On Saturday 25 April 2009, Steve Calfee wrote: > I haven't used it in awhile, but what does this change do to usbtest.c > and test.sh? That was the only widely known program that did user > level usb I/O Not exactly. What it does is an ioctl() to get at the various tests, none of which can be adequately issued from userspace. Greg ... comments? Do your patches keep that working? > and did it very well. Does the /sys/.... stuff > completely replace the capabilities of /proc/bus/usb/devices? Good question. Seems like maybe the drivers/usb/misc/usbtest.c code may be needing a new userspace interface, unless these patches preserve the ioctl functionality through debugfs. A new userspace interface might just be a simple chardev. The reason it didn't start that way was, as I recall, that hotplugging didn't support chardev creation back then ... so the only portable "find test devices" solution was browsing those /proc/bus/usb/BBB/DDD files. Now, there are actually two test modes. The main one I use is the one binding to a dedicated test device. But another mode is for control tests: "use *that* device". Since this is purely a test/development tool -- about 100% of our stress and negative test arsenal, in fact -- it's not a concern if its interface changes. But it is important that the tests continue to work. - Dave -- 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
