Worked in RELENG_5, fails in RELENG_6

[David Kelly]

On Nov 29, 2005, at 5:10 PM, Alex Zbyslaw wrote:

> David Kelly wrote:
>
>> Could I
>> boot a 5.4 CDROM to test ACPI?
> Yes you could.  I did this the other day on a machine that doesn't  
> even have a disk yet ;-)  acpidump is there.  Select fixit then  
> live file system (if memory serves) but you probably knew that !

Problem solved, but I've been working on it too hard not to share the  
story which might be useful:

Yes I knew how to launch the fixit disk but it didn't hurt one bit to  
remind me. The problem is that FreeBSD is so darn good I've almost  
never used the fixit mode in the past 10 years.

Have either been a bad boy and lost my 5.4 disks or a good boy and  
gave them to someone. So I hope 5.3 will do. From within 5.3 fixit:

Fixit# acpidump -d
iasl tmp file (read): No such file or directory

acpidump -t appears to work.

Forgot to bring a null modem cable for the serial console. OTOH I now  
have Kermit built on MacOS X.

Then I thought to boot my 6.0-RELEASE CD's. Dontchaknow ACPI worked?

Gave it some deep thought, moved my /boot/loader.conf out of the way  
and rebooted. ACPI worked. Conclusion: something in my loader.conf is  
breaking ACPI.

Many reboots later (reminded me of the last time I installed NT4, but  
that took 3 days) of selectively enabling lines in my /boot/ 
loader.conf I found the one line that broke ACPI, its commented out  
below:

geom_vinum_load=YES
if_ath_load=YES
snd_ich_load=YES
hw.ata.atapi_dma=1
kern.maxdsiz="2G"
kern.dfldsiz="2G"
#hw.physmem="2G"
kern.maxssiz="128M"

The reason hw.physmem="2G" was added was that I found FreeBSD running  
with only 1G of RAM rather than the full 2G installed and saw  
'#hw.physmem="1G"' in /boot/defaults/loader.conf. Added that line and  
noticed I had the full 2G once again, congratulated myself on solving  
that problem and moved on. Suspect that is when ACPI broke and I was  
simply moving too fast to notice that I had to disable ACPI to boot.

Without hw.physmem="2G" its now finding the full memory load so I  
don't know what was happening previously.

Sorry to bother you guys. Maybe there is something broken with  
hw.physmem with respect to ACPI so this wasn't wasted effort.

Meanwhile between my reboots I was testing P/R bin/89100 which still  
yesterday's chops files short as if a 32 bit length was all that was  
allowed, but a fresh copy of the same file made with "cp -p" fully  
makes it thru the sendfile(2) system call. But won't a couple of  
hours later.

--
David Kelly N4HHE, dkelly at HiWAAY.net
========================================================================
Whom computers would destroy, they must first drive mad.