strange corruption of module config when i try booting with acpi=off
by Joe Desbonnet
I apologise in advance for this rather ranty email, but I think its
relevant to the dev list because I believe hit on some bug.
I (foolishly) try to get FC4 to power manage my laptop (Thinkpad
T41p). I don't understand how to do "apm -s" with ACPI (can someone
tell me how? -- because I can't find it documented anywhere).
So I boot with "acpi=off" boot option so that I can use the more
intuitive apm system.
On boot system-config-display asks for root password (I've never seen
that before). I hit return a few times because I'm not interested in
reconfiguring my display (especially at boot time?!).
Boot goes bad -- lots of things don't load. In particular pcmcia stuff
-- so I don't have networking any more. Oh oh. Ok, no problem I just
reboot with default options. No good: my system is still hosed.
**Something has reconfigured/corrupted my system setup without my
permission.**
It has something to do with modues. Nothing unusual has changed in
/etc/* at the time of the incident, but all the modules.* files in
/lib/modues/2.6.11-1.1369_FC4 now have no data (just the header
comment line).
Is this a bug, or did I do something incredibly stupid that warranted
my system to be rendered unusable?
Joe.
18 years, 10 months
Bug: xm restore in unstable/FC4
by Michael Paesold
I have tried to do save/restore on FC 4. xm save now works (in the
latest development version, where the migrate/ directory is correctly
created by the RPM).
xm restore does not work, it fails in xc_linux_restore calling
xm_domain_create. Error message in xfrd log is "Could not create
domain. pfns=65536, 262144KB".
Debugging xfrd with gdb gave further information:
Inside xm_domain_create no error occurs at all. The result from the
do_dom_mem_op(xc_handle, MEMOP_increase_reservation,...) call is the
expected value of 65536 (mem_kb/4).
This is stored into the variable err, which is then returned.
Back in xc_linux_restore, the code now thinks that a result different then 0
is an error, when in fact -1 would indicate an error:
This seems to be wrong:
/* Create domain on CPU -1 so that it may auto load-balance in future.
*/
if ( xc_domain_create(xc_handle, nr_pfns * (PAGE_SIZE / 1024),
-1, 1, &dom) )
{
xcio_error(ioctxt, "Could not create domain. pfns=%d, %dKB",
nr_pfns,nr_pfns * (PAGE_SIZE / 1024));
goto out;
}
This should be probably be changed to xc_domain_create(..) == -1 as
condition for an error.
Well, I have just looked into the sources in the BitKeeper repository and
saw that this code has already changed substantially for some weeks now. Now
I don't really understand, what the FC4 package is really built from.
It says xen-2-20050530 -- xen 2 -- but booting xen says 3.0 devel.
Rik, could you clear this up for me, please? Thanks.
Best Regards,
Michael Paesold
18 years, 10 months
rawhide report: 20050619 changes
by Build System
Updated Packages:
selinux-policy-strict-1.23.18-14
--------------------------------
* Sat Jun 18 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-14
- Add Russell's patch for net_contexts
selinux-policy-targeted-1.23.18-14
----------------------------------
* Sat Jun 18 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-14
- Add Russell's patch for net_contexts
18 years, 10 months
Yum Not Excluding
by Janina Sajka
This morning I noticed my yum updates weren't taking because of two
obscure (to me) packages' missing dependencies. I tried to issue
--exclude=wl-xemacs --exclude=gda-postgres but this had no effect. I had
to rpm -e these two packages to get my 700Mb update rolling on my x86_64
box.
18 years, 11 months
Grub problem after FC4 (xfs) install ?
by mbneto
Hi,
I was wondering if anyone has a solution for a problem where after
installing fc4 (with xfs partition) grub fails to install.
In my case since it is a dual boot machine it simply enters directly
in the "primary" OS.
18 years, 11 months
Build 2.6.12 on FC4?
by Neal Becker
My wireless worked on FC3, but stopped working on FC4. I want to try a
stock 2.6.12 kernel. It doesn't compile the current FC4 gcc. I grabbed
the patch linux-2.6.11-compile-fixes.patch, and it compiles. But after
make, make modules; make modules_install; make install; it doesn't boot.
Just says can't open console (/dev/console IIRC).
