On 02/07/17 16:09, Rick Stevens wrote:
On 02/07/2017 12:54 PM, David A. De Graaf wrote:
The r8712u driver from the staging area has stopped working in Fedora
25.
I've monitored dmesg while plugging the DLink adapter in and it's
apparent that
1 - it is detected with proper vendor and product IDs.
2 - the r8712u module is selected and registered.
3 - the correct firmware is loaded from rtlwifi/rtl8712u.bin
4 - an endless stream of messages ensues:
[ 262.524809] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
and, of course, the new interface never becomes available to
NetworkManager.
What's gone wrong?
Why does the same driver work in Fedora 24 but not in Fedora 25?
How do I regain operability?
Disable IPV6 in NM? The message you're getting is IPV6-specific and
that may be causing NM issues (it keeps waiting on a valid IPV6 address
to become available, never does, and NM keeps waiting).
That's just a guess.
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital ricks@xxxxxxxxxxxxxx -
- AIM/Skype: therps2 ICQ: 226437340 Yahoo: origrps2 -
- -
- When in doubt, mumble. -
----------------------------------------------------------------------
Thank you, Rick Stevens, for your very rational guess.
However, I'm at a loss how exactly to disable ipv6 in NetworkManager
because NetworkManager has not yet recognized the new interface for
the freshly plugged in DLink USB adapter. That is, there's no ifcfg-*
file in /etc/sysconfig/network-scripts. In fact, there is no file
anywhere in that directory tree that contains the MAC address of the
adapter. So it appears that NetworkManager doesn't come into the
picture until AFTER <whatever> determines that the 'link is ready'.
In my retained /f24 filesystem there are several ifcfg-* files that
do contain that MAC address and had been used while Fedora 24 was
active. I copied one to the F25 filesystem. That was insufficient for
NetworkManager to see it and allow editing it. However, after
restarting NetworkManager, that new Network Connection became listed as
editable when I right-clicked on the nm-applet icon.
First, I used it unaltered - because it had worked perfectly in F24 -
eg, I simply plugged in the DLink USB adapter and watched 'dmesg -w'.
Second, I clicked on nm-applet and graphically edited that Network
Connection to set ipv6 settings to "Ignore".
Third, I manually edited the original ifcfg file to disable all the
ipv6 settings.
The output of dmesg was the same for all three cases.
There was NO CHANGE. To be specific and complete, here's the dmesg
output:
[82189.094547] usb 1-4: new high-speed USB device number 7 using xhci_hcd
[82189.265676] usb 1-4: New USB device found, idVendor=07d1, idProduct=3300
[82189.265686] usb 1-4: New USB device strings: Mfr=1, Product=2,
SerialNumber=3
[82189.265692] usb 1-4: Product: 11n Adapter
[82189.265697] usb 1-4: Manufacturer: Manufacturer Realtek
[82189.265702] usb 1-4: SerialNumber: 00e04c000001
[82189.269827] r8712u: register rtl8712_netdev_ops to netdev_ops
[82189.269839] usb 1-4: r8712u: USB_SPEED_HIGH with 4 endpoints
[82189.271387] usb 1-4: r8712u: Boot from EFUSE: Autoload OK
[82189.635092] usb 1-4: r8712u: CustomerID = 0x0006
[82189.635100] usb 1-4: r8712u: MAC Address from efuse = 28:10:7b:c2:09:ae
[82189.635105] usb 1-4: r8712u: Loading firmware from "rtlwifi/rtl8712u.bin"
[82189.686494] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
[82190.390956] r8712u 1-4:1.0 wlan1: 1 RCR=0x153f00e
[82190.391691] r8712u 1-4:1.0 wlan1: 2 RCR=0x553f00e
[82191.603127] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
[82193.122927] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
[82193.182252] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
[82194.707205] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
[82196.235123] IPv6: ADDRCONF(NETDEV_UP): wlan1: link is not ready
....
So, the IPv6 error messages recur endlessly and the link remains not
ready.
I have two other computers with slightly unusual USB wireless adapters
that are also behaving badly with Fedora 25. I will report these
observations when I've collected convincing data.
I cannot help but think there's a significant kernel bug...
Anyone with a plug-in USB wireless adapter should closely monitor its
behaviour.
--
David A. De Graaf DATIX, Inc. Hendersonville, NC
dad@xxxxxxxx www.datix.us
_______________________________________________
users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
