From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4.1) Gecko/20031114 Description of problem: The SMP kernel panics very early on while booting on a dual Opteron 240/Rioworks HDAMA system. (I have no scrollback or serial console to determine the point where the panic happens yet.) The UP kernel boots fine. Also tested with identical results on 2.4.22-1.2163. Version-Release number of selected component (if applicable): kernel-smp-2.4.22-1.2135 How reproducible: Always Steps to Reproduce: 1. On a fresh install of FC 0.96/x86_64, boot the SMP kernel.
At what point in the boot does this happen? Any chance of getting a serial console soon?
I hope to get to it soon. I'm working on the box, trying to resolve other showstoppers first.
The kernel panics right after initializing CPU 1 on this Dual Opteron. The call trace in the panic includes some symbol information, without the benefit of ksymoops. The panic occurs in: set_cpu_allowed+411. The next entries in the call trace are: free_uid+2, ksoftirq+77, child_rip+8, ksoftirq+0, child_rip+0 Code: 8b 4b 3c 48 69 c9 ... Still working on a serial console, to capture the whole thing, hopefully the above may be helpful in the meantime.
I have the serial console attached; however it looks like the SMP kernel does not panic consistently. I've been able to boot the SMP kernel a couple of times already, with the serial console attached. I will keep the serial console available, and update this bug when I capture a panic.
After some additional testing I cannot reproduce the bug after updating to kernel build 2.4.22-1.2166.nptlsmp and flashing the motherboard to the latest OEM BIOS. There does appear to be a residual problem with the aic79xx.o module (I can crash it fairly reliably), but that's going to be a different bug.
This bug can be closed.