flypig.co.uk

Before I get in to my diary entry today I want to first mention a change that will be happening in a week. Between Saturday 4th May and Sunday 12th May I'll be attending a conference, which means I'll be pretty tied up throughout. So as I've done a couple of times in the past already, I'll be taking a break from these dev diaries during that period. It'll basically be a break of a week and a couple of days (Saturday to Sunday), but I'll get back on track again straight after again.

With that announcement out of the way, let's get on with the entry for today.

I'm currently digging around in GLScreenBuffer::Swap(), the idea being to put in some code to print out pixel colours. I added some yesterday, but it returned only zeros. Looking at the code the reason became clear pretty quickly: the ReadPixels() method from SharedSurface was being called, which always fails and just returns false.

Even though the surface is shown in the debugger as a SharedSurface it's supposed to be a SharedSurface_EGLImage. But SharedSurface_EGLImage doesn't override the ReadPixels() method, so even if it were, it still wouldn't make any difference.

But although the code I wrote yesterday doesn't work I'm hoping it can still lay the groundwork for something that will. What I've done is implemented that missing override, so that now there's a ReadPixels() method in SharedSurface_EGLImage that looks like this:

bool SharedSurface_EGLImage::ReadPixels(GLint x, GLint y, GLsizei width, 
    GLsizei height,
                        GLenum format, GLenum type, GLvoid* pixels) {
  const auto& gl = GLContextEGL::Cast(mDesc.gl);

  gl->fBindTexture(LOCAL_GL_TEXTURE_2D, mProdTex);
  gl->fGetTexImage(LOCAL_GL_TEXTURE_2D, 0, format, type, pixels);
  return true;
}

The idea is that this will bind the texture associated with the surface and then read the pixels from the currently bound texture. Sadly when I run it, the app crashes with a backtrace that looks like this:

