Re: [RFC,resend] framebuffer rotation patch

[Michal Suchanek]

2010/2/20 Vladimir 'φ-coder/phcoder' Serbinenko <address@hidden>:
> Michal Suchanek wrote:
>> Hello,
>>
>> this is a framebuffer rotation patch rebased on top of current trunk.
>> Between the trunk commits and the actual rotation there are a few
>> fixes from my local branch most of which can be separated but the
>> scroll fix conflicts with the rotation patch.
>>
>> This works for me (including terminal scrolling) in qemu and I
>> compile-tested earlier version for efi as well. Only videotest and
>> gfxterm was tested.
>>
>> Most of the suggestions I got in the previous discussion are
>> incorporated in the new patch.
>>
> You have included some unrelated changes in your patch like useless
> grub-mkrescue.in change.

Yes, they are included as separate changesets, though.

> What is the reason for get_max_viewport function? Why not just transform
> mode_info?

Because get_max_viewport is a user function. Transforming mode_info in
every code using framebuffer would be error-prone, the users of the fb
code should not need to invoke the transform functions.

In the current users of get_max_viewport you can get away with
get_veiwport (and throwing away the first two values) because all such
uses happen just after graphics mode initialization and the current
viewport is the maximum one. This might not be always the case in the
future, though.

Thanks

Michal