[SOLVED] Huawei E1692 3G modem... cant get it to work

twilightning · 2010-03-21 18:40:58

I am really frustruted about this. I ve tried everything and I cant get my huawei modem to work.

My modem is supported by the option driver (modinfo option) according to the device ID found in /etc/usb_modeswitch.conf for my model.

After running usb_modeswitch my modem is recognized as E620 instead of E1692 with wrong device ID, dont know if that matters.

I ve configured wvdial, modprobe.conf, udev rules, ppp options, and all I get is a "NO CARRIER"!

What have I missed?

Network is WIND ADSM HELLAS (Greece)

Last edited by twilightning (2010-04-02 23:25:38)

jerro · 2010-03-23 13:45:01

I want to know how to let my 3G modem worked? I use ZTE AC580.

twilightning · 2010-04-02 23:26:51

Solved, I used Sakis3G script run through the terminal, debug option activated...

http://sakis.tel4u.gr/blog/

twilightning · 2010-04-03 00:14:05

Oh and I have made some addition to this wiki article to help other users in the future:

http://wiki.archlinux.org/index.php/USB_3G_Modem

Nixie · 2010-04-20 13:43:20

Thank you twilightning. I managed to get my T-Mobile modem working after finding your wiki addition.

The Sakis3G script really is a godsend. It's not too hard to add support for the countries not already covered in the script (like the UK)

twilightning · 2010-06-08 12:05:10

My addition to the Wiki was deleted by "Aedit" stating this:

(So-called "script" actually contains a binary. Actual source code not available (using binary to "disassemble" itself as the author suggests cannot be trusted). Not the Arch way in any case.)

The author of the script can be trusted, he has put a lot of effort in the script to make it run as universal for all users and situations, while at the same time it has small footprint comparing to other methods.

Aedit please read the faq before making any assumptions
http://wiki.sakis3g.org/wiki/index.php? … akis3G_faq

Last edited by twilightning (2010-06-08 12:06:03)

Aedit · 2010-06-09 13:19:11

I did check it carefully, and didn't make my comments lightly. A script containing a binary needs to be treated as a binary, but you suggested people just download it and run it as root. I am not saying that the author is trustworthy or not, but we should not _need_ to trust him.

With the source you could put together an AUR package. Unfortunately it is not available in any conventional manner. Self-extraction in general can't be trusted (it could lie), but after checking the shell parts and hexdumping it I found that you can extract a tarball containing scripts _and_binaries_ with "gunzip sakis3g.gz ; mkdir extract ; dd bs=1024 skip=15 if=sakis3g of=extract/sakis3g.tar.bz2". First of all _delete_ the included binaries. Also delete the unnecessary source of usb_modeswitch which is in [community].

The main problem then would be that sakis3g is just not the right way to do it. Arch users usually want the correct and minimal explanation of how to make things work. The main script is over 8400 lines long! It's clear the author put a lot of work into it, but making USB modems work just is not that hard. Just install usb_modeswitch, wvdial and write an /etc/wvdial.conf (perhaps compare settings with sakis3g/files/operators.db).

Sakis · 2010-06-09 21:48:51

Hi all,

thank you both for your time dedicated in my work. I admit, not supplying source code natively imposes risk, and is indeed a terrible, still unintentional, omission. I am already redesigning site, documentation, README files and man pages for this reason.

Sakis3G script, as a project is just a shell script (at least, I only coded shell commands). Indeed it can produce both:
- An all-in-one bundle (called "Full version") containing a pre-compiled usb_modeswitch binary for i386, able to recompile by itself if not actually being used on i386, or
- A so called binary free version, working unmodified in all architects, relying on distribution (or user) providing a descend usb_modeswitch version within PATH. (By the way, this version is only 120KB large).

I am supplying it in "Full version" form due to not all distributions providing an updated usb_modeswitch version (some do not even provide an official package). I was never intending to "promote" Full version for a distribution's package, and whenever I do not forget to do so, I do inform users of risk imposed if they do use it.

