drm/format-helper: Rework format-helper conversion functions

Move destination-buffer clipping from all format-helper conversion
functions into callers. Support destination-buffer pitch. Only
distinguish between system and I/O memory, but use same logic
everywhere.

Simply harmonize the interface and semantics of the existing code.
Not all conversion helpers support all combinations of parameters.
We have to add additional features when we need them.

v2:
	* fix default destination pitch in drm_fb_xrgb8888_to_gray8()
	  (Noralf)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Tested-by: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Noralf Trønnes <noralf@tronnes.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20211110103702.374-5-tzimmermann@suse.de

drivers/gpu/drm/drm_format_helper.c

@@ -165,7 +165,7 @@ void drm_fb_swab(void *dst, unsigned int dst_pitch, const void *src,
 }
 EXPORT_SYMBOL(drm_fb_swab);
 
-static void drm_fb_xrgb8888_to_rgb332_line(u8 *dbuf, __le32 *sbuf, unsigned int pixels)
+static void drm_fb_xrgb8888_to_rgb332_line(u8 *dbuf, const __le32 *sbuf, unsigned int pixels)
 {
 	unsigned int x;
 	u32 pix;
@@ -181,23 +181,24 @@ static void drm_fb_xrgb8888_to_rgb332_line(u8 *dbuf, __le32 *sbuf, unsigned int
 /**
  * drm_fb_xrgb8888_to_rgb332 - Convert XRGB8888 to RGB332 clip buffer
  * @dst: RGB332 destination buffer
+ * @dst_pitch: Number of bytes between two consecutive scanlines within dst
  * @src: XRGB8888 source buffer
  * @fb: DRM framebuffer
  * @clip: Clip rectangle area to copy
  *
  * Drivers can use this function for RGB332 devices that don't natively support XRGB8888.
- *
- * This function does not apply clipping on dst, i.e. the destination is a small buffer
- * containing the clip rect only.
  */
-void drm_fb_xrgb8888_to_rgb332(void *dst, void *src, struct drm_framebuffer *fb,
-			       struct drm_rect *clip)
+void drm_fb_xrgb8888_to_rgb332(void *dst, unsigned int dst_pitch, const void *src,
+			       const struct drm_framebuffer *fb, const struct drm_rect *clip)
 {
 	size_t width = drm_rect_width(clip);
 	size_t src_len = width * sizeof(u32);
 	unsigned int y;
 	void *sbuf;
 
+	if (!dst_pitch)
+		dst_pitch = width;
+
 	/* Use a buffer to speed up access on buffers with uncached read mapping (i.e. WC) */
 	sbuf = kmalloc(src_len, GFP_KERNEL);
 	if (!sbuf)
@@ -208,14 +209,14 @@ void drm_fb_xrgb8888_to_rgb332(void *dst, void *src, struct drm_framebuffer *fb,
 		memcpy(sbuf, src, src_len);
 		drm_fb_xrgb8888_to_rgb332_line(dst, sbuf, width);
 		src += fb->pitches[0];
-		dst += width;
+		dst += dst_pitch;
 	}
 
 	kfree(sbuf);
 }
 EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb332);
 
-static void drm_fb_xrgb8888_to_rgb565_line(u16 *dbuf, u32 *sbuf,
+static void drm_fb_xrgb8888_to_rgb565_line(u16 *dbuf, const u32 *sbuf,
 					   unsigned int pixels,
 					   bool swab)
 {
@@ -236,6 +237,7 @@ static void drm_fb_xrgb8888_to_rgb565_line(u16 *dbuf, u32 *sbuf,
 /**
  * drm_fb_xrgb8888_to_rgb565 - Convert XRGB8888 to RGB565 clip buffer
  * @dst: RGB565 destination buffer
+ * @dst_pitch: Number of bytes between two consecutive scanlines within dst
  * @vaddr: XRGB8888 source buffer
  * @fb: DRM framebuffer
  * @clip: Clip rectangle area to copy
@@ -243,13 +245,10 @@ static void drm_fb_xrgb8888_to_rgb565_line(u16 *dbuf, u32 *sbuf,
  *
  * Drivers can use this function for RGB565 devices that don't natively
  * support XRGB8888.
- *
- * This function does not apply clipping on dst, i.e. the destination
- * is a small buffer containing the clip rect only.
  */
-void drm_fb_xrgb8888_to_rgb565(void *dst, void *vaddr,
-			       struct drm_framebuffer *fb,
-			       struct drm_rect *clip, bool swab)
+void drm_fb_xrgb8888_to_rgb565(void *dst, unsigned int dst_pitch, const void *vaddr,
+			       const struct drm_framebuffer *fb, const struct drm_rect *clip,
+			       bool swab)
 {
 	size_t linepixels = clip->x2 - clip->x1;
 	size_t src_len = linepixels * sizeof(u32);
@@ -257,6 +256,9 @@ void drm_fb_xrgb8888_to_rgb565(void *dst, void *vaddr,
 	unsigned y, lines = clip->y2 - clip->y1;
 	void *sbuf;
 
