From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: To: 9fans@cse.psu.edu Subject: Re: [9fans] keyboard maps From: Fco.J.Ballesteros In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Date: Mon, 17 Nov 2003 19:41:30 +0100 Topicbox-Message-UUID: 8da60e7a-eacc-11e9-9e20-41e7f4b1d025 > this work you are doing with user-level USB is really interesting. It will > be neat if this turns into "usb outside the kernel" given the amazing > error-prone lockup nature of the usb chipsets we've seen. Supposing we can > push all this USB crud into user procs, do you think we'll be able to suck > up USB 2.0 at reasonable rates? I've only done a HID driver, all of the usb code was already there, and btw, all of the configuration is in usbd outside the kernel. The kernel seems to configure (almost) the bare minimum in usb(3). I'd like to keep the usb stack outside the kernel just while seeing errors; once it gets stable, I don't know if I wouldnt want to put it all back in the kernel just to avoid complexity (more file systems, more binds, more levels of indirection, etc.). Regarding 2.0, I'm in the mood of using usb for almost any new piece of hw we buy that is not integrated inside the PC. This means that sooner or later we'll have usb 2.0, it depends on who needs this earlier, if we or other 9fans.