Anyone intending to provide a package for a distribution, should pay a visit on this Wiki page which had been there since the very first day 0.2.0 was released. Reading at least Ideal setup section, could give a clue that I am still stepping on earth and what my intentions really are.

Another one part that I feel like mentioning is that it indeed does not target people wanting the minimal and politically correct way to do things. And actually, those people do not need wvdial either. Minimal way would be a fine tuned chat script to be used with pppd. And this is what Sakis3G is actually doing: it finally calls pppd with a specially crafted chat script, derived according to user's modem model and operator detected.

Ability to rely on wvdial as well, was only introduced due to some distributions providing pppd builds which restrict (as "privileged") some pppd options from being used, and because of another one restriction I wanted to apply on Sakis3G: not to touch anything on filesystem (It can perfectly work when all filesystems are mounted read-only). Creating a peer file within /etc/ppp/peers, just for being allowed to use privileges options (just like wvdial is made one on those distributions), was not an option for read-only-compliant Sakis3G.

Sakis3G script targets inexperienced users (immigrants from commercially available OSes) who expect their Linux distribution to handle their hardware, at least, equivalently to their previous OS (I know, Linux is not Windows, but they do not know that). This kind of people sometimes do a convention and sacrifice some of their gadget-oriented hardware, being disappointed by the amount of work required to make it work and them lacking the knowledge to do so.

However, inability to use a 3G modem in Linux, is capable to even make ME switch to another OS. Expensive services are usually paid by people to satisfy a need and not a hobby. An operating system unable to support a need, is immediately fdisked out of partition table. To those people it targets, and provides them with a single, self-contained, as-small-as-possible file which will make their 3G modem actually work, with them configuring nothing at all.

Poor Linux kernel actually has anything provided: it is GUI-oriented user-space which is not yet ready on reliably handling all devices yet (e.g. NetworkManager's modemmanager part being confused by many serial ports provided and using wrong one for attempting P-t-P connection, or gnome-ppp and kppp relying on user providing connection settings). Best fact to support this one statement is that a shell script is able to succeed where specialized binary user-space programs fail.

Those 8.4K lines of shell scripting, are actually logic ensuring it will work under all circumstances (kernel versions/distributions/available utilities/graphical environments/USB modems/Bluetooth modems/operators/APNs). Reason behind it being a shell script is because I do use C at work, and wanted some "unstructured" fun.

I've not yet decided the way that I will be distributing source, however, in order to justify things (until redesign is applied), I've created following permalinks which will be maintained in future releases as well, regardless of new structure:
http://www.sakis3g.org/versions/latest/ … rce.tar.gz (99KB)
http://www.sakis3g.org/versions/latest/ … ce.tar.bz2 (109KB)

Whole this discussion had been really constructive, bringing into my attention that I was supplying source code only via the wrong way. This is already changed, and will be better promoted in the upcoming days and revisions.

Please [pause] excuse me if I went offending or emotionally driven. None of them had been intentional, and some part may even rely on my English being "rude". I just want to have things cleared. Even if Sakis3G script is far away from "the Arch way", it is still a GNU GPL project which has already helped people to have their hardware working on their distribution. This one by itself is the only one (and trust me, adequate) reward for me.

Thank you both,

Sakis Dimopoulos

Edit: Fixxed source tarball links.

Last edited by Sakis (2010-06-10 11:46:07)

Aedit · 2010-06-10 10:43:26

Hi Sakis,
thank you for your long and interesting reply! It's very good that you are going to supply a source-only tar file (are those links broken at the moment by the way?). I understand your program is aimed at those who would prefer/expect an all-in-one package and its size reflects the difficulty of making it work in this way on all distros. Some networking tools are frustrating I agree. And the mobile networks and device manufacturers could do much better at supplying settings information.