+	if (!dst_pitch)
+		dst_pitch = dst_len;
+
 	/*
 	 * The cma memory is write-combined so reads are uncached.
 	 * Speed up by fetching one line at a time.
@@ -270,7 +272,7 @@ void drm_fb_xrgb8888_to_rgb565(void *dst, void *vaddr,
 		memcpy(sbuf, vaddr, src_len);
 		drm_fb_xrgb8888_to_rgb565_line(dst, sbuf, linepixels, swab);
 		vaddr += fb->pitches[0];
-		dst += dst_len;
+		dst += dst_pitch;
 	}
 
 	kfree(sbuf);
@@ -278,9 +280,9 @@ void drm_fb_xrgb8888_to_rgb565(void *dst, void *vaddr,
 EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb565);
 
 /**
- * drm_fb_xrgb8888_to_rgb565_dstclip - Convert XRGB8888 to RGB565 clip buffer
+ * drm_fb_xrgb8888_to_rgb565_toio - Convert XRGB8888 to RGB565 clip buffer
  * @dst: RGB565 destination buffer (iomem)
- * @dst_pitch: destination buffer pitch
+ * @dst_pitch: Number of bytes between two consecutive scanlines within dst
  * @vaddr: XRGB8888 source buffer
  * @fb: DRM framebuffer
  * @clip: Clip rectangle area to copy
@@ -288,37 +290,36 @@ EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb565);
  *
  * Drivers can use this function for RGB565 devices that don't natively
  * support XRGB8888.
- *
- * This function applies clipping on dst, i.e. the destination is a
- * full (iomem) framebuffer but only the clip rect content is copied over.
  */
-void drm_fb_xrgb8888_to_rgb565_dstclip(void __iomem *dst, unsigned int dst_pitch,
-				       void *vaddr, struct drm_framebuffer *fb,
-				       struct drm_rect *clip, bool swab)
+void drm_fb_xrgb8888_to_rgb565_toio(void __iomem *dst, unsigned int dst_pitch,
+				    const void *vaddr, const struct drm_framebuffer *fb,
+				    const struct drm_rect *clip, bool swab)
 {
 	size_t linepixels = clip->x2 - clip->x1;
 	size_t dst_len = linepixels * sizeof(u16);
 	unsigned y, lines = clip->y2 - clip->y1;
 	void *dbuf;
 
+	if (!dst_pitch)
+		dst_pitch = dst_len;
+
 	dbuf = kmalloc(dst_len, GFP_KERNEL);
 	if (!dbuf)
 		return;
 
 	vaddr += clip_offset(clip, fb->pitches[0], sizeof(u32));
-	dst += clip_offset(clip, dst_pitch, sizeof(u16));
 	for (y = 0; y < lines; y++) {
 		drm_fb_xrgb8888_to_rgb565_line(dbuf, vaddr, linepixels, swab);
 		memcpy_toio(dst, dbuf, dst_len);
 		vaddr += fb->pitches[0];
-		dst += dst_len;
+		dst += dst_pitch;
 	}
 
 	kfree(dbuf);
 }
-EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb565_dstclip);
+EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb565_toio);
 
