Hi Josua, let's say I'm newbie in kernel development. Actually I am definitely not focused in it. Therefore I can not give you an answer why in userspace or kernel. I just decided to do it kernel because I also have Huawei E220 modem and I saw that kernel driver can do it without any userspace activity. Kernel just prepares device to be used in userspace and this is what I enjoyed. So I also wanted that device to be put for userspace usage in the mode it has to be. Kernel will prepare it and afterwards pppd will work with it as a modem. Really, the only reason why I decided to create kernel patch was this idea, to give the device for use as it should be. And as I spent some time on it, I wanted to share it with the linux community. May be that you, kernel developers, will consider if it should be in the kernel, like those other are there, or not. If you want to know my opinion then it is that a kernel should provide devices in a state in which userspace application can use it. And many times those applications does not even have to know that device. For instance I have used the same ppp scripts with Huawei and this Toshiba. Because they are just modems which support standard set of AT commands. Regards, Peter Magdina On Tuesday 06 October 2009 15:38:06 Josua Dietze wrote: > Peter Magdina schrieb: > > I have created patch for linux kernel which handles Toshiba G450 > > as a HSDPA modem. > > Honestly, I'd rather see a principal decision to do all this > switching stuff either in the kernel or in the userspace. Up > to now every device could be handled by "udev" and > "usb_modeswitch". Even proper device identification (beyond > vendor and product ID, to avoid ambiguities) can be done > there. And lately the "option" driver is bound to switched > modems via the "new_id" facility, with no immediate need to > add them to the "option" ID list. > > I suggested earlier to include all mode switching in the > kernel and received no clear signal. So I went back to > maintaining the userspace mechanism. > > The situation is disturbing enough with Sierra and Option > devices handled by their own storage extension and Huawei > devices by "initializers.c". On the other hand, new devices > keep popping up and the userspace tool can be adapted quickly > once their switching command is known. > > With that mechanism in place, adding the switching command for > a specific device to usb-storage just creates confusion. > > > Josua Dietze > -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
