_t rgb; | ||
| rgb.val[0] = b8; | ||
| rgb.val[1] = g8; | ||
| rgb.val[2] = r8; |
There was a problem hiding this comment.
lv_rgb565_to_xrgb8888_8 returns a 4-channel vector but leaves the 4th (X/alpha) lane uninitialized, which vst4_u8 will store as garbage.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 1298:
<comment>lv_rgb565_to_xrgb8888_8 returns a 4-channel vector but leaves the 4th (X/alpha) lane uninitialized, which vst4_u8 will store as garbage.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ uint8x8x3_t rgb;
+ rgb.val[0] = b8;
+ rgb.val[1] = g8;
+ rgb.val[2] = r8;
+ return rgb;
+}
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| const uint8_t b = (((src_row[src_x] & 0x001F)) * 2106) >> 8; | ||
| dest_row[dest_x + 0] = b; | ||
| dest_row[dest_x + 1] = g; | ||
| dest_row[dest_x + 2] = r; |
There was a problem hiding this comment.
Scalar xrgb8888 path omits writing the 4th byte; the X/alpha channel remains uninitialized, leading to undefined pixel data.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 458:
<comment>Scalar xrgb8888 path omits writing the 4th byte; the X/alpha channel remains uninitialized, leading to undefined pixel data.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const uint8_t b = (((src_row[src_x] & 0x001F)) * 2106) >> 8;
+ dest_row[dest_x + 0] = b;
+ dest_row[dest_x + 1] = g;
+ dest_row[dest_x + 2] = r;
+ }
+ dest_buf_u8 = drawbuf_next_row(dest_buf_u8, dest_stride);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| uint32_t src_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc, | ||
| uint32_t dest_px_size, uint32_t src_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_argb888_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
Use consistent ARGB8888 naming across macros and prototypes to avoid confusion and potential mismatches.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Stylistic naming only (ARGB8888 in macros vs argb888 in function names). Common pattern elsewhere; no functional impact.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 135:
<comment>Use consistent ARGB8888 naming across macros and prototypes to avoid confusion and potential mismatches.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Stylistic naming only (ARGB8888 in macros vs argb888 in function names). Common pattern elsewhere; no functional impact.</comment>
<file context>
@@ -0,0 +1,154 @@
+ uint32_t src_px_size);
+lv_result_t lv_draw_sw_blend_neon_rgb888_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc,
+ uint32_t dest_px_size, uint32_t src_px_size);
+lv_result_t lv_draw_sw_blend_neon_argb888_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_argb888_premultiplied_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc,
+ uint32_t dest_px_size);
</file context>
[internal] Confidence score: 7/10
[internal] Posted by: General AI Review Agent
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
AL88 with_opa_mask prototype is missing dest_px_size; inconsistent with other blend variants and may break stride/offset calculations.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 119:
<comment>AL88 with_opa_mask prototype is missing dest_px_size; inconsistent with other blend variants and may break stride/offset calculations.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+
+lv_result_t lv_draw_sw_blend_neon_rgb565_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
AL88 with_mask prototype is missing dest_px_size; align with other blend variants to ensure proper destination pixel stride handling.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 118:
<comment>AL88 with_mask prototype is missing dest_px_size; align with other blend variants to ensure proper destination pixel stride handling.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
AL88 with_opa prototype is missing dest_px_size, inconsistent with other blend variants and likely hindering correct dest stride handling.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 117:
<comment>AL88 with_opa prototype is missing dest_px_size, inconsistent with other blend variants and likely hindering correct dest stride handling.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
There was a problem hiding this comment.