-static void drm_fb_xrgb8888_to_rgb888_line(u8 *dbuf, u32 *sbuf,
+static void drm_fb_xrgb8888_to_rgb888_line(u8 *dbuf, const u32 *sbuf,
 					   unsigned int pixels)
 {
 	unsigned int x;
@@ -333,24 +334,25 @@ static void drm_fb_xrgb8888_to_rgb888_line(u8 *dbuf, u32 *sbuf,
 /**
  * drm_fb_xrgb8888_to_rgb888 - Convert XRGB8888 to RGB888 clip buffer
  * @dst: RGB888 destination buffer
+ * @dst_pitch: Number of bytes between two consecutive scanlines within dst
  * @src: XRGB8888 source buffer
  * @fb: DRM framebuffer
  * @clip: Clip rectangle area to copy
  *
  * Drivers can use this function for RGB888 devices that don't natively
  * support XRGB8888.
- *
- * This function does not apply clipping on dst, i.e. the destination
- * is a small buffer containing the clip rect only.
  */
-void drm_fb_xrgb8888_to_rgb888(void *dst, void *src, struct drm_framebuffer *fb,
-			       struct drm_rect *clip)
+void drm_fb_xrgb8888_to_rgb888(void *dst, unsigned int dst_pitch, const void *src,
+			       const struct drm_framebuffer *fb, const struct drm_rect *clip)
 {
 	size_t width = drm_rect_width(clip);
 	size_t src_len = width * sizeof(u32);
 	unsigned int y;
 	void *sbuf;
 
+	if (!dst_pitch)
+		dst_pitch = width * 3;
+
 	/* Use a buffer to speed up access on buffers with uncached read mapping (i.e. WC) */
 	sbuf = kmalloc(src_len, GFP_KERNEL);
 	if (!sbuf)
@@ -361,7 +363,7 @@ void drm_fb_xrgb8888_to_rgb888(void *dst, void *src, struct drm_framebuffer *fb,
 		memcpy(sbuf, src, src_len);
 		drm_fb_xrgb8888_to_rgb888_line(dst, sbuf, width);
 		src += fb->pitches[0];
-		dst += width * 3;
+		dst += dst_pitch;
 	}
 
 	kfree(sbuf);
@@ -369,48 +371,48 @@ void drm_fb_xrgb8888_to_rgb888(void *dst, void *src, struct drm_framebuffer *fb,
 EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb888);
 
 /**
- * drm_fb_xrgb8888_to_rgb888_dstclip - Convert XRGB8888 to RGB888 clip buffer
+ * drm_fb_xrgb8888_to_rgb888_toio - Convert XRGB8888 to RGB888 clip buffer
  * @dst: RGB565 destination buffer (iomem)
- * @dst_pitch: destination buffer pitch
+ * @dst_pitch: Number of bytes between two consecutive scanlines within dst
  * @vaddr: XRGB8888 source buffer
  * @fb: DRM framebuffer
  * @clip: Clip rectangle area to copy
  *
  * Drivers can use this function for RGB888 devices that don't natively
  * support XRGB8888.
- *
- * This function applies clipping on dst, i.e. the destination is a
- * full (iomem) framebuffer but only the clip rect content is copied over.
  */
-void drm_fb_xrgb8888_to_rgb888_dstclip(void __iomem *dst, unsigned int dst_pitch,
-				       void *vaddr, struct drm_framebuffer *fb,
-				       struct drm_rect *clip)
+void drm_fb_xrgb8888_to_rgb888_toio(void __iomem *dst, unsigned int dst_pitch,
+				    const void *vaddr, const struct drm_framebuffer *fb,
+				    const struct drm_rect *clip)
 {
 	size_t linepixels = clip->x2 - clip->x1;
 	size_t dst_len = linepixels * 3;
 	unsigned y, lines = clip->y2 - clip->y1;
 	void *dbuf;
 
