On Fri, 18 Dec 2009, Filip Zyzniewski wrote: > Hi, > > I have found out a way to connect an USB receptacle to USB host pins on > SA1111 inside a HP Jornada 720. The modification is described at > http://jlime.com/mw4/index.php/Hardware_documentation_jornada700_usbhost . > > Unfortunately it still doesn't work as expected - I have dumped some dmesgs > here: http://filip.math.uni.lodz.pl/jornada/usb_host/ . > The most interesting ones are these where kernel can retrieve proper > identification numbers (optical mouse, card driver). But communication fails > afterwards. Devices aren't usable. > > Things that surprise me in these logs: > > [ 1126.593375] sa1111-ohci 0400: OHCI 1.0, NO legacy support registers > > I've googled dmesg dumps and usually it says "with legacy support registers". > OHCI 1.0 seems strange (SA1111 is 1.1 compatible), but i've seen that people > have it in their logs. The legacy support thing is a bug in ohci-dbg.c. Look at the first line of ohci_dump_status(): temp = ohci_readl (controller, ®s->revision) & 0xff; The "& 0xff" shouldn't be there. By contrast, the "OHCI 1.0" value comes directly from the hardware. As far as I know there is no OHCI-1.1 standard, so it's not surprising that the device only claims compatibility with OHCI-1.0. > [ 1127.025533] hub 1-0:1.0: no power switching (usb 1.0) > > SA1111 is able to control power switching (and I have implemented the circuitry > doing this with a TPS2043). This probably results from detecting it as > a 1.0 host. No. It results from the value of the NPS bit in the hardware HcRhDescriptorA register. If value of NPS is 1 then you get that "no power switching" message. If you want to enable power switching, you might be able to do it by editing ohci-sa1111.c: Make it clear the NPS bit. > I am posting this in hope that somebody could give me pointers about where > to begin digging to fix this. I think hardware part is ok, because id > numbers ARE somehow transferred. What do you think about it? > > I know this is an old hardware, but HP Jornadas are still widely used > (especially > by nerds) thanks to their unique balance of features (size, keyboard, > battery life, > slots etc.). There is an active users community at http://www.jlime.com/. You should post some usbmon traces. You should also get a stack trace (Alt-SysRq-T) to find out where rmmod hangs. And you should check out the contents of the debugging files under /sys/kernel/debug/usb/ohci/0400/. Alan Stern -- 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