**** BEGIN LOGGING AT Tue Feb 23 02:59:58 2010 Feb 23 15:03:18 denkenz: FYI: Huawei E156C -- SMS not supported by this modem. If this is in error please submit patches to support this hardware Feb 23 15:03:41 I can make it register to a network but can't establish a voicecall Feb 23 15:04:04 (and power, of course) Feb 23 16:08:32 jprvita: you'll have to give me an at trace so we can determine why it doesn't work Feb 23 16:08:51 jprvita: The voicecall on the USB sticks is generally impossible anyway, no audio transport Feb 23 16:36:39 denkenz: around? Feb 23 16:36:48 ygu_home: yep Feb 23 16:37:11 I have several questions on ofono and phonesim. Feb 23 16:37:24 First is about the patch for Huawei modem. Feb 23 16:37:52 I still think the way I use udev is correct. Feb 23 16:38:07 it isn't Feb 23 16:38:16 What I judge there is in the context of em770 modem. Feb 23 16:39:09 Still took the example I gave in mail list. Feb 23 16:39:16 looking at device '/devices/pci0000:00/0000:00:1a.7/usb1/1-3/1-3:1.2/ttyUSB2/tty/ttyUSB2': (device1) KERNEL=="ttyUSB2" SUBSYSTEM=="tty" DRIVER=="" Feb 23 16:39:42 For this modem, the kernel name is "ttyUSB2" Feb 23 16:40:05 This is not a name which can be changed by udev rules. Feb 23 16:40:19 Yes, but it can be changed by presence of other devices Feb 23 16:40:37 If you plug 3 EM770, USB2 is not a valid thing to choose always Feb 23 16:40:55 I tried to plug in another Huawei em770 modem. Feb 23 16:41:21 The path became /devices/pci0000:00/0000:00:1a.7/usb1/1-4/1-4:1.2/ttyUSB2/tty/ttyUSB2 Feb 23 16:41:33 The name of kernel is still ttyUSB2 Feb 23 16:42:43 I think if I plug in others, the kernel name would stay unchanged. Feb 23 16:43:51 What's your opinion? Feb 23 16:45:19 This string ("ttyUSB*") is not the thing we see in /dev. Feb 23 16:45:46 I don't know for sure, but I guess it's decided by driver. Feb 23 16:46:11 I don't like it, but if there's no other attribute we can select on I'm ok with this Feb 23 16:46:29 However, put a big fat comment in there explaining why your approach works Feb 23 16:46:37 Resubmit and let holtmann review Feb 23 16:47:03 Hold on here. The TTY layer works differently here. The second Huawei dongle will not get ttyUSB2. Feb 23 16:47:46 holtmann: You mean the KERNEL? Feb 23 16:48:19 Yes, the USB serial layer numbers the TTYs in order. There is no chance that when both are plugged in that both are ttyUSB2. Feb 23 16:49:09 If I plug in two Huawei em770 modem, I get two KERNEL=="ttyUSB2" Feb 23 16:49:21 Also if you plug in an Option modem first and then a Huawei, the TTY numbering is different. Feb 23 16:49:22 /devices/pci0000:00/0000:00:1a.7/usb1/1-3/1-3:1.2/ttyUSB2/tty/ttyUSB2 and /devices/pci0000:00/0000:00:1a.7/usb1/1-4/1-4:1.2/ttyUSB2/tty/ttyUSB2 Feb 23 16:49:42 That is impossible. If would clash in /sys/class/tty/ Feb 23 16:49:55 Are these plugged in at the same time. Feb 23 16:50:02 holtmann: This string is not the one we see in /dev Feb 23 16:50:37 That is irrelevant. What does /sys/class/tty/ tell you about the USB devices. Feb 23 16:50:43 For first modem, I see ttyUSB0-ttyUSB4. For the second one, I see ttyUSB5-ttyUSB9 Feb 23 16:51:02 What kernel version is this? Feb 23 16:51:24 I'm not at office now, and I don't have the modem at hand. Feb 23 16:51:53 It's ubuntu karmic, and I think the kernel is 2.6.31 Feb 23 16:51:57 What