Re: Fwd: [PATCH 1/2] Framebuffer split

[Michal Suchanek]

2009/8/13 Vladimir 'phcoder' Serbinenko <address@hidden>:
>> Considering that vbe.c and sdl.c currently aren't affected a lot
>> whether there is or there isn't encapsulation in place, I'm ok tih
>> encapsulating it but if any driver needs the breach of encapsulation
>> it will be broken and I'll post no opposition to it.
>> I'll do the encapsulation tomorrow.
> Patch attached. One incremental version (for review) and one complete
> (for patching)

The patch as presented here works for me.

However, I wonder some of the changes:

In the latest round of patches struct grub_video_fbrender_target
declaration was needlessly and somewhat illogically moved from
video_fb.h to fbfill.h.

I am not sure if this is to hide the structure from other code but it
logically belongs to video_fb.h.

In grub_video_vbe_setup function in vbe.c the default palette is
loaded before the created render target is set as active. I guess this
would be a problem if the palette was actually needed/supported.
Changing the order to create, set_active, set_palette makes the code
fail for me,though.

In this same code struct grub_video_mode_info mode_info is added in
the last round of patches but I cannot find where the changed code
touches the variable. This is somewhat odd.

In create_render_target_from_pointer the is_allocated is set to 0 twice.

Thanks

Michal