Any hints?
Oh, this is x86_64.
18 years, 11 months
gcj-dbtool xerces scriptlet fails
by sean darcy
Cleanup : xerces-j2
####################### [50/57]
gcj-dbtool: Manipulate gcj map database files
Usage:
gcj-dbtool -n file.gcjdb [size] - Create a new gcj
map database
gcj-dbtool -a file.gcjdb file.jar file.so
- Add the contents of file.jar to a gcj map
database
gcj-dbtool -f file.gcjdb file.jar file.so
- Add the contents of file.jar to a gcj map
database
gcj-dbtool -t file.gcjdb - Test a gcj map
database
gcj-dbtool -l file.gcjdb - List a gcj map
database
gcj-dbtool [-][-0] -m dest.gcjdb [source.gcjdb]...
- Merge gcj map databases into dest
Replaces dest
To add to dest, include dest in the list of
sources
If the first arg is -, read the list from stdin
If the first arg is -0, filenames separated
by nul
gcj-dbtool -p [LIBDIR] - Print default
database name
error: %postun(xerces-j2-2.6.2-4jpp_7fc.x86_64) scriptlet
failed, exit status 123
sean
18 years, 11 months
Audit / Netlink slowness
by Bernardo Innocenti
Hello,
on a server running kernel 2.6.11-1.1369_FC4, both ssh
and su where taking a longish amount of time (over >1.5 sec.)
Running "strace -r 2>strace.out su", I discovered that
netlink communication is the major cause of slowdown.
"su" connects to a NETLINK_AUDIT socket 3 or 4 times.
Each time it does 2 sendto() + recvfrom() operations,
with a latency of ~200ms. This adds up to 800ms wasted
time.
Disabling CONFIG_AUDIT in the kernel makes su and ssh
very fast again.
Is this behavior to be expected? CONFIG_AUDIT is enabled
by default, therefore many people are going to be hit by
this problem.
--
// Bernardo Innocenti - Develer S.r.l., R&D dept.
\X/ http://www.develer.com/
18 years, 11 months
[PATCH] initscripts for OpenIPMI
by Matt Domsch
More progress on the initscripts to automatically load the IPMI device
drivers. initscript and config file included in this patch. Please
review and incorporate into the OpenIPMI package.
Thanks,
Matt
--
Matt Domsch
Software Architect
Dell Linux Solutions linux.dell.com & www.dell.com/linux
Linux on Dell mailing lists @ http://lists.us.dell.com
--- /dev/null Thu Apr 11 09:25:15 2002
+++ ipmi/ipmi.init Fri Jun 17 12:10:09 2005
@@ -0,0 +1,248 @@
+#!/bin/sh
+#############################################################################
+#
+# ipmi: OpenIPMI Driver init script
+#
+# Authors: Matt Domsch <Matt_Domsch(a)dell.com>
+# Chris Poblete <Chris_Poblete(a)dell.com>
+#
+# chkconfig: 2345 04 96
+# description: OpenIPMI Driver init script
+#
+### BEGIN INIT INFO
+# Provides: ipmidrv
+# Required-Start: $localfs $remotefs $syslog
+# Required-Stop: $localfs $remotefs $syslog
+# Default-Start: 2 3 4 5
+# Default-Stop:
+# Short-Description: OpenIPMI Driver init script
+# Description: OpenIPMI Driver init script
+### END INIT INFO
+#
+#############################################################################
+# for log_success_msg and friends
+[ -r /lib/lsb/init-functions ] && . /lib/lsb/init-functions
+# source config info
+[ -r /etc/sysconfig/ipmi ] && . /etc/sysconfig/ipmi
+
+#############################################################################
