Output of "who -a" on 32-bit system is corrupt

This forum is for topics dealing with problems with software specifically in the ARMv7h repo.

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Mon Sep 05, 2022 2:46 pm

There are some bug reports about /usr/bin/who failing but they don't seem to be relevant.
https://forum.manjaro.org/t/who-coreutils-9-1-1-is-broken/113675/7
https://bugs.archlinux.org/task/58306

"Changing the GID of utmp to 20 and rebooting restores the functionality of libutempter again."

Before making that change my system has GID 996 for utmp.
$this->bbcode_second_pass_code('', '$ getent group utmp
utmp:x:996:
')

After changing (I edited /etc/group directly)
$this->bbcode_second_pass_code('', '$ getent group utmp
utmp:x:20:
')

File permissions look ok.
$this->bbcode_second_pass_code('', '$ ls -l /run/utmp /usr/lib/utempter/utempter
-rw-rw-r-- 1 root utmp 1536 Sep 5 08:00 /run/utmp
-rwxr-sr-x 1 root utmp 9596 Jun 14 21:14 /usr/lib/utempter/utempter
')
Then I rebooted. But there's no improvement. Still it shows that there may be issues in software outside of coreutils 1.9 and wget 1.21.3.
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Mon Sep 05, 2022 4:08 pm

Debian unstable for armhf is still using coreutils 8.32-4 and wget 1.21.3-1+b1
Notes on building a 32-bit Debian Sid armhf VM
https://gist.github.com/ambiamber/42d77db75fa8e8da461f57392ab6a002

I don't see any newer versions in Debian experimental
https://packages.debian.org/sid/allpackages

I was hoping to get support about this issue from Debian people since Arch people only support 64-bit systems.

Maybe Fedora rawhide...
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm

Re: Output of "who -a" on 32-bit system is corrupt

Postby robg » Mon Sep 05, 2022 6:33 pm

Arch ARM officially supports 32- and 64-bit systems (see "Platforms" on the main site.)

As we have managed to pinpoint in the thread, the issue is due missing / wrong / mismatched compiler options that affect how `time_t` behaves on 32-bit systems. This problem is being investigated. A solution is likely to arrive soon, just have a little patience.
robg
 
Posts: 194
Joined: Tue Jan 05, 2021 8:22 am

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Tue Sep 06, 2022 10:42 pm

$this->bbcode_second_pass_quote('', '.') This problem is being investigated.

Do you have any evidence that their remediation will improve coreutils?

Meanwhile, I have a working Fedora 36 armhfp VM and the bug is in it too.
$this->bbcode_second_pass_code('', '
$ who -a
system boot 1970-01-04 15:35
run-level 3 1970-01-08 05:11
LOGIN tty1 1970-01-07 20:01 482 id=tty1
user + ttyAMA0 1970-01-09 17:37 . 483
user + pts/0 504588700852756167 . 645 (192.168.X.X)
$ rpm -q coreutils
coreutils-9.0-8.fc36.armv7hl
')
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Wed Sep 07, 2022 12:08 am

Fedora deprecated armv7.
https://pagure.io/releng/issue/10387
Fedora 36 is the last release for 32-bit arm.

Hopefully my MochaBIN will show up soon.
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Wed Sep 07, 2022 1:18 pm

$this->bbcode_second_pass_quote('', 'A')rch ARM officially supports 32- and 64-bit systems (see "Platforms" on the main site.)

I didn't write "Arch Linux ARM does not support 32-bits".
I wrote that the Arch people do not support 32-bits. Go and post on bbs about 32-bit issues and they will get hostile.
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Wed Sep 07, 2022 1:57 pm

I'm exploring Alpine in my effort to find more support for these issues. but they have original ways of doing things. It's using musl C with optional glibc compatibility.

https://gist.github.com/ambiamber/4e70be9ae0c7ecbbdbd7880b36db59ca

The stable release does have coreutils-9.1.
utmp support is currently only available on edge AFAIK.
https://pkgs.alpinelinux.org/package/edge/main/armv7/utmps

A little history about my Arch ARM Linux adventure. In 2016 or 2017 I found out about https://www.krackattacks.com/
I went shopping for a router where I could keep the software on it up-to-date and I settled on a Clearfog Pro. I ordered the Clearfog Pro in December 2017 and it arrived in April 2018. I tried Debian armhf but I didn't like it and I tried Arch Linux ARM. I like Arch Linux ARM but there has been the occasional bump in the road. Things have been really smooth for a long time until upstream support has started become an issue. Now I have to make plans about what to do. I could go distro hopping or just wait for my 64-bit MochaBIN while rolling my own packages with y2038 mitigations disabled for the Clearfog Pro.
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm

Re: Output of "who -a" on 32-bit system is corrupt

Postby robg » Thu Sep 08, 2022 9:57 am

$this->bbcode_second_pass_quote('Mettacrawler', '')$this->bbcode_second_pass_quote('', 'T')his problem is being investigated.

Do you have any evidence that their remediation will improve coreutils?


The issue has been identified, so it is only a matter of time until it is resolved. As emphasized in the linked thread, any solution must be carefully examined because changes to `time_t` must likely be applied to all packages consistently.

Appendum: As a temporary workaround, you might try switching to Debian or Armbian. You may have to port over the kernel from ALARM.
robg
 
Posts: 194
Joined: Tue Jan 05, 2021 8:22 am
Top

Re: Output of "who -a" on 32-bit system is corrupt

Postby Mettacrawler » Thu Sep 08, 2022 11:33 am

$this->bbcode_second_pass_quote('', 'A')s a temporary workaround, you might try switching to Debian or Armbian.

In my previous comment I wrote "I tried Debian armhf but I didn't like it". A few comments before that I wrote "Debian unstable (Sid) for armhf is still using coreutils 8.32-4 and wget 1.21.3-1+b1".
$this->bbcode_second_pass_quote('', 'Y')ou may have to port over the kernel from ALARM.

I didn't have to that, see
https://gist.github.com/ambiamber/42d77db75fa8e8da461f57392ab6a002

I've assembled an Alpine edge armv7 VM although I've been informed of a better way to do it than my procedure. With wget 1.21.3, coreutils 9.1 and the utmps package it does not seem to have any time_t bugs. My first guess is that musl C makes a difference.
Mettacrawler
 
Posts: 56
Joined: Sun Mar 18, 2018 7:19 pm
Top

Re: Output of "who -a" on 32-bit system is corrupt

Postby robg » Thu Sep 08, 2022 1:21 pm

$this->bbcode_second_pass_quote('Mettacrawler', '')$this->bbcode_second_pass_quote('', 'A')s a temporary workaround, you might try switching to Debian or Armbian.

In my previous comment I wrote "I tried Debian armhf but I didn't like it". A few comments before that I wrote "Debian unstable (Sid) for armhf is still using coreutils 8.32-4 and wget 1.21.3-1+b1".


Sorry, I missed these! Odd to see Debian unstable so far behind...

$this->bbcode_second_pass_quote('Mettacrawler', 'I')'ve assembled an Alpine edge armv7 VM although I've been informed of a better way to do it than my procedure. With wget 1.21.3, coreutils 9.1 and the utmps package it does not seem to have any time_t bugs. My first guess is that musl C makes a difference.


Happy to read that you got it working!
robg
 
Posts: 194
Joined: Tue Jan 05, 2021 8:22 am
Top

PreviousNext

Return to ARMv7h

Who is online

Users browsing this forum: No registered users and 11 guests