I think for Arch it would be best to create an AUR package (perhaps you would like to maintain it? It's quite quick and easy). That way you get an up-to-date and Arch-specific installation which is integrated properly with the package management. It would help if you made individual tarballs available for each release (rather than just the latest) so the makepkg integrity does not fail when you release a new version. Also a git repo would help others follow the development and send patches.

Since Arch is a rolling distro, usb_modeswitch and other deps will stay current. I saw you suggested that it be packaged as 2 packages -- one for the program+udev and one for the device configuration files. In Arch these have been rolled into one package. I don't think it's a big issue, because it can be updated as frequently as necessary. (The version number could be the concatenation of the two subpackage version numbers though, but there haven't been any updates of either part since the split).

One question: would it make sense to dump out the chat script so those who want to could use it directly?

Last edited by Aedit (2010-06-10 12:55:36)

Sakis · 2010-06-10 12:55:36

Aedit wrote:

are those links broken at the moment by the way?

They had been suffering from a typo (on the domain part). They are now fixed.

Aedit wrote:

I think for Arch it would be best to create an AUR package (perhaps you would like to maintain it? It's quite quick and easy).

Project, for the time being, is rather time consuming (especially because I consider documentation an important factor). Providing specific-distro packages may result into more people (directly) requiring support which I am not yet capable to provide. While I remain one man, capacity is limited. Not to mention being the same man providing packages does not expand "trust". A distinct package maintainer also carries/implies the responsibility of having checked source code against malicious parts.

Aedit wrote:

Also a git repo would help others follow the development and send patches.

This would provide an answer why I am still one man. Project is currently 5 months old. Back in January, I only published a shell script I was using myself. Was not expecting people will like it that much to "force" me releasing updates. Place I am hosting it already suffers from bandwidth issues. If traffic keeps growing with the same rate, it will be soon transferred to SourceForce and many things will anyway change.

Aedit wrote:

Since Arch is a rolling distro, usb_modeswitch and other deps will stay current. I saw you suggested that it be packaged as 2 packages -- one for the program+udev and one for the device configuration files. In Arch these have been rolled into one package. I don't think it's a big issue, because it can be updated as frequently as necessary.

Exactly. Suggestion applies on distributions which run a specific cycle for releasing a new binary. Market penetration rate of those devices is currently high enough (new devices popping out every day, some of them being OEM re-branded models sporting different USB IDs for no apparent reason) requiring often updates into the configuration part of Usb-ModeSwitch.

Aedit wrote:

One question: would it make sense to dump out the chat script so those who want to could use it directly?

This is a really interesting suggestion. However, a chat script will be unable to incorporate logic (for example, does SIM need a PIN unlock?), which will as well lead into a different pppd peer file being required per situation (first time connecting or not), pretty much resembling the situation of wvdial configuration files ("wvdial dialisp" and "wvdial pin dialisp"). Not to mention that Sakis3G also setups factors which pppd, wvdial and networkmanager do not take into consideration, e.g. unbinds usb_storage from device if kernel keeps resetting device due to storage part misbehaving (see "avoid_reset_quirk" available in newer kernels but not in older ones).

Sakis

PS: "Latest" word on URLs, whenever appearing, is always a filesystem symlink. One can exchange it with version number and will still be working. "Latest" is only promoted so that instructions appearing around the web, do not turn obsolete.

Aedit · 2010-06-10 13:40:11

Sakis wrote:

Project, for the time being, is rather time consuming ... Not to mention being the same man providing packages does not expand "trust". A distinct package maintainer also carries/implies the responsibility of having checked source code against malicious parts.

I see your point. I'm sure someone else will do it. At least we are talking about human-readable files now. Everyone is _supposed_ to thoroughly check AUR PKGBUILDS before using them (and not blindly use yaourt...).

