I think I found a quite severe bug, what is the next step?
Howdy, I started this new topic, because I think we have gone from weird quirk to serious bug. Fenrir, in the console, uses /dev/vcsa and /dev/vcsu. Here is the output of ls -l for each of them: [storm@mjollnir dev] $ ls -1l vcsa* crw-rw---- 1 root tty 7, 128 Feb 15 18:30 vcsa crw-rw---- 1 root tty 7, 129 Feb 15 18:30 vcsa1 crw-rw---- 1 root tty 7, 130 Feb 15 18:30 vcsa2 crw-rw---- 1 root tty 7, 131 Feb 15 18:30 vcsa3 crw-rw---- 1 root tty 7, 132 Feb 15 18:30 vcsa4 crw-rw---- 1 root tty 7, 133 Feb 15 18:30 vcsa5 crw-rw---- 1 root tty 7, 134 Feb 15 18:30 vcsa6 crw-rw---- 1 root tty 7, 135 Feb 15 18:30 vcsa7 [storm@mjollnir dev] $ ls -1l vcsu* crw-rw---- 1 root tty 7, 64 Feb 15 18:30 vcsu crw-rw---- 1 root tty 7, 65 Feb 15 18:30 vcsu1 crw-rw---- 1 root tty 7, 66 Feb 15 18:30 vcsu2 crw-rw---- 1 root tty 7, 67 Feb 15 18:30 vcsu3 crw-rw---- 1 root tty 7, 68 Feb 15 18:30 vcsu4 crw-rw---- 1 root tty 7, 69 Feb 15 18:30 vcsu5 crw-rw---- 1 root tty 7, 70 Feb 15 18:30 vcsu6 crw-rw---- 1 root tty 7, 71 Feb 15 18:30 mcsu7 And now, things get interesting: [storm@mjollnir dev] $ sudo \cat /dev/vcsa cat: /dev/vcsa1: No such device or address Same for /dev/vcsu also including 1-7 for each of them. Other tests I tried. Importing them with python results the same for vcsa and vcsu: [root@mjollnir ~]# python Python 3.10.9 (main, Dec 19 2022, 17:35:49) [GCC 12.2.0] on linux Type "help", "copyright", "credits" or "license" for more information.
import os useVCSA = os.access('/dev/vcsa', os.R_OK) print(useVCSA) True vcsa= open("/dev/vcsa",'rb') d = vcsa.read() Traceback (most recent call last): File "<stdin>", line 1, in <module> OSError: [Errno 6] No such device or address
I hope I have provided enough information. I'm kind of stuck at this point. What should I do next? This happened yesterday after updates. I can add the pacman logs from yesterday if those are needed. Thanks, Storm
On 2/16/23 16:11, Storm Dragon wrote:
Howdy,
I started this new topic, because I think we have gone from weird quirk to serious bug. Fenrir, in the console, uses /dev/vcsa and /dev/vcsu. ...
[storm@mjollnir dev] $ sudo \cat /dev/vcsa
cat: /dev/vcsa1: No such device or address
Hi These devices are readable only by root or group tty. Try it as root see if you get any further. gene
On Thu, Feb 16, 2023 at 05:53:24PM -0500, Genes Lists wrote:
Hi
These devices are readable only by root or group tty. Try it as root see if you get any further.
gene
Howdy, Same output: cat: /dev/vcsa1: No such device or address
On 2/16/23 18:45, Storm Dragon wrote:
Howdy,
Same output: cat: /dev/vcsa1: No such device or address
Same for me - in my case I see gdm as owner of vcsa1 so perhaps there is no buffer to read in this case? Sorry I know nothing about these devices and man vcs didn't enlighten me - i have plain gnome and a bunch of text terminals - but they are all associated with the usual /dev/pty/xxx and many more of those than vcs ones.
Hi Storm,
$ sudo cat /dev/vcsa cat: /dev/vcsa1: No such device or address
That error is errno ENXIO, see errno(3). One view of the kernel source code has vcs_open() return that errno under some conditions, but I can't help more than that. https://lxr.linux.no/#linux+v6.0.9/drivers/tty/vt/vc_screen.c#L744 if (currcons && !vc_cons_allocated(currcons - 1)) ret = -ENXIO; -- Cheers, Ralph.
If makedev and the source for vcsa1 are already on the machine it may be necessary to build vcsa1 using makedev. Jude <jdashiel at panix dot com> "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) . On Sat, 18 Feb 2023, Ralph Corderoy wrote:
Hi Storm,
$ sudo cat /dev/vcsa cat: /dev/vcsa1: No such device or address
That error is errno ENXIO, see errno(3).
