Op 22-07-13 07:08, [email protected] schreef: > From: Christopher James Halse Rogers <[email protected]> > > This is missing too much functionality to be usefully appiled, but the > skeleton is here and the APIs it relies on are sufficiently stable. Sending > to the list for extra visibility and for preliminary comments. > > We're sufficiently different to Wayland to make sharing code difficult, > but I think there's probably some scope for common code in output handling, > and possibly in GLX's interaction with the underlying compositor. > > Signed-off-by: Christopher James Halse Rogers > <[email protected]> > --- > configure.ac | 11 ++ > hw/xfree86/Makefile.am | 9 +- > hw/xfree86/common/xf86Config.c | 12 ++ > hw/xfree86/common/xf86Globals.c | 3 + > hw/xfree86/common/xf86Init.c | 20 +++ > hw/xfree86/common/xf86Priv.h | 3 + > hw/xfree86/xmir/Makefile.am | 26 ++++ > hw/xfree86/xmir/xmir-output.c | 225 ++++++++++++++++++++++++++++++ > hw/xfree86/xmir/xmir-private.h | 84 +++++++++++ > hw/xfree86/xmir/xmir-thread-proxy.c | 115 +++++++++++++++ > hw/xfree86/xmir/xmir-window.c | 271 > ++++++++++++++++++++++++++++++++++++ > hw/xfree86/xmir/xmir.c | 217 +++++++++++++++++++++++++++++ > hw/xfree86/xmir/xmir.h | 92 ++++++++++++ > include/xorg-server.h.in | 3 + > test/Makefile.am | 9 +- > test/xmir-thread-proxy.c | 154 ++++++++++++++++++++ > 16 files changed, 1251 insertions(+), 3 deletions(-) > create mode 100644 hw/xfree86/xmir/Makefile.am > create mode 100644 hw/xfree86/xmir/xmir-output.c > create mode 100644 hw/xfree86/xmir/xmir-private.h > create mode 100644 hw/xfree86/xmir/xmir-thread-proxy.c > create mode 100644 hw/xfree86/xmir/xmir-window.c > create mode 100644 hw/xfree86/xmir/xmir.c > create mode 100644 hw/xfree86/xmir/xmir.h > create mode 100644 test/xmir-thread-proxy.c
A few quick notes.. - Please use the prime import calls in the drivers instead of duplicating the functionality. - Instead of overriding AuthMagic2 in every driver, this should be done in xmir. - The RFC patch for nouveau mir compositor support should be sent to xorg-devel too. - You added a NVHasMirSupport function, but the xserver should already know whether it has a drm fd for the pci-id before calling the probe function, so that check should not be in the ddx. - Any chance of getting rid of the separate mir thread and using xorg's polling mechanism? ~Maarten _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
