ROD's new TCP/IP stack.

I am also using v7.03 on my Pinebook Pro without any issues.

@Raik was it easy to setup the Huawei E3372 LTE USB Stick to obtain a portable wifi option?

Is an E3372H. Maybe it make any sense.
Is easy to use. You need EtherUSB 0.43 and USBDevSwp. Configure your Stick in USBDevSwp. Most is autodetect. If your Stick is configured, he will be swapped from “CD” to Etherdevice than ever you plug in.
Use the PineTools or an Obey to connect your Computer with the mobile internet.
But it is not WiFi! Is LTE 4G. You need a SIM card…

The Huawei E8372h-320 also works for me with EtherUSB 0.43. Its behaviour is sligthly different. It doesn’t requires USBDevSwp as it swaps itself when no driver tries to claim the offered USB class. The Huawei E8372h-320 additionally to LTE offers an inbuild WiFi Server (not a client!).

Until now I didn’t found an USB Wifi client adapter which makes usage of USB CDC ECM class. May be there had been such sticks a couple of years ago. The problem is that WiFi always needs configuration which USB CDC ECM doesn’t provide. So the only possibility to make usage of USB CDC ECM in WiFi or mobile communication adapters is to implement a Web based configuration. So both Huawei sticks are offering such a Web based configuration (Iris required or configuration must be done using a LINUX/Windows Browser!).

Huawei also offers HiLink Network access for limited control of such sticks but I didn’t manage to get this working under RISC OS until now.

Thanks Raik and Thomas I have a Huawei EAAD modem stick which USBDevSwp recognises and has a config file for. However, usbinfo identifies it as a CD alternating with a modem icon. I am not sure where to get EtherUSB 0.43, the latest ROOL download only has EtherUSB 0.42 and the new stack identifies as EtherUSB 7.03. Any ideas?

usbinfo identifies it as a CD alternating with a modem icon

Hi Norman, I think the goal of USBDevswap is to get the device out of the CD/mass storage (of windows driver) and into the modem mode. I think the device id changes if that is successful.
That should visible with something that reads ids. another method might by looking at the System variables which should update.
Good to hear of Thomas’s E8372h-320 that works out what to do by itself.

USBDevswap is to get the device out of the CD/mass storage (of windows driver) and into the modem mode.

Thats the point. For my old ZTE stick it was possible to switch in modemmode by dismount the CD via Iconbar icon.

Fair enough, is there possibly a usb/FS command that would assimilate the dismount click?

Fair enough, is there possibly a usb/FS command that would assimilate the dismount click?

What about

*-scsi-dismount

@ Ronald, happy to report that all is working well. My take is that the alternating between CD and Modem states simply indicates the modem stick is available to work in either mode. It just requires a suitable piece of software to claim one or other mode. EtherUSB 0.43 was the missing piece of software required to get it all working. Big thanks to Raik for helping me to get there.

No, not me.
Big thanks to Thomas!!!

hi guys i have a e3372 USB modem will this work with drivers above if so where do i get the software to try out on my Pinebook pro?

Norman. so just the EtherUSB 0.43 on it’s own is enough then. Sounds promising. I assume the looping between the two ID’s stops once EtherUSB takes it over as an internet interface? To the system it would be like rapidly changing usb devices physically and refreshing the status each time while it is looping.
Unless both CD and Modem use the same device ID?

…so where do i get the software to try out on my Pinebook pro?

One way is you contact me via riscos.berlin or you contact Thomas.
;-)

The product ID is changing after swap. My stick swaps and swaps back if the right EtherUSB is not present.

My Huawei modem stick is E8372h-320 and all I needed to get it talking to the PineBook Pro was USBDevSwp and the EtherUSB version 0.43. Both of which have been developed by Thomas Milius. The modem stick will also work with the PineBook.

Just a quick thank you Raik & Thomas the software work great, I am posting this from hillarys boat harbour Perth Western Australia on my now mobile Pinebook Pro
While i am here i would like to wish everone a great Christmas & a happy new year.

@Rob, great news another PineBook Pro goes mobile. BTW, great spot Hillary Boat harbour, I was there quite a few years ago but never made the trip over to Rottnest which is supposed to be great.

You know what it’s like when you live next to it you don’t visit enough but it is good you sit there and have an early morning coffee
Looking out over the Indian Ocean.
Much better than sunny Coventry where I was born.

Looks like the v7.03 stack is now out of beta – see https://groups.google.com/g/comp.sys.acorn.announce

Given it a whirl here on FOURtress this morning and it seems OK so far.

Has anyne tried to do any IPv6 sockets programming yet? I’m falling at the first hurdle. In BASIC, I thought I could create a socket like this:

SYS XSocket_Creat%, AF_INET6, SOCK_STREAM, 0 TO s%; flags%

where AF_INET6 = 28 and SOCK_STREAM = 1. But I get the error “Protocol not supported”. I’ve also tried with R2 = IPPROTO_IPV6, i.e. 41, and it makes no difference.

Internet 7.03 (15 Dec 2022)

I’m going to assume that XSocket_Creat% is actually the equivalent of “XSocket_Creat”, given what you’ve said, but… AF_INET6 is 24, not 28.

That’s why it’s not working – you’re requesting with the wrong constant.

Tested on Pyromaniac with the Internet stack:

>SYS "Socket_Creat", 28, 1, 0 TO s%
Issuing service call &00000006 (Error)

Protocol not supported
>SYS "Socket_Creat", 24, 1, 0 TO s%
>P.s%
         1
>

Thanks, Charles. Very interesting… Here’s a line from AcornC/C++.Export.APCS-32.Lib.TCPIPLibs.sys.h.socket on my system:

#define AF_INET6 28 /* IPv6 */

I thought I couldn’t have made so stupid an error as to copy the constant incorrectly.

In the same file, AF_SIP is defined as 24 (Simple Internet Protocol).

I guess that’s wrong then. Complain to your supplier.

The value used by the Internet 6 system (and RISC OS Pyromaniac Internet which supports IPv6, which has followed that set by the Internet 6 system, despite having been first to do so), is that AF_INET6 is 24.

As you were asking, there is an example of using the IPv6 calls to fetch content from a remote server in my RISC OS tests repository:

https://github.com/gerph/riscos-tests/blob/master/testcode/s/internet_http

with the constants in:

https://github.com/gerph/riscos-tests/blob/master/testcode/hdr/internet

These tests have been used to test by RISC OS Pyromaniac and the Internet 6 system (although the latter was some time back, and I’ve not re-tested them myself – I presume others will have done so, as I supplied them to the developers).

Thanks, Charles. Very interesting… Here’s a line from AcornC/C++.Export.APCS-32.Lib.TCPIPLibs.sys.h.socket on my system:
#define AF_INET6 28 /* IPv6 */

Doing a quick check around the web, I came across various values for AF_INET6, depending on the operating system. E.g. Linux uses 10, Apple 30, someone on a HP forum suggested 22 and I’ve seen 23 mentioned as well.
Apparently 24 is from OpenBSD and 28 from FreeBSD. Are we talking about the ROOL stack versus the ROD stack here? If so, we’re heading for a nice, new type of mess…

@Frank: I found the same problem. I think we’ve known about something similar for years: note that error numbers also vary across systems.

@Charles: thank you again. I have got the first signs of operation of an IPv6 server. It’s a nuisance that Firefox really, really doesn’t want to do non-secure HTTP, but I’ll just have to live with that for now.