[Josh]

Check again the output of "dmesg" (or "journalctl") after issuing the two commands.

Are there still those disconnects that LOM pointed to?

[chipsharpdotcom]

Looks like it....

Code: Select all

Apr 24 20:19:48 chipsharplnx kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 24 20:19:48 chipsharplnx kernel: usb 2-6: USB disconnect, device number 13
Apr 24 20:19:49 chipsharplnx kernel: usb 2-6: new high-speed USB device number 15 using xhci_hcd
Apr 24 20:19:49 chipsharplnx kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 24 20:19:49 chipsharplnx kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 24 20:19:49 chipsharplnx kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 24 20:19:49 chipsharplnx kernel: usb 2-6: Manufacturer: HP
Apr 24 20:19:49 chipsharplnx kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, be:9a:44:c2:21:8d
Apr 24 20:19:49 chipsharplnx kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 24 20:19:49 chipsharplnx kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready
Apr 24 20:19:56 chipsharplnx kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 24 20:19:56 chipsharplnx kernel: usb 2-6: USB disconnect, device number 15
Apr 24 20:19:56 chipsharplnx kernel: usb 2-6: new high-speed USB device number 16 using xhci_hcd
Apr 24 20:19:56 chipsharplnx kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 24 20:19:56 chipsharplnx kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 24 20:19:56 chipsharplnx kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 24 20:19:56 chipsharplnx kernel: usb 2-6: Manufacturer: HP
Apr 24 20:19:56 chipsharplnx kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, 42:ba:6d:3f:4e:7c
Apr 24 20:19:57 chipsharplnx kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 24 20:19:57 chipsharplnx kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready

[LOM]

Yes you have complete device disconnects from the bus when you echo the new configuration value and that is wrong, usb_modeswitch will not be able to select a new configuration as long as you can't do that manually by echoing the value.

Your linux version 4.12 is End-Of-Life from the kernel maintainers point of view so it does not get any updates from them and your kernel was built 8 month ago, the XHCI driver has had many updates and bug fixes since then. Update!!

You can also try to narrow down the problem by booting another distro than Fedora, there are many others available as live boot from a USB flash stick.

[chipsharpdotcom]

I switched to Antergos (which is where I started, I foolishly switched to fedora thinking there would be better HW support).

I'll rinse and repeat the testing. I'm on 4.16.3-1-ARCH for my kernel version now, so let's see what we get.