6 issues found across 4 files
Prompt for AI agents (all 6 issues)
Understand the root cause of the following 6 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
The mask buffer is advanced using `src_stride` instead of `mask_stride`, which can lead to reading incorrect mask data or memory corruption if the source image and mask have different stride values. This breaks the blending abstraction for this specific NEON-accelerated path.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1299">
Uninitialized alpha written in RGB565->XRGB8888: vector path doesn't set A lane; scalar tail doesn't write A byte</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:116">
Prototype declared for AL88→RGB888 NEON path but no implementation or macro wiring exists, leaving the fast path unused and causing dead declarations.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:117">
Prototype declared for AL88→RGB888 with opa but no implementation or macro wiring, so the NEON path cannot be used.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:118">
Prototype declared for AL88→RGB888 with mask but no implementation or macro wiring exists, leaving the NEON fast path unavailable.</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:119">
Prototype declared for AL88→RGB888 mix mask+opa but no implementation or macro wiring, so the NEON path cannot be used.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
The mask buffer is advanced using src_stride instead of mask_stride, which can lead to reading incorrect mask data or memory corruption if the source image and mask have different stride values. This breaks the blending abstraction for this specific NEON-accelerated path.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>The mask buffer is advanced using `src_stride` instead of `mask_stride`, which can lead to reading incorrect mask data or memory corruption if the source image and mask have different stride values. This breaks the blending abstraction for this specific NEON-accelerated path.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
Prototype declared for AL88→RGB888 mix mask+opa but no implementation or macro wiring, so the NEON path cannot be used.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 119:
<comment>Prototype declared for AL88→RGB888 mix mask+opa but no implementation or macro wiring, so the NEON path cannot be used.</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+
+lv_result_t lv_draw_sw_blend_neon_rgb565_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
Prototype declared for AL88→RGB888 with mask but no implementation or macro wiring exists, leaving the NEON fast path unavailable.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 118:
<comment>Prototype declared for AL88→RGB888 with mask but no implementation or macro wiring exists, leaving the NEON fast path unavailable.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
Prototype declared for AL88→RGB888 with opa but no implementation or macro wiring, so the NEON path cannot be used.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 117:
<comment>Prototype declared for AL88→RGB888 with opa but no implementation or macro wiring, so the NEON path cannot be used.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
Prototype declared for AL88→RGB888 NEON path but no implementation or macro wiring exists, leaving the fast path unused and causing dead declarations.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 116:
<comment>Prototype declared for AL88→RGB888 NEON path but no implementation or macro wiring exists, leaving the fast path unused and causing dead declarations.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: General AI Review Agent
| rgb.val[0] = b8; | ||
| rgb.val[1] = g8; | ||
| rgb.val[2] = r8; | ||
| return rgb; |
There was a problem hiding this comment.
Uninitialized alpha written in RGB565->XRGB8888: vector path doesn't set A lane; scalar tail doesn't write A byte
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 1299:
<comment>Uninitialized alpha written in RGB565->XRGB8888: vector path doesn't set A lane; scalar tail doesn't write A byte</comment>
<file context>
@@ -0,0 +1,1702 @@
+ rgb.val[0] = b8;
+ rgb.val[1] = g8;
+ rgb.val[2] = r8;
+ return rgb;
+}
+static inline uint8x8x4_t lv_rgb565_to_xrgb8888_8(const uint16_t * color)
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
There was a problem hiding this comment.
1 issue found across 4 files
Prompt for AI agents (all 1 issues)
Understand the root cause of the following 1 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
Wrong stride used for mask in rgb888→rgb888_with_opa_mask (uses src_stride instead of mask_stride).</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
Wrong stride used for mask in rgb888→rgb888_with_opa_mask (uses src_stride instead of mask_stride).
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>Wrong stride used for mask in rgb888→rgb888_with_opa_mask (uses src_stride instead of mask_stride).</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
There was a problem hiding this comment.
6 issues found across 4 files
Prompt for AI agents (all 6 issues)
Understand the root cause of the following 6 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:116">
`lv_draw_sw_blend_neon_al88_to_rgb888` is declared here but no definition exists anywhere in the tree, so any caller will fail to link. Please either provide the implementation or remove the declaration.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: No definition exists, but there are no usages or macro mappings to this symbol. Without callers, it causes no current link/runtime impact. Low priority; filter out.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:117">
`lv_draw_sw_blend_neon_al88_to_rgb888_with_opa` is declared without any definition in the codebase, which will break linking when used. Please add the implementation or drop the declaration.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Declared without definition and unused across the repo. No current build impact; minor maintainability issue. Filter out.</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:117">
The function signatures for NEON AL88 blending with opacity and/or masks are incorrect. They are missing the `dest_px_size` parameter, which is passed from the call site in `lv_draw_sw_blend_to_rgb888.c`, creating a mismatch between the calling code and the function implementation.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</violation>
<violation number="4" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:118">
`lv_draw_sw_blend_neon_al88_to_rgb888_with_mask` is declared here but never defined, so any caller will end up with an unresolved symbol at link time. Please either supply the implementation or remove the prototype.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Only declared, no implementation, and not referenced. No present effect on linking or functionality. Filter out.</violation>
<violation number="5" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:119">
`lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask` is declared without any definition in the project, so using it will produce an undefined reference at link time. Please add the implementation or remove the declaration.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Missing definition but unused throughout the codebase, so it doesn’t break linking. Low impact. Filter out.</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
