[PATCH 1/3] gpu: ipu-v3: image-convert: Fix input bytesperline width/height align

[PATCH 1/3] gpu: ipu-v3: image-convert: Fix input bytesperline width/height align

[Steve Longerbeam]

The output width and height alignment values were being used in the
input bytesperline calculation. Fix by separating local vars w_align
and h_align into w_align_in, h_align_in, w_align_out, and h_align_out.

Fixes: d966e23d61a2c ("gpu: ipu-v3: image-convert: fix bytesperline
adjustment")

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
---
 drivers/gpu/ipu-v3/ipu-image-convert.c | 32 +++++++++++++++++---------
 1 file changed, 21 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/ipu-v3/ipu-image-convert.c b/drivers/gpu/ipu-v3/ipu-image-convert.c
index 36e88434513a..36eb4c77ad91 100644
--- a/drivers/gpu/ipu-v3/ipu-image-convert.c
+++ b/drivers/gpu/ipu-v3/ipu-image-convert.c
@@ -1876,7 +1876,8 @@ void ipu_image_convert_adjust(struct ipu_image *in, struct ipu_image *out,
 			      enum ipu_rotate_mode rot_mode)
 {
 	const struct ipu_image_pixfmt *infmt, *outfmt;
-	u32 w_align, h_align;
+	u32 w_align_out, h_align_out;
+	u32 w_align_in, h_align_in;
 
 	infmt = get_format(in->pix.pixelformat);
 	outfmt = get_format(out->pix.pixelformat);
@@ -1908,22 +1909,31 @@ void ipu_image_convert_adjust(struct ipu_image *in, struct ipu_image *out,
 	}
 
 	/* align input width/height */
-	w_align = ilog2(tile_width_align(IMAGE_CONVERT_IN, infmt, rot_mode));
-	h_align = ilog2(tile_height_align(IMAGE_CONVERT_IN, infmt, rot_mode));
-	in->pix.width = clamp_align(in->pix.width, MIN_W, MAX_W, w_align);
-	in->pix.height = clamp_align(in->pix.height, MIN_H, MAX_H, h_align);
+	w_align_in = ilog2(tile_width_align(IMAGE_CONVERT_IN, infmt,
+					    rot_mode));
+	h_align_in = ilog2(tile_height_align(IMAGE_CONVERT_IN, infmt,
+					     rot_mode));
+	in->pix.width = clamp_align(in->pix.width, MIN_W, MAX_W,
+				    w_align_in);
+	in->pix.height = clamp_align(in->pix.height, MIN_H, MAX_H,
+				     h_align_in);
 
 	/* align output width/height */
-	w_align = ilog2(tile_width_align(IMAGE_CONVERT_OUT, outfmt, rot_mode));
-	h_align = ilog2(tile_height_align(IMAGE_CONVERT_OUT, outfmt, rot_mode));
-	out->pix.width = clamp_align(out->pix.width, MIN_W, MAX_W, w_align);
-	out->pix.height = clamp_align(out->pix.height, MIN_H, MAX_H, h_align);
+	w_align_out = ilog2(tile_width_align(IMAGE_CONVERT_OUT, outfmt,
+					     rot_mode));
+	h_align_out = ilog2(tile_height_align(IMAGE_CONVERT_OUT, outfmt,
+					      rot_mode));
+	out->pix.width = clamp_align(out->pix.width, MIN_W, MAX_W,
+				     w_align_out);
+	out->pix.height = clamp_align(out->pix.height, MIN_H, MAX_H,
+				      h_align_out);
 
 	/* set input/output strides and image sizes */
 	in->pix.bytesperline = infmt->planar ?
-		clamp_align(in->pix.width, 2 << w_align, MAX_W, w_align) :
+		clamp_align(in->pix.width, 2 << w_align_in, MAX_W,
+			    w_align_in) :
 		clamp_align((in->pix.width * infmt->bpp) >> 3,
-			    2 << w_align, MAX_W, w_align);
+			    2 << w_align_in, MAX_W, w_align_in);
 	in->pix.sizeimage = infmt->planar ?
 		(in->pix.height * in->pix.bytesperline * infmt->bpp) >> 3 :
 		in->pix.height * in->pix.bytesperline;
-- 
2.17.1

[PATCH 2/3] gpu: ipu-v3: image-convert: Fix input bytesperline for packed formats

[Steve Longerbeam]

The input bytesperline calculation for packed pixel formats was
incorrect. The min/max clamping values must be multiplied by the
packed bits-per-pixel. This was causing corrupted converted images
when the input format was RGB4 (probably also other input packed
formats).

