On Sat, Jun 27, 2009 at 12:35 AM, Dominic<dominic.r...@gmx.de> wrote:
Firstly I am not a Windows driver expert and I do not and can not code for live. But I know quite a bit of USB under Linux/Windows, especially libusb and libusb-win32. > Is the libusb-win32 1.0 branch compatible with libftdi (i.e. is it > compatible with libusb 0.1)? >From what I know, yes. It is a development branch so things can change. For Linux, libusb 1.0 is not API compatible with libusb-0.1. > Not sure if I understood you correctly here - the HID backend is for > HID-class devices, right? Specifically it is not for the FTDI chips? Right. It is not for FTDI chips. > > o the WinUSB backend brings a WHQL signed driver, which is good, because > > it runs on all windows versions that required signed drivers (Vista > > 64-bit, Windows 7) > > > > o The WinUSB How-To says that you need to get the whole driver package > > including the .inf file signed. Some postings to newsgroups suggest that > > this is not necessary. It might be that you get a warning, which > > shouldn't be that much of a problem. > > It is ok to load the driver even though there is a warning. Yes. I do not have XP/Vista 64 but the reports confirms this. >> You can do >> the same with libusb-win32 device driver if somebody can pay the >> digital signature (not so expensive). One company already got their >> libusb-win32 based device driver certified by WHQL but they >> renamed all the driver name to their own name. > > You seem to know the signature and WHQL stuff. As I understand it there are > at least two issues: > - WHQL certification for the driver (.sys) > - Signature for the driver package (.cab), including the .inf file that will > need to be modified whenever we add a new VID/PID Rather: 1) KMCS digital signature of the core driver 2) WHQL of the driver package (including the INF file) The WIndows built-in WinUSB and usbser.sys are examples of 1. You still need an INF file to load the driver. You still need to submit the packages for WHQL. > I'm not sure if WHQL certification and the driver signature are the same > thing? They are not. > Does digital signature mean for example a VeriSign issued signature? Would > Windows accept self-signed signatures? Can we create root certificates of > our own for this purpose? VeriSign is once of them. You can use only a few CAs -- you have to pay. http://www.microsoft.com/whdc/winlogo/drvsign/kmcs_walkthrough.mspx > I dislike the idea of paying for any of this because because some of these > costs might be recurring (digital signatures might expire?), and because it > would mean the project would have to cover costs for mostly commercial > dongles. It is said that you need to get the digital signature only once. This is good for libusb-win32 and many projects which use libusb-win32. WinUSB does not have this problem. >> Based on my experiences it is quite good. The filter driver can cause >> many problems (including serious BSOD and lost of all USB device) >> and is no longer recommended by the author. The device driver has >> never caused problem for me. > > The advantage of the filter driver is that it runs in parallel with other > installed drivers, e.g. FTDI's D2XX, right? Yes. But it is more difficult to get a proper filter driver. >> > o (Re-)writing libftdi with plain WinUSB should be possible and would >> > remove the uncertainties of libusb-win32, but would burden the OpenOCD >> > project with the task of keeping it stable and up to date. It would also >> > leave out Windows 2000 and before. >> >> You can always use libusb-win32+libftdi for Windows 2k. > > Having to maintain multiple different Windows drivers puts a pretty heavy > burden on the Windows packagers. > Maybe not that bad if a proper package can be created to give the user a selection. Hope this helps. -- Xiaofan http://mcuee.blogspot.com _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
