All, I've posted here earlier regarding a kernel crash we were seeing on kernel 2.4.13, on our embedded hardware... We have recently upgraded to 2.4.16, and are still seeing crash issues. We are running Vanilla 2.4.16 (including some of our own device driver modules) on a PowerPC 7410, with the root partition being EXT3. We have a test which creates a EXT3 partition, coppeis /usr/bin to it, reboots, deletes the files, reboots, and starts over.. When running this, it seems to run for a while and then we start getting these crashes: Oops: Exception in kernel mode, sig: 4 NIP: C0080424 XER: 00000000 LR: C0080384 SP: C09B1C80 REGS: c09b1bd0 TRAP: 0700 Not tainted MSR: 00089032 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11 TASK = c09b0000[90] 'init' Last syscall: 11 last math 00000000 last altivec 00000000 GPR00: C0080384 C09B1C80 C09B0000 CFD328E0 00001032 000000F0 00000001 00000000 GPR08: CFD32000 CFD320E0 00000039 00000700 00000080 01881934 00000000 00000000 GPR16: 00000000 00000000 01880000 C0029EB4 C09B1D88 00000001 C09B1D58 CFAF2E00 GPR24: 00001000 C020C320 00000014 000000F0 00000001 C01942E8 C089D000 00000000 Call backtrace: C0080384 C0076440 C006F998 C0055650 C0057394 C0029C18 C002A08C C0045FC4 C00466A8 C0046DEC C0006ED0 C0003DFC 0000000C 01812B7C 01801E80 018021A4 018031A0 01804C00 01804FDC 0180578C 00000000>>NIP; c0080424 <journal_blocks_per_page+148/e4f8> <====Trace; c0080384 <journal_blocks_per_page+a8/e4f8>Trace; c0076440 <journal_start+ac/3a8> Trace; c006f998 <grok_partitions+9358/fd54> Trace; c0055650 <__mark_inode_dirty+4c/948> Trace; c0057394 <update_atime+64/74> Trace; c0029c18 <do_generic_file_read+4d0/8b4> Trace; c002a08c <generic_file_read+90/8bc> Trace; c0045fc4 <kernel_read+58/354> Trace; c00466a8 <prepare_binprm+100/114> Trace; c0046dec <do_execve+17c/290> Trace; c0006ed0 <start_thread+190/334> Trace; c0003dfc <set_context+6c0/8f0> Trace; 0000000c Before first symbol Trace; 01812b7c Before first symbol Trace; 01801e80 Before first symbol Trace; 018021a4 Before first symbol Trace; 018031a0 Before first symbol Trace; 01804c00 Before first symbol Trace; 01804fdc Before first symbol Trace; 0180578c Before first symbol Trace; 00000000 Before first symbol Any help or suggestions would be appreciated.. Paul W.
Hi, Paul. Paul White wrote:> > All, > > I've posted here earlier regarding a kernel crash we were seeing on kernel 2.4.13, > on our embedded hardware... We have recently upgraded to 2.4.16, and are still seeing > crash issues. > ...> >>NIP; c0080424 <journal_blocks_per_page+148/e4f8> <====> Trace; c0080384 <journal_blocks_per_page+a8/e4f8> > Trace; c0076440 <journal_start+ac/3a8> > ...Something is wrong with this trace - journal_start doesn't call journal_blocks_per_page, and journal_blocks_per_page ain't 40kbytes long! Could you please check, redo, resend? Thanks.
Hi, On Thu, Dec 13, 2001 at 11:42:51AM -0800, Paul White wrote:> I've posted here earlier regarding a kernel crash we were seeing on kernel 2.4.13, > on our embedded hardware... We have recently upgraded to 2.4.16, and are still seeing > crash issues.As Andrew said, this just looks bizarre. However, could you please try the 2.4.17-rc kernel instead and see if that helps? There were some bitop manipulation bugs fixed in a set of ext3 patches that went into 2.4.17-pre, and those may well address corruptions on bigendian 64-bit machines. The bug was benign on LE or 32-bit machines, so I haven't been able to test that myself, but you may well see random instabilities disappear with the newer kernel. Cheers, Stephen