+	if (!dst_pitch)
+		dst_pitch = dst_len;
+
 	dbuf = kmalloc(dst_len, GFP_KERNEL);
 	if (!dbuf)
 		return;
 
 	vaddr += clip_offset(clip, fb->pitches[0], sizeof(u32));
-	dst += clip_offset(clip, dst_pitch, sizeof(u16));
 	for (y = 0; y < lines; y++) {
 		drm_fb_xrgb8888_to_rgb888_line(dbuf, vaddr, linepixels);
 		memcpy_toio(dst, dbuf, dst_len);
 		vaddr += fb->pitches[0];
-		dst += dst_len;
+		dst += dst_pitch;
 	}
 
 	kfree(dbuf);
 }
-EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb888_dstclip);
+EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb888_toio);
 
 /**
  * drm_fb_xrgb8888_to_gray8 - Convert XRGB8888 to grayscale
  * @dst: 8-bit grayscale destination buffer
+ * @dst_pitch: Number of bytes between two consecutive scanlines within dst
  * @vaddr: XRGB8888 source buffer
  * @fb: DRM framebuffer
  * @clip: Clip rectangle area to copy
@@ -424,16 +426,21 @@ EXPORT_SYMBOL(drm_fb_xrgb8888_to_rgb888_dstclip);
  *
  * ITU BT.601 is used for the RGB -> luma (brightness) conversion.
  */
-void drm_fb_xrgb8888_to_gray8(u8 *dst, void *vaddr, struct drm_framebuffer *fb,
-			       struct drm_rect *clip)
+void drm_fb_xrgb8888_to_gray8(void *dst, unsigned int dst_pitch, const void *vaddr,
+			      const struct drm_framebuffer *fb, const struct drm_rect *clip)
 {
 	unsigned int len = (clip->x2 - clip->x1) * sizeof(u32);
 	unsigned int x, y;
 	void *buf;
-	u32 *src;
+	u8 *dst8;
+	u32 *src32;
 
 	if (WARN_ON(fb->format->format != DRM_FORMAT_XRGB8888))
 		return;
+
+	if (!dst_pitch)
+		dst_pitch = drm_rect_width(clip);
+
 	/*
 	 * The cma memory is write-combined so reads are uncached.
 	 * Speed up by fetching one line at a time.
@@ -442,20 +449,22 @@ void drm_fb_xrgb8888_to_gray8(u8 *dst, void *vaddr, struct drm_framebuffer *fb,
 	if (!buf)
 		return;
 
+	vaddr += clip_offset(clip, fb->pitches[0], sizeof(u32));
 	for (y = clip->y1; y < clip->y2; y++) {
-		src = vaddr + (y * fb->pitches[0]);
-		src += clip->x1;
-		memcpy(buf, src, len);
-		src = buf;
+		dst8 = dst;
+		src32 = memcpy(buf, vaddr, len);
 		for (x = clip->x1; x < clip->x2; x++) {
-			u8 r = (*src & 0x00ff0000) >> 16;
-			u8 g = (*src & 0x0000ff00) >> 8;
-			u8 b =  *src & 0x000000ff;
+			u8 r = (*src32 & 0x00ff0000) >> 16;
+			u8 g = (*src32 & 0x0000ff00) >> 8;
+			u8 b =  *src32 & 0x000000ff;
 
 			/* ITU BT.601: Y = 0.299 R + 0.587 G + 0.114 B */
-			*dst++ = (3 * r + 6 * g + b) / 10;
-			src++;
+			*dst8++ = (3 * r + 6 * g + b) / 10;
+			src32++;
 		}
+
+		vaddr += fb->pitches[0];
+		dst += dst_pitch;
 	}
 
 	kfree(buf);
@@ -502,15 +511,15 @@ int drm_fb_blit_rect_dstclip(void __iomem *dst, unsigned int dst_pitch,
 
