Some random notes about several topics
This project is maintained by jgbarah
Finally I have my Katana II. I ordered it with Debian, but after receiving it, and testing that it seemed to work well with Debian 9.0, the version that had installed, I decided to move to Debian testing, which is my usual system.
These are some notes on installing Debian testing on this Katana II.
Since Debian 10 is close (I’m writing in March 2019), I used the installers that are being prepared for it. I used the network installed, and it worked like a charm. No specific problem, except for needing the one supporting proprietary firmware (for the WiFi card).
Everything seems to work: USB, WiFi, sound, camera…
Just before tring to mount the encrypted volume, I see this message, produced by the kernel:
[Firmware Bug] TSC_DEADLINE disabled due to errata: please update
microcode to version: 0x52 (or later)
Fortunately, the fix is easy:
% sudo apt-get install intel-microcode
I found out about Slimbook Battery 3. I installed it, and it seems the battery like is increased by about 25%-30% (not exact accounts) wiht the default values for the balanced working model.
Installing it in Debian buster is a bit tricky, because python-appindicator is no longer available as a package. So, I used two packages from Debian strech: python-appindicator_0.4.92-4_amd64.deb and libappindicator1_0.4.92-4_amd64.deb.
So, I downloaded both of them. Then, I added the Slimbook PPA repository to my list of apt repositories. And finally, I apt-installed some stuff that are dependencies for the libappindicator1 and python-appindicator packages, those packages themselves (via dpkg) and finally slimbookbattery:
% sudo add-apt-repository "deb http://ppa.launchpad.net/slimbook/slimbook/ubuntu disco main"
% sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys BE80F1EEB3838E61E42091B378A22399981017FC
% sudo apt-get update
% sudo apt-get install libdbusmenu-gtk4 libindicator7 python-gobject
% sudo dpkg -i libappindicator1_0.4.92-4_amd64.deb python-appindicator_0.4.92-4_amd64.deb
% sudo apt-get install slimbookbattery
It is important to follow this order, otherwise you can get inconsistent states in installed packages (that can be easily solved with sudo apt --fix-broken install, but anyway…).
To get the SlimbookBattery icon in the top icon bar, you also need to install the corresponding GNOME Shell extension, which is not installed by default:
% sudo apt-get install gnome-shell-extension-appindicator
After this, I run the configuration GUI, by searching the apps, and launching the one with the SlimbookBattery and the gear wheel. The only option I change is, in Balanced Mode, “Wi-Fi disabled when not in use”. This seems to prevent that sometimes, when booting after hibernation, WiFi was not active.
After saving configuration (OK button), I see errors in the console like:
sh: 1: prime-select: not found
but it seems that’s only a warning because I don’t have NVIDIA graphics cards in my Katana II. Which is fine.
Now, I’m seeing an increasy in battery life of about 25%-30%. Nice.
To get the icon in the top bar, I still had to reboot the laptop (maybe rebooting the GNOME Shell would be enough), and run the GNOME Tweak Tool upon reboot to activate the opton for getting Appindicator notifications: menu “Extensions”, option “KStatusNotifierItem/AppIndicator”.
[ It seems this is no longer needed, after I installed and configured
slimbookbattery, see above. It includes exactly the same file
described below ;-) However, I’m letting the instructions here, just in case. ]
When suspending, batteries seem to drain rather quickly. This can be fixed by making the laptop hibernate after staying some time in suspension. I decide to do that after three hours in suspension, which seems a good trade between battery duration and booting quickly.
The process is described in SystemdSuspendSedation in the Debian Wiki, and is easy. First, create this file:
# /etc/systemd/system/suspend-sedation.service
[Unit]
Description=Hibernate after suspend
Documentation=https://bbs.archlinux.org/viewtopic.php?pid=1420279#p1420279
Documentation=https://bbs.archlinux.org/viewtopic.php?pid=1574125#p1574125
Documentation=https://wiki.archlinux.org/index.php/Power_management
Documentation=http://forums.debian.net/viewtopic.php?f=5&t=129088
Documentation=https://wiki.debian.org/SystemdSuspendSedation
Conflicts=hibernate.target hybrid-sleep.target
Before=sleep.target
StopWhenUnneeded=true
[Service]
Type=oneshot
RemainAfterExit=yes
Environment="ALARM_SEC=10800"
Environment="WAKEALARM=/sys/class/rtc/rtc0/wakealarm"
ExecStart=/usr/sbin/rtcwake --seconds $ALARM_SEC --auto --mode no
ExecStop=/bin/sh -c '\
ALARM=$(cat $WAKEALARM); \
NOW=$(date +%%s); \
if [ -z "$ALARM" ] || [ "$NOW" -ge "$ALARM" ]; then \
echo "suspend-sedation: Woke up - no alarm set. Hibernating..."; \
systemctl hibernate; \
else \
echo "suspend-sedation: Woke up before alarm - normal wakeup."; \
/usr/sbin/rtcwake --auto --mode disable; \
fi \
'
[Install]
WantedBy=sleep.target
Then, just enable it:
% sudo systemctl enable suspend-sedation
The background noise I hear when speaking to the microphone can be removed (in part) by activating the ALSA echo cacelling module. I follow these instructions, and (in part) these recommendations.
