Hi Anthony,
"Anthony PERARD" <anthony.per...@vates.tech> writes:
> On Sat, Apr 05, 2025 at 03:30:49AM +0000, Volodymyr Babchuk wrote:
>> --- a/xen/Rules.mk
>> +++ b/xen/Rules.mk
>> @@ -133,18 +133,19 @@ $(filter %.init.o,$(obj-y) $(obj-bin-y) $(extra-y)):
>> CFLAGS-y += -DINIT_SECTIONS
>>
>> non-init-objects = $(filter-out %.init.o, $(obj-y) $(obj-bin-y) $(extra-y))
>>
>> -ifeq ($(CONFIG_COVERAGE),y)
>
> This removes an "ifeq ()", so you probably need to remove and "endif"
> somewhere else, which doesn't appear in this snippet.
Yes, I'm sorry, it just didn't got into the snippet. I wanted to discuss
approach only, so this is not the final version.
>> ifeq ($(CONFIG_CC_IS_CLANG),y)
>> - COV_FLAGS := -fprofile-instr-generate -fcoverage-mapping
>> + cov-flags-$(CONFIG_COVERAGE) := -fprofile-instr-generate
>> -fcoverage-mapping
>
> If you do this assignment like that, it would be better to make sure
> $(cov-flags-y) is initialised properly, that is have a:
>
> cov-flags-y :=
>
> before the first conditional assignment, then have all conditional
> assignment be +=.
Sure.
>
>> else
>> - COV_FLAGS := -fprofile-arcs -ftest-coverage
>> + cov-flags-$(CONFIG_COVERAGE) := -fprofile-arcs -ftest-coverage
>> + cov-flags-$(CONFIG_CONDITION_COVERAGE) += -fcondition-coverage
>
> What happen if CONFIG_CONDITION_COVERAGE=y but CONFIG_COVERAGE=n ?
Kconfig ensures that this is impossible:
config CONDITION_COVERAGE
bool "Condition coverage support"
depends on COVERAGE && CC_HAS_MCDC
I believe, this is enough, and we don't need a separate check on
Makefile level.
--
WBR, Volodymyr
