The officially official Devuan Forum!

You are not logged in.

#1 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-13 07:36:22

Did you try nvidia drivers from backports / non-free?
See: https://pkginfo.devuan.org/cgi-bin/pack … 10~bpo10+1

And there are other packages for that from backports/non-free. That seems to be supported by kernel 5.10.

#2 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-12 20:32:16

Yes, here it shuts down properly...

Last patch:
https://geki.selfhost.eu/hacks/1005-e10 … tach.patch

On a second thought, if your system entered sleep and e1000e called the pm_freeze function things could be corrupted, indeed. And next shutdown and start may go boom?! Another wild guess. big_smile

And for WoL, too tricky to patch actually. I just recommend you to set the ethtool disable WoL command early in boot process and not before shutdown. cool e heh. Just because we want it to stay put from the beginning. roll

And yes, with this I am out of thoughts....

#3 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-11 07:27:42

You got that PCI REMOVE prints once out of dmesg? A normal shutdown would not enter S3? Though, I am out with that. :-D

I patch WOL and PM FREEZE detach thingie and that's it from me.  :-)

#4 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-10 21:17:17

A ha, and the dmesg output shows what for the freeze?

#5 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-10 21:12:45

Since you got no pm tooling, I wonder who enters S3. Only one visible for now is /sbin/halt with its unsetted NETDOWN. If not that, it must be the kernel on its own, then?! roll

Anyone here knows about sleep state handling? smile

#6 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-10 20:58:58

Well, you want to put NETDOWN=no into /etc/default/halt to disable WoL handling of /sbin/halt. Quite strange parameter to what it is doing in the end. big_smile

Though, you still want to disable WoL via ethtool. The more disabling, the better. cool

And yes, we want to disable WoL, too. We can get a patch to forcefully disable WoL by module parameter. I already spotted the flag in code, which activates WoL. We can, hypothetically for now, reset that in src/param.c, e heh. big_smile But even with that hypothetical patch, We want to disable WoL from /sbin/halt. And ethtool will be superfluous, then.

#7 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-10 19:02:16

For disabling CONFIG_PM, you have to build your own kernel. big_smile There seems to be any pm tooling installed. roll Otherwise, I just can remind you not to use a 4.x kernel and neither kernel version < 5.5. Unless we see some NIC hang, I am mostly out of thoughts. big_smile

A last idea is to check /etc/init.d/halt for hddown= and netdown=, which you can disable by respective configuration settings from /etc/default/halt. In my box, netdown and hdown are set. You may set that configuration parameters, so that they are disabled. I put

        read -p "Press enter to halt ($netdown $poweroff $hddown)" reply

before

        halt -d -f $netdown $poweroff $hddown

to see what is set.

#8 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-09 22:18:11

Just for the fun of it, check kernel commandline parameter[0] pcie_port_pm=off. Maybe that helps disabling sleep states for your pci express slots and NIC. There is also the parameter apm, I wonder. Maybe there is some suspend software installed in /etc/pm, /etc/apm or /etc/acpi and check for such in /etc/default. big_smile

[0] https://www.kernel.org/doc/html/v4.19/a … eters.html

#9 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-09 22:03:08

Oh, and I remembered just now, if you got to the frozen "reboot: Power down", press Alt + SysRq (Print Screen key) + o for shutdown
See: http://blog.kember.net/articles/reisub- … x-restart/

#10 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-09 21:46:21

Well, the /sys/power stuff belongs to Kernel CONFIG_PM I guess. Feel free to disable that. smile You may also try shutdown -h -P to halt and power off.

Just to verify:

# cat /etc/default/halt
# Default behaviour of shutdown -h / halt. Set to "halt" or "poweroff".
HALT=poweroff

#11 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-09 15:01:19

That does not sound like an NIC (driver) issue. At least, if boot already goes noisy.

#12 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-06 06:25:22

Then, we need the dmesg output to have a complete view. cool

#14 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-05 06:24:21

Then I better use the other print function. That one hopeefully prints something on pci remove.

#15 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-04 19:35:03

Humm, I got confused. big_smile It is just right, that that message pops up there.... Though, your last screenshot of shutdown did not show any of the "PCI REMOVE" debug messages. Is that screenshot taken with the very latest patched moule build?

#16 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-04 09:11:38

Humm, I'll have to check where  that EEE TX LPI TIMER message comes from.... not from the pci shutdown callback.

#17 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-04 06:16:12

JFYI, that  warning is not mine. big_smile And that EEE TX LPI TIMER message is actually good. NULL means nothing active. cool That is just a debug message, e heh, they did not remove yet.

#19 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-03 05:54:29

Yes, therefore, I added the numbering prefix to the patches. That is the order to apply. big_smile

#20 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-02 22:39:55

Yah, faster than for my own good. cool

Now, better naming and to have a complete list, these files may be applied:
https://geki.selfhost.eu/hacks/0001-e10 … bled.patch
https://geki.selfhost.eu/hacks/0002-e10 … s_v3.patch
https://geki.selfhost.eu/hacks/0003-e10 … eeze.patch

As a reminder for me:
- Patch 0001 needs default disabled globally, default enabled for EEE featured devices.
- In the end, remove all debug messages again, e heh.

#21 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-02 22:12:22

Altoid wrote:
geki wrote:

... run a newer Kernel >= 5.5.0.
Just found: https://www.spinics.net/lists/stable/msg443520.html
Another netdev resource locking issue fixed.

Well ...

That's interesting.
I wonder if these patches will get backported to Beowulf?

Just use the kernel from beowulf-backports and you are good. smile

#22 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-02 22:00:49

Sadly, I overlooked one function that is called in the shutdown process... in src/netdev.c. There are these functions involved: e1000e_close (netdev callback seen early in your screen capture), e1000_remove and e1000_shutdown (pci device callbacks). In e1000_shutdown, it seems that the call to e1000e_pm_freeze is just superfluous. Other shutdown callbacks handle it; without that funny unprotected call to the netdev detach function. It seems that this call can safely be removed. I will do a V3 for debug messages, tomorrow. big_smile

#23 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-02 21:14:49

And remember to run a newer Kernel >= 5.5.0. Just found: https://www.spinics.net/lists/stable/msg443520.html
Another netdev resource locking issue fixed. big_smile

#24 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-02 21:10:27

Well, if it is not happening always, it is most likely an issue with concurrent access to one resource. In this case the netdev resource on shutdown.
This may help, also mentioned in that mailing list thread: https://www.kernel.org/doc/htmldocs/net … -lock.html

#25 Re: Installation » Linux e1000e module removal and e1000e EEE timer - Part II » 2021-05-02 18:20:17

Actually, nothing more to test. It works but the confused initialization and shutdown. Let's see where it hangs in the end. I wonder, if the NETIF CLOSE callback is run in another thread than the PM FREEZE callback. And that they sporadically block each other with that stray detach call, which is obviously wrong in that scope. big_smile

Board footer

Forum Software