On 2012-03-05 7:20 AM, Philip Prindeville wrote: > On 3/4/12 5:45 PM, Felix Fietkau wrote: >> On 2012-03-05 1:19 AM, Felix Fietkau wrote: >>> On 2012-03-04 11:31 PM, Felix Fietkau wrote: >>>> On 2012-03-01 5:18 AM, Adam Gensler wrote: >>>>> For what its worth I spent some time this evening putting together a >>>>> Fedora Core 16 virtual machine to test with. I pulled a fresh copy of >>>>> trunk and built the alix2 target and got the same traces on boot up. >>>> Please run gdb on the vmlinux file again, but ignore the whole kernel >>>> module issue, I think we probably won't be needing that. >>>> To show the source file context, run this: >>>> >>>> (gdb) l *0xc105a93e >>>> >>>> (example address from the trace) >>>> >>>> Please show the source code context for the first 2-3 trace addresses >>>> from the first stack strace. >>> Never mind, I managed to reproduce the issue in qemu and I'm going to >>> take care of it. >> Bug fixed in r30813 (needs kernel tree clean or manual patch apply) - >> the issue was simply an empty module init section leading to vmalloc(0) >> calls in the kernel, which generate a non-fatal warning + stack trace. >> I've fixed the issue by pushing the size check to the callsite before >> module_alloc() is called. >> >> - Felix > > Just out of curiosity, which was the misbehaving module? There was no misbehaving module. There were several modules without module_init() (which is fine), and the kernel didn't deal with that properly.
- Felix _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