Thread 37 "Compositor" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f178ee7e0 (LWP 2120)]
0x0000000000000000 in ?? ()
(gdb) bt
#0  0x0000000000000000 in ?? ()
#1  0x0000007ff112e5f0 in mozilla::gl::GLContext::fGetTexImage (
    img=0x7ed81ea5d0, type=5121, format=6408, level=0, target=3553, 
    this=0x7ed819aa50)
    at ${PROJECT}/gecko-dev/gfx/gl/GLContext.h:1331
#2  mozilla::gl::SharedSurface_EGLImage::ReadPixels (this=<optimized out>, 
    x=<optimized out>, y=<optimized out>, width=<optimized out>, 
    height=<optimized out>, format=6408, type=5121, pixels=0x7ed81ea5d0)
    at ${PROJECT}/gecko-dev/gfx/gl/SharedSurfaceEGL.cpp:178
#3  0x0000007ff1101a34 in mozilla::gl::GLScreenBuffer::ReadPixels (
    this=this@entry=0x7ed4002e00, x=x@entry=540, y=y@entry=1260, 
    width=width@entry=1, 
    height=height@entry=1, format=format@entry=6408, type=type@entry=5121, 
    pixels=pixels@entry=0x7ed81ea5d0)
    at ${PROJECT}/gecko-dev/gfx/gl/GLScreenBuffer.cpp:462
#4  0x0000007ff11027e4 in mozilla::gl::GLScreenBuffer::Swap (
    this=this@entry=0x7ed4002e00, size=...)
    at ${PROJECT}/gecko-dev/gfx/gl/GLScreenBuffer.cpp:531
#5  0x0000007ff3663f48 in mozilla::gl::GLScreenBuffer::PublishFrame (size=..., 
    this=0x7ed4002e00)
    at ${PROJECT}/obj-build-mer-qt-xr/dist/include/GLScreenBuffer.h:229
#6  mozilla::embedlite::EmbedLiteCompositorBridgeParent::
    PresentOffscreenSurface (this=0x7fc4b422d0)
    at ${PROJECT}/gecko-dev/mobile/sailfishos/embedthread/
    EmbedLiteCompositorBridgeParent.cpp:191
#7  0x0000007ff367d3ac in mozilla::embedlite::nsWindow::PostRender (
    this=0x7fc4b09280, aContext=<optimized out>)
    at ${PROJECT}/gecko-dev/mobile/sailfishos/embedshared/nsWindow.cpp:248
#8  0x0000007ff2a62664 in mozilla::widget::InProcessCompositorWidget::
    PostRender (this=0x7fc4d0eb40, aContext=0x7f178ed7e8)
    at ${PROJECT}/gecko-dev/widget/InProcessCompositorWidget.cpp:60
#9  0x0000007ff128d1c8 in mozilla::layers::LayerManagerComposite::Render (
    this=this@entry=0x7ed81b7160, aInvalidRegion=..., aOpaqueRegion=...)
    at ${PROJECT}/obj-build-mer-qt-xr/dist/include/mozilla/layers/Compositor.h:
    575
#10 0x0000007ff128d644 in mozilla::layers::LayerManagerComposite::
    UpdateAndRender (this=this@entry=0x7ed81b7160)
    at ${PROJECT}/gecko-dev/gfx/layers/composite/LayerManagerComposite.cpp:657
#11 0x0000007ff128d9f4 in mozilla::layers::LayerManagerComposite::
    EndTransaction (this=this@entry=0x7ed81b7160, aTimeStamp=..., 
    aFlags=aFlags@entry=mozilla::layers::LayerManager::END_DEFAULT)
    at ${PROJECT}/gecko-dev/gfx/layers/composite/LayerManagerComposite.cpp:572
#12 0x0000007ff12cf174 in mozilla::layers::CompositorBridgeParent::
    CompositeToTarget (this=0x7fc4b422d0, aId=..., aTarget=0x0, 
    aRect=<optimized out>)
    at ${PROJECT}/obj-build-mer-qt-xr/dist/include/mozilla/RefPtr.h:313
#13 0x0000007ff3663c48 in mozilla::embedlite::EmbedLiteCompositorBridgeParent::
    CompositeToDefaultTarget (this=0x7fc4b422d0, aId=...)
    at ${PROJECT}/gecko-dev/mobile/sailfishos/embedthread/
    EmbedLiteCompositorBridgeParent.cpp:165
#14 0x0000007ff12b48d0 in mozilla::layers::CompositorVsyncScheduler::Composite (
    this=0x7fc4c9c050, aVsyncEvent=...)
    at ${PROJECT}/gecko-dev/gfx/layers/ipc/CompositorVsyncScheduler.cpp:256
#15 0x0000007ff12acd50 in mozilla::detail::RunnableMethodArguments<mozilla::
    VsyncEvent>::applyImpl<mozilla::layers::CompositorVsyncScheduler, void (
    mozilla::layers::CompositorVsyncScheduler::*)(mozilla::VsyncEvent const&), 
    StoreCopyPassByConstLRef<mozilla::VsyncEvent>, 0ul> (args=..., m=<optimized 
    out>, 
    o=<optimized out>) at ${PROJECT}/obj-build-mer-qt-xr/dist/include/
    nsThreadUtils.h:887
[...]
#27 0x0000007ff6a0289c in thread_start () at ../sysdeps/unix/sysv/linux/aarch64/
    clone.S:78
(gdb) 

Although the crash isn't ideal it does tell me one useful thing: the method is getting called, which means that the surface really is a SharedSurface_EGLImage after all. That's at least one part confirmed. The reason for the crash is less clear, given that all the inputs to the fGetTexImage() method look sensible. The reason must be because the fGetTexImage() function pointer is null. And indeed it is:

(gdb) frame 1
#1  0x0000007ff112e5f0 in mozilla::gl::GLContext::fGetTexImage (
    img=0x7ed81ea5d0, type=5121, format=6408, level=0, target=3553, 
    this=0x7ed819aa50)
    at ${PROJECT}/gecko-dev/gfx/gl/GLContext.h:1331
1331    in ${PROJECT}/gecko-dev/gfx/gl/GLContext.h
(gdb) p mSymbols.fGetTexImage
$5 = (void (*)(GLenum, GLint, GLenum, GLenum, GLvoid *)) 0x0
(gdb) 

Okay. I'm not sure what I can do about that. It could be that the reason is because there's no fGetTexImage() in the GLContext::InitImpl() method, which is used to set up all of these function pointers. It's worth noting that the call to fBindTexture() that happens right before the call to fGetTexImage() works just fine and fBindTexture() can be found in a table at the top of the GLContext.cpp file, so that adds weight to the claim. If that's the case, maybe I can add fGetTexImage in?

Unfortunately I don't have time to do that this evening, but it does at least give me a very clear plan for what I can be working on tomorrow morning.

If you'd like to read any of my other gecko diary entries, they're all available on my Gecko-dev Diary page.