2.2.9 kernel deadlocks at boot

Mark Hull-Richter (markh@procom.com)
Thu, 27 May 1999 09:07:11 -0700

[Second transmission - the first does not seem to have come through...]

I am running an AlphaServer 1200 (rawhide, dual 21164, .5Gb memory).
When I try to boot the 2.2.9 kernel, I get a deadlock that looks like

Linux version 2.2.9 (markh@pisces) (gcc version egcs-2.90.29 980515
(egcs-1.0.3 release)) #1 SMP Wed May 26 17:11:30 PDT 1999
Booting on Rawhide variation Tincup using machine vector Rawhide
Command line: bootdevice=sda2 bootfile=vmlinux.gz
mcpcia_init_arch: found 2 hoses
setup_smp: 2 CPUs probed, cpu_present_map 0x3, boot_cpu_id 0
HWRPB cycle frequency bogus. Estimated 532798504 Hz
Console: colour VGA+ 80x25
Calibrating delay loop... 528.48 BogoMIPS
Memory: 512656k available
VFS: Diskquotas version dquot_6.4.0 initialized
POSIX conformance testing by UNIFIX
Entering SMP Mode.
secondary_console_message: on 0 from 1 HALT_REASON 0x0 FLAGS 0x1ee
secondary_console_message: on 0 message is 'P01>>>START P01>>>'
smp_boot_cpus: Total of 2 Processors activated (156.96 BogoMIPS)
start_secondary: commencing CPU 1 current fffffc002fff0000
spinlock stuck at fffffc0000321c00(0) owner swapper at fffffc00003212c0
spinlock stuck at fffffc0000321c00(1) owner swapper at fffffc00003212c0

This happens on every boot of 2.2.9, and I have to go back to 2.2.7 to
get the machine to come back to life again. I am running a fairly stock
configuration with the only modifications to the standard config
questions to hard link in the drivers for my SCSI and network adapters
(NCR53C800, QLogic ISP and DEC Tulip), plus Rawhide, SMP and
experimental drivers (not in that order).

Any suggestions? I typed this in from craning my neck to look at the
screen about 3 feet away, so I hope it looks familiar to someone....

Thanks (in advance).