+# GLOBALS
+#############################################################################
+MODULE_NAME="ipmi"
+INTF_NUM=0
+
+IPMI_SMB_MODULE_NAME="ipmi_smb"
+IPMI_SI_MODULE_NAME="ipmi_si"
+kernel=`uname -r | cut -d. -f1-2`
+if [ "${kernel}" == "2.4" ]; then
+ IPMI_SMB_MODULE_NAME="ipmi_smb_intf"
+ IPMI_SI_MODULE_NAME="ipmi_si_drv"
+fi
+
+MODULES_INTERFACES="ipmi_imb ipmi_devintf"
+MODULES_FEATURES="ipmi_watchdog ipmi_poweroff"
+MODULES_HW="${IPMI_SMB_MODULE_NAME} ${IPMI_SI_MODULE_NAME}"
+MODULES_BASE="ipmi_msghandler"
+MODULES="${MODULES_INTERFACES} ${MODULES_FEATURES} ${MODULES_HW} ${MODULES_BASE}"
+
+
+RETVAL=0
+LOCKFILE=/var/lock/subsys/ipmi
+
+
+#############################################################################
+load_si()
+{
+ if [ "${IPMI_SI}" = "1" ]; then
+ modprobe ${IPMI_SI_MODULE_NAME} || RETVAL=1
+ fi
+}
+
+load_smb()
+{
+ if [ "${IPMI_SMB}" = "1" ]; then
+ modprobe ${IPMI_SMB_MODULE_NAME} || RETVAL=1
+ fi
+}
+
+load_hw_modules()
+{
+ load_si
+ load_smb
+}
+
+start_watchdog()
+{
+ if [ "${IPMI_WATCHDOG}" = "1" ]; then
+ load_hw_modules
+ modprobe ipmi_watchdog ${IPMI_WATCHDOG_OPTIONS} || RETVAL=2
+ if [ ! -x /sbin/udev -a ! -e /dev/watchdog ]; then
+ mknod -m 0600 /dev/watchdog 10 130 || RETVAL=2
+ fi
+ fi
+}
+
+stop_watchdog()
+{
+ modprobe -q -r ipmi_watchdog
+ [ ! -x /sbin/udev ] && rm /dev/watchdog
+}
+
+start_powercontrol()
+{
+ local poweroff_opts=""
+ if [ "${IPMI_POWEROFF}" = "1" ]; then
+ load_hw_modules
+ [ "${IPMI_POWERCYCLE}" == "1" ] && poweroff_opts="poweroff_control=2"
+ modprobe ipmi_poweroff "${poweroff_opts}" || RETVAL=2
+ fi
+}
+
+stop_powercontrol()
+{
+ modprobe -q -r ipmi_poweroff
+}
+
+#############################################################################
+unload_all_ipmi_modules()
+{
+ [ ! -x /sbin/udev ] && rm -f "/dev/ipmi${INTF_NUM}"
+ stop_watchdog
+ for m in ${MODULES}; do
+ modprobe -q -r ${m}
+ done
+}
+
+unload_ipmi_modules_leave_features()
+{
+ [ ! -x /sbin/udev ] && rm -f "/dev/ipmi${INTF_NUM}"
+ for m in ${MODULES_INTERFACES}; do
+ modprobe -q -r ${m}
+ done
+ lsmod | egrep -q "ipmi_poweroff|ipmi_watchdog"
+ if [ "$?" -ne "0" ]; then
+ stop_watchdog
+ for m in ${MODULES}; do
+ modprobe -q -r ${m}
+ done
+ fi
+}
+
+
+#############################################################################
+load_ipmi_modules ()
+{
+ modprobe ipmi_msghandler || RETVAL=1
+ load_hw_modules
+ [ "${RETVAL}" = "1" ] && unload_all_ipmi_modules && return
+
+ start_watchdog
+ start_powercontrol
+
+ if [ "${DEV_IPMI}" = "1" ]; then
+ modprobe ipmi_devintf || RETVAL=2
+ if [ "${RETVAL}" != "2" ]; then
+ if [ ! -x /sbin/udev ]; then
+ DEVMAJOR=`cat /proc/devices | awk '/ipmidev/{print $1}'`
+ mknod -m 0600 /dev/ipmi${INTF_NUM} c ${DEVMAJOR} 0 || RETVAL=2
+ fi
+ fi
+ fi
+
+ if [ "${IPMI_IMB}" = "1" ]; then
+ modprobe ipmi_imb || RETVAL=2
+ if [ "${RETVAL}" != "2" ]; then
+ DEVMAJOR=`cat /proc/devices | awk '/imb/{print $1}'`
+ mknod -m 0600 /dev/imb c ${DEVMAJOR} 0 || RETVAL=2
+ fi
+ fi
+
+ # Per Corey Minyard, essentially no one uses ipmi_radisys
+ # and we don't want to encourage its further use
+ # so it won't be handled here.
