You are not logged in.
Pages: 1
Hi, I'm having random freezes in Arch Linux around 2 days, before that i had no issues, I'm facing this random freeze (happens in both Gnome Desktop and i3wm that i have) which makes system completely unresponsive, I even cannot enter TTYs usinng CTRL + ALT + Fx, or any other key combination
It was working fine, since i had 3 changes in system, first change was i changed my swap partition, i deleted swap from my main ssd and set new swap on external usb hdd and updated my fstab accordingly,
The Second change is that i used to connect to internet using LAN, but I'm using Wifi now, note to mention that i have 2 WLAN interfaces, one is the PCI that laptop has, the second is 802.11 Usb wireless card which i installed 8188fu-kelebek333-dkms-git to get it to work
The Third change is that i installed nvidia-390xx modules and deletetd nouveau, now it is using nvidia (i have hybrid, intel and nvidia, even though i installed it and it is in use, but nvidia-smi gives error)
I suppose anything that causes freeze is related to these 3 changes i mentioned, cause i had no issues before, i also have not updated any kernel driver or etc since that time
also here is my journal -b-1 (freeze happened around 23:20)
I tried installing intel-ucode as microcode and regenerated my grub.cfg and rebooted, still not working, i also tried to unplug the 802.11 Usb Wireless card and reboot and it didn't work either
This is my uname -a :
^^>>> uname -a 23:35:56
Linux erfan 5.9.13-arch1-1 #1 SMP PREEMPT Tue, 08 Dec 2020 12:09:55 +0000 x86_64 GNU/LinuxLast edited by erfanjoker (2020-12-15 12:07:53)
Offline
i deleted swap from my main ssd and set new swap on external usb hdd
So was your system out of memory when you experienced the freeze? If so then earlyoom might help (or add more memory).
i installed nvidia-390xx modules and deletetd nouveau
If in doubt blame the blob ![]()
nvidia-smi gives error
https://bbs.archlinux.org/viewtopic.php?id=57855
here is my journal -b-1
That shows a core dump for gnome-software, to investigate it see https://wiki.archlinux.org/index.php/De … ing_Traces
I tried installing intel-ucode as microcode and regenerated my grub.cfg and rebooted, still not working
Have you confirmed that the µcode is actually loaded? The ArchWki can tell you how to do that.
Jin, Jîyan, Azadî
Offline
According to the journal you're running on the intel chip, so it's not necessarily surprising that nvidia-smi can't use it (because it's possibly powered down) and less likely that the nvida blob is the cause of your freezes.
swap over usb (2?) sounds more like it, notably if the external drive is "victim" usb autosuspense.
Online
Hi, thanks for your response,
I have to mention that i had no freezes since yesterday (i updated linux kernel, also i changed my connection from wifi to LAN again, just don't know which did the trick)
So was your system out of memory when you experienced the freeze?
Nah ! I have 12GB of ram and it was only using 2GB of it, also swap usage was 0 all the time (before freeze, cause in the moment of freeze i have no access to system monitor to see if swap is using or not)
If in doubt blame the blob big_smile
I would like to, but it is not working ![]()
Also for nvidia-smi problem, i don't know which log should i attach, if u can guide me to, i would be really happy to,
Im gonna make a new thread for my nvidia problem
Have you confirmed that the µcode is actually loaded? The ArchWki can tell you how to do that.
^^>>> dmesg | grep microcode 15:35:27
[ 0.104646] SRBDS: Vulnerable: No microcode
[ 0.764700] microcode: sig=0x306a9, pf=0x10, revision=0x21
[ 0.764746] microcode: Microcode Update Driver: v2.2.and the final question, now that im not having the freeze, do i have to keep microcode ? is it any good at all when system is acting fine ?
Offline
According to the journal you're running on the intel chip, so it's not necessarily surprising that nvidia-smi can't use it (because it's possibly powered down) and less likely that the nvida blob is the cause of your freezes.
swap over usb (2?) sounds more like it, notably if the external drive is "victim" usb autosuspense.
I have :
^^>>> lspci | grep VGA 15:38:28
00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)
01:00.0 VGA compatible controller: NVIDIA Corporation GF119M [GeForce 610M] (rev a1)Installed nvidia-390xx and deleted nouveau, i suppose nvidia-smi should give me info and nvidia should work fine
Offline
I have 12GB of ram and it was only using 2GB of it
Ftr, a memory leak can easily waste that in a split second, forcing a swapout of anon pages.
i suppose nvidia-smi should give me
Not if the chip is powered down.
What's the actual nvidia-smi output as requested by HoaS?
Though the journal randomy has
Dec 13 23:19:45 erfan bumblebeed[389]: modprobe: FATAL: Module bbswitch not found in directory /lib/modules/5.9.13-arch1-1
Dec 13 23:19:45 erfan bumblebeed[383]: [ 5.142113] [ERROR]Module bbswitch could not be loaded (timeout?)
Dec 13 23:19:45 erfan bumblebeed[383]: [ 5.142400] [WARN]No switching method available. The dedicated card will always be on.
Dec 13 23:19:45 erfan bumblebeed[383]: [ 5.144483] [INFO]/usr/bin/bumblebeed 3.2.1 startedyou should verify the power state, https://wiki.archlinux.org/index.php/Bu … management
Sidenote: please don't use the ubuntu pastebin - it's horrible and requires one to register in order to download the plain text.
Online
Ftr, a memory leak can easily waste that in a split second, forcing a swapout of anon pages.
So you mean that it could be because of memory leak ? Im surely aware of how fast RAM could be filled, but memory leak of what ? 802.11 driver ? now that i have it off it seems to working fine
Not if the chip is powered down.
What's the actual nvidia-smi output as requested by HoaS?
I have created a topic for this, i would be thankful if you attend
Last edited by erfanjoker (2020-12-15 13:23:27)
Offline
idk, but if there *is* a mem leak, any notion of "I've a lot of RAM" is void to argue that the swap cannot be used.
Fyi, "802.11" is not a model or driver but https://en.wikipedia.org/wiki/IEEE_802.11 - the thing you're using is a realtek rtl8188fu chip - the ot-of-tree module certainly has the capacity to crash the kernel - see https://wiki.archlinux.org/index.php/Kdump on how to debug that.
Online
idk, but if there *is* a mem leak, any notion of "I've a lot of RAM" is void to argue that the swap cannot be used.
Fyi, "802.11" is not a model or driver but https://en.wikipedia.org/wiki/IEEE_802.11 - the thing you're using is a realtek rtl8188fu chip - the ot-of-tree module certainly has the capacity to crash the kernel - see https://wiki.archlinux.org/index.php/Kdump on how to debug that.
The driver i used was 8188fu-kelebek333-dkms-git , I also have to ask whats the difference between 8188fu-kelebek333-dkms-git and 8188fu-dkms-git ? could the unstability be related to kelebek fork of this driver ?
Offline
Pinned comment: https://aur.archlinux.org/packages/8188 … ned-778662
Original code is https://github.com/ulli-kroll/rtl8188fu (and Kroll seems to work on many realtek drivers)
Online
Pages: 1