render/gles2: fix y-inverted output when rendering to buffer
This commit is contained in:
parent
6136fe87d1
commit
1245730ea2
|
@ -214,8 +214,12 @@ static void gles2_scissor(struct wlr_renderer *wlr_renderer,
|
|||
push_gles2_debug(renderer);
|
||||
if (box != NULL) {
|
||||
struct wlr_box gl_box;
|
||||
wlr_box_transform(&gl_box, box, WL_OUTPUT_TRANSFORM_FLIPPED_180,
|
||||
renderer->viewport_width, renderer->viewport_height);
|
||||
if (renderer->current_buffer != NULL) {
|
||||
memcpy(&gl_box, box, sizeof(gl_box));
|
||||
} else {
|
||||
wlr_box_transform(&gl_box, box, WL_OUTPUT_TRANSFORM_FLIPPED_180,
|
||||
renderer->viewport_width, renderer->viewport_height);
|
||||
}
|
||||
|
||||
glScissor(gl_box.x, gl_box.y, gl_box.width, gl_box.height);
|
||||
glEnable(GL_SCISSOR_TEST);
|
||||
|
@ -225,6 +229,12 @@ static void gles2_scissor(struct wlr_renderer *wlr_renderer,
|
|||
pop_gles2_debug(renderer);
|
||||
}
|
||||
|
||||
static const float flip_180[9] = {
|
||||
1.0f, 0.0f, 0.0f,
|
||||
0.0f, -1.0f, 0.0f,
|
||||
0.0f, 0.0f, 1.0f,
|
||||
};
|
||||
|
||||
static bool gles2_render_subtexture_with_matrix(
|
||||
struct wlr_renderer *wlr_renderer, struct wlr_texture *wlr_texture,
|
||||
const struct wlr_fbox *box, const float matrix[static 9],
|
||||
|
@ -257,10 +267,15 @@ static bool gles2_render_subtexture_with_matrix(
|
|||
abort();
|
||||
}
|
||||
|
||||
float gl_matrix[9];
|
||||
if (renderer->current_buffer != NULL) {
|
||||
wlr_matrix_multiply(gl_matrix, flip_180, matrix);
|
||||
} else {
|
||||
memcpy(gl_matrix, matrix, sizeof(gl_matrix));
|
||||
}
|
||||
// OpenGL ES 2 requires the glUniformMatrix3fv transpose parameter to be set
|
||||
// to GL_FALSE
|
||||
float transposition[9];
|
||||
wlr_matrix_transpose(transposition, matrix);
|
||||
wlr_matrix_transpose(gl_matrix, gl_matrix);
|
||||
|
||||
push_gles2_debug(renderer);
|
||||
|
||||
|
@ -271,7 +286,7 @@ static bool gles2_render_subtexture_with_matrix(
|
|||
|
||||
glUseProgram(shader->program);
|
||||
|
||||
glUniformMatrix3fv(shader->proj, 1, GL_FALSE, transposition);
|
||||
glUniformMatrix3fv(shader->proj, 1, GL_FALSE, gl_matrix);
|
||||
glUniform1i(shader->invert_y, texture->inverted_y);
|
||||
glUniform1i(shader->tex, 0);
|
||||
glUniform1f(shader->alpha, alpha);
|
||||
|
@ -309,15 +324,20 @@ static void gles2_render_quad_with_matrix(struct wlr_renderer *wlr_renderer,
|
|||
struct wlr_gles2_renderer *renderer =
|
||||
gles2_get_renderer_in_context(wlr_renderer);
|
||||
|
||||
float gl_matrix[9];
|
||||
if (renderer->current_buffer != NULL) {
|
||||
wlr_matrix_multiply(gl_matrix, flip_180, matrix);
|
||||
} else {
|
||||
memcpy(gl_matrix, matrix, sizeof(gl_matrix));
|
||||
}
|
||||
// OpenGL ES 2 requires the glUniformMatrix3fv transpose parameter to be set
|
||||
// to GL_FALSE
|
||||
float transposition[9];
|
||||
wlr_matrix_transpose(transposition, matrix);
|
||||
wlr_matrix_transpose(gl_matrix, gl_matrix);
|
||||
|
||||
push_gles2_debug(renderer);
|
||||
glUseProgram(renderer->shaders.quad.program);
|
||||
|
||||
glUniformMatrix3fv(renderer->shaders.quad.proj, 1, GL_FALSE, transposition);
|
||||
glUniformMatrix3fv(renderer->shaders.quad.proj, 1, GL_FALSE, gl_matrix);
|
||||
glUniform4f(renderer->shaders.quad.color, color[0], color[1], color[2], color[3]);
|
||||
|
||||
glVertexAttribPointer(renderer->shaders.quad.pos_attrib, 2, GL_FLOAT, GL_FALSE,
|
||||
|
@ -337,10 +357,15 @@ static void gles2_render_ellipse_with_matrix(struct wlr_renderer *wlr_renderer,
|
|||
struct wlr_gles2_renderer *renderer =
|
||||
gles2_get_renderer_in_context(wlr_renderer);
|
||||
|
||||
float gl_matrix[9];
|
||||
if (renderer->current_buffer != NULL) {