Code: Select all

Apr 25 20:44:28 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 25 20:44:28 ChipSharpLNX kernel: usb 2-6: USB disconnect, device number 10
Apr 25 20:44:29 ChipSharpLNX kernel: usb 2-6: new high-speed USB device number 11 using xhci_hcd
Apr 25 20:44:29 ChipSharpLNX kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 25 20:44:29 ChipSharpLNX kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 25 20:44:29 ChipSharpLNX kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 25 20:44:29 ChipSharpLNX kernel: usb 2-6: Manufacturer: HP
Apr 25 20:44:29 ChipSharpLNX kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, 9e:c2:eb:64:78:6d
Apr 25 20:44:29 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 25 20:44:29 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready
Apr 25 20:44:29 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 25 20:44:29 ChipSharpLNX kernel: usb 2-6: USB disconnect, device number 11
Apr 25 20:44:30 ChipSharpLNX kernel: usb 2-6: new high-speed USB device number 12 using xhci_hcd
Apr 25 20:44:30 ChipSharpLNX kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 25 20:44:30 ChipSharpLNX kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 25 20:44:30 ChipSharpLNX kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 25 20:44:30 ChipSharpLNX kernel: usb 2-6: Manufacturer: HP
Apr 25 20:44:30 ChipSharpLNX kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, ea:33:dd:03:02:ee
Apr 25 20:44:30 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 25 20:44:30 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready
Apr 25 20:45:21 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 25 20:45:21 ChipSharpLNX kernel: usb 2-6: USB disconnect, device number 12
Apr 25 20:45:21 ChipSharpLNX kernel: usb 2-6: new high-speed USB device number 13 using xhci_hcd
Apr 25 20:45:21 ChipSharpLNX kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 25 20:45:21 ChipSharpLNX kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 25 20:45:21 ChipSharpLNX kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 25 20:45:21 ChipSharpLNX kernel: usb 2-6: Manufacturer: HP
Apr 25 20:45:21 ChipSharpLNX kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, 06:94:4d:28:98:98
Apr 25 20:45:22 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 25 20:45:22 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready
Apr 25 20:45:22 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 25 20:45:22 ChipSharpLNX kernel: usb 2-6: USB disconnect, device number 13
Apr 25 20:45:23 ChipSharpLNX kernel: usb 2-6: new high-speed USB device number 14 using xhci_hcd
Apr 25 20:45:23 ChipSharpLNX kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 25 20:45:23 ChipSharpLNX kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 25 20:45:23 ChipSharpLNX kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 25 20:45:23 ChipSharpLNX kernel: usb 2-6: Manufacturer: HP
Apr 25 20:45:23 ChipSharpLNX kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, 82:a5:37:76:ca:67
Apr 25 20:45:23 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 25 20:45:23 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready
Apr 25 20:45:23 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready
Apr 25 20:45:23 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_CHANGE): enp0s20u6c2: link becomes ready
Apr 25 20:45:27 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: unregister 'cdc_ether' usb-0000:00:14.0-6, CDC Ethernet Device
Apr 25 20:45:27 ChipSharpLNX kernel: usb 2-6: USB disconnect, device number 14
Apr 25 20:45:28 ChipSharpLNX kernel: usb 2-6: new high-speed USB device number 15 using xhci_hcd
Apr 25 20:45:28 ChipSharpLNX kernel: usb 2-6: New USB device found, idVendor=03f0, idProduct=911d
Apr 25 20:45:28 ChipSharpLNX kernel: usb 2-6: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Apr 25 20:45:28 ChipSharpLNX kernel: usb 2-6: Product: HP lt4211 Gobi 4G Module
Apr 25 20:45:28 ChipSharpLNX kernel: usb 2-6: Manufacturer: HP
Apr 25 20:45:28 ChipSharpLNX kernel: cdc_ether 2-6:2.0 usb0: register 'cdc_ether' at usb-0000:00:14.0-6, CDC Ethernet Device, 32:69:86:5a:f4:48
Apr 25 20:45:28 ChipSharpLNX kernel: cdc_ether 2-6:2.0 enp0s20u6c2: renamed from usb0
Apr 25 20:45:28 ChipSharpLNX kernel: IPv6: ADDRCONF(NETDEV_UP): enp0s20u6c2: link is not ready

So....still a bunch of disconnecting nonsense. I got errors when I tried to repeat the echo 0, echo 3...

Code: Select all

[root@ChipSharpLNX ~]# echo "0" > /sys/bus/usb/devices/2-6/bConfigurationValue 
-bash: echo: write error: Broken pipe

However when I cat bConfigurationValue:

Code: Select all

[root@ChipSharpLNX ~]# cat /sys/bus/usb/devices/2-6/bConfigurationValue 
2

So can someone bottom line this for me? Am I out of luck with this card, or is there another package forum I need to go harass and then come back here later?

--C#

[LOM]

Does it make any difference if you echo the values without "" around them?
I guess it shouldn't matter..

I can think of a few reasons why you are getting those disconnects that we haven't seen before for a config select, one is a buggy modem firmware and the other one is a buggy USB HCD, either hardware or software driver.

[chipsharpdotcom]

Quotes, no quotes, same results. Do you have any other suggestions, or is this the end of the line?

[Josh]

If you had annother Linux machine to try (different hardware), you could determine if there is a host-side problem (as opposed to faulty modem firmware).

There is a (weak) chance the behaviour is different when using a USB hub (with external power, to be on the safe side).

[chipsharpdotcom]

Sorry for the radio silence. My travel schedule has been a bit hectic.