Sakis wrote:

... a chat script will be unable to incorporate logic (for example, does SIM need a PIN unlock?), which will as well lead into a different pppd peer file being required per situation (first time connecting or not), pretty much resembling the situation of wvdial configuration files ("wvdial dialisp" and "wvdial pin dialisp"). Not to mention that Sakis3G also setups factors which pppd, wvdial and networkmanager do not take into consideration, e.g. unbinds usb_storage from device if kernel keeps resetting device due to storage part misbehaving (see "avoid_reset_quirk" available in newer kernels but not in older ones).

Sure a generating a chat script (or a wvdial.conf) wouldn't provide all the functionality but I was thinking of advanced users who know how to deal with other issues like the PIN (e.g. disable it) and usb_storage but have trouble finding the correct modem or network settings. Also for communication -- this thread is marked as solved yet nobody has posted a wvdial.conf or otherwise stated what really made the devices work.

Sakis · 2010-06-10 14:36:55

Aedit wrote:

...At least we are talking about human-readable files now...

When it comes to source, yes. Even source tarball still includes png icons (I am yet uncomfortable with svg although I know they can contribute into minimizing precompiled bundle's size even further).

Aedit wrote:

Sure a generating a chat script (or a wvdial.conf) wouldn't provide all the functionality but I was thinking of advanced users who know how to deal with other issues like the PIN (e.g. disable it) and usb_storage but have trouble finding the correct modem or network settings. Also for communication

Such users can:
- Rely on latest Usb-ModeSwitch which provides a symlink (/dev/gsmmodem) on modem port which is suitable for use (that's why I am listed there),
- Contact their operator's customer support for determining APN, APN username and APN password. While those details look too much technical, casual customer support technician knows how to answer those details because many smartphone users contact them for the same reason. All he has to do, is to avoid using the "Linux" word, or else technician might get seriously damaged.

Regarding intra-Sakis3G-user-to-Sakis3G-user communication, this is the way to do it because it (usually) works out-of-the-box and indeed there is no further "advice" a user could give.

Aedit wrote:

-- this thread is marked as solved yet nobody has posted a wvdial.conf or otherwise stated what really made the devices work.

Seems like threadstarter indeed enjoyed a zero-configuration situation . On the other hand, this one thread has gone badly abused and out of topic.

Sakis

twilightning · 2010-06-10 15:29:35

I am willing to create the initial AUR package of the Sakis3G script, but I would really like someone to help me in maintaining it.

Aedit · 2010-06-11 10:48:34

Sakis wrote:

Such users can:
- Rely on latest Usb-ModeSwitch which provides a symlink (/dev/gsmmodem) on modem port

Nice! Thanks.

Sakis wrote:

- Contact their operator's customer support

(picks self up from floor) Life is too short, really. "APN? Could you spell that please?"

twilightning wrote:

I am willing to create the initial AUR package of the Sakis3G script, but I would really like someone to help me in maintaining it.

Great! I have no modem hardware I can test with at the moment unfortunately. If you get it started I'm sure people will point out any problems.

twilightning · 2010-06-11 19:11:46

EDIT:// post cancelled!

Last edited by twilightning (2011-04-26 14:02:30)

Vamp898 · 2010-06-14 18:53:09

//EDIT Nevermind -.-

Last edited by Vamp898 (2010-06-14 19:08:50)

Arch Linux

#1 2010-03-21 18:40:58

[SOLVED] Huawei E1692 3G modem... cant get it to work

#2 2010-03-23 13:45:01

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#3 2010-04-02 23:26:51

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#4 2010-04-03 00:14:05

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#5 2010-04-20 13:43:20

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#6 2010-06-08 12:05:10

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#7 2010-06-09 13:19:11

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#8 2010-06-09 21:48:51

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#9 2010-06-10 10:43:26

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#10 2010-06-10 12:55:36

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#11 2010-06-10 13:40:11

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#12 2010-06-10 14:36:55

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#13 2010-06-10 15:29:35

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#14 2010-06-11 10:48:34

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#15 2010-06-11 19:11:46

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

#16 2010-06-14 18:53:09

Re: [SOLVED] Huawei E1692 3G modem... cant get it to work

Board footer