Re: [PATCH v4 05/30] bpftool: Avoid adding EXTRA_CFLAGS to HOST_CFLAGS

[Quentin Monnet <qmo@kernel.org>]

2026-03-13 16:24 UTC+0000 ~ Leo Yan <leo.yan@arm.com>
> On Fri, Mar 13, 2026 at 04:07:27PM +0000, Quentin Monnet wrote:
> 
> [...]
> 
>>>>> +
>>>>> +# This must be done before appending EXTRA_CFLAGS to CFLAGS to avoid
>>>>> +# including flags that are not applicable to the host compiler.
>>>>> +HOST_CFLAGS := $(subst -I$(LIBBPF_INCLUDE),-I$(LIBBPF_BOOTSTRAP_INCLUDE),\
>>>>> +               $(subst $(CLANG_CROSS_FLAGS),,$(CFLAGS)))
>>>>> +
>>>>>  ifneq ($(EXTRA_CFLAGS),)
>>>>>  CFLAGS += $(EXTRA_CFLAGS)
>>>>>  endif
>>>>> @@ -88,8 +94,6 @@ ifneq ($(EXTRA_LDFLAGS),)
>>>>>  LDFLAGS += $(EXTRA_LDFLAGS)
>>>>>  endif
>>>>>
>>>>> -HOST_CFLAGS := $(subst -I$(LIBBPF_INCLUDE),-I$(LIBBPF_BOOTSTRAP_INCLUDE),\
>>>>> -               $(subst $(CLANG_CROSS_FLAGS),,$(CFLAGS)))
>>>>>  HOST_LDFLAGS := $(LDFLAGS)
>>>>
>>>> Should HOST_LDFLAGS be similarly moved? Otherwise, it seems target
>>>> flags in EXTRA_LDFLAGS will be added for the host.
>>>
>>> Though this series is irrelevant to LDFLAGS, I am fine to move
>>> HOST_LDFLAGS together.
>>
>>
>> If we move the HOST_LDFLAGS earlier so that they no longer receive the
>> EXTRA_LDFLAGS, then we lose the possibility to use any EXTRA variable to
>> pass additional flags to HOST_LDFLAGS. So OK to move, but maybe add a
>> "HOST_LDFLAGS += $(HOST_EXTRALDFLAGS)" to it for consistency with the
>> processing of HOST_CFLAGS?
> 
> Adding HOST_EXTRALDFLAGS seems a bit over designed to me, as there is
> no use case for it currently.  To avoid complexity, let's keep this
> patch as it is, we can add HOST_EXTRALDFLAGS if later have requirement.
> 
> P.s. I saw that you already replied the same comment in the previous
> version.  I should have confirmed at that time, sorry for the duplicate
> discussion.


That's fine :) - And agreed, it's probably best to leave it as-is for
now. Thanks!

Quentin