stable/current bug reports

PartiZan · 03-21-2022, 07:44 PM

Quote:

Originally Posted by Petri Kaukasoina

Maybe the OP is using a different version of slackpkg than slackpkg-15.0.10-noarch-1? Or maybe there is also slackpkg+ involved?

No. I took 15.0 DVD iso file, then installed minimal version to boot. After that I noticed this strange behavior of slackpkg

chrisretusn · 03-23-2022, 04:41 AM

Quote:

Originally Posted by PartiZan

Ok, could you try this:
- slackpkg update
- find some unneeded patched package and remove it
- remove corresponding files from /var/log/pkgtools/removed_{packages,scripts}
- slackpkg install full_name_of_patched_package

Observe the dialog window

I just did this on with slackware64-15.0 full and clean install in a VirtualBox VM.

Using 'slackpkg remove openssl' I ensured only openssl-1.1.1n-x86_64-1_slack15.0 was selected and then removed it.

I removed the openssl-1.1.1n-x86_64-1 and openssl-1.1.1n-x86_64-1_slack15.0 files in removed_packages and removed_scripts.

I then ran 'slackpkg install openssl-1.1.1n-x86_64-1_slack15.0' and got the expected result, only openssl-1.1.1n-x86_64-1_slack15.0 was available for installation.

Code:

# slackpkg -dialog=off install openssl-1.1.1n-x86_64-1_slack15.0

Looking for openssl-1.1.1n-x86_64-1_slack15.0 in package list. Please wait... DONE

openssl-1.1.1n-x86_64-1_slack15.0.txz

Total package(s): 1

Do you wish to install selected packages (Y/n)?

steve4lqdo · 03-23-2022, 11:23 AM

Slackware64-current on ThinkPad T495.
Suspend fails after kernel 5.17.0 (ok with 5.16.15).

Reference:

https://lkml.org/lkml/2022/3/23/248

brobr · 03-23-2022, 03:21 PM

Quote:

Originally Posted by steve4lqdo

Slackware64-current on ThinkPad T495.
Suspend fails after kernel 5.17.0 (ok with 5.16.15).

is this in a VM? I have current running on the box as main OS and suspend is ok with 5.17.0 (clevo laptop)

steve4lqdo · 03-23-2022, 03:48 PM

Not a VM - bare metal. Lenovo ThinkPad T495 has AMD Ryzen 7 PRO 3700U w/ Radeon Vega Mobile Gfx AuthenticAMD. Back to running 5.16.15 for now, as no suspend is a deal breaker for me.

marav · 03-23-2022, 03:58 PM

Quote:

Originally Posted by steve4lqdo

Not a VM - bare metal. Lenovo ThinkPad T495 has AMD Ryzen 7 PRO 3700U w/ Radeon Vega Mobile Gfx AuthenticAMD. Back to running 5.16.15 for now, as no suspend is a deal breaker for me.

with elilo ?

steve4lqdo · 03-23-2022, 04:21 PM

Actually, using Grub with Slackware and Mint (and Win10 - never used ;-). Spousal unit uses Mint, so able to help out from time to time. Slackware since about 1993 or 1994?

marav · 03-23-2022, 04:26 PM

Quote:

Originally Posted by steve4lqdo

Actually, using Grub with Slackware and Mint (and Win10 - never used ;-). Spousal unit uses Mint, so able to help out from time to time.

Ok, so not elilo (I had an issue with elilo and hibernation)

Quote:

Slackware since about 1993 or 1994?

Me ?
No, only since 2018 ;-)

steve4lqdo · 03-23-2022, 04:38 PM

Started with SLS (1993?), kernel 0.99pl13, and Slackware ever since (may have inadvertently posted this twice).

PartiZan · 03-26-2022, 01:05 PM

Quote:

Originally Posted by PartiZan

When I explicitly install package from patches using "slackpkg install", the dialog shows me both versions, old and patched new one.

For example:
slackpkg install openssl-1.1.1n-x86_64-1_slack15.0

The dialog shows me two packages:
openssl-1.1.1m-x86_64-1
openssl-1.1.1n-x86_64-1_slack15.0

I would expect to see the only new patched version, which I explicitly specified in "slackpkg install"

This problem existed, but as I continued adding more packages to the system, this issue somehow magically disappeared. All is good now.

PartiZan · 03-26-2022, 09:27 PM

Quote:

Originally Posted by PartiZan

This problem existed, but as I continued adding more packages to the system, this issue somehow magically disappeared. All is good now.

Still not good. Another, but similar weird thing.

Click image for larger version

Name: Screenshot_2022-03-26_21-24-30.png
Views: 67
Size: 101.8 KB
ID: 38671

bassmadrigal · 03-27-2022, 01:26 AM

Quote:

Originally Posted by PartiZan

Still not good. Another, but similar weird thing.

Attachment 38671

This is by design.

slackpkg info $ARG will pass $ARG through a sed command to escape all periods and pluses. Then it will pass the escaped $ARG to the cutpkg function, which will first remove any extensions before passing it through the cutpkg.awk script, which will extract just the package name.

It then uses awk to check the PACKAGES.TXT to show you the info for any packages on the mirror that share that name from all available folders.

xunilq · 03-27-2022, 08:18 PM

Report of wireshark behavior over Kernel 5.17.0

Wireshark doesn't show packets over Monitor Mode on wlan0

/var/log/syslog no messages related

/var/log/messages:

Code:

kernel: device wlan0 left promiscuous mode

Kernel 5.15.27, works fine on monitor mode, difference on log:

Code:

kernel: device wlan0 entered promiscuous mode
kernel: device wlan0 left promiscuous mode

My Wlan device:

Code:

AR9565 Wireless Network Adapter (rev 01)

-----------------------------

PS for possible googlers, the monitor checkbox on wireshark doesn't work. i have set monitor mode like this:

Code:

ip link set wlan0 down
iw dev wlan0 set type monitor
ip link set wlan0 up

steve4lqdo · 03-28-2022, 06:19 PM

So kernel 5.17.1 is still "broken" regarding suspend/resume for an AMD based Lenovo T495 ThinkPad laptop. Back to kernel 5.16.15 where all is "well". Looking at 5.17.x various kernel discussions, there seem to be some regressions in the 5.17.x kernel with AMD based GPUs, etc. Oh, well, it's "not stable", so "it is what it is". ;-)

PartiZan · 04-01-2022, 11:25 PM

Quote:

Originally Posted by bassmadrigal

This is by design.

slackpkg info $ARG will pass $ARG through a sed command to escape all periods and pluses. Then it will pass the escaped $ARG to the cutpkg function, which will first remove any extensions before passing it through the cutpkg.awk script, which will extract just the package name.

It then uses awk to check the PACKAGES.TXT to show you the info for any packages on the mirror that share that name from all available folders.

Very poor excuse for wrong behavior. I would say it's even funny, like saying that this is not a bug, this is feature.