Iain -- I'm somewhat familiar with these sorts of embedded USB flash modules from my day job. You may want to see if you can get a small adaptor board which would allow you to connect your module into a normal Type A socket, like those on the outside of your chassis. That would open up a lot more configurations for testing. Many vendors of these modules make an apropriate adaptor board. You should ask your vendor. Matt On Fri, Sep 23, 2011 at 11:57 AM, Iain Barker <IBarker@xxxxxxxxxxxxxx> wrote: > I'll add a few more details since you seem to be somewhat interested in our specific application. > > The HP DL120 server has an EHCI-only HCD, which connects to an embedded USB 2.0 hub chip on the motherboard itself. There is no OHCI nor UHCI companion controller as far as I can tell. USB physical ports connect via motherboard traces from that embedded hub to the USB type-A sockets on the front and rear of the chassis, and also to an internal USB header. > > The internal header is intended for use with fixed USB storage modules. We are using an industrial-grade module similar to this one http://www.atpinc.com/p2-4a.php?sn=00000533 > > The signal traces on the motherboard from the embedded USB socket to the USB controller are fairly short and direct, but I noticed the USB bus traces running from the embedded hub out to the external chassis connectors pass very close to the CPU. A keyboard and mouse are connected to those external connectors on the rear of the chasiss. > > So I thought: maybe the problem was a physical issue with EMI or noise, and perhaps de-rating the transfer speed on those signal traces to FS could mitigate the problem. I hoped for a way to tell the HCD to do that, which is where this thread started... > > In terms of physical workarounds, as the USB storage module resides within the server chassis itself, there's nowhere to insert a FS hub unfortunately. So any fix or workaround is going to have to be in software. Unfortunately from Alan's responses it seems there isn't a way to slow down the HDC to BUS connection in software, as that is hard-coded at HS. > > I am testing now; if I can get a debug trace (the problem is very intermittent) I'll post a followup shortly. > > -- > 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 > -- Matthew Dharm Maintainer, USB Mass Storage driver for Linux -- 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
