Sorry, forgot to copy the list on this. If anyone has a Mach64 GX or CX in their machine, please give it a go! - Dave Begin forwarded message: > From: David Riley <fraveydank%gmail.com@localhost> > Date: April 27, 2011 8:06:18 PM EDT > To: Michael <macallan%netbsd.org@localhost> > Subject: Re: machfb (sort of) working on Mach64 GX > > On Apr 27, 2011, at 12:26 PM, Michael wrote: > >> Ah, so machfb doesn't initialize the display start register and blindly >> assumes it's always pointing at the beginning of the video memory? I've seen >> firmware leaving the first couple KBytes unused ( for a cursor image or >> something? ) on different video controllers. > > What seems to be happening is that the mode doesn't really get set at > startup, which is fine, but for whatever reason there's an X offset register > that OpenFirmware configures to offset us by 0x400 (in the CRTC_OFF_PITCH > MSBs). The framebuffer driver uses it to set a virtual X location, but that > should be set to 0 for a static console. I just had it set the default > acquired mode on boot, which takes care of it. > > It's odd that OpenFirmware configures this because the hardware cursor seems > to be hard-mapped to the *end* of the memory range, not the start (which is > the normal practice; I've corrupted plenty of cursors in my day by writing > off the end of the buffer). > >> >>> The good news, in any case, is that that's working now as a console driver. >>> I'm not terribly familiar with the process with NetBSD... how do I submit >>> a patch to get integrated into the mainline? I have a netatalk kernel fix >>> as well. >> >> Please post it here and I'll deal with it. > > Here we go! It's against the 5.0.2 kernel, but this isn't exactly a driver > which changes often. If the patch blows up, let me know and I'll port it > over. > > How about the netatalk patch? I have that, too; Appletalk was pretty much > unusable before and I don't want this one to get lost. I filed a PR a while > ago, but never got any response. > > > > >
Attachment:
machfb_patch.diff
Description: Binary data
>