On Mon, Nov 3, 2014 at 9:06 PM, Benjamin Bellec <b.bel...@gmail.com> wrote:
> Here is an example trace, if somebody wants to give a look : > > https://drive.google.com/file/d/0B7D2Y0QXFND2ODF2QlExc2w0Y0U/view?usp=sharing > > Thanks. I took a quick look. This call draws the weapon's geometry to depth-map 329154 glDrawRangeElements(mode = GL_TRIANGLES, start = 0, end = 7581, count = 18585, type = GL_UNSIGNED_SHORT, indices = NULL) and then later on on this call the weapon is lit: 333371 glDrawRangeElements(mode = GL_TRIANGLES, start = 0, end = 7581, count = 18585, type = GL_UNSIGNED_SHORT, indices = NULL) In both cases I'm seeing weird geometry on the weapon. AFAICT, all the state being bound on this calls is pretty standard stuff -- nothing that could cause problems to apitrace. The only explanation I can think of so far is that some glMapBufferRange(GL_MAP_WRITE_BIT) done by the application are in write-combining memory, and that when apitrace's wrapper tries to copy that memory into the trace file it doesn't get the data that the application wrote, but something wrong, which then gets baked into the trace file, appearing on subsequent replays. I'll need to investigate this more. Jose
_______________________________________________ apitrace mailing list apitrace@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/apitrace
