swrast: Convert color glReadPixels slow path to using MapRenderbuffer.

This may be a bit slower than before because we're switching from
per-format compiled loops in GetRow to
_mesa_unpack_rgba_block_unpack's loop around a callback to unpack a
pixel.  The solution there would be to make _mesa_unpack_rgba_block
fold the span loop into the format handlers.

(On the other hand, function call overhead will hardly matter if
MapRenderbuffer means the driver gets the data into cacheable memory
instead of uncached).

The adjust_colors code should no longer be required, since the unpack
function does the 565 to float conversion in a single pass instead of
converting it (poorly) through 8888 as apparently happened in the
past.

Reviewed-by: Brian Paul <brianp@vmware.com>

diff --git a/src/mesa/swrast/s_readpix.c b/src/mesa/swrast/s_readpix.c
index b2585998306cdd0c3dc1ffa8bf51d36ebabccf7a..2c047908ae6da80b4e9121306459d2ac8303b0ea 100644 (file)
--- a/src/mesa/swrast/s_readpix.c
+++ b/src/mesa/swrast/s_readpix.c
@@ -311,50 +311,41 @@ fast_read_rgba_pixels_memcpy( struct gl_context *ctx,
    return GL_TRUE;
 }
 
-
-/**
- * When we're using a low-precision color buffer (like 16-bit 5/6/5)
- * we have to adjust our color values a bit to pass conformance.
- * The problem is when a 5 or 6-bit color value is converted to an 8-bit
- * value and then a floating point value, the floating point values don't
- * increment uniformly as the 5 or 6-bit value is incremented.
- *
- * This function adjusts floating point values to compensate.
- */
-static void
-adjust_colors(const struct gl_framebuffer *fb, GLuint n, GLfloat rgba[][4])
+static GLboolean
+slow_read_rgba_pixels( struct gl_context *ctx,
+                      GLint x, GLint y,
+                      GLsizei width, GLsizei height,
+                      GLenum format, GLenum type,
+                      GLvoid *pixels,
+                      const struct gl_pixelstore_attrib *packing,
+                      GLbitfield transferOps )
 {
-   const GLuint rShift = 8 - fb->Visual.redBits;
-   const GLuint gShift = 8 - fb->Visual.greenBits;
-   const GLuint bShift = 8 - fb->Visual.blueBits;
-   GLfloat rScale = 1.0F / (GLfloat) ((1 << fb->Visual.redBits  ) - 1);
-   GLfloat gScale = 1.0F / (GLfloat) ((1 << fb->Visual.greenBits) - 1);
-   GLfloat bScale = 1.0F / (GLfloat) ((1 << fb->Visual.blueBits ) - 1);
-   GLuint i;
-
-   if (fb->Visual.redBits == 0)
-      rScale = 0;
-   if (fb->Visual.greenBits == 0)
-      gScale = 0;
-   if (fb->Visual.blueBits == 0)
-      bScale = 0;
-
-   for (i = 0; i < n; i++) {
-      GLint r, g, b;
-      /* convert float back to ubyte */
-      CLAMPED_FLOAT_TO_UBYTE(r, rgba[i][RCOMP]);
-      CLAMPED_FLOAT_TO_UBYTE(g, rgba[i][GCOMP]);
-      CLAMPED_FLOAT_TO_UBYTE(b, rgba[i][BCOMP]);
-      /* using only the N most significant bits of the ubyte value, convert to
-       * float in [0,1].
-       */
-      rgba[i][RCOMP] = (GLfloat) (r >> rShift) * rScale;
-      rgba[i][GCOMP] = (GLfloat) (g >> gShift) * gScale;
-      rgba[i][BCOMP] = (GLfloat) (b >> bShift) * bScale;
+   struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
+   GLfloat rgba[MAX_WIDTH][4];
+   GLubyte *dst, *map;
+   int dstStride, stride, j;
+
+   dstStride = _mesa_image_row_stride(packing, width, format, type);
+   dst = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
+                                          format, type, 0, 0);
+
+   ctx->Driver.MapRenderbuffer(ctx, rb, x, y, width, height, GL_MAP_READ_BIT,
+                              &map, &stride);
+
+   for (j = 0; j < height; j++) {
+      _mesa_unpack_rgba_row(_mesa_get_srgb_format_linear(rb->Format),
+                           width, map, rgba);
+      _mesa_pack_rgba_span_float(ctx, width, rgba, format, type, dst,
+                                packing, transferOps);
+
+      dst += dstStride;
+      map += stride;
    }
-}
 
+   ctx->Driver.UnmapRenderbuffer(ctx, rb);
 
+   return GL_TRUE;
+}
 
 /*
  * Read R, G, B, A, RGB, L, or LA pixels.
@@ -366,7 +357,6 @@ read_rgba_pixels( struct gl_context *ctx,
                   GLenum format, GLenum type, GLvoid *pixels,
                   const struct gl_pixelstore_attrib *packing )
 {
-   SWcontext *swrast = SWRAST_CONTEXT(ctx);
    GLbitfield transferOps = ctx->_ImageTransferState;
    struct gl_framebuffer *fb = ctx->ReadBuffer;
    struct gl_renderbuffer *rb = fb->_ColorReadBuffer;
@@ -393,37 +383,8 @@ read_rgba_pixels( struct gl_context *ctx,
       }
    }
 
-   /* width should never be > MAX_WIDTH since we did clipping earlier */
-   ASSERT(width <= MAX_WIDTH);
-
-   {
-      const GLint dstStride
-         = _mesa_image_row_stride(packing, width, format, type);
-      GLfloat (*rgba)[4] = swrast->SpanArrays->attribs[FRAG_ATTRIB_COL0];
-      GLint row;
-      GLubyte *dst
-         = (GLubyte *) _mesa_image_address2d(packing, pixels, width, height,
-                                             format, type, 0, 0);
-
-      for (row = 0; row < height; row++, y++) {
-
-         /* Get float rgba pixels */
-         _swrast_read_rgba_span(ctx, rb, width, x, y, GL_FLOAT, rgba);
-
-         /* apply fudge factor for shallow color buffers */
-         if ((fb->Visual.redBits < 8 && fb->Visual.redBits != 0) ||
-             (fb->Visual.greenBits < 8 && fb->Visual.greenBits != 0) ||
-            (fb->Visual.blueBits < 8 && fb->Visual.blueBits != 0)) {
-            adjust_colors(fb, width, rgba);
-         }
-
-         /* pack the row of RGBA pixels into user's buffer */
-         _mesa_pack_rgba_span_float(ctx, width, rgba, format, type, dst,
-                                    packing, transferOps);
-
-         dst += dstStride;
-      }
-   }
+   slow_read_rgba_pixels(ctx, x, y, width, height,
+                        format, type, pixels, packing, transferOps);
 }
 
 /**