What finally got my existing Win10 guest VM (originally upgraded from Win7)
to see a TPM2 device was the following process:
1) Make a copy of the raw disk image. This isn't really necessary, but I did
this for backup purposes.
2) In virt-manager I created a new VM, pointing to the existing raw disk
image, and directed it to use the canned "Windows 10" operating system
configuration, and also selecting the manual configuration option, and then
using that to add the TPM2 module. Google searches seemed to suggest that
the "TIS" model is the one to use, so that's what I selected.
3) This creates a new VM with the emulated "Q35" chipset, rather thanthe
existing i440fx chipset, that Win7 was originally configured for. This is
the trick, apparently.
When I booted the new VM, Windows 10 went through some noticable setup and
reconfiguration, but it did survive the transplant. The only result was it
required me sign into my Microsoft account (I recommend that the Win10 seat
be registered to a Microsoft account before doing this, this appears to be
the simplest way to avoid license/activation problems). Over the next couple
of minutes Windows10 also popped up occasional prompts about setting up this
PCI device, or that PCI device. But nothing seemed to indicate a problem
with the new VM.
And it now sees the TPM2 device, however it does show a "Device health
attestion isn't available" because "Your device does not support this
feature". After a few minutes it offered me the option to "Clear TPM" to fix
this issue (initially the button was disabled, but it became enabled a few
minutes after the boot). However that made no difference, this status
remained after the "Clear TPM" and the reboot.
I have another Win10 license which I'll try, later, with the other "CRB"
emulated TPM model, to see if that works fully. It's also possible that this
is it's way of expressing that it knows it's running in a VM. And it now
sees the TPM2 device, however it does show a "Device health attestion isn't
available" status because "Your device does not support this feature".
Is anyone else getting this error in the "Security processor
troubleshooting" (21H1, with all updates installed)?
I have a stream of messages on the console, as follows:
Message from syslogd@Bree at Sep 23 03:56:24 ...
kernel:watchdog: BUG: soft lockup - CPU#0 stuck for 4802s! [kworker/0:0:323591]
This is F34, fully updated. I've not seen this before.
I have been looking at the command line options
and I am confused.
I need to create a live usb with the following:
8 GB flash drive (I will do a dd /dev/zero wipe first)
safe boot if possible
Mate (I have the iso)
small persistent /home directory
I need to install xrdp on it and have xrdp persist
Any idea what the command line should be?
Some told me they thought that browsers talked directly to the
With upgrading from F33 (X/XFCE4/ALSA) to F34 (Wayland/Sway/PipeWire) I
have been thrashing the system a bit to see how it goes and found trying
to serially record stuff online from three different browsers and also
recording stuff with a number of apps causes me to lose access to the
default mic for ALL inputs . .
Since I don't have any trouble PLAYING from as many sources as I want to
but ONLY recording, I have turned OFF access to the mic for all the
browsers (C, FF, B) and will try to only use stand-alone apps that need
the mic (using default) ie audacity, jisti, obs, zoom, arecord etc - I
will see how that approach goes for a while.
Also, because of problems in the early days with PulseAudio, I got into
the habit of routinely uninstalling it and just using ALSA plus my
.asoundrc file - now I thought PW was going to completely replace PA but
I see there are still a number of RPMs installed:
Any info on how this stuff is supposed to work / hang together would be
appreciated! - a diagram somewhere would be great!
PO Box 896
Cowra NSW 2794
I have fedora34 installed on my desktop with three monitors; two of
them are older 27" Samsungs, and the third is a new 34" Samsung
S34J55x. Occasionally, the 34" seems to shut off for a second, then
comes back on and continues to work normally.
Has anyone seen this type of behavior before? The other two monitors
have never done this. However, when it happens, the other two seem to
follow with shutting off for a second. It always seems to be driven by
the new 34" monitor.
The problem is that, when this happens, all of the windows on the 34"
monitor move to the 27" monitor on the right once the monitors turn on
This is not a power problem, as the entire computer is connected to
the same power strip, backed by a 2200W UPS.
My video card is also new - it's a Radeon RX570. All three monitors
are connected via displayport.
