On Wed, Mar 24, 2021 at 11:23:11PM +0000, Stuart Henderson wrote: > In my ongoing search to find a umb(4) that will actually work that > isn't the one in my laptop (since my EM7305 has been a failure), > I picked up a Huawei ME906s-158 off ebay. It attaches to cdce and ugen > and fails to work: > > cdce0 at uhub2 port 4 configuration 2 interface 0 "Huawei Technologies Co., > Ltd. HUAWEI Mobile" rev 2.10/1.02 addr 3 > cdce0: could not find data bulk in > ugen0 at uhub2 port 4 configuration 2 "Huawei Technologies Co., Ltd. HUAWEI > Mobile" rev 2.10/1.02 addr 3 > > Any information I can find for it suggests that it does MBIM, and > indeed if I disable cdce in the kernel it picks up on a different > config: > > umb0 at uhub2 port 4 configuration 3 interface 0 "Huawei Technologies Co., > Ltd. HUAWEI Mobile" rev 2.10/1.02 addr 3 > ugen0 at uhub2 port 4 configuration 3 "Huawei Technologies Co., Ltd. HUAWEI > Mobile" rev 2.10/1.02 addr 3 > > and after I figured out which of the APU2's mPCIe slots is routed to > the SIM slot (the middle one) it actually negotiates with the network > > umb0: flags=8855<UP,DEBUG,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1500 > index 11 priority 6 llprio 3 > roaming disabled registration home network > state up cell-class LTE rssi -83dBm speed 143Mbps up 143Mbps down > SIM initialized PIN valid (3 attempts left) > subscriber-id 234xxxxxxxxxxxx ICC-id 8944xxxxxxxxxxxxxxx provider 3 > device ML1ME906SM IMEI 867xxxxxxxxxxxx firmware 11.617.04.00.00 > phone# +44xxxxxxxxxx APN 3internet > dns 217.171.132.0 217.171.135.0 > groups: egress > status: active > inet 94.197.84.2 --> 94.197.84.1 netmask 0xfffffffc > > It doesn't seem to be working properly yet (packets transmitted over it > don't arrive at the destination) but I'm not 100% convinced that it's > not the network yet, I'll find some more SIMs to try. > > In the meantime, any suggestions how to knock it out from attaching > to cdce? Is there a way to drop out in the attach routine (e.g. if there's > no data bulk in) to allow another driver to attempt attaching, > or is it committed to a specific driver once it has matched? > > If the latter, would it make sense for cdce to look for MBIM on > another configuration and skip matching in that case? Or do (move > or replicate) some of cdce_attach()'s checks in cdce_match so it > can skip attaching if cdce isn't going to work? > > There are loads of these showing up (probably from laptops broken for > parts) for about 20 $currency_units on ebay/similar now so it would > be quite nice to get them working. A few similar devices showed up > on the lists before but I haven't noticed anyone trying to disable > cdce on them yet. > https://www.google.com/search?q=huawei+%22cdce0:+could+not+find+data+bulk+in%22+site:openbsd-archive.7691.n7.nabble.com > > If anyone else is thinking of getting one to poke at, to use in an APU > they also need a M.2 -> mPCIe adapter (aka NGFF -> mPCIe) with 'B' key > (doesn't need a sim carrier on the adapter), plus whatever u.fl pigtails > and antennas (the proper multiband LTE antennas usually have SMA > connectors rather than the RP-SMA common with wifi). > > Descriptors from lsusb below (seach for "HUAWEI Mobile Broadband Module" > for the 'right' one).
Without having looked at anything, it might be worth looking at the most recent mail in this thread: 'Re: [PATCH] umb(4) fix for X20 (DW5821e) in Dell Latitude 7300'
