Em Qui, 2011-01-06 às 22:03 +0100, Peter Stuge escreveu: > > The MSO-19 uses a usb-serial bridge, and except for hardware > > identification, all the communication is done in the serial layer. > > That is incredibly stupid device design.
Indeed, I think they did this to keep firmware compatibility, as they have/had products that use real serial ports. > The only correct way to identify devices is using the serial number, > as stored in a string descriptor referenced in the iSerial field in > the device descriptor. On USB identifying device instances is not an issue, even without iSerial: Bus/Device numbers are unique. > > and not support multiple devices at all? > > No. Use the serial number to distinguish devices. The device uses > vendor specific interface so it is easy enough to program directly > using libusb. I'm not sure that it really makes sense to use a kernel > driver for this device. You mean.. Replicating standard usb-serial, supported by kernel mode drivers on virtually all platforms with an userspace implementation? I don't think it is worth, it would be easier to support multiple instances on linux(with libudev), and just a single instance on other platforms(with libusb). libudev can be a linux-only dependency, and the code would definitely be smaller than a full usb-serial userspace implementation. -- Daniel Ribeiro
signature.asc
Description: Esta é uma parte de mensagem assinada digitalmente
------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________ sigrok-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/sigrok-devel