I do see lines in my Xorg.0.log file periodically along the lines off
[992023.990] (EE) event3 - Logitech Gaming Mouse G900: client bug:
event processing lagging behind by 16ms, your system is too slow
[994674.766] (EE) event6 - Logitech Gaming Keyboard G910: client bug:
event processing lagging behind by 29ms, your system is too slow
But are they related? Is this also an issue? The system itself is a
12-core with 64GB of RAM, so certainly not slow.
Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
Please let me know what more I can do to troubleshoot this before
contacting Samsung directly.
I've been doing testing on a setup for my G4L project to
get both legacy and UEFI booting working on a single
setup. Have that working just fine now with help from
this list. Been testing with my Dell 9020 machine that has
Windows 10 as the hard disks OS that it came with. My 5
other machines have Fedora 33. Everything works fine,
except. Sometimes I miss hitting the F12 key to select the
flash for UEFI or standard USB boot from flash. Thus the
machine boots to windows. Problem, if I do a restart after
the windows boots and go to the my projects linux, it
loads fine, but it will not get IP from dhcpd server of my
wifi router. Have to do a full power off on system. Then it
will boot ok. Found a web site that seemed to identify
problem as windows doing something to the nic on
shutdown that causes the issue. The sites solution was
the echo a 1 to the reset value for the PCI device for the
nic. Came up with a process that does that easy, but
doesn't seem to work. Might be after some time, the issue
resolves, but thought the reset would be a quick solution.
Shutting down the interface with ifconfig doesn't reset it
either? Wonder if there is a command line process that
someone knows of? I can manually set an ipaddress, and
that works fine, but you have to know the network range.
Don't have network manager in the build, so right now
solution is to do a poweroff, and then power back on.
Thanks. Got the UEFI booting just fine, and have regular
boot running with grub4dos just fine. Systems using same
exact kernel and ramdisk on the flash.
Thanks again. Have a great day.
Latest version of Thunderbird does integrate OpenPGP and does provide end-to-end encryption.
There is, IMO, a security flaw. The "Generate Key" function they supply does not have the
option to create a private key with a pass phrase.
This means if you get an encrypted message, walk away from you system, and forget to secure it
anyone can click on an encrypted message it will be displayed.
It should allow for pass phrase. And, there should be a "timeout" setting for how long a
vaild pass phrase has been entered for viewing messages.
Nothing to see here.
This morning when I botted my laptop (Dell G15) and the login didn' appear, just
the mouse arrow. I thougt it was a sddm problem so I switched to gdm and it
worked and Gnome started but not KDE(X11), not even with a new user. I installed
LXDE and that too worked without problems but it gave me an error about
KDEconnect(crashed) and the same when I tried to start Konsole, while the LXDE
Journalctl reports a lot of kdesomething that dumped core,the first seems to be
It seems that KDE is, for some reason hosed. Is it possible to do a reinstall of
KDE or reinstalling F34 is the only solution.
GianPiero Puccioni |Istituto dei Sistemi Complessi-CNR
gianpiero.puccioni(a)isc.cnr.it |Via Madonna del Piano, 10
T:+39 0555226682 |50019 Sesto F. (Firenze) ITALY
I installed mpich
and set the path to
however, it seems that I need
Checking if MPI is available
checking for mpif90... mpif90
checking for MPI_Init... no
checking for MPI_Init in -lmpi... no
checking for MPI_Init in -lmpich... no
configure: error: MPI not found: openmpi
Do you know how I could get this MPI_init ?
Patrick DUPRÉ | | email: pdupre(a)gmx.com
Laboratoire interdisciplinaire Carnot de Bourgogne
9 Avenue Alain Savary, BP 47870, 21078 DIJON Cedex FRANCE
Tel: +33 (0)380395988 | | Room# D114A
My external network connection was set up before the days of NetworkManager,
configured in a /etc/sysconfig/network-scripts ifcfg file.
It has a bunch of IPADDR settings, for the public IPs.
"nmcli connection show" shows the connection just fine. "ipv4.addresses"
lists the IP addresses that get fished out of the ifcfg file, for example.
There's nothing in /etc/NetworkManager/system-connections so I take that
NetworkManager is reading the initscripts directly.
I need to change these IP addresses. What's the least painful way to do that?
Just edit the ifcfg file, and restart Network Manager?
Attempt to update them via nmcli (I have to work out the syntax)? Will that
work, and update the ifcfg files?