Quoting Andreas Robinson: > > I've just improved viafb from DirectFB CVS: > > Hmm, I have not received any messages on the CVS list for a while.
Sorry for that, our admins transfered the list to an external server, obviously doing some mistakes ;( I've forwarded your mail to them. At least the archive is working: http://www.directfb.org/mailinglists/directfb-cvs/2003/11-2003/maillist.html > > - revived all pci code, taking physical addresses and sizes from pci > > - export MMIO region and accelerator ID for user space > > - default to 720x576 PAL overscan (may be discussed) > > - fixed warnings > > - rewrote Makefile stuff > > > > cle266vgaio is obsolete now! > > But the DirectFB still supports it as a fallback for vesafb users. > > Nice work! > > Have any additions to the kernel (eg #define FB_ACCEL_VIA_UNICHROME) been > passed on to the maintainer(s) yet? No, at least not yet. I'm sure there will be continuous development of the driver within our CVS. Maybe it's better to have the maintainer(s) working on our tree, too. Or even take over maintenance of the driver completely. > > Running df_dok (at 720x576 16bit) surprised me, some values are twice as > high as > > stated in unichrome.c, others are a bit higher and a few are lower. Maybe > df_dok > > ran at 32 bit before? > > Yes. I used 32 bits for testing alpha blending. Alpha blending is not limited to 32 bit destination formats ;) Matrox even has hardware accelerated blending in 8 bit (RGB332). -- Best regards, Denis Oliver Kropp .------------------------------------------. | DirectFB - Hardware accelerated graphics | | http://www.directfb.org/ | "------------------------------------------" Convergence GmbH -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe directfb-users" as subject.
