[haiku-commits] Re: r41089 - in haiku/trunk/src/add-ons/accelerants/radeon_hd: . atombios atombios/includes
- From: Alexander von Gluck IV <kallisti5@xxxxxxxxxxx>
- To: haiku-commits@xxxxxxxxxxxxx
- Date: Wed, 23 Mar 2011 21:08:43 -0500
On Wed, 23 Mar 2011 08:53:21 +0100
Philippe Houdoin <philippe.houdoin@xxxxxxxxx> wrote:
> > +SubInclude HAIKU_TOP src add-ons accelerants radeon_hd atombios ;
> > +
> > SetSubDirSupportedPlatformsBeOSCompatible ;
>
> Usually, we put SubInclude at Jamfiles bottom, not at top.
Fixed, sorry. My Jamfile experience is a little limited.
> > +#define DISABLE_EASF
>
> To avoid touching this "vendor" file, you can instead define
> DISABLE_EASF symbol in atombios/Jamfile :
>
Fixed by reverting original file and making Jamfile adjustments.
>
> > +++
> > haiku/trunk/src/add-ons/accelerants/radeon_hd/atombios/includes/CD_Common_Types.h
> > 2011-03-22 23:00:27 UTC (rev 41089) @@ -37,34 +37,11 @@ #ifndef
> > _COMMON_TYPES_H_ #define _COMMON_TYPES_H_
> >
> > - #ifndef LINUX
>
> Why removing FGL_LINUX, !LINUX and other conditional code here?
> You could keep this code and instead use a #if defined(__HAIKU__)
> #else <everything else> #endif.
> The more compact and less impacted your changes are in vendor code,
> the easier it is to upgrade new release in the future.
You are right here, I was in too much of a hurry doing this. Restored
original file and added a HAIKU ifdef
> > -#include <regsdef.h> //This important file is dynamically
> > generated based on the ASIC!!!! +#include "regsdef.h" //This
> > important file is dynamically generated based on the ASIC!!!!
>
> IIRC, this could also be avoid by forcing via Jamfile to search an
> explicit include file locally.
> But I may remember wrongly, as I can't quickly find how.
This seemed incorrect anyway. For now I'll leave this change unless
there are any arguments.
> > -#if (PARSER_TYPE==DRIVER_TYPE_PARSER)
> > +#if (PARSER_TYPE==DRIVER_TYPE_PARSER) && !defined(__HAIKU__)
> > #pragma pack(pop)
> > #endif
>
> Why this was causing issue without !__HAIKU_ check?
No clue:
" #pragma pack(push[, id], <n>) is not supported on this target "
I think the pack(push), pack(1), and pack(pop) should work.. but it is
reported as not being supported on this target.. maybe my static
library target is causing it?
> Anyway, thanks for working on Radeon HD support.
> Yes, I'm biased while saying that ;-) (Radeon HD4970 owner).
Thanks for the help!
-- Alex
Other related posts:
