DXR will be turned off on Tuesday, December 29th. It will redirect to Searchfox.
See the announcement on Discourse.

DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Line Code
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557
/* -*- Mode: c++; c-basic-offset: 4; indent-tabs-mode: nil; tab-width: 40; -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#include "GLUploadHelpers.h"

#include "GLContext.h"
#include "mozilla/gfx/2D.h"
#include "mozilla/gfx/Tools.h"  // For BytesPerPixel
#include "nsRegion.h"

namespace mozilla {

using namespace gfx;

namespace gl {

/* These two techniques are suggested by "Bit Twiddling Hacks"
 */

/**
 * Returns true if |aNumber| is a power of two
 * 0 is incorreclty considered a power of two
 */
static bool
IsPowerOfTwo(int aNumber)
{
    return (aNumber & (aNumber - 1)) == 0;
}

/**
 * Returns the first integer greater than |aNumber| which is a power of two
 * Undefined for |aNumber| < 0
 */
static int
NextPowerOfTwo(int aNumber)
{
#if defined(__arm__)
    return 1 << (32 - __builtin_clz(aNumber - 1));
#else
    --aNumber;
    aNumber |= aNumber >> 1;
    aNumber |= aNumber >> 2;
    aNumber |= aNumber >> 4;
    aNumber |= aNumber >> 8;
    aNumber |= aNumber >> 16;
    return ++aNumber;
#endif
}

static unsigned int
DataOffset(const nsIntPoint &aPoint, int32_t aStride, SurfaceFormat aFormat)
{
  unsigned int data = aPoint.y * aStride;
  data += aPoint.x * BytesPerPixel(aFormat);
  return data;
}

static GLint GetAddressAlignment(ptrdiff_t aAddress)
{
    if (!(aAddress & 0x7)) {
       return 8;
    } else if (!(aAddress & 0x3)) {
        return 4;
    } else if (!(aAddress & 0x1)) {
        return 2;
    } else {
        return 1;
    }
}

// Take texture data in a given buffer and copy it into a larger buffer,
// padding out the edge pixels for filtering if necessary
static void
CopyAndPadTextureData(const GLvoid* srcBuffer,
                      GLvoid* dstBuffer,
                      GLsizei srcWidth, GLsizei srcHeight,
                      GLsizei dstWidth, GLsizei dstHeight,
                      GLsizei stride, GLint pixelsize)
{
    unsigned char *rowDest = static_cast<unsigned char*>(dstBuffer);
    const unsigned char *source = static_cast<const unsigned char*>(srcBuffer);

    for (GLsizei h = 0; h < srcHeight; ++h) {
        memcpy(rowDest, source, srcWidth * pixelsize);
        rowDest += dstWidth * pixelsize;
        source += stride;
    }

    GLsizei padHeight = srcHeight;

    // Pad out an extra row of pixels so that edge filtering doesn't use garbage data
    if (dstHeight > srcHeight) {
        memcpy(rowDest, source - stride, srcWidth * pixelsize);
        padHeight++;
    }

    // Pad out an extra column of pixels
    if (dstWidth > srcWidth) {
        rowDest = static_cast<unsigned char*>(dstBuffer) + srcWidth * pixelsize;
        for (GLsizei h = 0; h < padHeight; ++h) {
            memcpy(rowDest, rowDest - pixelsize, pixelsize);
            rowDest += dstWidth * pixelsize;
        }
    }
}

// In both of these cases (for the Adreno at least) it is impossible
// to determine good or bad driver versions for POT texture uploads,
// so blacklist them all. Newer drivers use a different rendering
// string in the form "Adreno (TM) 200" and the drivers we've seen so
// far work fine with NPOT textures, so don't blacklist those until we
// have evidence of any problems with them.
bool
CanUploadSubTextures(GLContext* gl)
{
    if (!gl->WorkAroundDriverBugs())
        return true;

    // There are certain GPUs that we don't want to use glTexSubImage2D on
    // because that function can be very slow and/or buggy
    if (gl->Renderer() == GLRenderer::Adreno200 ||
        gl->Renderer() == GLRenderer::Adreno205)
    {
        return false;
    }

    // On PowerVR glTexSubImage does a readback, so it will be slower
    // than just doing a glTexImage2D() directly. i.e. 26ms vs 10ms
    if (gl->Renderer() == GLRenderer::SGX540 ||
        gl->Renderer() == GLRenderer::SGX530)
    {
        return false;
    }

    return true;
}

static void
TexSubImage2DWithUnpackSubimageGLES(GLContext* gl,
                                    GLenum target, GLint level,
                                    GLint xoffset, GLint yoffset,
                                    GLsizei width, GLsizei height,
                                    GLsizei stride, GLint pixelsize,
                                    GLenum format, GLenum type,
                                    const GLvoid* pixels)
{
    gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                     std::min(GetAddressAlignment((ptrdiff_t)pixels),
                              GetAddressAlignment((ptrdiff_t)stride)));
    // When using GL_UNPACK_ROW_LENGTH, we need to work around a Tegra
    // driver crash where the driver apparently tries to read
    // (stride - width * pixelsize) bytes past the end of the last input
    // row. We only upload the first height-1 rows using GL_UNPACK_ROW_LENGTH,
    // and then we upload the final row separately. See bug 697990.
    int rowLength = stride/pixelsize;
    gl->fPixelStorei(LOCAL_GL_UNPACK_ROW_LENGTH, rowLength);
    gl->fTexSubImage2D(target,
                       level,
                       xoffset,
                       yoffset,
                       width,
                       height-1,
                       format,
                       type,
                       pixels);
    gl->fPixelStorei(LOCAL_GL_UNPACK_ROW_LENGTH, 0);
    gl->fTexSubImage2D(target,
                       level,
                       xoffset,
                       yoffset+height-1,
                       width,
                       1,
                       format,
                       type,
                       (const unsigned char *)pixels+(height-1)*stride);
    gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);
}

static void
TexSubImage2DWithoutUnpackSubimage(GLContext* gl,
                                   GLenum target, GLint level,
                                   GLint xoffset, GLint yoffset,
                                   GLsizei width, GLsizei height,
                                   GLsizei stride, GLint pixelsize,
                                   GLenum format, GLenum type,
                                   const GLvoid* pixels)
{
    // Not using the whole row of texture data and GL_UNPACK_ROW_LENGTH
    // isn't supported. We make a copy of the texture data we're using,
    // such that we're using the whole row of data in the copy. This turns
    // out to be more efficient than uploading row-by-row; see bug 698197.
    unsigned char *newPixels = new unsigned char[width*height*pixelsize];
    unsigned char *rowDest = newPixels;
    const unsigned char *rowSource = (const unsigned char *)pixels;
    for (int h = 0; h < height; h++) {
            memcpy(rowDest, rowSource, width*pixelsize);
            rowDest += width*pixelsize;
            rowSource += stride;
    }

    stride = width*pixelsize;
    gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                     std::min(GetAddressAlignment((ptrdiff_t)newPixels),
                              GetAddressAlignment((ptrdiff_t)stride)));
    gl->fTexSubImage2D(target,
                       level,
                       xoffset,
                       yoffset,
                       width,
                       height,
                       format,
                       type,
                       newPixels);
    delete [] newPixels;
    gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);
}

static void
TexSubImage2DHelper(GLContext *gl,
                    GLenum target, GLint level,
                    GLint xoffset, GLint yoffset,
                    GLsizei width, GLsizei height, GLsizei stride,
                    GLint pixelsize, GLenum format,
                    GLenum type, const GLvoid* pixels)
{
    if (gl->IsGLES2()) {
        if (stride == width * pixelsize) {
            gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                             std::min(GetAddressAlignment((ptrdiff_t)pixels),
                                      GetAddressAlignment((ptrdiff_t)stride)));
            gl->fTexSubImage2D(target,
                               level,
                               xoffset,
                               yoffset,
                               width,
                               height,
                               format,
                               type,
                               pixels);
            gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);
        } else if (gl->IsExtensionSupported(GLContext::EXT_unpack_subimage)) {
            TexSubImage2DWithUnpackSubimageGLES(gl, target, level, xoffset, yoffset,
                                                width, height, stride,
                                                pixelsize, format, type, pixels);

        } else {
            TexSubImage2DWithoutUnpackSubimage(gl, target, level, xoffset, yoffset,
                                              width, height, stride,
                                              pixelsize, format, type, pixels);
        }
    } else {
        // desktop GL (non-ES) path
        gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                         std::min(GetAddressAlignment((ptrdiff_t)pixels),
                                  GetAddressAlignment((ptrdiff_t)stride)));
        int rowLength = stride/pixelsize;
        gl->fPixelStorei(LOCAL_GL_UNPACK_ROW_LENGTH, rowLength);
        gl->fTexSubImage2D(target,
                           level,
                           xoffset,
                           yoffset,
                           width,
                           height,
                           format,
                           type,
                           pixels);
        gl->fPixelStorei(LOCAL_GL_UNPACK_ROW_LENGTH, 0);
        gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);
    }
}

static void
TexImage2DHelper(GLContext *gl,
                 GLenum target, GLint level, GLint internalformat,
                 GLsizei width, GLsizei height, GLsizei stride,
                 GLint pixelsize, GLint border, GLenum format,
                 GLenum type, const GLvoid *pixels)
{
    if (gl->IsGLES2()) {

