Hi Mark! On 2/8/21 2:45 PM, Mark Cave-Ayland wrote: >> # nvsetenv boot-device "/ht@0,f2000000/pci@5/k2-sata-root@c/@-1/@0:2,\\:grub" > ^^^ > This path looks wrong - according to the aliases listing at > https://www.mail-archive.com/misc@\ > openbsd.org/msg164291.html the unit address of the k2-sata device which is a > child of the > k2-sata-root device should be 0, whereas above it is given as -1? Perhaps > something in ofpathname > is subtracting 1 from the unit address of 0 when it shouldn't. Does the > following alternative work?
I would be surprised to see that ofpathname returns wrong paths on PowerMac G5s as we explicitly tested the tool on these machines as well. I will soon have access to my G5 machines again so I verify that myself. There might be a bug hiding somewhere. > nvsetenv boot-device "/ht@0,f2000000/pci@5/k2-sata-root@c/@0/@0:2,\grub" > > Once your support for blessed files has been merged then there is no need to > specify the path to the > grub binary itself as it can be found like this: > > nvsetenv boot-device "/ht@0,f2000000/pci@5/k2-sata-root@c/@0/@0:2,\\:tbxi" > > where \\:tbxi is the magic which tells the firmware to search the filesystem > on the given device for the > blessed file and execute that rather than needing a full path. This is how > OpenBIOS locates the bootloader > under QEMU and so if you can boot an installation on one of the QEMU Mac > machines then everything is working > as it should. Yes, but we're not there yet. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
