<div dir="ltr">Intel have the plan to opensource the PCIe driver. I would put it here once the Intel opensource it. <div>But the driver do not offer a cdc_mbim port, just offer AT port. In MM, it would manage with the base modem. And the VLAN ID map is defined by Intel, so I do not think the driver could be use in other modem chips.</div><div>As the link that Daniele put shows, MHI seems like a usual solution.</div></div> <div class="gmail_quote"><div dir="ltr" class="gmail_attr">Bjørn Mork <<a href="mailto:bjorn@mork.no">bjorn@mork.no</a>> 于2019年3月15日周五下午4:28写道： </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Now you made me really curious. That sounds a lot like the cdc_mbim userspace ABI. And I didn't expect anyone to copy that since I don't think anyone is happy with it. Including me. Or are they simply creating a usb mbim device? That would make sense. But then it should have worked with mm, so I guess not... Bjørn On March 15, 2019 8:59:04 AM GMT+01:00, "王道之" <<a href="mailto:lingzangwuhen@gmail.com" target="_blank">lingzangwuhen@gmail.com</a>> wrote: >Hey >The Intel offer a PCIe device driver to suit my situation. The driver >could >grab a virtual net port from the PCIe modem, then the user like me >could >use VLAN to divide the virtual net port in several VLAN ports, the VLAN >port can be mapped by socat as AT port and trace port. The map is >defined >by Intel. >But the MM could not add those virtual AT ports and virtual net port. >To do >that, I change some filter logic that would not influence the >architecture >now in use. And I add two UDEV flag to set the subsystem and port name >of >the virtual ports, flags named ID_MM_VIRTUAL_SUBSYSTEMS and >ID_MM_VIRTUAL_NAME, the way that the two new flags work just like the >already existing UDE flag D_MM_PHYSDEV_UID did. In this way, the MM >would >support modem devices with virtual ports. > >Bjørn Mork <<a href="mailto:bjorn@mork.no" target="_blank">bjorn@mork.no</a>> 于2019年3月9日周六下午8:09写道： > >> Daniele Palmas <<a href="mailto:dnlplm@gmail.com" target="_blank">dnlplm@gmail.com</a>> writes: >> > Il giorno lun 10 dic 2018 alle ore 14:25 Bjørn Mork <<a href="mailto:bjorn@mork.no" target="_blank">bjorn@mork.no</a>> >ha >> scritto: >> >> >> >> As for Qualcomm based modems: Haven't yet seen any docs for the >PCIe >> >> interface there either. So I'll assume it is similar to the Intel >one. >> > >> > related to Qualcomm I guess this development >> > >> > >> ><a href="http://archive.lwn.net:8080/linux-kernel/1524795811-21399-1-git-send-email-sdias@codeaurora.org/" rel="noreferrer" target="_blank">http://archive.lwn.net:8080/linux-kernel/1524795811-21399-1-git-send-email-sdias@codeaurora.org/</a> >> > >> > should be related, but unfortunately it seems that the attempt for >> > upstream inclusion stopped. >> >> Thanks for the pointer! Really interesting and encouraging. >> >> >> Bjørn >> </blockquote></div>