I request a 3.3 core context and call glEnable(GL_FRAMEBUFFER_SRGB), create textures using GL_SRGB8 or GL_SRGB8_ALPHA8 as internal format and pow 2.2 on colors sent to gpu with the uniform functions. This looks fine on the intel drivers for windows and on nvidia drivers on windows and linux. If I record with apitrace it looks correct when replaying the trace.
Please attach a test case and/or screenshots.
Sorry I forgot. This is what I believe it's supposed to look like (nvidia gtx 970 on windows 10): https://dl.dropboxusercontent.com/u/244444/mesa%20srgb/srgb%20nvidia%20windows.png This is what it looks like on mesa: https://dl.dropboxusercontent.com/u/244444/mesa%20srgb/with%20srgb.png This is what it looks like if I call glDisable(GL_FRAMEBUFFER_SRGB) (it looks even darker) https://dl.dropboxusercontent.com/u/244444/mesa%20srgb/disable%20gl_framebuffer_srgb.png This is without any of the steps described in the OP: https://dl.dropboxusercontent.com/u/244444/mesa%20srgb/no%20%20srgb.png This is the apitrace recorded on linux: https://dl.dropboxusercontent.com/u/244444/mesa%20srgb/luajit.trace There are some other issues that I haven't looked into yet such as there being a black border around the text and the snow being very dark so ignore those as they might be a user error.
@CapsAdmin is this related to this bug? https://bugs.freedesktop.org/show_bug.cgi?id=92759
Created attachment 119375 [details] attachment-1672-0.html Possibly. I don't have a setup to test the latest intel drivers with patches. On Tue, Nov 3, 2015 at 1:16 PM, <bugzilla-daemon@freedesktop.org> wrote: > *Comment # 3 <https://bugs.freedesktop.org/show_bug.cgi?id=91817#c3> on > bug 91817 <https://bugs.freedesktop.org/show_bug.cgi?id=91817> from Deve > <deveee@gmail.com> * > > @CapsAdmin is this related to this bug?https://bugs.freedesktop.org/show_bug.cgi?id=92759 > > ------------------------------ > You are receiving this mail because: > > - You reported the bug. > >
It seems very likely that this bug is the same as bug 92759. Although the original description says that it's rendering to GL_SRGB8 textures which wouldn't have the bug, the trace enables GL_FRAMEBUFFER_SRGB throughout the entire the trace so presumably it would be enabled for the final writes to the framebuffer as well. The calls to glXChooseVisual don't include the GLX_ALPHA_SIZE parameter so it seems quite likely that it's picking up an RGB format without sRGB capabilities. I haven't been able to test it with the trace because glretrace seems to ignore the visual from glXChooseVisual and always requests an RGBA visual which means it would end up with an sRGB-capable visual regardless. Thanks for the bug report. *** This bug has been marked as a duplicate of bug 92759 ***
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.