Red Hat Bugzilla – Bug 150123
[PATCH] pciReadDrivers doesn't fallback and read kernel modules.pcimap if needed
Last modified: 2014-03-16 22:52:33 EDT
Description of problem:
ICBW, but I don't see anything using the id_file_name field of pacc in pciProbe.
More importantly, pciReadDrivers returns in two places I believe it should
instead jump to the pcimap label. This makes all the difference when kudzu is
run without /usr mounted. This happens when /sbin/kmodule is called early in
/etc/rc.sysinit with /usr on a separate partition and since kudzu can't read
/usr/share/hwdata/pci.ids it only generates a partial list. Attached is a patch
to kudzu to correct this behavior.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Run kmodule from initscripts without /usr mounted, *OR*
2. move pcitable and pci.ids out of /usr/share/hwdata/
3. run kudzu -s -p
Multiple devices have "unknown" driver
Devices have drivers as specified in /lib/modules/`uname -r`/modules.pcimap
This is another part of the same problem (see Issue Tracker ref) that caused me
to reopen bug 130350. So after fixing the test for /dev/hw_random in rc.sysinit,
and with this fix and a rebuild of initscripts (/sbin/kmodule statically links
probeDevices from kudzu) the rngd daemon is started at bootup. Unfortunately, on
my laptop /sbin/rngd hangs trying to read from /dev/hw_random. I guess I'll be
opening a bug there against kernel-utils.
Created attachment 111587 [details]
read /lib/modules/`uname -r`/modules/pcimap if /usr/share/hwdata/pci* not available
Fixed in 184.108.40.206-1.
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.