[libvirt PATCH] meson: Check usability of linux/kvm.h

[Michal Prívozník]

On 4/27/23 11:41, Andrea Bolognani wrote:
> This fixes cross-building in some scenarios.
> 
> Specifically, when building for armv7l on x86_64, has_header()
> will see the x86_64 version of the header and consider it usable.
> Later, when an attempt is made to actually use it, the compiler
> will quickly realize that things can't quite work.
> 
> The reason why we haven't hit this in our CI is that we only ever
> install the foreign version of header files. When building the
> Debian package, however, some of the Debian-specific tooling will
> bring in the native version of the Linux headers in addition to
> the foreign one, causing Meson to misreport the header's
> availability status.
> 
> Checking for its actual usability, as opposed to mere presence,
> is enough to make things work correctly in all cases.
> 
> https://bugs.debian.org/1024504
> 
> Suggested-by: Helmut Grohne <helmut at subdivi.de>
> Signed-off-by: Andrea Bolognani <abologna at redhat.com>
> ---
>  meson.build | 13 ++++++++++++-
>  1 file changed, 12 insertions(+), 1 deletion(-)
> 
> diff --git a/meson.build b/meson.build
> index d35d5e076b..21a4bd5b37 100644
> --- a/meson.build
> +++ b/meson.build
> @@ -614,7 +614,6 @@ headers = [
>    'asm/hwcap.h',
>    'ifaddrs.h',
>    'libtasn1.h',
> -  'linux/kvm.h',
>    'mntent.h',
>    'net/ethernet.h',
>    'net/if.h',
> @@ -635,12 +634,24 @@ if host_machine.system() == 'freebsd'
>    headers += 'libutil.h'
>  endif
>  
> +# headers for which we need to check actual usability. in most
> +# cases, checking for presence is enough (and it's way faster)
> +check_headers = [
> +  'linux/kvm.h',
> +]
> +
>  foreach name : headers
>    if cc.has_header(name)
>      conf.set('WITH_ at 0@'.format(name.underscorify().to_upper()), 1)
>    endif
>  endforeach
>  
> +foreach name : check_headers
> +  if cc.check_header(name)
> +    conf.set('WITH_ at 0@'.format(name.underscorify().to_upper()), 1)
> +  endif
> +endforeach
> +
>  # check for kernel header required by src/util/virnetdevbridge.c
>  if host_machine.system() == 'linux'
>    if not cc.has_header('linux/sockios.h')


One could argue, that the semantics of has_header() is broken then. What
good is there to check if the host has a header file when compiler
rejects it subsequently?

But leaving meson aside, shouldn't we just use check_header() for every
header file then? I mean, this fixes this particular instance of the
problem, but can't we hit it again with say ifaddrs.h or any other
header file on the list?

Michal