[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <57C2DB15.3050103@bfs.de>
Date: Sun, 28 Aug 2016 14:37:41 +0200
From: walter harms <wharms@....de>
To: Mikko Rapeli <mikko.rapeli@....fi>
CC: linux-kernel@...r.kernel.org, linux-hams@...r.kernel.org
Subject: Re: [PATCH v05 69/72] uapi rose.h: glibc netrose/rose.h header file
compatibility fixes
Am 28.08.2016 07:59, schrieb Mikko Rapeli:
> On Fri, Aug 26, 2016 at 05:38:00PM +0200, walter harms wrote:
>> perhaps this not tested snipped would make sure that
>> you have included linux/rose.h ?
>>
>> #ifndef ROSE_KERNEL_H
>> #include <linux/rose.h>
>> #endif
>>
>> #include <netrose/rose.h>
>
> Sorry, I did not quite get this.
>
> <linux/rose.h> has conflicting definitions with glibc <netrose/rose.h>.
> The patches fixes the uapi headers <linux/rose.h> so that it hides
> definitions if <netrose/rose.h> from glibc was already included.
>
> -Mikko
no problem,
so far i understand the conflict arise because you need to include
linux/rose.h before netrose/rose.h
My suggestion was to add a check for linux/rose.h (ROSE_KERNEL_H)
and include it if not already done. (basicly this should work out
of the box but it seems a problem here).
I have no idea if you need to cover other cases but in my
understanding userspace programms should not include kernel headers.
re,
wh
Powered by blists - more mailing lists