On Sat, Oct 18, 2008 at 5:04 AM, Finn Thain <[EMAIL PROTECTED]> wrote: > > On Fri, 17 Oct 2008, Finn Thain wrote: > >> >> I'll see if I can come up with a kernel having a suitable initramfs. > > OK, here it is. Boot this and hopefully it will tell you if you CPU is > buggy. > > http://www.telegraphics.com.au/~fthain/vmlinux-040-test.gz
I'll try this on my powerbook540 (and maybe on my 520 too). As I said before the 2.6.18 and 20 kernels would panic (I think on fsck) before mounting root, and I am pretty sure that the emile rescue disk did too (which neither mounts a file system nor even creates the device nodes). So, I am not optimistic, but I will try (just deleted my old sarge partition on the 540 to make room for some macos music/art stuff). Brian p.s. sorry for my false claim that 605/475 would run linux; only if replace the cpu before which I did but forgot I did. Its socketed unlike the books. > > No command line arguments are needed. No initrd is needed either since it > uses an initramfs. The test program runs upon boot, and no interaction is > necessary (or possible). > > It quickly detects an FPU emulation issue on my LC 475 (68LC040 mask set > 2E23G). Whereas, even after 10 minutes or so, no problem was detected with > my PowerBook 190cs (68LC040). I tested a Quadra (68040) and it passed this > test too. That's good news about the 190. Wonder if you can get enough RAM in it to do much. (520/540 takes up to 36MB) > > The source code is available*, in case anyone wants to look through it and > confirm that it really tests for the bug in question, and my results are > not due to some quirk of the LC 475 or whatever else. > > This kernel only boots Macs because I don't know what other machines might > be affected. I don't think it has any buggy drivers. > > HTH. > > Finn > > * http://www.telegraphics.com.au/~fthain/sw/040test/ > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
