Re: [PATCH] xsm: removing inclusion of byteorder.h

["Daniel P. Smith" <dpsmith@xxxxxxxxxxxxxxxxxxxx>]

On 8/17/23 08:17, Jason Andryuk wrote:
> On Wed, Aug 16, 2023 at 3:05 PM Daniel P. Smith
> <dpsmith@xxxxxxxxxxxxxxxxxxxx> wrote:
> > This is to complement patch 'common: move Linux-inherited fixed width type
> > decls to common header' by removing the unnecessary include of
> > 'asm/byteorder.h'. In the process of removing these includes, the ordering was
> > corrected to comply with current coding style.
> >
> > Signed-off-by: Daniel P. Smith <dpsmith@xxxxxxxxxxxxxxxxxxxx>
> > ---
> >   xen/xsm/flask/ss/avtab.c       | 3 +--
> >   xen/xsm/flask/ss/conditional.c | 9 ++++-----
> >   xen/xsm/flask/ss/ebitmap.c     | 8 ++++----
> >   xen/xsm/flask/ss/policydb.c    | 9 ++++-----
> >   4 files changed, 13 insertions(+), 16 deletions(-)
> These four files all call le32_to_cpu(), so AFAICT, you are trading an
> explicit dependency for an implicit one.  Is that what you want?
In fact, no. You comment here caused me to go back and try to see why 
Jan said they needed to move. I realized there was a misunderstanding 
about his commit message. He was not saying later, at a point in the 
future, they need to be moved, but that as a result of his change the 
inclusion ordering had to be changed. The former is what I understood 
and thus attempted a build without them to see if it would work. I was 
not expecting that conversions would have been able to be implicitly 
resolved.
Since I already rebased to latest stable, I can resubmit a v2 with 
correcting the inclusion order but with byteoder.h inclusion remaining. 
Though that would render this down to simply a style change patch. Would 
that still be acceptable Jan?
v/r,
dps