Red Hat Bugzilla – Bug 101793
Using an SMP kernel, the system hangs when the configuration space of the PCI device is actively communicating.
Last modified: 2007-03-27 00:08:44 EDT
All SMP kernels in RedHat 7.3, 8.0, 9.0
STEPS TO REPRODUCE:
0) List the PCI devices on your system:
The output will be a long list of devices, in the following format:
00:07.0 VGA compatible controller: Chips an...
1) run the following command (which reads the configuration space):
$while [ 1 ] ; do cat /proc/bus/pci/00/07.0 > /dev/null ; done &
Notice how the lspci output is used in this command - 00:07.0
becomes /proc/bus/pci/00/07.0 (feel to use a different device or more than one)
2) start and stop X windows (using xinit) between 3 and 15 times
The system enters a completely unresponsive state and must be rebooted.
Interaction with the PCI configuration space should not alter system stability
and should not cause a system crash.
This problem has been identified and tested in numerous hardware environments
(dual PIII & dual XEON, different motherboards). In all of these situations
(using an SMP kernel), activity in the configuration space of the PCI bus,
along with the initialization of X, caused the system to crash.
This test was also run on single processor machines and the symptoms were NOT
Thanks for the bug report. However, Red Hat no longer maintains this version of
the product. Please upgrade to the latest version and open a new bug if the problem
The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases,
and if you believe this bug is interesting to them, please report the problem in
the bug tracker at: http://bugzilla.fedora.us/