First, edit /etc/pulse/default.pa, adding at the end of the file:
#load-module module-echo-cancel source_name=noiseless aec_method=webrtc aec_args="analog_gain_control=0\ digital_gain_control=0"
load-module module-echo-cancel source_name=noiseless aec_method=webrtc
...
set-default-source noiseless
To activate the new configuration:
$ pulseaudio -k
This issues were a problem, but got solved just by upgrading to new versions of Debian over time.
I see restarts of the WiFi driver from time to time (apparently, after periods of intense WiFi use). The sympthom is continuous “ieee80211 phy0: Hardware reset was requested” messages:
[ 1098.438897] ieee80211 phy0: Hardware restart was requested
[ 1102.911225] iwlwifi 0000:02:00.0: Error sending STATISTICS_CMD: time out after 2000ms.
[ 1102.911240] iwlwifi 0000:02:00.0: Current CMD queue read_ptr 93 write_ptr 94
[ 1102.911404] iwlwifi 0000:02:00.0: HW error, resetting before reading
[ 1102.918383] iwlwifi 0000:02:00.0: Start IWL Error Log Dump:
[ 1102.918408] iwlwifi 0000:02:00.0: Status: 0x00000100, count: -560800355
[ 1102.918421] iwlwifi 0000:02:00.0: Loaded firmware version: 38.755cfdd8.0
[ 1102.918434] iwlwifi 0000:02:00.0: 0x98CB9419 | ADVANCED_SYSASSERT
...
[ 1102.919062] ieee80211 phy0: Hardware restart was requested
...
In some cases, this leads to the WiFi driver stop working, and to kernel panics.
This seems to be a known problem affecting many computers with the same WiFi driver, Wireless-AC 9260.
Apparently, it is fixed in kernel 4.20 (current one in testing is 4.19),
with firmware -41.ucode (current one for testing are -41.ucode
and -38.ucode, in package firmware-iwlwifi).
I think my only chance is, for now, to wait until 4.20 is in Debian…
Maybe I’m having some other problem. The trace I get with dmesg is maybe a bit different from what the issue mentioned above found:
[ 77.233370] iwlwifi 0000:02:00.0: firmware: direct-loading firmware iwlwifi-9260-th-b0-jf-b0-38.ucode
[ 77.234358] iwlwifi 0000:02:00.0: loaded firmware version 38.755cfdd8.0 op_mode iwlmvm
[ 77.276269] Adding 16691196k swap on /dev/mapper/librito--vg-swap_1. Priority:-2 extents:1 across:16691196k SSFS
[ 77.287562] iwlwifi 0000:02:00.0: Detected Intel(R) Dual Band Wireless AC 9260, REV=0x324
[ 77.336160] iwlwifi 0000:02:00.0: base HW address: 18:56:80:ee:67:22
...
[ 83.325684] iwlwifi 0000:02:00.0: Unhandled alg: 0x707
...
[ 262.735955] iwlwifi 0000:02:00.0: Error sending SCAN_CFG_CMD: time out after 2000ms.
[ 262.735973] iwlwifi 0000:02:00.0: Current CMD queue read_ptr 194 write_ptr 195
[ 262.736138] iwlwifi 0000:02:00.0: HW error, resetting before reading
[ 262.743138] iwlwifi 0000:02:00.0: Start IWL Error Log Dump:
[ 262.743154] iwlwifi 0000:02:00.0: Status: 0x00000100, count: -560800353
[ 262.743162] iwlwifi 0000:02:00.0: Loaded firmware version: 38.755cfdd8.0
[ 262.743171] iwlwifi 0000:02:00.0: 0x98CB941A | ADVANCED_SYSASSERT
...
[ 262.743703] iwlwifi 0000:02:00.0: 0x48E61756 | isr status reg
[ 262.743754] ieee80211 phy0: Hardware restart was requested
[ 266.323698] iwlwifi 0000:02:00.0: Queue 0 is inactive on fifo 2 and stuck for 2500 ms. SW [113, 114] HW [162, 162] FH TRB=0x0a5a5a5a2
[ 266.324963] iwlwifi 0000:02:00.0: Hardware error detected. Restarting.
[ 266.325127] iwlwifi 0000:02:00.0: HW error, resetting before reading
[ 266.332181] iwlwifi 0000:02:00.0: Start IWL Error Log Dump:
...
This seems to happen only when conneted to WiFi networks using the 5 GHz band, so a temporary fix is to connect only to 2.4 GHz bands… To know that the proble is not appearing, just run:
sudo dmesg | grep restart
If no new line appears, this problem is now showing up.
Note from my future self: Now, well above 4.20, I’ve never experienced these problems again.