FreeBSD 10-STABLE CD PowerMac3, 4 installation stuck repeating CAM status: Command timeout

[Mark Millard]

This is a report for: CDs made from FreeBSD-10.0-STABLE-powerpc-20140528-r266807-disc1.iso (md5 verified)
Also for: CDs made from FreeBSD-11.0-CURRENT-powerpc-20140525-r266655-disc1.iso (nearly identical behavior)

I have recently been experimenting with basic installation and configuring of FreeBSD and some Linux variants on some old PowerMacs and one old Pentium D PC.

I will note up front that Mac OS X 10.4 installation media and the Linux media that I've tried all boot the configurations listed below just fine, can do installs, and so on. So the hardware appears to be okay.


1st Context: PowerMac3,4 466MHz (Early 2001) with one ATA "disk" drive (a SSD with an adapter), a Hitachi DVD-ROM GD-7000 drive on a separate ATA bus, and 1.5GBytes of RAM, ATY,Rage6.

Booting from either CD eventually goes silent for a bit and then starts reporting the following. Most of the lines I'm reporting below start with "(aprobe0:ata1:0:1:0):" so I'll not repeat that part of the lines.

ATAPI_IDENTIFY ACB: a1 00 00 00 00 40 00 00 00 00 00 00
CAM status: Command timeout
Retrying command
ATAPI_IDENTIFY ACB: a1 00 00 00 00 40 00 00 00 00 00 00
CAM status: Command timeout
Error 5, Retries exhausted
run_interrupt_driven_hooks: Still waiting after 60 seconds for xpt_config

Such lines repeat and another run_interrupt_driven_hooks notice is put out for 120 seconds. After that there are no more examples of that particular line are generated that I saw. It seems to be indefinitely retrying, generating those other messages as it goes.

http://lists.freebsd.org/pipermail/freebsd-stable/2014-January/077091.html reports something similar against "10-STABLE r261177" back on 2014-Jan-27 for a Thinkpad T42p. There seem to be a historical list of reports of this sort of thing going way back.

Side note: During the sequence above does find the SSD on "ada0 at ata0 bus 0 scbus 0 target 0 lun 0" and puts out a correct description of it. I've not listed those lines. The 11.0 CD does have a difference here. Just after these ada0 messages it has an additional line:

WARNING: Witness option enabled, expect reduced performance.

But other than the additional text I could not tell any difference from deadlock/cycle checking.


2nd Context: PowerMac3,4 733MHz (Early 2001) with one ATA "disk" drive (a SSD with an adapter), a DVD-ROM drive of the same model as above (again on the separate ATA bus), and 1.5GBytes of RAM, GeForce4 Ti 4600.

Both CDs appear to stop at what appears to be the same point where those messages started above but no messages are produced, not even for the SSD on ata0. I've let it sit for 1hr+ and nothing new was displayed. There was no witness option enabled massage from 11.0 either.


Unfortunately I'm going to be away from the PowerMacs for a notable time so for now I will not being trying burns of older 10.0 or 11.0 .iso's or 9.3/9.2 ones or other experiments. May be next month.

===
Mark Millard
markmi at dsl-only.net