On Tue, May 16, 2017 at 05:04:37PM +0200, Guido.Kiener@xxxxxxxxxxxxxxxxx wrote: > > Sounds good. I think if we decide to extend the usbtmc driver then we will > hopefully find some driver developers within the companies who are > experienced :-) If you all are maintaining a Windows driver, I'm sure some of the developers involved will be more than willing to work on a sane operating system and help you out :) > > Where exactly are things not going fast enough? Have you found any > > specific bottlenecks? How fast are you needing to go? What type of > > interface do you expect userspace to have to handle high rates of data? > > > > I'm not sure for 100%, but I assume that reading the IN pipe could be > setup asynchronously (e.g. with usb_submit_urb(..)) just before sending > send_request_dev_dep_msg_in(..). > USBTMC_SIZE_IOBUFFER = 2kB is a bottleneck when dealing with data size > > 1MB (Need of Scatter/Gather). Please test and send us patches for what you find. You all have the hardware, which is the hardest thing for people to get to test this type of thing. > > Have you evaluated both and found any differences? > > The Windows USBTMC driver is maintained by the IVI Foundation. The main > difference is that protocol specific headers are filled up or analyzed in > user mode. This allows VISA vendors to be more flexible to maintain their > applications. The specification is here: > http://ivifoundation.org/downloads/Protocol%20Specifications/Ivi-6%202_USBTMC_2010-03-23.pdf That puts most of the device specific control into the applications, forcing everyone to duplicate the same logic over and over in all programs. Are you sure that is really very wise? Anyway, please test out the Linux driver, and if you find problems, please send patches, that's all we can offer here. > At the moment I'm still not sure about the right way to go. The goal is to > avoid troubles between different USBTMC applications. So we could do one > of the following: > 1. Drop out or disable the usbtmc driver. The IVI Foundation agrees to use > only libusb from now and in the future. I'm not going to delete the in-kernel driver, as obviously it is being used by people today, unless _everyone_ agrees that it should be done in userspace instead, and we have a working solution there that everyone can use (i.e. properly licensed.) > 2. Replace or extend the usbtmc driver with a new one. For current > distributions the libusb is an alternative. Again, please test what we have now, and send changes for what you have issues with. That's the way all Linux development happens, nothing special here for just this tiny driver :) good luck! greg k-h -- 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
