Booting the kernel on the livecd 1.4_rc2 results in a very reproducable oops on my G4 cube (450 with 1.1G memory). I am able to boot a 2.4.20-ben1 (and 19-ben0) kernel with the live.img boot file using the in-kernel vlinitrd loader and run the livecd from there. Oops: kernel access of bad area, sig: 11 NIP: C04CF470 XER: 20000000 LR: 00140198 SP: DFFEDD70 REGS: dffedcc0 TRAP: 0300 Not tainted MSR: 00009032 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11 DAR: E1143000, DSISR 42000000 TASK = dffec000[1] 'swapper' Last syscall: 120 last math 00000000 last altivec 00000000 CPU: 0 GPR00: 00000020 DFFEDD70 DFFEC000 DFFD2000 0000003F 00000190 C04DCEBD 000000FF GPR08: E1143000 E1003000 00000000 E1000000 22004042 00000000 00000008 C057F700 GPR16: E1003000 00000500 00140000 00140000 C05A5E3C 00000000 C0565D18 C0565928 GPR24: C04DCDBC C04DCCBC C04DCBBC C0480000 00140190 00000000 0000018F 000000E0 Call backtrace: C024C85C C011DD24 C0122248 C0249128 C04D2DB4 C04D29E4 C04D2728 C04CEBEC C04C0E60 C04AD5D8 C04AD624 C00040BC C0009710 Kernel panic: Attempted to kill init! <0>Rebooting in 180 seconds.. using the System.map from /boot ksymoops gives: Warning (Oops_read): Code line not seen, dumping what data is available >>EIP; c04cf470 <nvramData+98c/8000> <===== Trace; c024c85c <pcmcia_deregister_erase_queue+50/ac> Trace; c011dd24 <xfs_da_do_buf+3a8/654> Trace; c0122248 <xfs_dir2_block_removename+13c/1a4> Trace; c0249128 <pcmcia_replace_cis+64/b8> Trace; c04d2db4 <nvramData+42d0/8000> Trace; c04d29e4 <nvramData+3f00/8000> Trace; c04d2728 <nvramData+3c44/8000> Trace; c04cebec <nvramData+108/8000> Trace; c04c0e60 <__res+1224/73c4> Trace; c04ad5d8 <pci_vendor_list+4b88/4cbc> Trace; c04ad624 <pci_vendor_list+4bd4/4cbc> Trace; c00040bc <do_linuxrc+160/184> Trace; c0009710 <__ashldi3+1c/24>
WOuld you be able to try the kernel from here on your machine and see if results are any better? http://ppckernel.org/kernel.php?id=28 If you can let me know it would be great.
the ben kernel occasionally hung after TCIC cardbus probe with 242 bad address the bk kernel was ok. However neither had initrd. latest config still missing loop. had a few hangs after detecting SCSI (have a sbp-2 disk that doesn't show active) on ben kernel but otherwise stable.
Yes thanks for the catch (again). I just uploaded the proper ./config file and triggered builds, should be more within a couple hours to test. Thanks for your efforts here miltonm
ok the ben kernel works, although i did have 1 boot in 10 that hung after the scsi subsystem banner. The bk kernel did not get recompiled with loop yet.
Yes, in fact there is an upstream issue with the bk tree (and the kernel.org 2.4.21-pre2). Seems marcelo may have comitted some things he shouldn't have (assuming that th dir named 'dontdiff' he used wan't for commit in his diffs). He broke the Makefile for iee1394 support, so once that gets resolved there shoud also be a bk kernel to test. If the benh works for most people I would lean towards that one as it has some hardare support that bk currently doesnt. Thanks for testing stuff Milton.
this is old, newer kernel should be great Milton: Can you test latest release ISO on your cube?
closing for staleness