08-02-2020, 09:06 PM
Hello, folks,
I have this exact problem, and after blowing a weekend on it, I think it's a problem with the 5.7 kernel.
The problem is that the Pinephone would not show up at all in dmesg, nor in the lsusb listings, of the host PC, when it's plugged into USB. The usb0 interface does show up on the phone side, can be configured with an ip address, and the interface can be set to "up", but "ip addr" and ifconfig will still show it as either "state DOWN" or "NO-CARRIER" no matter what.
My Pine Phone is a ubports CE - hardware rev 1.2 - shipped with Ubuntu Touch
My host PC is an amd64 system running Ubuntu Bionic.
There are two ways I //have// gotten this to work with a 5.7 kernel:
* The jumpdrive image works perfectly, as advertised. This uses a configfs implementation, which is completely different than the "g_multi" module most distros seem to be using.
* booting the Ubuntu Touch installed on emmc using volume up + power, which boots the phone into a rescue mode. it looks an awful lot like the Jump drive distro, so it's probably a modification of that.
What hasn't worked:
* postmarketos - with phosh, plasma mobile, fbkeyboard, and shelli interfaces (I was distro-hopping, since I just got my phone last week).
* Mobian image from July 16, using the scripts from the pinephone-devtools package
* Manjaro Alpha1 image from early July
* Attempting to backport the jumpdrive's USB configfs implementation into a Mobian system. It would not produce errors, and would create the usb0 interface, but it would still always be "state DOWN". I was able to blacklist the g_multi kernel module, but that did not help.
The Ubuntu touch distro that shipped with the phone happens to have a 5.6 kernel, but frankly, that distro is so alien to me that I have given even getting an openssh server working on it.
Since a few of you folks thought this used to work, I backtracked and pulled down the Mobian image from Jun 11th. It had the 5.6 kernel. The usb networking worked perfectly!
I'm pleased to know that this is definitely not a hardware issue. I'll leave my results here, and hope more adept kernel hackers than I can make use of it.
I have this exact problem, and after blowing a weekend on it, I think it's a problem with the 5.7 kernel.
The problem is that the Pinephone would not show up at all in dmesg, nor in the lsusb listings, of the host PC, when it's plugged into USB. The usb0 interface does show up on the phone side, can be configured with an ip address, and the interface can be set to "up", but "ip addr" and ifconfig will still show it as either "state DOWN" or "NO-CARRIER" no matter what.
My Pine Phone is a ubports CE - hardware rev 1.2 - shipped with Ubuntu Touch
My host PC is an amd64 system running Ubuntu Bionic.
There are two ways I //have// gotten this to work with a 5.7 kernel:
* The jumpdrive image works perfectly, as advertised. This uses a configfs implementation, which is completely different than the "g_multi" module most distros seem to be using.
* booting the Ubuntu Touch installed on emmc using volume up + power, which boots the phone into a rescue mode. it looks an awful lot like the Jump drive distro, so it's probably a modification of that.
What hasn't worked:
* postmarketos - with phosh, plasma mobile, fbkeyboard, and shelli interfaces (I was distro-hopping, since I just got my phone last week).
* Mobian image from July 16, using the scripts from the pinephone-devtools package
* Manjaro Alpha1 image from early July
* Attempting to backport the jumpdrive's USB configfs implementation into a Mobian system. It would not produce errors, and would create the usb0 interface, but it would still always be "state DOWN". I was able to blacklist the g_multi kernel module, but that did not help.
The Ubuntu touch distro that shipped with the phone happens to have a 5.6 kernel, but frankly, that distro is so alien to me that I have given even getting an openssh server working on it.
Since a few of you folks thought this used to work, I backtracked and pulled down the Mobian image from Jun 11th. It had the 5.6 kernel. The usb networking worked perfectly!
I'm pleased to know that this is definitely not a hardware issue. I'll leave my results here, and hope more adept kernel hackers than I can make use of it.