There's absolutely no chance it's a bad USB hub. This is an internal card. not a (un)pluggable dongle. It works perfectly under Windows, and the WLAN adapter on the same USB bus works perfectly in Windows and every one of the 3 linux distributions I've tried (Antergos/Arch, Fedora, Ubuntu). I dare say the flaky firmware is more accurate. I just need a brief understanding of the nature of its flakiness, so I can go to the manufacturer and see if they can get it squared away. I'll also go back to the Windows half of my laptop and see if I can find a firmware update/upgrade for the device.

Absent that, a short description of what we need the firmware to do that it's currently not, would be very helpful.

--C#

[chipsharpdotcom]

*bump*

[Josh]

I don't know how we can help in any more ways than we did. There are no firmware specialists around here.
It's obvious that the device should not do what it's doing now, that is, permanently resetting. The details are up to the manufacturer.

You could also open a bug for your Linux distribution, or ask the "linux-usb" mailing list.

[chipsharpdotcom]

That's exactly what I was looking for. Thanks!

--C#

[LOM]

There's absolutely no chance it's a bad USB hub. This is an internal card. not a (un)pluggable dongle.

There is a chance that it is a bad usb controller or root hub, a hardware or software bug affecting devices with multiple configurations ofwhich you probably don't have any other in the laptop.
This is more likely than a card firmware bug but it would be good to single out one of the possible reasons for the current behavior.

The card is not glued to your motherboard so of course it can be tested on another computer, I will actually have an LT4211 card in a week or two and will put it in an external NGFF->USB box thereby making it pluggable Lets see then how it behaves when switching config.

[LOM]

It arrived quicker than expected, postman brought it an hour ago..

First veryfying that adapter and test method is correct, this
is from an Ericsson N5321 NGFF card with 4 configs:

Code: Select all

$ echo 3 > /sys/bus/usb/devices/1-1.1/bConfigurationValue
$ cat /sys/bus/usb/devices/1-1.1/bConfigurationValue
3

$ echo 2 > /sys/bus/usb/devices/1-1.1/bConfigurationValue
$ cat /sys/bus/usb/devices/1-1.1/bConfigurationValue
2

Worked as expected, now the LT4211 :

Code: Select all

$ echo 3 > /sys/bus/usb/devices/1-1.1/bConfigurationValue
sh: write error: Broken pipe
$ cat  /sys/bus/usb/devices/1-1.1/bConfigurationValue
2

$ echo 2 > /sys/bus/usb/devices/1-1.1/bConfigurationValue
$ cat  /sys/bus/usb/devices/1-1.1/bConfigurationValue
2

$ echo 1 > /sys/bus/usb/devices/1-1.1/bConfigurationValue
sh: write error: Broken pipe
$ cat  /sys/bus/usb/devices/1-1.1/bConfigurationValue
2

$ echo 0 > /sys/bus/usb/devices/1-1.1/bConfigurationValue
sh: write error: Broken pipe
$ cat  /sys/bus/usb/devices/1-1.1/bConfigurationValue
2

The result speaks for itself, the firmware of this card does not accept
configuration changes. It may be intentionally factory locked to config #2
otherwise it is a firmware bug.

[chipsharpdotcom]

Thank you for confirming my result. It wasn't my intention that you would have to invest money in hardware to get a reliable result but I appreciate your help! I'll submit this back to HP and hope that they can get a resolution from the manufacturer. If I can get a productive submission channel with HP (which is to say, I get the issue past the level 1 support desk and up to someone in engineering) I would appreciate if you could follow my submission with one of your own so as to show a "popular need" and not a one-off user. If you don't want this to be a part of your life anymore, I completely understand that as well.

I'll report back any result I get from HP.

--C#

[LOM]

The function to select configuration externally has been disabled by the manufacturer.
The cards firmware does detect the way MS Win probes the card and will switch to cfg#1 if probed by Win7 and will switch to cfg#3 if probed by Win8/10, leaving cfg#2 as the default for the case when it can not detect any MS Win probing.

I'll submit this back to HP and hope that they can get a resolution from the manufacturer.

The card works with the MS Win versions that are supported by HP for this computer, they will tell you that there is no linux support for it.