On Thu, 6 Aug 2015, Greg Andrzejewski wrote: > Greetings, > > After finally getting my Mac SE/30 working again, I set about trying to > get a modern version of Linux installed on the little fellow. Early > experiments with 3.14 kernels were successful
BTW, are you using mac_scsi on the SE/30? > and when a trio of Quadra 950s appeared on the local craigslist, I > picked them up, looking forward to a more powerful 68k machine. Problem > is I can't get any recent kernel to boot. I haven't booted a Q950 for some years, but it should work. Using Penguin, I did boot Linux 3.1 on a Quadra 700 which is somewhat similar hardware. The only other report I found on the list about a Q950 was for Linux 2.6.20. If you would like, I can send you a current kernel binary that should boot any Mac (that is, SE/30 or Quadra 950). Also, any debian kernel binary from 3.x or 4.x should give useful results. (Kernel modules aren't relevant to isolating the cause of an early crash.) > > I've tried using Penguin-19 on MacOS 7.1 and 7.5.3 with identical > results; the machine just hangs on the "Bootling Linux" message. The > screen never clears, nothing even comes across the serial port with > earlyprintk. Penguin has a known bug in its zlib code that can cause kernel decompression to fail like that. Is this a large debian kernel binary? I suggest you try booting the vmlinux, e.g. # gzip -dc < vmlinuz > vmlinux I use Penguin with MacOS 7.5.3 without any issues; however, you might want to try with extensions disabled (hold down <shift> when you hear the chime) to avoid a possible unhandled slot interrupt during early boot. > I installed MacsBug in hopes of finding something useful in __log_buf on > reboot, but the entire buffer is empty (zeros). I'd suspect the > bootloader is at fault, but Penguin successfully boots a 4.0.0 kernel on > my SE/30. Penguin log is attached, in case anyone's interested. I don't know anything about __log_buf. I suspect you'd need to avoid the POST memory test for that to work. > > Still not 100% confident in Penguin, I tried booting with an EMILE > rescue disk. EMILE reads the kernel from disk and shortly thereafter the > chimes of death play (!!!!). Is this something the kernel can > intentionally do or is it more likely sort sort of triple fault-like > situation? I've never tried EMILE on my Q950. Perhaps Laurent can speak to that. > > I've done a touch of kernel debugging, but this was on x86 and never > this early in the boot process. What next steps can I take to further > debug this issue? The first thing you should see is the output from the early boot code: "ABCFGHIJK". If you don't see that then likely causes are an unhandled early interrupt or bootloader bug. However, it is also possible that a recent commit has messed up the OSS driver: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/arch/m68k/mac?id=b24f670b7f5b2058b95370caa9f104b3cefb9f1d But this first appeared in v4.1, so it probably isn't relevant. > I've glanced at the early arch code, but all I really got out of it was > a few chuckles from the comments venting about Apple's, uh, peculiar > hardware design. I suspect that the resentment of kernel developers would have been minimal had Apple made more documentation public. Unlike the authors of those comments, I don't blame Apple engineers for its corporate policies. Finn > > Thanks, > Greg > -- To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/alpine.LNX.2.00.1508071448280.11075@nippy.intranet