 	} else if (dst_format == DRM_FORMAT_RGB565) {
 		if (fb_format == DRM_FORMAT_XRGB8888) {
-			drm_fb_xrgb8888_to_rgb565_dstclip(dst, dst_pitch,
-							  vmap, fb, clip,
-							  false);
+			dst += clip_offset(clip, dst_pitch, fb->format->cpp[0]);
+			drm_fb_xrgb8888_to_rgb565_toio(dst, dst_pitch, vmap, fb, clip,
+						       false);
 			return 0;
 		}
 	} else if (dst_format == DRM_FORMAT_RGB888) {
 		if (fb_format == DRM_FORMAT_XRGB8888) {
-			drm_fb_xrgb8888_to_rgb888_dstclip(dst, dst_pitch,
-							  vmap, fb, clip);
+			dst += clip_offset(clip, dst_pitch, fb->format->cpp[0]);
+			drm_fb_xrgb8888_to_rgb888_toio(dst, dst_pitch, vmap, fb, clip);
 			return 0;
 		}
 	}

drivers/gpu/drm/drm_mipi_dbi.c

@@ -216,7 +216,7 @@ int mipi_dbi_buf_copy(void *dst, struct drm_framebuffer *fb,
 			drm_fb_memcpy(dst, 0, src, fb, clip);
 		break;
 	case DRM_FORMAT_XRGB8888:
-		drm_fb_xrgb8888_to_rgb565(dst, src, fb, clip, swap);
+		drm_fb_xrgb8888_to_rgb565(dst, 0, src, fb, clip, swap);
 		break;
 	default:
 		drm_err_once(fb->dev, "Format is not supported: %p4cc\n",

drivers/gpu/drm/gud/gud_pipe.c

@@ -74,7 +74,7 @@ static size_t gud_xrgb8888_to_r124(u8 *dst, const struct drm_format_info *format
 	if (!buf)
 		return 0;
 
-	drm_fb_xrgb8888_to_gray8(buf, src, fb, rect);
+	drm_fb_xrgb8888_to_gray8(buf, 0, src, fb, rect);
 	pix8 = buf;
 
 	for (y = 0; y < height; y++) {
@@ -190,13 +190,13 @@ retry:
 				goto end_cpu_access;
 			}
 		} else if (format->format == DRM_FORMAT_R8) {
-			drm_fb_xrgb8888_to_gray8(buf, vaddr, fb, rect);
+			drm_fb_xrgb8888_to_gray8(buf, 0, vaddr, fb, rect);
 		} else if (format->format == DRM_FORMAT_RGB332) {
-			drm_fb_xrgb8888_to_rgb332(buf, vaddr, fb, rect);
+			drm_fb_xrgb8888_to_rgb332(buf, 0, vaddr, fb, rect);
 		} else if (format->format == DRM_FORMAT_RGB565) {
-			drm_fb_xrgb8888_to_rgb565(buf, vaddr, fb, rect, gud_is_big_endian());
+			drm_fb_xrgb8888_to_rgb565(buf, 0, vaddr, fb, rect, gud_is_big_endian());
 		} else if (format->format == DRM_FORMAT_RGB888) {
-			drm_fb_xrgb8888_to_rgb888(buf, vaddr, fb, rect);
+			drm_fb_xrgb8888_to_rgb888(buf, 0, vaddr, fb, rect);
 		} else {
 			len = gud_xrgb8888_to_color(buf, format, vaddr, fb, rect);
 		}

drivers/gpu/drm/tiny/cirrus.c

@@ -329,14 +329,12 @@ static int cirrus_fb_blit_rect(struct drm_framebuffer *fb, const struct dma_buf_
 		drm_fb_memcpy_toio(dst, fb->pitches[0], vmap, fb, rect);
 
 	} else if (fb->format->cpp[0] == 4 && cirrus->cpp == 2) {
-		drm_fb_xrgb8888_to_rgb565_dstclip(cirrus->vram,
-						  cirrus->pitch,
-						  vmap, fb, rect, false);
+		dst += drm_fb_clip_offset(cirrus->pitch, fb->format, rect);
+		drm_fb_xrgb8888_to_rgb565_toio(dst, cirrus->pitch, vmap, fb, rect, false);
 
 	} else if (fb->format->cpp[0] == 4 && cirrus->cpp == 3) {
-		drm_fb_xrgb8888_to_rgb888_dstclip(cirrus->vram,
-						  cirrus->pitch,
-						  vmap, fb, rect);
+		dst += drm_fb_clip_offset(cirrus->pitch, fb->format, rect);
+		drm_fb_xrgb8888_to_rgb888_toio(dst, cirrus->pitch, vmap, fb, rect);
 
