[PATCH] include/compat: produce stubs for headers not otherwise generated

Jan Beulich posted 1 patch 1 year, 3 months ago
Failed in applying to current master (apply log)
[PATCH] include/compat: produce stubs for headers not otherwise generated
Posted by Jan Beulich 1 year, 3 months ago
Public headers can include other public headers. Such interdependencies
are retained in their compat counterparts. Since some compat headers are
generated only in certain configurations, the referenced headers still
need to exist. The lack thereof was observed with hvm/hvm_op.h needing
trace.h, where generation of the latter depends on TRACEBUFFER=y. Make
empty stubs in such cases (as generating the extra headers is relatively
slow and hence better to avoid). Changes to .config and incrementally
(re-)building is covered by the respective .*.cmd then no longer
matching the command to be used, resulting in the necessary re-creation
of the (possibly stub) header.

Reported-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
There may be differing views on which commit this actually fixes, hence
I'd prefer to omit a Fixes: tag here. The issue was exposed by
4c5edd2449bc ("xen: Drop $ARCH/trace.h").

--- a/xen/include/Makefile
+++ b/xen/include/Makefile
@@ -34,6 +34,8 @@ headers-$(CONFIG_TRACEBUFFER) += compat/
 headers-$(CONFIG_XENOPROF) += compat/xenoprof.h
 headers-$(CONFIG_XSM_FLASK) += compat/xsm/flask_op.h
 
+headers-n := $(filter-out $(headers-y),$(headers-n) $(headers-))
+
 cppflags-y                := -include public/xen-compat.h -DXEN_GENERATING_COMPAT_HEADERS
 cppflags-$(CONFIG_X86)    += -m32
 
@@ -43,13 +45,16 @@ public-$(CONFIG_X86) := $(wildcard $(src
 public-$(CONFIG_ARM) := $(wildcard $(srcdir)/public/arch-arm/*.h $(srcdir)/public/arch-arm/*/*.h)
 
 .PHONY: all
-all: $(addprefix $(obj)/,$(headers-y))
+all: $(addprefix $(obj)/,$(headers-y) $(headers-n))
 
 quiet_cmd_compat_h = GEN     $@
 cmd_compat_h = \
     $(PYTHON) $(srctree)/tools/compat-build-header.py <$< $(patsubst $(obj)/%,%,$@) >>$@.new; \
     mv -f $@.new $@
 
+quiet_cmd_stub_h = GEN     $@
+cmd_stub_h = echo '/* empty */' >$@
+
 quiet_cmd_compat_i = CPP     $@
 cmd_compat_i = $(CPP) $(filter-out -Wa$(comma)% -include %/include/xen/config.h,$(XEN_CFLAGS)) $(cppflags-y) -o $@ $<
 
@@ -69,6 +74,13 @@ targets += $(headers-y)
 $(obj)/compat/%.h: $(obj)/compat/%.i $(srctree)/tools/compat-build-header.py FORCE
 	$(call if_changed,compat_h)
 
+# Placeholders may be needed in case files in $(headers-y) include files we
+# don't otherwise generate.  Real dependencies would need spelling out explicitly,
+# for them to appear in $(headers-y) instead.
+targets += $(headers-n)
+$(addprefix $(obj)/,$(headers-n)): FORCE
+	$(call if_changed,stub_h)
+
 .PRECIOUS: $(obj)/compat/%.i
 targets += $(patsubst %.h, %.i, $(headers-y))
 $(obj)/compat/%.i: $(obj)/compat/%.c FORCE
Re: [PATCH] include/compat: produce stubs for headers not otherwise generated
Posted by Anthony PERARD 1 year, 3 months ago
On Thu, Jan 12, 2023 at 10:17:47AM +0100, Jan Beulich wrote:
> Public headers can include other public headers. Such interdependencies
> are retained in their compat counterparts. Since some compat headers are
> generated only in certain configurations, the referenced headers still
> need to exist. The lack thereof was observed with hvm/hvm_op.h needing
> trace.h, where generation of the latter depends on TRACEBUFFER=y. Make
> empty stubs in such cases (as generating the extra headers is relatively
> slow and hence better to avoid). Changes to .config and incrementally
> (re-)building is covered by the respective .*.cmd then no longer
> matching the command to be used, resulting in the necessary re-creation
> of the (possibly stub) header.
> 
> Reported-by: Andrew Cooper <andrew.cooper3@citrix.com>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>

This patch also takes care of "removing" compat headers that are no
longer needed due to change in .config, which is good to have (even if
this only remove the content in it).

Thanks,

-- 
Anthony PERARD