Mask stride typo: uses src_stride instead of mask_stride in rgb888→rgb888 with opa+mask.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask is declared without any definition in the project, so using it will produce an undefined reference at link time. Please add the implementation or remove the declaration.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Missing definition but unused throughout the codebase, so it doesn’t break linking. Low impact. Filter out.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 119:
<comment>`lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask` is declared without any definition in the project, so using it will produce an undefined reference at link time. Please add the implementation or remove the declaration.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Missing definition but unused throughout the codebase, so it doesn’t break linking. Low impact. Filter out.</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+
+lv_result_t lv_draw_sw_blend_neon_rgb565_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
</file context>
[internal] Confidence score: 7/10
[internal] Posted by: General AI Review Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
lv_draw_sw_blend_neon_al88_to_rgb888_with_mask is declared here but never defined, so any caller will end up with an unresolved symbol at link time. Please either supply the implementation or remove the prototype.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Only declared, no implementation, and not referenced. No present effect on linking or functionality. Filter out.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 118:
<comment>`lv_draw_sw_blend_neon_al88_to_rgb888_with_mask` is declared here but never defined, so any caller will end up with an unresolved symbol at link time. Please either supply the implementation or remove the prototype.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Only declared, no implementation, and not referenced. No present effect on linking or functionality. Filter out.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+
</file context>
[internal] Confidence score: 7/10
[internal] Posted by: General AI Review Agent
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
lv_draw_sw_blend_neon_al88_to_rgb888_with_opa is declared without any definition in the codebase, which will break linking when used. Please add the implementation or drop the declaration.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Declared without definition and unused across the repo. No current build impact; minor maintainability issue. Filter out.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 117:
<comment>`lv_draw_sw_blend_neon_al88_to_rgb888_with_opa` is declared without any definition in the codebase, which will break linking when used. Please add the implementation or drop the declaration.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Declared without definition and unused across the repo. No current build impact; minor maintainability issue. Filter out.</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 7/10
[internal] Posted by: General AI Review Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
lv_draw_sw_blend_neon_al88_to_rgb888 is declared here but no definition exists anywhere in the tree, so any caller will fail to link. Please either provide the implementation or remove the declaration.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: No definition exists, but there are no usages or macro mappings to this symbol. Without callers, it causes no current link/runtime impact. Low priority; filter out.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 116:
<comment>`lv_draw_sw_blend_neon_al88_to_rgb888` is declared here but no definition exists anywhere in the tree, so any caller will fail to link. Please either provide the implementation or remove the declaration.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: No definition exists, but there are no usages or macro mappings to this symbol. Without callers, it causes no current link/runtime impact. Low priority; filter out.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 7/10
[internal] Posted by: General AI Review Agent
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc); |
There was a problem hiding this comment.
The function signatures for NEON AL88 blending with opacity and/or masks are incorrect. They are missing the dest_px_size parameter, which is passed from the call site in lv_draw_sw_blend_to_rgb888.c, creating a mismatch between the calling code and the function implementation.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 117:
<comment>The function signatures for NEON AL88 blending with opacity and/or masks are incorrect. They are missing the `dest_px_size` parameter, which is passed from the call site in `lv_draw_sw_blend_to_rgb888.c`, creating a mismatch between the calling code and the function implementation.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*</comment>
<file context>
@@ -0,0 +1,154 @@
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
Mask stride typo: uses src_stride instead of mask_stride in rgb888→rgb888 with opa+mask.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>Mask stride typo: uses src_stride instead of mask_stride in rgb888→rgb888 with opa+mask.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: Functional Bugs Agent
There was a problem hiding this comment.
4 issues found across 4 files
Prompt for AI agents (all 4 issues)
Understand the root cause of the following 4 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:96">
Potential out-of-bounds write/read: missing validation that stride >= w * pixel_size (and mask_stride >= w) across functions (CWE-787/CWE-125).
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Omitting stride validation here aligns with established LVGL patterns (non-NEON paths also rely on valid descriptors and don’t check stride). Without evidence that descriptors are attacker-controlled or inconsistent at call sites, labeling this as CWE-787/125 is speculative and low-value for this PR. Filter to avoid a false positive.</violation>
<violation number="2" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
Mask stride incorrectly set to src_stride; causes mask row misalignment in rgb888_with_opa_mask.</violation>
<violation number="3" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:1582">
4-byte blend paths zero alpha for partial mixes; XRGB alpha should be preserved or set to 0xFF.</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:116">
These AL88 NEON blend functions are never selected—there is no `LV_DRAW_SW_AL88_BLEND_*_TO_RGB888` macro in this header, so the dispatcher keeps calling the non-NEON path and the new SIMD code is dead. Please add the missing macro overrides (as done for the RGB565 header) so these functions become reachable.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| uint8_t * dest_buf_8 = dsc->dest_buf; | ||
|
|
||
| if(dest_px_size == 3) { | ||
| const int32_t byte_w = w * 3; |
There was a problem hiding this comment.