 	} else {
 		WARN_ON_ONCE("cpp mismatch");

drivers/gpu/drm/tiny/repaper.c

@@ -560,7 +560,7 @@ static int repaper_fb_dirty(struct drm_framebuffer *fb)
 	if (ret)
 		goto out_free;
 
-	drm_fb_xrgb8888_to_gray8(buf, cma_obj->vaddr, fb, &clip);
+	drm_fb_xrgb8888_to_gray8(buf, 0, cma_obj->vaddr, fb, &clip);
 
 	drm_gem_fb_end_cpu_access(fb, DMA_FROM_DEVICE);
 

drivers/gpu/drm/tiny/st7586.c

@@ -73,7 +73,7 @@ static void st7586_xrgb8888_to_gray332(u8 *dst, void *vaddr,
 	if (!buf)
 		return;
 
-	drm_fb_xrgb8888_to_gray8(buf, vaddr, fb, clip);
+	drm_fb_xrgb8888_to_gray8(buf, 0, vaddr, fb, clip);
 	src = buf;
 
 	for (y = clip->y1; y < clip->y2; y++) {

include/drm/drm_format_helper.h

@@ -20,21 +20,21 @@ void drm_fb_memcpy_toio(void __iomem *dst, unsigned int dst_pitch, const void *v
 void drm_fb_swab(void *dst, unsigned int dst_pitch, const void *src,
 		 const struct drm_framebuffer *fb, const struct drm_rect *clip,
 		 bool cached);
-void drm_fb_xrgb8888_to_rgb332(void *dst, void *vaddr, struct drm_framebuffer *fb,
-			       struct drm_rect *clip);
-void drm_fb_xrgb8888_to_rgb565(void *dst, void *vaddr,
-			       struct drm_framebuffer *fb,
-			       struct drm_rect *clip, bool swab);
-void drm_fb_xrgb8888_to_rgb565_dstclip(void __iomem *dst, unsigned int dst_pitch,
-				       void *vaddr, struct drm_framebuffer *fb,
-				       struct drm_rect *clip, bool swab);
-void drm_fb_xrgb8888_to_rgb888(void *dst, void *src, struct drm_framebuffer *fb,
-			       struct drm_rect *clip);
-void drm_fb_xrgb8888_to_rgb888_dstclip(void __iomem *dst, unsigned int dst_pitch,
-				       void *vaddr, struct drm_framebuffer *fb,
-				       struct drm_rect *clip);
-void drm_fb_xrgb8888_to_gray8(u8 *dst, void *vaddr, struct drm_framebuffer *fb,
-			      struct drm_rect *clip);
+void drm_fb_xrgb8888_to_rgb332(void *dst, unsigned int dst_pitch, const void *vaddr,
+			       const struct drm_framebuffer *fb, const struct drm_rect *clip);
+void drm_fb_xrgb8888_to_rgb565(void *dst, unsigned int dst_pitch, const void *vaddr,
+			       const struct drm_framebuffer *fb, const struct drm_rect *clip,
+			       bool swab);
+void drm_fb_xrgb8888_to_rgb565_toio(void __iomem *dst, unsigned int dst_pitch,
+				    const void *vaddr, const struct drm_framebuffer *fb,
+				    const struct drm_rect *clip, bool swab);
+void drm_fb_xrgb8888_to_rgb888(void *dst, unsigned int dst_pitch, const void *src,
+			       const struct drm_framebuffer *fb, const struct drm_rect *clip);
+void drm_fb_xrgb8888_to_rgb888_toio(void __iomem *dst, unsigned int dst_pitch,
+				    const void *vaddr, const struct drm_framebuffer *fb,
+				    const struct drm_rect *clip);
+void drm_fb_xrgb8888_to_gray8(void *dst, unsigned int dst_pitch, const void *vaddr,
+			      const struct drm_framebuffer *fb, const struct drm_rect *clip);
 
 int drm_fb_blit_rect_dstclip(void __iomem *dst, unsigned int dst_pitch,
 			     uint32_t dst_format, void *vmap,