|
||||
wlr_matrix_multiply(gl_matrix, flip_180, matrix);
|
||||
} else {
|
||||
memcpy(gl_matrix, matrix, sizeof(gl_matrix));
|
||||
}
|
||||
// OpenGL ES 2 requires the glUniformMatrix3fv transpose parameter to be set
|
||||
// to GL_FALSE
|
||||
float transposition[9];
|
||||
wlr_matrix_transpose(transposition, matrix);
|
||||
wlr_matrix_transpose(gl_matrix, gl_matrix);
|
||||
|
||||
static const GLfloat texcoord[] = {
|
||||
1, 0, // top right
|
||||
|
@ -352,7 +377,7 @@ static void gles2_render_ellipse_with_matrix(struct wlr_renderer *wlr_renderer,
|
|||
push_gles2_debug(renderer);
|
||||
glUseProgram(renderer->shaders.ellipse.program);
|
||||
|
||||
glUniformMatrix3fv(renderer->shaders.ellipse.proj, 1, GL_FALSE, transposition);
|
||||
glUniformMatrix3fv(renderer->shaders.ellipse.proj, 1, GL_FALSE, gl_matrix);
|
||||
glUniform4f(renderer->shaders.ellipse.color, color[0], color[1], color[2], color[3]);
|
||||
|
||||
glVertexAttribPointer(renderer->shaders.ellipse.pos_attrib, 2, GL_FLOAT,
|
||||
|
@ -472,16 +497,32 @@ static bool gles2_read_pixels(struct wlr_renderer *wlr_renderer,
|
|||
if (pack_stride == stride && dst_x == 0 && flags != NULL) {
|
||||
// Under these particular conditions, we can read the pixels with only
|
||||
// one glReadPixels call
|
||||
glReadPixels(src_x, renderer->viewport_height - height - src_y,
|
||||
width, height, fmt->gl_format, fmt->gl_type, p);
|
||||
*flags = WLR_RENDERER_READ_PIXELS_Y_INVERT;
|
||||
|
||||
uint32_t y = src_y;
|
||||
if (renderer->current_buffer == NULL) {
|
||||
y = renderer->viewport_height - height - src_y;
|
||||
}
|
||||
|
||||
glReadPixels(src_x, y, width, height, fmt->gl_format, fmt->gl_type, p);
|
||||
|
||||
if (renderer->current_buffer != NULL) {
|
||||
*flags = 0;
|
||||
} else {
|
||||
*flags = WLR_RENDERER_READ_PIXELS_Y_INVERT;
|
||||
}
|
||||
} else {
|
||||
// Unfortunately GLES2 doesn't support GL_PACK_*, so we have to read
|
||||
// the lines out row by row
|
||||
for (size_t i = 0; i < height; ++i) {
|
||||
glReadPixels(src_x, renderer->viewport_height - src_y - i - 1, width, 1, fmt->gl_format,
|
||||
uint32_t y = src_y + i;
|
||||
if (renderer->current_buffer == NULL) {
|
||||
y = renderer->viewport_height - src_y - i - 1;
|
||||
}
|
||||
|
||||
glReadPixels(src_x, y, width, 1, fmt->gl_format,
|
||||
fmt->gl_type, p + i * stride + dst_x * fmt->bpp / 8);
|
||||
}
|
||||
|
||||
if (flags != NULL) {
|
||||
*flags = 0;
|
||||
}
|
||||
|
@ -510,6 +551,7 @@ static bool gles2_blit_dmabuf(struct wlr_renderer *wlr_renderer,
|
|||
goto restore_context_out;
|
||||
}
|
||||
|
||||
// TODO: get inverted_y right when current_buffer != NULL
|
||||
// This is to take into account y-inversion on both buffers rather than
|
||||
// just the source buffer.
|
||||
bool src_inverted_y =
|
||||
|
@ -517,9 +559,12 @@ static bool gles2_blit_dmabuf(struct wlr_renderer *wlr_renderer,
|
|||
bool dst_inverted_y =
|
||||
!!(dst_attr->flags & WLR_DMABUF_ATTRIBUTES_FLAGS_Y_INVERT);
|
||||
struct wlr_gles2_texture *gles2_src_tex = gles2_get_texture(src_tex);
|
||||
// The result is negated because wlr_matrix_projection y-inverts the
|
||||
// texture.
|
||||
gles2_src_tex->inverted_y = !(src_inverted_y ^ dst_inverted_y);
|
||||
gles2_src_tex->inverted_y = src_inverted_y ^ dst_inverted_y;
|
||||
if (renderer->current_buffer == NULL) {
|
||||
// The result is negated because wlr_matrix_projection y-inverts the
|
||||
// texture.
|
||||
gles2_src_tex->inverted_y = !gles2_src_tex->inverted_y;
|
||||
}
|
||||
|
||||
if (!wlr_egl_make_current(renderer->egl, EGL_NO_SURFACE, NULL)) {
|
||||
goto texture_destroy_out;
|
||||
|
|
Loading…
Reference in New Issue