xf86-video-ati: Branch 'master' - 9 commits

[Alex Deucher]

On 8/23/07, Michel Dänzer <michel at tungstengraphics.com> wrote:
> On Thu, 2007-08-23 at 09:49 -0400, Alex Deucher wrote:
> > On 8/23/07, Michel Dänzer <michel at tungstengraphics.com> wrote:
> > >
> > > radeon_crtc.c: In function 'radeon_crtc_mode_set':
> > > radeon_crtc.c:762: warning: 'output' may be used uninitialized in this function
> > >
> > > I don't understand the logic of that function so I'm not sure how to fix
> > > it.
> >
> > We need to get the output type connected to the crtc in question in
> > order to do some fix ups to the crtc regs for tv-out and in order to
> > avoid odd dividers in the plls for DFPs.
>
> Here's what I don't understand:
>
>     for (i = 0; i < xf86_config->num_output; i++) {
>         output = xf86_config->output[i];
>         radeon_output = output->driver_private;
>
>         if (output->crtc == crtc) {
>             montype = radeon_output->MonType;
>         }
>     }
>
> At the end of this loop, 'output' will be the last output, and 'montype'
> will be the monitor type of the last output assigned to the crtc in
> question.
>
>     if (montype == MT_STV || montype == MT_CTV) {
>         switch (radeon_crtc->crtc_id) {
>         case 0:
>             RADEONAdjustCrtcRegistersForTV(pScrn, &info->ModeReg, adjusted_mode, output);
>             RADEONAdjustPLLRegistersForTV(pScrn, &info->ModeReg, adjusted_mode, output);
>
> So AFAICT, 'montype' and 'output' here don't necessarily correspond to
> the same output, in fact 'output' may not even correspond to the crtc in
> question.
>
> Am I missing something?

Whoops.  Sorry, I missed that last part.  yeah, we should break when
we get montype.

Alex