Re: Panic at boot on 4/330 with cgfour and cgsix
Michael <[email protected]> Mon, 14 Apr 2025 23:04:28 -0400
| Newsgroups | gmane.os.netbsd.ports.sparc |
|---|---|
| Message-ID | <20250414230428.781c57a3@glenmorangie> |
Hello, On Tue, 15 Apr 2025 12:11:10 +1000 matthew green <[email protected]> wrote: > > ... and probably did this as well - I didn't check the code when I had > > a working cg4. However, looking at the current source tree, we don't > > have either Xsun or xf86-video-suncg4 (hmm, did the latter exist?) so > > we'll have to put this together ourselves to work with modern X. > > maybe we use xf86-video-wsfb for this? it is at least true for > xf86-video-suncg3 (that has an upstream driver we don't use or > apparently want.) Since cg3 is a dumb frame buffer without even a hardware cursor, there isn't much to be gained from a dedicated driver - we'd do colour map updates by talking to the DAC instead of using ioctl()s, that's it. I would assume the same applies to cg4. All we need to get wsfb working is: - mmap() support through ttyE* - a couple ioctl()s to read fb layout, colour depth etc., and write to the colour map If we really wanted to we could use the overlay / enable plane thing to simulate a cursor sprite, I'll probably do this for cg12 at some point. ( the assumption here is that the xserver would spend a lot more time deleting and redrawing the cursor every time something under or around it changes, than we would if we just draw the cursor into the overlay plane every time it moves ) have fun Michael