One view of the kernel source code has vcs_open() return that errno under some conditions, but I can't help more than that.
https://lxr.linux.no/#linux+v6.0.9/drivers/tty/vt/vc_screen.c#L744 if (currcons && !vc_cons_allocated(currcons - 1)) ret = -ENXIO;
Howdy, On Sat, Feb 18, 2023 at 03:37:36PM +0000, Ralph Corderoy wrote:
Hi Storm,
$ sudo cat /dev/vcsa cat: /dev/vcsa1: No such device or address
That error is errno ENXIO, see errno(3).
One view of the kernel source code has vcs_open() return that errno under some conditions, but I can't help more than that.
https://lxr.linux.no/#linux+v6.0.9/drivers/tty/vt/vc_screen.c#L744 if (currcons && !vc_cons_allocated(currcons - 1)) ret = -ENXIO;
-- Cheers, Ralph.
Fortunately a change has been pushed to Fenrir's master branch that works around the issue. The vcs* devices show the contents, then error out. I'm guessing that's why I saw some output on a couple of cats, but then none on most of them. I was probably too slow reviewing the screen. At this point, I'm about 99.99 percent sure we have a kernel bug on our hands. I have never reported any issues with the kernel, and I'm not sure what the process is for this. I guess it's time for a bit of research. The only thing that makes me wonder if it really is a kernel bug is, I'm using LTS, and the other people who experience it that I have spoken with are using the latest kernel, 6.something. LTS is 5.15.94. So why did we all start noticing this at the same time? Seems like they should have found it first, and I would have gotten to it eventually if it weren't fixed.
The Linux kernel has its own email list kernel@vger.org if memory serves. Jude <jdashiel at panix dot com> "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) . On Sat, 18 Feb 2023, Storm Dragon wrote:
Howdy,
On Sat, Feb 18, 2023 at 03:37:36PM +0000, Ralph Corderoy wrote:
Hi Storm,
$ sudo cat /dev/vcsa cat: /dev/vcsa1: No such device or address
That error is errno ENXIO, see errno(3).
One view of the kernel source code has vcs_open() return that errno under some conditions, but I can't help more than that.
https://lxr.linux.no/#linux+v6.0.9/drivers/tty/vt/vc_screen.c#L744 if (currcons && !vc_cons_allocated(currcons - 1)) ret = -ENXIO;
-- Cheers, Ralph.
Fortunately a change has been pushed to Fenrir's master branch that works around the issue. The vcs* devices show the contents, then error out. I'm guessing that's why I saw some output on a couple of cats, but then none on most of them. I was probably too slow reviewing the screen.
At this point, I'm about 99.99 percent sure we have a kernel bug on our hands. I have never reported any issues with the kernel, and I'm not sure what the process is for this. I guess it's time for a bit of research.
The only thing that makes me wonder if it really is a kernel bug is, I'm using LTS, and the other people who experience it that I have spoken with are using the latest kernel, 6.something. LTS is 5.15.94. So why did we all start noticing this at the same time? Seems like they should have found it first, and I would have gotten to it eventually if it weren't fixed.
Hi Storm,
Fortunately a change has been pushed to Fenrir's master branch that works around the issue.
Probably https://github.com/chrys87/fenrir/commit/cd421712f5b4d0a8ddb1deba115dee1b87b... for those following the topic.
I have never reported any issues with the kernel, and I'm not sure what the process is for this.
See https://kernel.org/category/faq.html#how-do-i-report-a-problem-with-the-kern... Its first suggestion is to start with your distro. -- Cheers, Ralph.
On Sun, Feb 19, 2023 at 10:13:40AM +0000, Ralph Corderoy wrote:
Hi Storm,
Fortunately a change has been pushed to Fenrir's master branch that works around the issue.
Probably https://github.com/chrys87/fenrir/commit/cd421712f5b4d0a8ddb1deba115dee1b87b... for those following the topic.
I have never reported any issues with the kernel, and I'm not sure what the process is for this.
See https://kernel.org/category/faq.html#how-do-i-report-a-problem-with-the-kern... Its first suggestion is to start with your distro.
-- Cheers, Ralph.
I have created a bug at: https://bugs.archlinux.org/task/77575 Hopefully I did it right lol. Thanks Ralph and everyone for all the help with this. Thanks, Storm
participants (4)
-
Genes Lists
-
Jude DaShiell
-
Ralph Corderoy
-
Storm Dragon