driver is claiming this device? option? Feb 23 16:52:29 This is huawei em770 modem. Feb 23 16:52:38 What driver is claiming it? Feb 23 16:53:07 We don't care about Karmic kernel. That is off upstream anyway. Redo testing with a 2.6.33 kernel. Feb 23 16:53:42 OK. I will update my kernel tomorrow. Feb 23 16:53:49 And get back to you for the result. Feb 23 16:54:28 Include dumps from udev database. I need to know what is going on here. If we get consistent port numbers, that would be great, but I have never seen it so far. Feb 23 16:55:08 OK. Feb 23 16:55:51 By the way, if this port number is not stable, how to judge if it's ttyUSB2, rather than ttyUSB0? Feb 23 16:56:10 May I use ATTRS{bInterfaceNumber}=="02"? Feb 23 16:57:15 I just looked at the udev info I ever got for this modem: Feb 23 16:57:33 looking at device '/devices/pci0000:00/0000:00:1a.7/usb1/1-3/1-3:1.2/ttyUSB2/tty/ttyUSB2': (device1) KERNEL=="ttyUSB2" SUBSYSTEM=="tty" DRIVER=="" looking at parent device '/devices/pci0000:00/0000:00:1a.7/usb1/1-3/1-3:1.2/ttyUSB2': (device2) KERNELS=="ttyUSB2" SUBSYSTEMS=="usb-serial" DRIVERS=="option1" ATTRS{port_number}=="0" looking at parent device '/devices/pci0000:00/0000:00:1a.7/usb1/1-3/1-3:1.2': Feb 23 16:57:55 Oh, the format is bad. Feb 23 16:58:07 The driver there is "option". Feb 23 16:58:54 device1-4 were added by me. Feb 23 17:00:47 I ever tried to plugin two em770 modems at office, and the info for two modems are almost identical. Feb 23 17:01:05 Except the first one is 1-3, the second one is 1-4 Feb 23 17:01:44 However, I will update my kernel to 2.6.33 to see what will happen. Feb 23 17:02:39 denkenz: I have another question on phonesim. Feb 23 17:03:19 Current phonesim would not work if "-gui" option is missed. Feb 23 17:04:09 denkenz: I looked into the code, and found it is a bug in src/phonesim.cpp. Feb 23 17:04:42 We issue "+CMGL=4" to phonesim, but phonesim would return no feedback. Feb 23 17:11:21 denkenz: I have another question on phonesim. Feb 23 17:12:17 I've honestly never tested the non-gui version, what do you need it for? Feb 23 17:12:38 It should work without gui Feb 23 17:13:02 And the only problem now is about SMS part. Feb 23 17:13:41 Maybe it should, but I've never tested it ;) Feb 23 17:13:56 Feel free to send patches to ofono to make phonesim work Feb 23 17:15:54 Fine. I don't quite understand the design why SMSList binds to the gui. However, I will try to fix it. Feb 23 17:16:18 denkenz: By the way, do you know how to compile phonesim with O0? Feb 23 17:16:39 export CFLAGS=-O0 and reconfigure? Feb 23 17:16:47 We don't have --disable-optimization. Feb 23 17:17:09 I will try your way. Feb 23 17:18:27 don't try to understand phonesim, you'll go insane ;) Feb 23 18:00:37 holtmann: it must be the usb_serial driver if that's what you mean about the huawei modem Feb 23 18:01:02 I think the Huawei is driven by the option driver. Feb 23 18:01:04 anyway, ATTRS{bInterfaceNumber} is the only thing you can match against (plus the configuration number or name, iConfiguration or similar) Feb 24 02:36:07 holtmann: denkenz: I tried to plug two em770 modems. This time I got two different port numbers. There must be some wrong with my test several days ago. I will modify the patch to use ATTRS{bInterfaceNumber}, and submit it ML. **** ENDING LOGGING AT Wed Feb 24 03:00:02 2010