I played a while with atyfb in LinuxBIOS. move the xl_init.c into LinuxBIOS.
there is one patch call xlinit.c that can be used even ati fb is not inited in BIOS to make kernel still can use atyfb. I wonder if James put that in mainstream, he already sent one patch for 2.6.5.... please refer to http://www.linuxbios.org/pipermail/linuxbios/2004-May/007734.html I guess the mips fw already execute the ati option rom via x86 emulator... YH On 8/12/05, Daniël Mantione <[EMAIL PROTECTED]> wrote: > > > Op Fri, 12 Aug 2005, schreef Jim Ramsay: > > > I have the following issue. I am trying to get an ATI Rage XL chip > > working on a MIPS-based processor, with a 2.6.11-based kernel from > > linux-mips.org. Now, I know that this was working with a 2.4.25-based > > kernel previously. > > Okay, the 2.4 driver is more intrusive, it programs the chip from start as > much as possible, while the 2.6 driver tries to depend on Bios settings. I > haven't checked out the 2.6 driver enough to see if it is still possible > to program from scratch. > > > I seem to get intermittent strange issues, such as the machine > > freezing from time to time, but in general I get the following in my > > dmesg when I load the atyfb module: > > > > atyfb: using auxiliary register aperture > > atyfb: 3D RAGE XL (Mach64 GR, PCI-33MHz) [0x4752 rev 0x27] > > atyfb(aty_valid_pll_ct): pllvclk=50 MHz, vclk=25 MHz > > atyfb(aty_dsp_gt): dsp_config 0x307c0001, dsp_on_off 0x14fffff0 > > < Sometimes it will hang here > > > atyfb: 512K RESV, 29.498928 MHz XTAL, 230 MHz PLL, 83 Mhz MCLK, 63 MHz XCLK > > atyfb: Unsupported xclk source: 7. > > > I'm assuming that most of my issues are due to the "Unsupported xclk > > source" message. Any ideas what I can do about this, or where I can > > go to learn more about how to make this thing work? > > Yes, according to my register data sheet a 7 means the memory clock > frequency is derived from DLLCLK. Unfortunately I don't know what this > DLLCLK is. I think it means the chip isn't properly initialized yet and it > clocks the memory from a safe clock source to allow the computer to start. > > However, we most likely have no way to find out the speed of this DLLCLK. > > The memory clock frequency is important for the driver to be able to set a > display mode; it needs to program a memory reload frequency into the chip > which depends on the memory frequency. > > Daniël > > - > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to [EMAIL PROTECTED] > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/ > - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/