Re: [PATCH v2 01/15] meson: *-user: only descend into *-user when configured

[Kyle Evans]

On Mon, Oct 18, 2021 at 12:02 AM Warner Losh <imp@bsdimp.com> wrote:
>
>
>
> On Sun, Oct 17, 2021 at 10:29 PM Warner Losh <imp@bsdimp.com> wrote:
>>
>>
>>
>> On Sun, Oct 17, 2021 at 9:43 PM Kyle Evans <kevans@freebsd.org> wrote:
>>>
>>> On Fri, Oct 8, 2021 at 6:15 PM Warner Losh <imp@bsdimp.com> wrote:
>>> >
>>> > To increase flexibility, only descend into *-user when that is
>>> > configured. This allows *-user to selectively include directories based
>>> > on the host OS which may not exist on all hosts. Adopt Paolo's
>>> > suggestion of checking the configuration in the directories that know
>>> > about the configuration.
>>> >
>>> > Message-Id: <20210926220103.1721355-2-f4bug@amsat.org>
>>> > Message-Id: <20210926220103.1721355-3-f4bug@amsat.org>
>>> > Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>> > Signed-off-by: Warner Losh <wlosh@bsdimp.com>
>>> > Acked-by: Paolo Bonzini <pbonzinni@redhat.com>
>>> >
>>> > Sponsored by:           Netflix
>>> > ---
>>> >  bsd-user/meson.build   | 4 ++++
>>> >  linux-user/meson.build | 4 ++++
>>> >  meson.build            | 3 +--
>>> >  3 files changed, 9 insertions(+), 2 deletions(-)
>>> >
>>> > diff --git a/bsd-user/meson.build b/bsd-user/meson.build
>>> > index 0369549340..243fb78930 100644
>>> > --- a/bsd-user/meson.build
>>> > +++ b/bsd-user/meson.build
>>> > @@ -1,3 +1,7 @@
>>> > +if not config_target.has_key('CONFIG_BSD_USER')
>>> > +   subdir_done()
>>> > +endif
>>> > +
>>> >  bsd_user_ss.add(files(
>>> >    'bsdload.c',
>>> >    'elfload.c',
>>> > diff --git a/linux-user/meson.build b/linux-user/meson.build
>>> > index 9549f81682..602255a3d6 100644
>>> > --- a/linux-user/meson.build
>>> > +++ b/linux-user/meson.build
>>> > @@ -1,3 +1,7 @@
>>> > +if not config_target.has_key('CONFIG_LINUX_USER')
>>> > +   subdir_done()
>>> > +endif
>>> > +
>>> >  linux_user_ss.add(files(
>>> >    'elfload.c',
>>> >    'exit.c',
>>> > diff --git a/meson.build b/meson.build
>>> > index 99a0a3e689..1f2da5f7d9 100644
>>> > --- a/meson.build
>>> > +++ b/meson.build
>>> > @@ -2303,10 +2303,9 @@ subdir('ebpf')
>>> >
>>> >  common_ss.add(libbpf)
>>> >
>>> > -bsd_user_ss.add(files('gdbstub.c'))
>>> >  specific_ss.add_all(when: 'CONFIG_BSD_USER', if_true: bsd_user_ss)
>>> >
>>> > -linux_user_ss.add(files('gdbstub.c', 'thunk.c'))
>>> > +linux_user_ss.add(files('thunk.c'))
>>> >  specific_ss.add_all(when: 'CONFIG_LINUX_USER', if_true: linux_user_ss)
>>> >
>>> >  # needed for fuzzing binaries
>>> > --
>>> > 2.32.0
>>> >
>>>
>>> I don't understand the gdbstub.c removal  here; don't we still want to
>>> be compiling it in, just only if the appropriate
>>> CONFIG_{BSD,LINUX}_USER knob is set? I note that it doesn't appear to
>>> be added in individual *-user/meson.build, I assume it's uncommon to
>>> add in ../foo.c in meson-land...
>>
>>
>> It's added to specific_ss at line 2536
>> specific_ss.add(files('cpu.c', 'disas.c', 'gdbstub.c'), capstone)
>>
>> so we don't need to add it again here.
>
>
> I've also confirmed that it's built as both 
> libqemu-i386-bsd-user.fa.p/gdbstub.c.o
> and libqemu-x86_64-bsd-user.fa.p/gdbstub.c.o, which is what I'd expect given
> the current upstream supported architectures are only i386 and x86_64.
>
> Warner

Ah, ok, thanks! So that looks like a kind-of tangential cleanup, but
related enough that it makes sense.

Reviewed-by: Kyle Evans <kevans@FreeBSD.org>