>
> I have just bought a new laptop which has a ATI Radeon Mobility LY display
> adapter. The 4.1.0 XFree86 release does not include support for this card,
> but the CVS source tree does. I've downloaded cvs tree and compiled it. The
> server will configure itself correctly with a XFree86 -configure command, but
> the server will not run. The 'relevant' lines from XFree86 appear to be
I noticed this issue recently. Hui Yu said in email that he'd be submitting a patch
"in the near future" for it, and said in the meantime to force the panel to the
appropriate size as I was doing.
Here's the relevant stuff from my post (msg id 20011009000155.E3F923E9D):
> Hi all,
> XFree86 4.1.99.1 from yesterday's CVS doesn't probe the flat panel
> correctly on an HP Omnibook 6100 with a Radeon Mobility LY (chip id
> 4c59).
> In the function RadeonGetDFPInfo (in radeon_driver.c), the test of
> info-> HasEDID is false, and the remaining code (around line 2007)
> does this:
> r = INREG(RADEON_FP_VERT_STRETCH);
> r &= 0x00fff000;
> info->PanelYRes = (unsigned short)(r >> 0x0c) + 1;
> However, 'r' is zero after the INREG(RADEON_FP_VERT_STRETCH) op, so
> info-> PanelYRes is 1 on entry to this code:
> switch(info->PanelYRes)
> {
> case 480: info->PanelXRes = 640;
> break;
> case 600: info->PanelXRes = 800;
> break;
> case 768: info->PanelXRes = 1024;
> break;
> case 1024: info->PanelXRes = 1280;
> break;
> case 1050: info->PanelXRes = 1400;
> break;
> case 1200: info->PanelXRes = 1600;
> break;
> default:
> xf86DrvMsg(pScrn->scrnIndex, X_ERROR,
> "Failed to detect the panel size.\n");
> return FALSE;
> }
> Entering a bogus case to set PanelXRes = 1400 & PanelYRes = 1050 if
> the entry value was 1 allows X to start and run.
Cheers,
AS
PGP signature
