[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 Brian Paul bri...@vmware.com changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #9 from Brian Paul bri...@vmware.com --- Fixed by commit e2091f64cb9ea79f3b51c353ed9facc03ec5690a -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 --- Comment #2 from Brian Paul bri...@vmware.com --- Created attachment 75059 -- https://bugs.freedesktop.org/attachment.cgi?id=75059action=edit llvmpipe texture size patch Can you test the attached patch? -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 --- Comment #3 from Roland Scheidegger srol...@vmware.com --- (In reply to comment #0) The following(ish) code produces an assertion failure using llvmpipe libGL from Mesa 9.0.2: Near as I can tell, the call responsible for storing that state with the pbuffer is this code: http://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/state_trackers/glx/ xlib/xm_api.c?id=369e46888904c6d379b8b477d9242cff1608e30e#n460 A call to get_drawable_size retrieves the size of the pbuffer drawable into local variables width and height, but these values never make it into the XMesaBuffer structure. This isn't llvmpipe specific right? That call you mention indeed looks a bit odd. The function comment explicitly states Width/Height of the new buffer will be 0 so I don't know why it calls get_drawable_size() there in the first place (probably some leftover from older code). It looks like for pixmaps/windows that information will be filled out later at MakeCurrent time, which will eventually call xmesa_check_buffer_size() which will fill it in. However, for pbuffers this is a noop. My guess is it should be filled out in XMesaCreatePBuffer() instead since pbuffers have fixed size (?). -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 --- Comment #4 from Brian Paul bri...@vmware.com --- Created attachment 75065 -- https://bugs.freedesktop.org/attachment.cgi?id=75065action=edit Initialize the buffer's size in create_xmesa_buffer() Here's another patch to try. It uses the results of get_drawable_size() to initialize the buffer's dimensions. I've run other test programs and it seems to be OK. I think my previous patch for llvmpipe is also needed since it's possible to create a pbuffer of size 0 x 0. -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 --- Comment #6 from Brian Paul bri...@vmware.com --- (In reply to comment #5) Created attachment 75068 [details] [review] another attempt to fix pbuffer initialization Hmm is it legal to use XGetGeometry() with pbuffers? Not normally, but in the glx/xlib code we create a dummy pixmap for each pbuffer so that we have an XID that we can pass around. I think something like this patch would be better. That would be fine too. It's what I first tried. Not sure if guarding against zero-sized buffers in drivers is needed. Might be but there are other instances where we hack up such windows to have width/height of 1 for that reason so we don't have to do it in drivers. I hacked up a test for a 0x0 surface. Softpipe worked but the llvmpipe assertion failed. I guess I'd consider the llvmpipe change to be a defensive coding check. One less way for llvmpipe to fail is good thing. -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 --- Comment #7 from Roland Scheidegger srol...@vmware.com --- (In reply to comment #6) (In reply to comment #5) Created attachment 75068 [details] [review] [review] another attempt to fix pbuffer initialization Hmm is it legal to use XGetGeometry() with pbuffers? Not normally, but in the glx/xlib code we create a dummy pixmap for each pbuffer so that we have an XID that we can pass around. Ah ok then that should be fine too. I think though in this case the function comment should be updated too (which is why I was thinking this function shouldn't really set up size for whatever reason). I think something like this patch would be better. That would be fine too. It's what I first tried. Not sure if guarding against zero-sized buffers in drivers is needed. Might be but there are other instances where we hack up such windows to have width/height of 1 for that reason so we don't have to do it in drivers. I hacked up a test for a 0x0 surface. Softpipe worked but the llvmpipe assertion failed. I guess I'd consider the llvmpipe change to be a defensive coding check. One less way for llvmpipe to fail is good thing. Yeah probably. Though zero-sized resources are a pretty nasty thing. -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [Bug 61012] alloc_layout_array tx * ty assertion failure when making pbuffer current
https://bugs.freedesktop.org/show_bug.cgi?id=61012 --- Comment #1 from Brian Crowell freedesk...@fluggo.com --- Created attachment 75010 -- https://bugs.freedesktop.org/attachment.cgi?id=75010action=edit Test case Hopefully minimal test case. Use the llvmpipe libGL. -- You are receiving this mail because: You are the assignee for the bug. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
