On Tuesday, 2017-12-12 11:58:14 -0800, Dylan Baker wrote: > Quoting Eric Engestrom (2017-12-12 06:15:58) > > On Monday, 2017-12-11 11:50:01 -0800, Dylan Baker wrote: > > > Quoting Eric Engestrom (2017-12-11 07:55:30) > > > > On Friday, 2017-12-08 16:27:21 -0800, Dylan Baker wrote: > > > > > Signed-off-by: Dylan Baker <dylanx.c.ba...@intel.com> > > > > > > > > Should come after the current 4/4, but > > > > Reviewed-by: Eric Engestrom <eric.engest...@imgtec.com> > > > > > > > > > --- > > > > > src/gallium/drivers/r600/meson.build | 7 +++++-- > > > > > 1 file changed, 5 insertions(+), 2 deletions(-) > > > > > > > > > > diff --git a/src/gallium/drivers/r600/meson.build > > > > > b/src/gallium/drivers/r600/meson.build > > > > > index 2132dbb33ad..5899518a2e8 100644 > > > > > --- a/src/gallium/drivers/r600/meson.build > > > > > +++ b/src/gallium/drivers/r600/meson.build > > > > > @@ -113,12 +113,15 @@ egd_tables_h = custom_target( > > > > > capture : true, > > > > > ) > > > > > > > > > > -# TODO: compute defines > > > > > +r600_c_args = [] > > > > > +if with_gallium_opencl > > > > > + r600_c_args += '-DHAVE_OPENCL' > > > > > +endif > > > > > > > > > > libr600 = static_library( > > > > > 'r600', > > > > > [files_r600, egd_tables_h], > > > > > - c_args : [c_vis_args], > > > > > + c_args : [c_vis_args, r600_c_args], > > > > > cpp_args : [cpp_vis_args], > > > > > include_directories : [ > > > > > inc_src, inc_include, inc_gallium, inc_gallium_aux, > > > > > inc_amd_common, > > > > > -- > > > > > 2.15.1 > > > > > > > > > > > r600 needs this to work with clover at all. Since it's so minimal, how > > > would you > > > feel about just squashing this into 4/4? > > > > You mean r600 would be broken after 4/4 if 3/4 wasn't applied? > > I'm OK with squashing them if so. > > > > My issue was just that with this order, the build would be broken > > between 3/4 and 4/4 because `with_gallium_opencl` doesn't exist yet. > > Right, the current order is broken. Unless I've misunderstood the code without > this patch the code will compile, but trying to run OpenCL workloads with r600 > will fail. Maybe it's fine to put that in a follow up patch?
Sure, squash with 4/4 or swap them, both work for me, with a small preference on the latter ;) _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