+ return
+}
+
+#############################################################################
+start()
+{
+ echo -n $"Starting ${MODULE_NAME} drivers: "
+ load_ipmi_modules
+ [ "${RETVAL}" = "1" ] && log_failure_msg && return
+ [ "${RETVAL}" = "2" ] && touch ${LOCKFILE} && log_warning_msg
+ [ "${RETVAL}" = "0" ] && touch ${LOCKFILE} && log_success_msg
+}
+
+#############################################################################
+stop()
+{
+ echo -n $"Stopping ${MODULE_NAME} drivers: "
+ unload_ipmi_modules_leave_features
+ rm -f ${LOCKFILE}
+ log_success_msg
+}
+
+stop_all()
+{
+ echo -n $"Stopping ${MODULE_NAME} drivers: "
+ unload_all_ipmi_modules
+ rm -f ${LOCKFILE}
+ log_success_msg
+}
+
+#############################################################################
+restart()
+{
+ stop_all
+ start
+}
+
+#############################################################################
+status ()
+{
+ for m in ${MODULES}; do
+ if /sbin/lsmod | grep $m >/dev/null 2>&1 ; then
+ echo "$m module loaded"
+ else
+ echo "$m module not loaded"
+ fi
+ done
+}
+
+usage ()
+{
+ echo $"Usage: $0 {start|stop|status|restart|condrestart|" 1>&2
+ echo $" start-watchdog|stop-watchdog|" 1>&2
+ echo $" start-powercontrol|stop-powercontrol|" 1>&2
+ echo $" stop-all}" 1>&2
+ RETVAL=1
+}
+
+condrestart ()
+{
+ [ -e ${LOCKFILE} ] && restart
+}
+
+#############################################################################
+# MAIN
+#############################################################################
+case "$1" in
+ start) start ;;
+ stop) stop ;;
+ restart) restart ;;
+ status) status ;;
+ condrestart) condrestart ;;
+ start-watchdog) start_watchdog ;;
+ stop-watchdog) stop_watchdog ;;
+ start-powercontrol) start_powercontrol ;;
+ stop-powercontrol) stop_powercontrol ;;
+ stop-all) stop_all ;;
+ *) usage ;;
+esac
+
+exit ${RETVAL}
+
+#############################################################################
+# end of file
+#############################################################################
+
--- /dev/null Thu Apr 11 09:25:15 2002
+++ ipmi/ipmi.sysconf Fri Jun 17 12:10:15 2005
@@ -0,0 +1,36 @@
+# Enable standard hardware interfaces (KCS, BT, SMIC)
+# You probably want this enabled.
+IPMI_SI=1
+
+# Enable nonstandard interfaces (SMB via i2c)
+# IPMI_SMB=1
+
+# Enable /dev/ipmi0 interface, used by ipmitool, ipmicmd,
+# and other userspace IPMI-using applications.
+# You probably want this enabled.
+DEV_IPMI=1
+
+# Enable IPMI_WATCHDOG if you want the IPMI watchdog
+# to reboot the system if it hangs
+# IPMI_WATCHDOG=1
+#
+# Watchdog options - modinfo ipmi_watchdog for details
+# watchdog timeout value in seconds
+# as there is no userspace ping application that runs during shutdown,
+# be sure to give it enough time for any device drivers to
+# do their cleanup (e.g. megaraid cache flushes)
+# without the watchdog triggering prematurely
+IPMI_WATCHDOG_OPTIONS="timeout=60"
+
+# Enable IPMI_POWEROFF if you want the IPMI
+# poweroff module to be loaded.
+# IPMI_POWEROFF=1
+
+# Enable IPMI_POWERCYCLE if you want the system to be power-cycled (power
+# down, delay briefly, power on) rather than power off, on systems
+# that support such. IPMI_POWEROFF=1 is also required.
+# IPMI_POWERCYCLE=1
+
+# Enable "legacy" interfaces for applications
+# Intel IMB driver interface
+# IPMI_IMB=1
18 years, 11 months