Re: netboot loader not closing devices

[David Brownlee <abs%absd.org@localhost>]

On 6 June 2015 at 01:41, Charles Dickman <chd%chdickman.com@localhost> wrote:
> On Fri, Jun 5, 2015 at 8:20 AM, Felix Deichmann <m4j0rd0m0%gmail.com@localhost> wrote:
>>
>> It would be good if Charles could test a fixed loader with his machine (and
>> qe) which showed the real problem mentioned in his posts.
>> I dropped the patched -current loader with debug printf here:
>>
>>   http://www.flxd.de/netbsd/vax/boot
>>
>> Should show something like "XXX qeclose() XXX" just before the copyright
>> notice starts.
>
> I have not had the problem since I changed from 6.1.4 to -current AND
> I don't have the problem on 6.1.4 anymore!
>
> I feel that it was triggered by something on my network. Since it
> isn't happening now, its tough to track down. It was a very
> frustrating problem, because it would come and go and for the longest
> time I thought it was caused by my changes to the kernel.

I wonder if a flood ping of broadcast packets would cause any issue?
Just as root on another box: ping -f 255.255.255.255

> Also, the boot at the link above doesn't work for me because its load
> address is in the part of memory covered by the portion of the Qbus
> map occupied by the QVSS board in that machine. I have to make a
> special boot that loads at 0x3b0000 instead of 0x3f0000 because
> 0x3c0000 - 0x3fffff are occupied by the frame buffer.

(General question to anyone & everyone) We seem to be having recurrent
problems with finding a safe place for the bootloader. Is there a
location guaranteed to be viable for a 4MB uVaxII and a much bigger
Vax with 100s of MB?