| gnarface | seems possible but i thought there was | 00:01 |
|---|---|---|
| gnarface | i do remember something at one point in time regarding a bug where you needed to suppres one of the *ci modules to let the right one load but i also thought that was well before beowulf (again, not sure though) | 00:02 |
| gnarface | is the bios itself up-to-date? | 00:02 |
| gnarface | hmm, google claims usb3 wasn't on by default in the mainline kernel until 2018 | 00:05 |
| gnarface | that doesn't mean debian might not have enabled it by default earlier | 00:08 |
| gnarface | it does seem like the type of thing that would get enabled in the backports kernel too | 00:08 |
| devuser | I will have to upgrade this host system anyway so this is as good a time as any. Unless you are interested in a more "scientific" answer here... ;) | 02:07 |
| devuser | I wonder if I can simply load a module to see if usb3 starts working... | 02:08 |
| devuser | xhci-pci and xhci-hcd are loaded... | 02:12 |
| devuser | I mean xhci_pci and xhci_hcd | 02:12 |
| noobaroo | Hey all. Not trying to be negative in anyway, just looking to learn new things. Why do you guys choose Devuan over Artix? | 06:51 |
| onefang | noobaroo: That'll be a question for #devuan-offtopic, since this channel is for support. | 07:28 |
| noobaroo | onefang, that's unfortunate... Only 60 nicks in there, and nobody is trying to get help right now | 11:23 |
| systemdlete | after some rest, I decided to take one more look at the usb3 support on my two PCs. One thing I have discovered, despite both running beowulf with the latest kernels on both, is that the main box detects the xhci ports but deregisters them (not sure why yet). The testbox detects them but does not deregister them. | 13:22 |
| systemdlete | They have different main boards, and I can see that my main box does encounter some IO errors while doing this during the boot | 13:23 |
| systemdlete | The testbox lsusb output does show a usb 3.0 controller, which helps explain why it is able to use the device, even if maybe it is not really running at 3.0 speeds. | 13:23 |
| systemdlete | Whereas the main box does not show any usb 3.0 controller(s). | 13:24 |
| systemdlete | Not sure where I am hoping to take this, but I find it interesting. I have downloaded chimaera on both boxes and can probably start the install in the next day or so. I have a few other things to do so it might be a bit before I get going on this. | 13:25 |
| systemdlete | OK, so the kernel log shows lots of IO PAGE FAULTS during the boot (and I saw these on the console too, now that I think of it) It complains about some sort of random read, which fails, and then it decides to pick up its marbles and go home. The boot continues, but I am left with inoperative USB 3 ports. | 13:28 |
| u-amarsh04 | systemdlete - what is the output from lspci|grep USB | 14:18 |
| systemdlete | u-amarsh04, it lists all the ehci and ohci as well as | 14:28 |
| systemdlete | 02:00.0 USB controller: VIA Technologies, Inc. VL805 USB 3.0 Host Controller (rev 01) | 14:28 |
| systemdlete | So it does see it, but for whatever reason, it doesn't work | 14:28 |
| systemdlete | (and thanks for stepping in) | 14:28 |
| systemdlete | Also, dmidecode clearly shows 3 usb controllers, one for each of 1 2 and 3 | 14:31 |
| u-amarsh04 | I bought a couple of USB 3 cards and they both work but I run bleeding edge kernels | 14:31 |
| u-amarsh04 | systemdlete: | 14:38 |
| u-amarsh04 | Seen on a VLI VL805 PCIe to USB controller. For non-stream endpoints | 14:38 |
| u-amarsh04 | at least, if the xHC halts on a particular TRB due to an error then | 14:38 |
| u-amarsh04 | the DCS field in the Out Endpoint Context maintained by the hardware | 14:38 |
| u-amarsh04 | is not updated with the current cycle state. | 14:38 |
| systemdlete | well that is why I am thinking that upgrading to chimaera might help | 14:39 |
| u-amarsh04 | from the 5.14 release candidate stage of the kernel git log | 14:39 |
| u-amarsh04 | so 5.14 or later should have the fix | 14:39 |
| systemdlete | hmmm. Chimaera is 5.10.0, so I'd have to upgrade the kernel anyway. | 14:40 |
| systemdlete | But thanks for letting me know I'll need around 5.14 | 14:41 |
| systemdlete | Is there a way to determine whether the kernel is detecting a USB 3.0 device as running 3.0 versus only 2.x ? | 14:42 |
| systemdlete | (Or, at least thinks it is anyway) | 14:43 |
| u-amarsh04 | in dmesg output I see for my mouse: usb 5-1: new low-speed USB device number 2 using ohci-pci | 14:44 |
| u-amarsh04 | and for a USB3 hard disk: usb 4-2: new SuperSpeed USB device number 2 using xhci_hcd | 14:44 |
| systemdlete | ah. ok | 14:44 |
| systemdlete | interesting! I do see the kernel does recognize my usb-ethernet adapter as xhci (same message as yours) | 14:46 |
| systemdlete | but it says "gen 1" so I wonder if that makes a difference | 14:46 |
| tarzeau | so devuan doesn't have networkctl ? | 21:06 |
| PhotoJim | no systemd, no networkctl | 21:07 |
| tarzeau | sounds good - so i need to avoid using networkctl in portable scripts | 21:07 |
| tarzeau | what about ifconfig vs ip? both available, one default? | 21:07 |
| gnarface | they should both be there | 21:08 |
| gnarface | not sure if they're both always installed | 21:10 |
| gnarface | but both packages are available | 21:11 |
| PhotoJim | ifconfig is considered deprecated, but can be installed if not installed by default | 21:11 |
| PhotoJim | ip is considered the contemporary way of doing things | 21:11 |
| gnarface | i think a minimal install might still give you ifconfig by default but not ip | 21:11 |
| gnarface | not sure about that | 21:11 |
| PhotoJim | easy enough to find out | 21:11 |
| furrymcgee | busybox has both | 21:43 |
| fsmithred | ifconfig is in the net-tools package, and that has been devuanized to raise the priority to important, so that it does get installed. | 22:05 |
| gnarface | ah, good to know | 22:07 |
| fsmithred | yeah, I still haven't learned ip other than 'ip a' | 22:08 |
| furrymcgee | I think bsd use ifconfig too | 22:21 |
| rwp | Yes. *BSD uses ifconfig. For Linux kernels the problem is that there are new configurations which just don't map through ifconfig anymore and are invisible there. So "ip" is preferred. | 22:40 |
| rwp | Honestly I find "ip" more logically cohesive than "ifconfig". I have converted over to it. | 22:40 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!