[BusyBox] asm/posix_types.h (for arm) causing problems in loo p.c

[Deacon, Keith]

I have the latest BusyBox changes as of today around 1pm my time - I'll
resync just to make sure nothing got missed.  I have the latest loop.c and
associated Makefile.in that he changed 3 days ago.

Or, are you saying Erik updated bits/kernel_types.h in uClibc to fix the
problem?  (Been a week or so since I synced that). I'll go look.

Otherwise, the bits/kernel_types.h (at least in my uClibc toolchain) is
overriding asm/posix_types.h and causing __kernel_old_dev_t to not be
defined.

...Keith
  

-----Original Message-----
From: Manuel Novoa III [mailto:mjn3 at codepoet.org] 
Sent: Monday, February 09, 2004 4:06 PM
To: Deacon, Keith
Cc: 'busybox at mail.busybox.net'
Subject: Re: [BusyBox] asm/posix_types.h (for arm) causing problems in
loop.c

On Mon, Feb 09, 2004 at 03:59:43PM -0600, Deacon, Keith wrote:
> Greetings,
> Trying to get busybox compiled for ARM with 2.6 headers.  I have working
> compiler built with 2.6 headers for ARM with uClibc.  While compiling
> loop.c, I got the error about __kernel_old_dev_t not being defined.
Looking
> into the busybox CVS tree, I saw a patch that looked like it would fix the
> problem.  I pulled it in, but it didn't work.  After tracing it out, it
> appears that the #include <stdio.h> is eventually including
> bits/kernel_types.h from the toolchain - which has the following info in
the
> header:
> 
> /* Note that we use the exact same include guard #define names
>  * as asm/posix_types.h.  This will avoid gratuitous conflicts
>  * with the posix_types.h kernel header, and will ensure that
>  * our private content, and not the kernel header, will win.
>  *  -Erik
>  */
> 
> However, bits/kernel_types.h doesn't define __kernel_old_dev_t and is
> included before asm/posix_types.h.  Because of the include guard, the
> asm/posix_types.h is basically ignored.  May be a problem with how the
> uClibc tool chain goes together or how busybox tries to find the proper
> definition.  I coded around it for the moment and will revisit it later
when
> I get a chance.  Anyone else hit this?
> 
> Thanks
> ...Keith
> 

Erik fixed that in cvs 3 days ago.

Manuel