On Aug 19, 2016 18:05, "Eric Engestrom" <<a href="mailto:eric.engestrom@imgtec.com">eric.engestrom@imgtec.com</a>> wrote: > > On Fri, Aug 19, 2016 at 09:04:14AM -0700, Jason Ekstrand wrote: > > Not providing a path allows the ICD to work on multi-arch systems but > > breaks it if you install anywhere other than /usr/lib. Given that users > > may be installing locally in .local or similar, we probably do want to > > provide a filename. Distros can carry a revert of this commit if they want > > an intel_icd.json file without the path. > > > > Signed-off-by: Jason Ekstrand <<a href="mailto:jason@jlekstrand.net">jason@jlekstrand.net</a>> > > Cc: Mark Janes <<a href="mailto:mark.a.janes@intel.com">mark.a.janes@intel.com</a>> > > I have one question (below), but this patch is good regardless: > Reviewed-by: Eric Engestrom <<a href="mailto:eric.engestrom@imgtec.com">eric.engestrom@imgtec.com</a>> > > > --- > > src/intel/vulkan/.gitignore | 1 + > > src/intel/vulkan/Makefile.am | 7 ++++++- > > src/intel/vulkan/{intel_icd.json => <a href="http://intel_icd.json.in">intel_icd.json.in</a>} | 2 +- > > 3 files changed, 8 insertions(+), 2 deletions(-) > > rename src/intel/vulkan/{intel_icd.json => <a href="http://intel_icd.json.in">intel_icd.json.in</a>} (59%) > > > > diff --git a/src/intel/vulkan/.gitignore b/src/intel/vulkan/.gitignore > > index bde5cd8..a099ff6 100644 > > --- a/src/intel/vulkan/.gitignore > > +++ b/src/intel/vulkan/.gitignore > > @@ -3,3 +3,4 @@ > > /anv_entrypoints.h > > /anv_timestamp.h > > /dev_icd.json > > +/intel_icd.json > > diff --git a/src/intel/vulkan/Makefile.am b/src/intel/vulkan/Makefile.am > > index ad0148d..9fef960 100644 > > --- a/src/intel/vulkan/Makefile.am > > +++ b/src/intel/vulkan/Makefile.am > > @@ -141,7 +141,7 @@ anv_timestamp.h: > > $(AM_V_GEN) echo "#define ANV_TIMESTAMP \"$(TIMESTAMP_CMD)\"" > $@ > > > > BUILT_SOURCES = $(VULKAN_GENERATED_FILES) > > -CLEANFILES = $(BUILT_SOURCES) dev_icd.json > > +CLEANFILES = $(BUILT_SOURCES) dev_icd.json intel_icd.json > > EXTRA_DIST = \ > > $(top_srcdir)/include/vulkan/vk_icd.h \ > > anv_entrypoints_gen.py \ > > @@ -170,6 +170,11 @@ dev_icd.json : <a href="http://dev_icd.json.in">dev_icd.json.in</a> > > -e "s#@build_libdir@#${abs_top_builddir}/${LIB_DIR}#" \ > > < $(srcdir)/<a href="http://dev_icd.json.in">dev_icd.json.in</a> > $@ > > > > +intel_icd.json : <a href="http://intel_icd.json.in">intel_icd.json.in</a> > > + $(AM_V_GEN) $(SED) \ > > + -e "s#@install_libdir@#${libdir}#" \ > > + < $(srcdir)/<a href="http://intel_icd.json.in">intel_icd.json.in</a> > $@ > > I think I may have already asked when dev_icd.json was added, but why > use a relative path for the dependency and a full path when reading it? > Why not use the full path for the dep and read `< $<` instead? No good reason that I know of other than copy+paste of the one above. We could probably make that change. Emil? > > + > > # Libvulkan with dummy gem. Used for unit tests. > > libvulkan_test_la_SOURCES = $(VULKAN_GEM_STUB_FILES) > > libvulkan_test_la_LIBADD = $(VULKAN_LIB_DEPS) -lX11-xcb > > diff --git a/src/intel/vulkan/intel_icd.json b/src/intel/vulkan/<a href="http://intel_icd.json.in">intel_icd.json.in</a> > > similarity index 59% > > rename from src/intel/vulkan/intel_icd.json > > rename to src/intel/vulkan/<a href="http://intel_icd.json.in">intel_icd.json.in</a> > > index 277c14e..d9b363a 100644 > > --- a/src/intel/vulkan/intel_icd.json > > +++ b/src/intel/vulkan/<a href="http://intel_icd.json.in">intel_icd.json.in</a> > > @@ -1,7 +1,7 @@ > > { > > "file_format_version": "1.0.0", > > "ICD": { > > - "library_path": "libvulkan_intel.so", > > + "library_path": "@install_libdir@/libvulkan_intel.so", > > "abi_versions": "1.0.3" > > } > > } > > -- > > 2.5.0.400.gff86faf