Re: -current panics on cubietruck during daily/nightly activity and some network activity

[Matt Thomas <matt%3am-software.com@localhost>]

> On Nov 7, 2014, at 11:18 PM, Markus Kilbinger <mk%kilbi.de@localhost> wrote:
> 
> Hi!
> 
> It's now the second night that it happened:
> 
> NetBSD 7.99.1 (CUBIETRUCK) #0: Fri Nov  7 19:55:26 MET 2014
>        root@q:/usr/u/NetBSD/HEAD/src/
> sys/arch/evbarm/compile/CUBIETRUCK
> total memory = 2048 MB
> avail memory = 2022 MB
> sysctl_createv: sysctl_create(machine_arch) returned 17
> kern.module.path=/stand/evbarm/7.99.1/modules
> timecounter: Timecounters tick every 10.000 msec
> mainbus0 (root)
> cpu0 at mainbus0 core 0: 960 MHz Cortex-A7 r0p4 (Cortex V7A core)
> cpu0: DC enabled IC enabled WB disabled EABT branch prediction enabled
> [...]
> boot device: dk1
> root on dk1 dumps on dk2
> root file system type: ffs
> panic: kernel diagnostic assertion "pg != NULL" failed: file
> "../../../../arch/arm/arm32/pmap.c", line 1291
> cpu1: Begin traceback...
> 0xb279fdb4: netbsd:db_panic+0xc
> 0xb279fde4: netbsd:vpanic+0x1b0
> 0xb279fdfc: netbsd:__udivmoddi4
> 0xb279fe3c: netbsd:pmap_create+0x1bc
> 0xb279fe64: netbsd:uvmspace_init+0x6c
> 0xb279fe8c: netbsd:uvmspace_alloc+0x44
> 0xb279feac: netbsd:uvmspace_spawn+0x24
> 0xb279ff4c: netbsd:execve_runproc+0xac
> 0xb279ffac: netbsd:spawn_return+0x2bc
> cpu1: End traceback...
> 
> dump to dev 168,2 not possible
> Skipping crash dump on recursive panic
> panic: delay: arg 5000000 overflow, limit is 4473924 usec
> 
> cpu1: Begin traceback...
> 0xb279fd04: netbsd:db_panic+0xc
> 0xb279fd34: netbsd:vpanic+0x1b0
> 0xb279fd4c: netbsd:snprintf
> 0xb279fd7c: netbsd:cprng_fast_seed
> 0xb279fdb4: netbsd:dumpsys+0x18
> 0xb279fde4: netbsd:vpanic+0x1bc
> 0xb279fdfc: netbsd:__udivmoddi4
> 0xb279fe3c: netbsd:pmap_create+0x1bc
> 0xb279fe64: netbsd:uvmspace_init+0x6c
> 0xb279fe8c: netbsd:uvmspace_alloc+0x44
> 0xb279feac: netbsd:uvmspace_spawn+0x24
> 0xb279ff4c: netbsd:execve_runproc+0xac
> 0xb27Unified
> 
> My cubietruck (ct) is running actual -current (with SMP) with stock
> CUBIETRUCK kernel and userland.
> It's running some rsync data synchronization (over net) all the time
> (at the moment). The panic happened during additional
> /etc/daily-activity.
> 
> Does anybody see what's going wrong here?
> 
> Tell me if more info is needed.

I just committed a possible fix for this.