At 10:51 PM +0900 04.3.19, Kazuyuki Inanaga wrote:

>"Bad address"
>
>My kernels still have a strange "Bad address" error.
>It happens (not always) at ftp get with a big size file such as comp.tgz.
>My kernels also get 'freeze' at ftp get sometime.
>
>I don't think that is a hard drive problem because it happens on
>the other '030 Macs (hard drives) too.
>I also don't think that is a config problem because it happens on
>GENERIC kernel.
>I'm sure the file system is clean, and I don't see any error at ping,
>small data transfer such as ntp, telnet, or on MacOS/AppleTalk.
>
>I'm in doubt about netbsd-1-6 source has anything bug or "not '030-
>friendly" for the SCSI support ... ?
>Now I'm going to check the file 'scsiconf.c', but it's not easy for me,
>and it seems not make a sense.


netbsd-1-6-2 has 'scsiconf.c, v 1.184.4.3 2003/06/02'.
I swapped it down to 'v 1.184.4.1 2002/08/02'.
This makes a sense. I don't see those errors such as Bad address nor
freeze any more at the 'ftp get'. (^^)

I tried it before twice, but it didn't make a sense at that time.
I've also changed some options. I guess the problem is related not only
scsiconf.c but also other options.

Yes, Sysinst 1.6.2(INST030FPE) DOES work now on PB165 with
DaynaPORT SCSI/Link-T + PB165.


But alas, within 030FPE, at the 'ftp put', a lot of error appears
repeatedly.

>dse0: not queued, error 12

It also appears sometimes (not so often, not repeatedly) on the kernel
which has scsiconf.c, v 1.184.4.3.

I've seen this "not queued, error 12" before on Quadra 650 + Cabletron
EA412 SCSI to Ethernet adapter.
It happens in the several kinds of kernels including Hiroshi Noguchi's
nice 1.6G kernels. It appears on Quadra 650, when I connect EA412 to
LCIII+ (with optional FPU), I don't see "not queued, error 12".


Swapping down to  v 1.184.4.1 is not the answer. I think I should check
what is changed between v 1.184.4.1 and v 1.184.4.3.
How can I find what does "error 12" mean?

Thanks,
Kazu Inanaga