        NS_ASSERTION(format == (GLenum)internalformat,
                    "format and internalformat not the same for glTexImage2D on GLES2");

        if (!CanUploadNonPowerOfTwo(gl)
            && (stride != width * pixelsize
            || !IsPowerOfTwo(width)
            || !IsPowerOfTwo(height))) {

            // Pad out texture width and height to the next power of two
            // as we don't support/want non power of two texture uploads
            GLsizei paddedWidth = NextPowerOfTwo(width);
            GLsizei paddedHeight = NextPowerOfTwo(height);

            GLvoid* paddedPixels = new unsigned char[paddedWidth * paddedHeight * pixelsize];

            // Pad out texture data to be in a POT sized buffer for uploading to
            // a POT sized texture
            CopyAndPadTextureData(pixels, paddedPixels, width, height,
                                  paddedWidth, paddedHeight, stride, pixelsize);

            gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                             std::min(GetAddressAlignment((ptrdiff_t)paddedPixels),
                                      GetAddressAlignment((ptrdiff_t)paddedWidth * pixelsize)));
            gl->fTexImage2D(target,
                            border,
                            internalformat,
                            paddedWidth,
                            paddedHeight,
                            border,
                            format,
                            type,
                            paddedPixels);
            gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);

            delete[] static_cast<unsigned char*>(paddedPixels);
            return;
        }

        if (stride == width * pixelsize) {
            gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                             std::min(GetAddressAlignment((ptrdiff_t)pixels),
                                      GetAddressAlignment((ptrdiff_t)stride)));
            gl->fTexImage2D(target,
                            border,
                            internalformat,
                            width,
                            height,
                            border,
                            format,
                            type,
                            pixels);
            gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);
        } else {
            // Use GLES-specific workarounds for GL_UNPACK_ROW_LENGTH; these are
            // implemented in TexSubImage2D.
            gl->fTexImage2D(target,
                            border,
                            internalformat,
                            width,
                            height,
                            border,
                            format,
                            type,
                            nullptr);
            TexSubImage2DHelper(gl,
                                target,
                                level,
                                0,
                                0,
                                width,
                                height,
                                stride,
                                pixelsize,
                                format,
                                type,
                                pixels);
        }
    } else {
        // desktop GL (non-ES) path

        gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT,
                         std::min(GetAddressAlignment((ptrdiff_t)pixels),
                                  GetAddressAlignment((ptrdiff_t)stride)));
        int rowLength = stride/pixelsize;
        gl->fPixelStorei(LOCAL_GL_UNPACK_ROW_LENGTH, rowLength);
        gl->fTexImage2D(target,
                        level,
                        internalformat,
                        width,
                        height,
                        border,
                        format,
                        type,
                        pixels);
        gl->fPixelStorei(LOCAL_GL_UNPACK_ROW_LENGTH, 0);
        gl->fPixelStorei(LOCAL_GL_UNPACK_ALIGNMENT, 4);
    }
}

SurfaceFormat
UploadImageDataToTexture(GLContext* gl,
                         unsigned char* aData,
                         int32_t aStride,
                         SurfaceFormat aFormat,
                         const nsIntRegion& aDstRegion,
                         GLuint& aTexture,
                         bool aOverwrite,
                         bool aPixelBuffer,
                         GLenum aTextureUnit,
                         GLenum aTextureTarget)
{
    bool textureInited = aOverwrite ? false : true;
    gl->MakeCurrent();
    gl->fActiveTexture(aTextureUnit);

    if (!aTexture) {
        gl->fGenTextures(1, &aTexture);
        gl->fBindTexture(aTextureTarget, aTexture);
        gl->fTexParameteri(aTextureTarget,
                           LOCAL_GL_TEXTURE_MIN_FILTER,
                           LOCAL_GL_LINEAR);
        gl->fTexParameteri(aTextureTarget,
                           LOCAL_GL_TEXTURE_MAG_FILTER,
                           LOCAL_GL_LINEAR);
        gl->fTexParameteri(aTextureTarget,
                           LOCAL_GL_TEXTURE_WRAP_S,
                           LOCAL_GL_CLAMP_TO_EDGE);
        gl->fTexParameteri(aTextureTarget,
                           LOCAL_GL_TEXTURE_WRAP_T,
                           LOCAL_GL_CLAMP_TO_EDGE);
        textureInited = false;
    } else {
        gl->fBindTexture(aTextureTarget, aTexture);
    }

    nsIntRegion paintRegion;
    if (!textureInited) {
        paintRegion = nsIntRegion(aDstRegion.GetBounds());
    } else {
        paintRegion = aDstRegion;
    }

    GLenum format = 0;
    GLenum internalFormat = 0;
    GLenum type = 0;
    int32_t pixelSize = BytesPerPixel(aFormat);
    SurfaceFormat surfaceFormat = gfx::SurfaceFormat::UNKNOWN;

