Is there an implementation of Sun nd(4) (Network Disk) server other
than the one in NetBSD, by Matt Fredette?
Long story: Few days ago I fried SCSI controller on Indigo 2, and I'm
not gonna risk the same with my Sun 3/60. Few months ago I got SunOS
3.5 kernel to netboot using modified NetBSD ndbootd(8); it goes as far
as mounting rootfs. Then, just after printing the following line...
using 95 buffers containing 778240 bytes of main memory
... it just retries reading block number 16 over and over. I guess
the ND protocol implementation is wrong somehow. The whole transaction
goes like this; the only thing that changes is the "seq" field.
RECV: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 0 currcount
0 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 0 currcount
1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 1024
currcount 1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 2048
currcount 1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 3072
currcount 1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 4096
currcount 1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 5120
currcount 1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 6144
currcount 1024 op 0x01<READ >
XMIT: minor 0x00 error 0 vers 0 seq 29 blk 16 bcount 8192 resbcount 0 curroff 7168
currcount 1024 op 0x11<READ DONE >
Any ideas?
--
If you cut off my head, what would I say? Me and my head, or me and my body?