Potential out-of-bounds write/read: missing validation that stride >= w * pixel_size (and mask_stride >= w) across functions (CWE-787/CWE-125).
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: Omitting stride validation here aligns with established LVGL patterns (non-NEON paths also rely on valid descriptors and don’t check stride). Without evidence that descriptors are attacker-controlled or inconsistent at call sites, labeling this as CWE-787/125 is speculative and low-value for this PR. Filter to avoid a false positive.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 96:
<comment>Potential out-of-bounds write/read: missing validation that stride >= w * pixel_size (and mask_stride >= w) across functions (CWE-787/CWE-125).
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: Omitting stride validation here aligns with established LVGL patterns (non-NEON paths also rely on valid descriptors and don’t check stride). Without evidence that descriptors are attacker-controlled or inconsistent at call sites, labeling this as CWE-787/125 is speculative and low-value for this PR. Filter to avoid a false positive.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ uint8_t * dest_buf_8 = dsc->dest_buf;
+
+ if(dest_px_size == 3) {
+ const int32_t byte_w = w * 3;
+ const uint8x16x3_t vec_16 = lv_color_to_rgb888_16(&dsc->color);
+ const uint8x8x3_t vec_8 = lv_color_to_rgb888_8(&dsc->color);
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: Security Vulnerabilities Agent
| uint32x4_t blend_g = vshrq_n_u32(vaddq_u32(vmulq_u32(src_g, mix), vmulq_u32(dst_g, inv_mix)), 8); | ||
| uint32x4_t blend_b = vshrq_n_u32(vaddq_u32(vmulq_u32(src_b, mix), vmulq_u32(dst_b, inv_mix)), 8); | ||
|
|
||
| uint32x4_t blended = vorrq_u32(vorrq_u32(vshlq_n_u32(blend_r, 16), vshlq_n_u32(blend_g, 8)), blend_b); |
There was a problem hiding this comment.
4-byte blend paths zero alpha for partial mixes; XRGB alpha should be preserved or set to 0xFF.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 1582:
<comment>4-byte blend paths zero alpha for partial mixes; XRGB alpha should be preserved or set to 0xFF.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ uint32x4_t blend_g = vshrq_n_u32(vaddq_u32(vmulq_u32(src_g, mix), vmulq_u32(dst_g, inv_mix)), 8);
+ uint32x4_t blend_b = vshrq_n_u32(vaddq_u32(vmulq_u32(src_b, mix), vmulq_u32(dst_b, inv_mix)), 8);
+
+ uint32x4_t blended = vorrq_u32(vorrq_u32(vshlq_n_u32(blend_r, 16), vshlq_n_u32(blend_g, 8)), blend_b);
+ uint32x4_t result = vbslq_u32(zero_mask, dst, blended);
+ return vbslq_u32(full_mask, src, result);
</file context>
[internal] Confidence score: 8/10
[internal] Posted by: Functional Bugs Agent
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
Mask stride incorrectly set to src_stride; causes mask row misalignment in rgb888_with_opa_mask.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>Mask stride incorrectly set to src_stride; causes mask row misalignment in rgb888_with_opa_mask.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
|
|
||
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
These AL88 NEON blend functions are never selected—there is no LV_DRAW_SW_AL88_BLEND_*_TO_RGB888 macro in this header, so the dispatcher keeps calling the non-NEON path and the new SIMD code is dead. Please add the missing macro overrides (as done for the RGB565 header) so these functions become reachable.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 116:
<comment>These AL88 NEON blend functions are never selected—there is no `LV_DRAW_SW_AL88_BLEND_*_TO_RGB888` macro in this header, so the dispatcher keeps calling the non-NEON path and the new SIMD code is dead. Please add the missing macro overrides (as done for the RGB565 header) so these functions become reachable.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 7/10
[internal] Posted by: General AI Review Agent
There was a problem hiding this comment.
2 issues found across 4 files
Prompt for AI agents (all 2 issues)
Understand the root cause of the following 2 issues and fix them.
