Net booting current snapshot on openrd and sheevaplug

Jeff Penn jeff at jrpenn.demon.co.uk
Sat Mar 30 21:40:24 UTC 2013


I haven't had alot of spare time to make much progress on this.  I'm
now at a point where the openrd and sheevaplug are failing at the same
point of the boot process; when trying to extract the root path from the
DHCP request similar to this:

http://lists.freebsd.org/pipermail/freebsd-arm/2013-March/005325.html

The openrd also drops into the debugger every few boots as mentioned in 
my previous mail:

Sending DHCP Request packet from interface mge0 (f0:ad:4e:00:61:58)
Received DHCP Ack packet on mge0 from 192.168.0.10 (accepted) (got root path)
mge0 at 192.168.0.16 server 192.168.0.10 boot file openrd/kernel.bin
subnet mask 255.255.255.0 router 192.168.0.1 rootfs 192.168.0.10:/mnt/work/nfsroot/arm-8-le hostname openrd
Adjusted interface mge0
krpc_call: sosend: 64
krpc_call: sosend: 64
panic: nfs_boot: mountd root, error=64
KDB: enter: panic
[ thread pid 0 tid 100000 ]
Stopped at      kdb_enter+0x48: ldrb    r15, [r15, r15, ror r15]!
db>

I'm still using the same version of current, and haven't seen a repeat
of the RPC timeouts with the sheevaplug mentioned in my first mail.

I tried building a copy of i386 current to test diskless booting, but
ran into problems with buildkernel:

aicasm_gram.c:1539: warning: no previous prototype for 'yyparse'
*** [aicasm_gram.o] Error code 1

Stop in /mnt/work/obj/mnt/work/src/sys/GENERIC.

The best approach is probably to test arm kernels built with a fixed
ROOTDEVNAME.  I hope to find the time to have a go at this on Monday.

Jeff



More information about the freebsd-arm mailing list