Fixes: d966e23d61a2c ("gpu: ipu-v3: image-convert: fix bytesperline
adjustment")

Reported-by: Harsha Manjula Mallikarjun <Harsha.ManjulaMallikarjun@in.bosch.com>
Suggested-by: Harsha Manjula Mallikarjun <Harsha.ManjulaMallikarjun@in.bosch.com>
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
---
 drivers/gpu/ipu-v3/ipu-image-convert.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/ipu-v3/ipu-image-convert.c b/drivers/gpu/ipu-v3/ipu-image-convert.c
index 36eb4c77ad91..4dfdbd1adf0d 100644
--- a/drivers/gpu/ipu-v3/ipu-image-convert.c
+++ b/drivers/gpu/ipu-v3/ipu-image-convert.c
@@ -1933,7 +1933,9 @@ void ipu_image_convert_adjust(struct ipu_image *in, struct ipu_image *out,
 		clamp_align(in->pix.width, 2 << w_align_in, MAX_W,
 			    w_align_in) :
 		clamp_align((in->pix.width * infmt->bpp) >> 3,
-			    2 << w_align_in, MAX_W, w_align_in);
+			    ((2 << w_align_in) * infmt->bpp) >> 3,
+			    (MAX_W * infmt->bpp) >> 3,
+			    w_align_in);
 	in->pix.sizeimage = infmt->planar ?
 		(in->pix.height * in->pix.bytesperline * infmt->bpp) >> 3 :
 		in->pix.height * in->pix.bytesperline;
-- 
2.17.1

[PATCH 3/3] gpu: ipu-v3: image-convert: Fix image downsize coefficients

[Steve Longerbeam]

The output of the IC downsizer unit in both dimensions must be <= 1024
before being passed to the IC resizer unit. This was causing corrupted
images when:

input_dim > 1024, and
input_dim / 2 < output_dim < input_dim

Some broken examples were 1920x1080 -> 1024x768 and 1920x1080 ->
1280x1080.

Fixes: 70b9b6b3bcb21 ("gpu: ipu-v3: image-convert: calculate per-tile
resize coefficients")

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
---
 drivers/gpu/ipu-v3/ipu-image-convert.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/ipu-v3/ipu-image-convert.c b/drivers/gpu/ipu-v3/ipu-image-convert.c
index 4dfdbd1adf0d..e744f3527ce1 100644
--- a/drivers/gpu/ipu-v3/ipu-image-convert.c
+++ b/drivers/gpu/ipu-v3/ipu-image-convert.c
@@ -400,12 +400,14 @@ static int calc_image_resize_coefficients(struct ipu_image_convert_ctx *ctx,
 	if (WARN_ON(resized_width == 0 || resized_height == 0))
 		return -EINVAL;
 
-	while (downsized_width >= resized_width * 2) {
+	while (downsized_width > 1024 ||
+	       downsized_width >= resized_width * 2) {
 		downsized_width >>= 1;
 		downsize_coeff_h++;
 	}
 
-	while (downsized_height >= resized_height * 2) {
+	while (downsized_height > 1024 ||
+	       downsized_height >= resized_height * 2) {
 		downsized_height >>= 1;
 		downsize_coeff_v++;
 	}
-- 
2.17.1

Re: [PATCH 3/3] gpu: ipu-v3: image-convert: Fix image downsize coefficients

[Philipp Zabel]

On Tue, 2019-06-11 at 18:16 -0700, Steve Longerbeam wrote:
> The output of the IC downsizer unit in both dimensions must be <= 1024
> before being passed to the IC resizer unit. This was causing corrupted
> images when:
> 
> input_dim > 1024, and
> input_dim / 2 < output_dim < input_dim
> 
> Some broken examples were 1920x1080 -> 1024x768 and 1920x1080 ->
> 1280x1080.
> 
> Fixes: 70b9b6b3bcb21 ("gpu: ipu-v3: image-convert: calculate per-tile
> resize coefficients")
> 
> Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>

All applied on the imx-drm/fixes branch.

regards
Philipp