    MOZ_ASSERT(gl->GetPreferredARGB32Format() == LOCAL_GL_BGRA ||
               gl->GetPreferredARGB32Format() == LOCAL_GL_RGBA);
    switch (aFormat) {
        case SurfaceFormat::B8G8R8A8:
            if (gl->GetPreferredARGB32Format() == LOCAL_GL_BGRA) {
              format = LOCAL_GL_BGRA;
              surfaceFormat = SurfaceFormat::R8G8B8A8;
              type = LOCAL_GL_UNSIGNED_INT_8_8_8_8_REV;
            } else {
              format = LOCAL_GL_RGBA;
              surfaceFormat = SurfaceFormat::B8G8R8A8;
              type = LOCAL_GL_UNSIGNED_BYTE;
            }
            internalFormat = LOCAL_GL_RGBA;
            break;
        case SurfaceFormat::B8G8R8X8:
            // Treat BGRX surfaces as BGRA except for the surface
            // format used.
            if (gl->GetPreferredARGB32Format() == LOCAL_GL_BGRA) {
              format = LOCAL_GL_BGRA;
              surfaceFormat = SurfaceFormat::R8G8B8X8;
              type = LOCAL_GL_UNSIGNED_INT_8_8_8_8_REV;
            } else {
              format = LOCAL_GL_RGBA;
              surfaceFormat = SurfaceFormat::B8G8R8X8;
              type = LOCAL_GL_UNSIGNED_BYTE;
            }
            internalFormat = LOCAL_GL_RGBA;
            break;
        case SurfaceFormat::R5G6B5:
            internalFormat = format = LOCAL_GL_RGB;
            type = LOCAL_GL_UNSIGNED_SHORT_5_6_5;
            surfaceFormat = SurfaceFormat::R5G6B5;
            break;
        case SurfaceFormat::A8:
            internalFormat = format = LOCAL_GL_LUMINANCE;
            type = LOCAL_GL_UNSIGNED_BYTE;
            // We don't have a specific luminance shader
            surfaceFormat = SurfaceFormat::A8;
            break;
        default:
            NS_ASSERTION(false, "Unhandled image surface format!");
    }

    nsIntRegionRectIterator iter(paintRegion);
    const nsIntRect *iterRect;

    // Top left point of the region's bounding rectangle.
    nsIntPoint topLeft = paintRegion.GetBounds().TopLeft();

    while ((iterRect = iter.Next())) {
        // The inital data pointer is at the top left point of the region's
        // bounding rectangle. We need to find the offset of this rect
        // within the region and adjust the data pointer accordingly.
        unsigned char *rectData =
            aData + DataOffset(iterRect->TopLeft() - topLeft, aStride, aFormat);

        NS_ASSERTION(textureInited || (iterRect->x == 0 && iterRect->y == 0),
                     "Must be uploading to the origin when we don't have an existing texture");

        if (textureInited && CanUploadSubTextures(gl)) {
            TexSubImage2DHelper(gl,
                                aTextureTarget,
                                0,
                                iterRect->x,
                                iterRect->y,
                                iterRect->width,
                                iterRect->height,
                                aStride,
                                pixelSize,
                                format,
                                type,
                                rectData);
        } else {
            TexImage2DHelper(gl,
                             aTextureTarget,
                             0,
                             internalFormat,
                             iterRect->width,
                             iterRect->height,
                             aStride,
                             pixelSize,
                             0,
                             format,
                             type,
                             rectData);
        }

    }

    return surfaceFormat;
}

SurfaceFormat
UploadSurfaceToTexture(GLContext* gl,
                       DataSourceSurface *aSurface,
                       const nsIntRegion& aDstRegion,
                       GLuint& aTexture,
                       bool aOverwrite,
                       const nsIntPoint& aSrcPoint,
                       bool aPixelBuffer,
                       GLenum aTextureUnit,
                       GLenum aTextureTarget)
{
    unsigned char* data = aPixelBuffer ? nullptr : aSurface->GetData();
    int32_t stride = aSurface->Stride();
    SurfaceFormat format = aSurface->GetFormat();
    data += DataOffset(aSrcPoint, stride, format);
    return UploadImageDataToTexture(gl, data, stride, format,
                                    aDstRegion, aTexture, aOverwrite,
                                    aPixelBuffer, aTextureUnit,
                                    aTextureTarget);
}

bool
CanUploadNonPowerOfTwo(GLContext* gl)
{
    if (!gl->WorkAroundDriverBugs())
        return true;

    // Some GPUs driver crash when uploading non power of two 565 textures.
    return gl->Renderer() != GLRenderer::Adreno200 &&
           gl->Renderer() != GLRenderer::Adreno205;
}

}
}