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c:904">
Incorrect stride used for mask buffer, leading to memory access errors. The mask stride should be `dsc->mask_stride`, but `dsc->src_stride` is used instead, breaking the established pattern and the data structure's contract.</violation>
</file>
<file name="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h">
<violation number="1" location="src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h:116">
Declares lv_draw_sw_blend_neon_al88_to_rgb888* functions without any implementation, leading to unresolved symbols when the NEON path tries to use them.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: The header does declare several lv_draw_sw_blend_neon_al88_to_rgb888* functions without providing implementations, but no code in the repository references these symbols. The associated LV_DRAW_SW_AL88_* macros remain defined to LV_RESULT_INVALID fallbacks, so the NEON path never attempts to call the undeclared functions and no unresolved symbols occur. Without actual usages, the declarations do not cause build or linkage failures, making the reported violation inaccurate.</violation>
</file>
React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.
| const int32_t h = dsc->dest_h; | ||
| const int32_t dest_stride = dsc->dest_stride; | ||
| const int32_t src_stride = dsc->src_stride; | ||
| const int32_t mask_stride = dsc->src_stride; |
There was a problem hiding this comment.
Incorrect stride used for mask buffer, leading to memory access errors. The mask stride should be dsc->mask_stride, but dsc->src_stride is used instead, breaking the established pattern and the data structure's contract.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.c at line 904:
<comment>Incorrect stride used for mask buffer, leading to memory access errors. The mask stride should be `dsc->mask_stride`, but `dsc->src_stride` is used instead, breaking the established pattern and the data structure's contract.</comment>
<file context>
@@ -0,0 +1,1702 @@
+ const int32_t h = dsc->dest_h;
+ const int32_t dest_stride = dsc->dest_stride;
+ const int32_t src_stride = dsc->src_stride;
+ const int32_t mask_stride = dsc->src_stride;
+ const uint8_t opa = dsc->opa;
+ const uint8_t * mask_buf_8 = dsc->mask_buf;
</file context>
[internal] Confidence score: 10/10
[internal] Posted by: System Design Agent
| const int32_t mask_stride = dsc->src_stride; | |
| const int32_t mask_stride = dsc->mask_stride; |
|
|
||
| lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); | ||
|
|
||
| lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size); |
There was a problem hiding this comment.
Declares lv_draw_sw_blend_neon_al88_to_rgb888* functions without any implementation, leading to unresolved symbols when the NEON path tries to use them.
DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• GPT-5: The header does declare several lv_draw_sw_blend_neon_al88_to_rgb888* functions without providing implementations, but no code in the repository references these symbols. The associated LV_DRAW_SW_AL88_* macros remain defined to LV_RESULT_INVALID fallbacks, so the NEON path never attempts to call the undeclared functions and no unresolved symbols occur. Without actual usages, the declarations do not cause build or linkage failures, making the reported violation inaccurate.
Prompt for AI agents
Address the following comment on src/draw/sw/blend/neon/lv_draw_sw_blend_neon_to_rgb888.h at line 116:
<comment>Declares lv_draw_sw_blend_neon_al88_to_rgb888* functions without any implementation, leading to unresolved symbols when the NEON path tries to use them.
*DEV MODE: This violation would have been filtered out by screening filters. Failing filters: compilerGuaranteedChecks.*
DEV MODE: This violation would have been filtered out by GPT-5.
Reasoning:
• **GPT-5**: The header does declare several lv_draw_sw_blend_neon_al88_to_rgb888* functions without providing implementations, but no code in the repository references these symbols. The associated LV_DRAW_SW_AL88_* macros remain defined to LV_RESULT_INVALID fallbacks, so the NEON path never attempts to call the undeclared functions and no unresolved symbols occur. Without actual usages, the declarations do not cause build or linkage failures, making the reported violation inaccurate.</comment>
<file context>
@@ -0,0 +1,154 @@
+
+lv_result_t lv_draw_sw_blend_neon_l8_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888(lv_draw_sw_blend_image_dsc_t * dsc, uint32_t dest_px_size);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_opa(lv_draw_sw_blend_image_dsc_t * dsc);
+lv_result_t lv_draw_sw_blend_neon_al88_to_rgb888_with_mask(lv_draw_sw_blend_image_dsc_t * dsc);
</file context>
[internal] Confidence score: 9/10
[internal] Posted by: Functional Bugs Agent
RGB888 Overall 29% faster
XRGB8888 Overall 33% faster
Not a neon expert and the code could be better but performance improvements are here and code can be made better
I will run more benchmarks without the ARGB8888 implementation as it seems to produce worst results, also I didn't implement ARGB8888->RGB888 using neon code as that was a bit messy (basically I can load 2 or 4 pixels of ARGB8888 but can only load 8 or 16 pixels of RGB888 (properly). ARGB8888->XRGB8888 is implemented tho
cc @GorgonMeducer @terry0012
Implementation was tested against the SW renderer using the render demo and both implementations produce the same results
Summary by cubic
Add ARM NEON fast paths for RGB888 and XRGB8888 in the software renderer. Improves fills and blits by ~29% (RGB888) and ~33% (XRGB8888) on a Raspberry Pi 3B+.
New Features
Bug Fixes