Re: [PATCH v16 3/4] tools/tests: introduce unit tests for domain ID allocator

[dmkhn@xxxxxxxxx]

On Mon, Aug 25, 2025 at 03:47:39PM +0200, Anthony PERARD wrote:

Thanks for review!

Will address in the next revision.
Please see some responses below.

> On Tue, Aug 12, 2025 at 10:30:50PM +0000, dmkhn@xxxxxxxxx wrote:
> > diff --git a/tools/tests/domid/.gitignore b/tools/tests/domid/.gitignore
> > new file mode 100644
> > index 000000000000..0e02715159c2
> > --- /dev/null
> > +++ b/tools/tests/domid/.gitignore
> > @@ -0,0 +1,3 @@
> > +*.o
> 
> "*.o" is already in the .gitignore at the root of the project. I don't
> think it's useful here.

Ack.

> 
> > +generated
> > +test-domid
> > diff --git a/tools/tests/domid/Makefile b/tools/tests/domid/Makefile
> > new file mode 100644
> > index 000000000000..0a124a8bfc76
> > --- /dev/null
> > +++ b/tools/tests/domid/Makefile
> > @@ -0,0 +1,84 @@
> > +# SPDX-License-Identifier: GPL-2.0-only
> > +#
> > +# Unit tests for domain ID allocator.
> > +#
> > +# Copyright 2025 Ford Motor Company
> > +
> > +XEN_ROOT=$(CURDIR)/../../..
> > +include $(XEN_ROOT)/tools/Rules.mk
> > +
> > +TESTS := test-domid
> > +
> > +strip-list = $(sort $(strip $(foreach x,$(1),$(strip $(x)))))
> 
> What's that macro for? Also, what's a "list"?
> 
> > +define list-c-headers
> > +$(shell sed -n -r \
> 
> Could you use "-E" instead of "-r"? (-r might not work on FreeBSD)

re: FreeBSD: I've found there's a dedicated pipeline for Xen on FreeBSD
(cirrus CI), but I did not figure how to trigger it, will appreciate help
with that.

> 
> > +    's/^[ \t]*# *include[ \t]*[<"]([^">]+)[">].*/\1/p' $(1) 2>/dev/null)
> > +endef
> > +
> > +define emit-harness-nested-rule
> > +$(1): $(CURDIR)/harness.h
> > +   mkdir -p $$(dir $$@)
> 
> You can use $(@D) instead of $(dir $@). The only difference is a /
> not present at the end.
> 
> > +   ln -sf $$^ $$@
> 
> This should use $<, I don't think the command is going to work if
> there's multiple prerequisite.
> 
> > +endef
> > +
> > +define emit-harness-rules
> > +ifneq ($(strip $(3)),)
> 
> How many time do you need to call $(strip) ?
> Also, I think I would prefer to have $(if $(strip $(3)), [the rest])
> rather than actually evaluating code and generating code that we already
> know is isn't going to be executed.
> 
> > +$(foreach h,$(3),$(call emit-harness-nested-rule,$(CURDIR)/generated/$(h)))
> > +vpath $(1) $(2)
> > +$(1:.c=.o): $(addprefix $(CURDIR)/generated/,$(3))
> > +endif
> > +endef
> 
> This macro fails if there's more than one "#include" in "domid.c".
> 
> And if there's no "#include" in "domid.c", then Make doesn't know how to
> make "domid.o" for "test-domid".
> 
> > +
> > +define vpath-with-harness-deps
> > +$(call emit-harness-rules,$(1),$(2),\
> > +    $(call strip-list,$(call list-c-headers,$(2)$(1))))
> > +endef
> > +
> > +.PHONY: all
> > +all: $(TESTS)
> > +
> > +.PHONY: run
> > +run: $(TESTS)
> > +   $(foreach t,$(TESTS),./$(t);)
> 
> This recipe doesn't work as expected. You need `set -e` or only the last
> tests count.
> 
> > +
> > +.PHONY: clean
> > +clean:
> > +   $(RM) -rf $(CURDIR)/generated
> 
> $(RM) already contain the '-f' option, no need to add it a second time.
> 
> Also, we expected Make to run all commands in recipe from $(CURDIR), so
> adding $(CURDIR) is unnecessary, could potentially be an issue.
> 
> > +   $(RM) -- *.o $(TESTS) $(DEPS_RM)
> > +
> > +.PHONY: distclean
> > +distclean: clean
> > +   $(RM) -- *~
> > +
> > +.PHONY: install
> > +install: all
> > +   $(INSTALL_DIR) $(DESTDIR)$(LIBEXEC)/tests
> > +   $(INSTALL_PROG) test-domid $(DESTDIR)$(LIBEXEC)/tests
> > +
> > +.PHONY: uninstall
> > +uninstall:
> > +   $(RM) -- $(DESTDIR)$(LIBEXEC)/tests/test-domid
> > +
> > +CFLAGS += -D__XEN_TOOLS__
> > +# find-next-bit.c
> > +CFLAGS += '-DEXPORT_SYMBOL(x)=' \
> > +          -Dfind_first_bit \
> > +          -Dfind_first_zero_bit \
> > +          -Dfind_next_bit \
> > +          -Dfind_next_bit_le \
> > +          -Dfind_next_zero_bit_le
> > +CFLAGS += $(APPEND_CFLAGS)
> > +CFLAGS += $(CFLAGS_xeninclude)
> > +CFLAGS += -I./generated/
> > +
> > +LDFLAGS += $(APPEND_LDFLAGS)
> > +
> > +vpath find-next-bit.c $(XEN_ROOT)/xen/lib/
> > +# Ubuntu {16,18}.04 need single eval at the call site.
> 
> I'd rather see a comment about what's the macro is about rather that a
> comment some Linux distribution. Our target is GNU Make 3.80, without
> regards to a particular distribution. (Also I don't think it's useful to
> point out that `eval` is needed for older version of Make, at least in
> our project.)

Ack.

> 
> > +$(eval $(call vpath-with-harness-deps,domid.c,$(XEN_ROOT)/xen/common/))
> > +
> > +test-domid: domid.o find-next-bit.o test-domid.o
> > +   $(CC) $^ -o $@ $(LDFLAGS)
> > +
> > +-include $(DEPS_INCLUDE)
> > diff --git a/tools/tests/domid/test-domid.c b/tools/tests/domid/test-domid.c
> > new file mode 100644
> > index 000000000000..51a88a6a9550
> > --- /dev/null
> > +++ b/tools/tests/domid/test-domid.c
> > @@ -0,0 +1,93 @@
> > +/* SPDX-License-Identifier: GPL-2.0-only */
> > +/*
> > + * Unit tests for domain ID allocator.
> > + *
> > + * Copyright 2025 Ford Motor Company
> > + */
> > +
> > +#include "harness.h"
> > +
> > +#define verify(exp, fmt, args...) do { \
> > +    if ( !(exp) ) \
> > +        printf(fmt, ## args); \
> > +    assert(exp); \
> 
> Relying on assert() for the test isn't wise. It's useful for developing
> and debugging because it calls abort(), but they can easily be get rid of,
> by simply building with -DNDEBUG. Could you maybe replace it with exit()
> since you already check the condition?

Yep, will do.

> 
> 
> Thanks,
> 
> --
> Anthony PERARD
>