This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: Again Big endian ARM HAL_WRITE_UINT8 problem
- From: Andrew Lunn <andrew at lunn dot ch>
- To: Roland Ca?ebohm <roland dot cassebohm at visionsystems dot de>
- Cc: ecos-discuss at sources dot redhat dot com
- Date: Sun, 12 Oct 2003 14:05:58 +0200
- Subject: Re: [ECOS] Again Big endian ARM HAL_WRITE_UINT8 problem
- References: <200310101827.58265.roland.cassebohm@visionsystems.de>
On Fri, Oct 10, 2003 at 06:27:58PM +0200, Roland Ca?ebohm wrote:
> Hi,
>
> there is still the problem, that the HAL_WRITE_UINT8() macro
> on bigendian ARM hardware does not work.
>
> In the following message is a solution suggested, which would
> be good IMO.
> http://sources.redhat.com/ml/ecos-discuss/2002-07/msg00415.html
>
> I don't know which platform have to use the macros for bigendian
> like they are, maybe pid target? I uses the macro and can be set
> to bigendian.
>
> I would like to make a patch, but if I use a flag like
> HAL_IO_MACROS_ADDRESS_MUNGING a platform who needs the macros as
> they are have to define the flag not to be broken.
> Or should I define a flag like HAL_IO_MACROS_NO_ADDRESS_MUNGING
> for platforms like snds which don't need the address munging?
We always try to keep the default behavior. In this case you could
break a few targets without knowing it which is always bad.
I would use the flag HAL_IO_MACROS_NO_ADDRESS_MUNGING and it the HALs
you know need it.
Andrew
--
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss