[F25] help: semi-broken wifi on new Dell laptop
by Sherman Grunewagen
I just finished configuring F25 on a new Dell precision 7520
(essentially same as 7510) with Samsung's PM961 1TB PCIe M.2 SSD
"disk". Wonderously fast! Boots in like 2 seconds. KDE comes up in 5.
HOWEVER:
In the boot messages (/var/log/messages) I see these lines:
kernel: Intel(R) Wireless WiFi driver for Linux
...
kernel: iwlwifi 0000:02:00.0: Direct firmware load for iwlwifi-8265-28.ucode failed with error -2
...
kernel: iwlwifi 0000:02:00.0: capa flags index 3 larger than supported by driver
kernel: iwlwifi 0000:02:00.0: loaded firmware version 27.455470.0 op_mode iwlmvm
...
kernel: iwlwifi 0000:02:00.0: Detected Intel(R) Dual Band Wireless AC 8265, REV=0x230
kernel: iwlwifi 0000:02:00.0: L1 Enabled - LTR Enabled
kernel: iwlwifi 0000:02:00.0: L1 Enabled - LTR Enabled
kernel: iwlwifi 0000:02:00.0 wlp2s0: renamed from wlan0
...
There's a few more iwlwifi related messages when NetworkManager comes
up. No errors. Things appear to work fine. Then sometime later
(could be minutes or hours, in the middle of a large download,
the transfer rate goes to nearly 0 and a boatload of messages like this
dump into /var/log/messages:
kernel: iwlwifi 0000:02:00.0: Microcode SW error detected. Restarting 0x82000000.
kernel: iwlwifi 0000:02:00.0: CSR values:
kernel: iwlwifi 0000:02:00.0: (2nd byte of CSR_INT_COALESCING is CSR_INT_PERIODIC_REG)
kernel: iwlwifi 0000:02:00.0: CSR_HW_IF_CONFIG_REG: 0X00489000
kernel: iwlwifi 0000:02:00.0: CSR_INT_COALESCING: 0X00000040
kernel: iwlwifi 0000:02:00.0: CSR_INT: 0X00000000
kernel: iwlwifi 0000:02:00.0: CSR_INT_MASK: 0X00000000
kernel: iwlwifi 0000:02:00.0: CSR_FH_INT_STATUS: 0X00000000
kernel: iwlwifi 0000:02:00.0: CSR_GPIO_IN: 0X00000010
kernel: iwlwifi 0000:02:00.0: CSR_RESET: 0X00000000
...
After this, the "Microcode SW ..." message appear every few
seconds or minutes with the boatload of CSR values.
I have to reboot to get things right. Then it (intermittently)
happens again on some future download.
Is this a hardware bug that I'll be stuck with forever? If so,
I have only a short time to return the Dell before "it's mine".
Your help will be greatly appreciated!
By the way: "lspci | fgrep -i wireless" returns
02:00.0 Network controller: Intel Corporation Wireless 8265 / 8275 (rev 78)
-Sherman
6 years, 9 months
F26: particular keystrokes cause Xorg to freeze momentarily,
looking for debugging suggestions
by Daniel Miranda
Hello fellow Fedora users,
I've been using the same Fedora installation on my desktop for the last 4 years, but for the first time I'm completely stumbled with a bug after upgrading from Fedora Workstation 25 to 26.
When I press the slash/question-mark (on a PT-BR keyboard) key while focused into a window, or when pressing backspace to erase that slash immediately afterwards, X11 will freeze for about a second consuming 100% CPU. No other key causes the same behavior, not even the slash key from the numeric keypad. I can observe the issue in GTK2, GTK3, Qt and custom applications (such as Chrome and Firefox), but I cannot reproduce the behavior when sending inputs to `xev` for observation. Even GDM is afflicted when typing my user password, before ever having logged in.
Given the bizarre and extremely specific nature of the issue, my first suspects are input-related software such as ibus, libinput or Xorg itself. Using different ibus settings, and even attempting to disable it (by setting XMODIFIERS=@im=none) didn't seem to have any effect. libinput doesn't seem like a good candidate, as raw events captured by `xev` don't trigger the issue.
I profiled Xorg using `perf` but could not see any clues to what the problem is. I generated a flame-graph [1] (source data at [2]), but most of the symbols seem to be missing, probably courtesy of the NVIDIA proprietary driver.
I'm not even sure to which software to report this bug. Is anyone experiencing anything similar, or has any suggestions for ways to debug to get to the core of the problem?
[1]: https://linux.ime.usp.br/~danielqm/f26-x11-perf/xorg.svg
[2]: https://linux.ime.usp.br/~danielqm/f26-x11-perf/xorg.folded
--
Thanks in advance,
Daniel
6 years, 9 months
Blocking root login via PAM
by CLOSE Dave
If I run ">/etc/securetty" and add "auth [user_unknown=ignore success=ok
ignore=ignore default=bad] pam_securetty.so" to various /etc/pam.d/*
files (as described in
<https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/4/...>),
root login on the console is blocked as it should be. But the user gets
no hint of the problem, merely a refreshed login screen. Is there a way
to display a message in this case?
--
Dave Close
"Age is a very high price to pay for maturity." -- Tom Stoppard
6 years, 9 months
Only one problem after F25 -> F26
by Frank Elsner
Hello,
my update from version 25 to 26 of Fedora went (nearly) painless
but I have one annoying problem:
After succesfull login two windows come up. One asking for the
admin password to configure network, the other for using rfkill.
If I cancel both the system still works perfect.
What't behind? How to stop this behaviour?
Kind regards, Frank Elsner
6 years, 9 months
Linux anti-virus any good?
by William
Good evening,
A few years ago, I found a web site "https://www.av-test.org" when
trying to find comparisons of windows-7 anti-virus software. I more
recently re-visited that site, and found an article on Linux and
anti-virus software for Linux. It's here:
"https://www.av-test.org/en/news/news-single-view/linux-16-security-packag...".
I have a few questions for this list:
1. How independent and objective is the group doing these tests,
comparisons, and evaluations? This is important for knowing just how
much weight to give what they say.
2. If you've had any experience with any of the anti-virus packages
reported on in that article, especially those that received high scores
for workstations, what is your review of that/those packages?
I do see that this article is nearly 2 years old. And I do realize that
nothing gives me 100% protection or detection.
thanks,
Bill.
6 years, 9 months
routes not being applied to interfaces
by InvalidPath
Pre-F26 upgrade I used a route file to apply work related routes while
docked, for some reason since the upgrade when I look via route or ip route
I am no longer getting the ones from the file applied.
I did verify the name of my interface is correct.. the name changed once
after an update so that's always the first thing I check.
I also have the same routes in the Route tab in Network Manager, and
looking in journalctl I see:
Jul 17 08:16:53 Vostok audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295
ses=4294967295 msg='unit=NetworkManager-dispatcher comm="systemd"
Jul 17 08:16:53 Vostok dhclient[3662]: bound to 10.0.20.186 -- renewal in
1122048 seconds.
Jul 17 08:16:53 Vostok nm-dispatcher[3754]: req:1 'pre-up' [enp62s0u1u2]:
new request (1 scripts)
Jul 17 08:16:53 Vostok NetworkManager[1158]: <info> [1500301013.5560]
device (enp62s0u1u2): state change: ip-check -> secondaries (reason 'n
Jul 17 08:16:53 Vostok NetworkManager[1158]: <info> [1500301013.5563]
device (enp62s0u1u2): state change: secondaries -> activated (reason '
Jul 17 08:16:53 Vostok dnsmasq[1477]: reading /etc/resolv.conf
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.20.19#53
Jul 17 08:16:53 Vostok NetworkManager[1158]: <info> [1500301013.5650]
device (enp62s0u1u2): Activation: successful, device activated.
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.20.1#53
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.22.1#53
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.20.20#53
Jul 17 08:16:53 Vostok nm-dispatcher[3754]: req:2 'up' [enp62s0u1u2]: new
request (7 scripts)
Jul 17 08:16:53 Vostok nm-dispatcher[3754]: req:2 'up' [enp62s0u1u2]: start
running ordered scripts...
Jul 17 08:16:53 Vostok systemd[1]: iscsi.service: Unit cannot be reloaded
because it is inactive.
Jul 17 08:16:53 Vostok NetworkManager[1158]: <info> [1500301013.7452]
policy: set 'USB-c Ethernet' (enp62s0u1u2) as default for IPv4 routing
Jul 17 08:16:53 Vostok dnsmasq[1477]: reading /etc/resolv.conf
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.20.19#53
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.20.20#53
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.20.1#53
Jul 17 08:16:53 Vostok dnsmasq[1477]: using nameserver 10.0.22.1#53
Any thoughts?
Thanks
6 years, 9 months
rpmbuild
by Patrick Dupre
Hello,
I am trying to patch gnuplot with
https://sourceforge.net/p/gnuplot/patches/694/
However, it fails for several reasons,
The first one is that the patch does not apply to last version
of tabulate.c (maybe to a previous one).
Then a fixed directly the tabulate.c file
How can generate the package directly from the SOURCES?
with rpmbuild
I try to modify the .spc file by removing the
#%patch0 -p1 -b .refto
#%patch1 -p1 -b .font
#%patch3 -p1 -b .plot-sigsegv
#%patch4 -p1 -b .isinglethread
#%patch5 -p1 -b .checkint
#%patch6 -p1 -b .tabulate # This fails
But it does not work.
because I guess that it does not compile my new sources but the old ones.
Actually, %patch6 -p1 -b .tabulate
would fails
patch -p0 gnuplot-5.0.6a-tabulate.patch (manually)
would be OK, but it fails from the .spec file
Thank for your help.
===========================================================================
Patrick DUPRÉ | | email: pdupre(a)gmx.com
Laboratoire de Physico-Chimie de l'Atmosphère | |
Université du Littoral-Côte d'Opale | |
Tel. (33)-(0)3 28 23 76 12 | | Fax: 03 28 65 82 44
189A, avenue Maurice Schumann | | 59140 Dunkerque, France
===========================================================================
6 years, 9 months
F26 and blurry fonts
by Heinz Diehl
Hi,
for those who love razorsharp fonts just like me but got the new v40 interpreter
coming with freetype-2.7 enforced on them when installing F26: the good old v35
interpreter is still there (and I hope it will be forever). You can
use it by setting the environment variable
FREETYPE_PROPERTIES=truetype:interpreter-version=35
If you also want subpixel rendering, which now is disabled by default,
you can dig into the foption.h file in the config directory of the
freetype sourcecode, uncomment the option and recompile.
I have never been able to understand why people like the blurry and
muddy Micro$oft-like font rendering... gaah.. Now we have to hack the
source to get rid of that crap.
https://tinyurl.com/ybf3fkwv
Cheers,
Heinz
6 years, 9 months