<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hello Jonas,<br>
<br>
I tested your program and on my system (Fedora 19) on an Evergreen
(Radeon HD 5850) and I have the same issue indeed.<br>
Here is my result :<br>
<br>
Mesa 9.2.4 (from F19 repo) => Data does NOT match up!<br>
Mesa 9.2.5 => Data does NOT match up!<br>
Mesa 10.0.5 => Data does NOT match up!<br>
Mesa 10.1.0 => Data matches.<br>
<br>
So this is fixed in newer version.<br>
That said, Mesa 9.2 is not supported anymore and I really don't know
if there will be a new Mesa 10.0.x release given the imminence of
Mesa 10.2.<br>
If yes, I can bisect and you can open a bug.<br>
<br>
Mesa-dev, any new 10.0.x release planned ?<br>
<br>
Regards.<br>
<br>
Benjamin<br>
<br>
<div class="moz-cite-prefix">Le 06/05/2014 13:40, Jonas Kulla a
écrit :<br>
</div>
<blockquote
cite="mid:CAK0i_738fq3jD3Z36B=bVjbJ5JPt2tHNn7wQz0+oUpJMYieweA@mail.gmail.com"
type="cite">
<div dir="ltr">Hello list,
<div><br>
</div>
<div>after about 3 days of debugging, I was able to isolate a
rather weird behavior in Mesa GL.</div>
<div>The gist of it is the following: When I create a buffer
object and allocate uninitilaized</div>
<div>memory for it (glBufferData() with nullptr as 'data'),
then glCopyBufferSubData() data into</div>
<div>it from another buffer object, then subsequently fill a
part of it with glBufferSubData(),</div>
<div>this new data isn't visible to the buffer object. In fact,
it seems that the SubData'ed bytes</div>
<div>are completely lost. Any further data uploads however work
as expected. I will attach</div>
<div>a small C test case below that demonstrates this behavior.</div>
<div><br>
</div>
<div>I realize that I am working with an old Mesa release (on
Fedora 19), but I'm afraid of</div>
<div>upgrading my system to the newest distro release as I might
break my working environment.</div>
<div>That's why I would like to kindly ask if someone could
verify that this problem still persists</div>
<div>on the newest Mesa code, in which case I would go ahead and
file a bug report. At the</div>
<div>same time, maybe someone could spot a critical mistake in
my code that would explain</div>
<div>this strange behavior I'm seeing. I think the code paths
I'm hitting here in the driver are</div>
<div>sufficiently obscure though.</div>
<div><br>
</div>
<div>I should probably mention that my card is a Mobility Radeon
HD 3650 (ie. r600).</div>
<div><br>
</div>
<div>Here's the code sample (you can replace the GL setup code
with your own):</div>
<div><br>
</div>
<div>
<div>#include <stdlib.h></div>
<div>#include <string.h></div>
<div> </div>
<div>#include <GL/glew.h></div>
<div>#include <SDL2/SDL.h></div>
<div> </div>
<div>static SDL_Window *win;</div>
<div>static SDL_GLContext *ctx;</div>
<div> </div>
<div>void setupGL()</div>
<div>{</div>
<div> SDL_Init(SDL_INIT_VIDEO);</div>
<div> win = SDL_CreateWindow("CopyBufferBug",
SDL_WINDOWPOS_UNDEFINED, SDL_WINDOWPOS_UNDEFINED, 64, 64,
SDL_WINDOW_OPENGL);</div>
<div> ctx = SDL_GL_CreateContext(win);</div>
<div> glewInit();</div>
<div>}</div>
<div> </div>
<div>static void teardownGL()</div>
<div>{</div>
<div> SDL_GL_DeleteContext(ctx);</div>
<div> SDL_DestroyWindow(win);</div>
<div> </div>
<div> SDL_Quit();</div>
<div>}</div>
<div> </div>
<div>int main(int argc, char *argv[])</div>
<div>{</div>
<div> setupGL();</div>
<div> </div>
<div> /* These don't matter I think */</div>
<div>
#define BLOCK_SIZE 128</div>
<div> #define BUFFER1_SIZE BLOCK_SIZE</div>
<div> #define BUFFER2_SIZE BLOCK_SIZE</div>
<div> #define BUFFER1_TARGET GL_COPY_READ_BUFFER</div>
<div> #define BUFFER2_TARGET GL_COPY_WRITE_BUFFER</div>
<div> #define BUFFER1_USAGE GL_DYNAMIC_DRAW</div>
<div> #define BUFFER2_USAGE GL_DYNAMIC_DRAW</div>
<div> </div>
<div> GLuint buffers[2];</div>
<div> glGenBuffers(2, buffers);</div>
<div> </div>
<div>
/* We allocate both buffers with undefined memory */</div>
<div> glBindBuffer(BUFFER1_TARGET, buffers[0]);</div>
<div> glBufferData(BUFFER1_TARGET, BUFFER1_SIZE, 0,
BUFFER1_USAGE);</div>
<div> </div>
<div>
glBindBuffer(BUFFER2_TARGET, buffers[1]);</div>
<div> glBufferData(BUFFER2_TARGET, BUFFER2_SIZE, 0,
BUFFER2_USAGE);</div>
<div> </div>
<div> /* Then copy (undefined) bytes from the first
into the second buffer */</div>
<div> /* Note: If I comment this line out, everything
works */</div>
<div> glCopyBufferSubData(BUFFER1_TARGET,
BUFFER2_TARGET, 0, 0, BUFFER1_SIZE);</div>
<div> </div>
<div> /* Generate random string */</div>
<div> FILE *rand = fopen("/dev/urandom", "r");</div>
<div> char data[BLOCK_SIZE];</div>
<div> fread(data, 1, sizeof(data), rand);</div>
<div> fclose(rand);</div>
<div> </div>
<div>
/* We fill the second buffer with defined data */</div>
<div> /* Note: If I execute this call twice (just copy
paste the line), everything works */</div>
<div> glBufferSubData(BUFFER2_TARGET, 0, sizeof(data),
data);</div>
<div> </div>
<div> /* Then download it again to compare its contents
against our test string */</div>
<div> char data2[BLOCK_SIZE];</div>
<div> glGetBufferSubData(BUFFER2_TARGET, 0,
sizeof(data2), data2);</div>
<div> </div>
<div> if (memcmp(data, data2, sizeof(data)))</div>
<div> printf("Data does NOT match up!\n");</div>
<div> else</div>
<div> printf("Data matches.\n");</div>
<div> </div>
<div> glDeleteBuffers(2, buffers);</div>
<div> </div>
<div> teardownGL();</div>
<div> </div>
<div> return 0;</div>
<div>}</div>
</div>
<div><br>
</div>
<div>Thank you very much for your time.</div>
<div>Jonas</div>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
mesa-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:mesa-users@lists.freedesktop.org">mesa-users@lists.freedesktop.org</a>
<a class="moz-txt-link-freetext" href="http://lists.freedesktop.org/mailman/listinfo/mesa-users">http://lists.freedesktop.org/mailman/listinfo/mesa-users</a>
</pre>
</blockquote>
<br>
</body>
</html>