I (and Estel, thanks) found issue with enumerating USB Keyboards, please check here in extkbd thread for log + details.

Having some problems. Active usb-hub connected to n900 and tried to enumerate hub and charge n900.
Kernel: 2.6.28.10-power49.

$ dmesg
[35922.641937] bq2415x-charger 2-006b: failed setting hook function, automode not supported
[35922.642791] bq2415x-charger 2-006b: driver registred
[35929.941802] bq2415x-charger 2-006b: mode: Host/HUB charger
[35976.746795] bq2415x-charger 2-006b: mode: Boost
[35988.928466] Forced hostmode error: no device attached
[35988.929168] usb usb1: usb resume
[35988.948455] hub 1-0:1.0: hub_resume
[35988.948486] hub 1-0:1.0: port 1: status 0101 change 0001
[35989.057952] hub 1-0:1.0: state 7 ports 1 chg 0002 evt 0000
[35989.057983] hub 1-0:1.0: port 1, status 0101, change 0000, 12 Mb/s
[35989.120300] Forced hostmode error: a full/low-speed device attached but high-speed mode selected
[35989.182952] usb 1-1: new full speed USB device using musb_hdrc and address 12
[35994.182983] usb 1-1: khubd timed out on ep0in len=0/64
[35999.182983] usb 1-1: khubd timed out on ep0in len=0/64
[36004.182952] usb 1-1: khubd timed out on ep0in len=0/64
[36004.245452] Forced hostmode error: a full/low-speed device attached but high-speed mode selected
[36004.307952] usb 1-1: device descriptor read/64, error -110

Commands used:

$ stop bme
$ modprobe power_supply
$ insmod bq2415x_charger.ko
$ echo host > /sys/class/power_supply/bq24150-0/mode
$ echo hosth > /sys/devices/platform/musb_hdrc/mode
$ echo boost > /sys/class/power_supply/bq24150-0/mode
$ echo F > /proc/driver/musb_hdrc

This is question for @joerg_rw:
How to correct set-up host mode with charging?

Everyone working on this, thanks for the continuing work on h-e-n, something I really value since I don't need to drag a netbook everywhere, and if I read right we can expect soon full OTG!
Please push this as a H-E-N update or new package soon.

Sort of OT but upthread there was mention that 50-80% was optimal battery charge. Can this somehow be set as a soft charge limit profile in some config file for people who are at home or near power and want to preserve the life of a battery at the expense of long run times? If I am going on an airplane or train I could edit it back for 100% to dead. Ideally it would be something you could click in the battery status applet.

Hi Pali,

there's no setup for charging hostmode. It's simply you don't run booston but rather charge script (please refer to my prev post that explains how to do that). And of course user is then responsible to apply VBUS +5V the right moment when normally H-E-N GUI would start booston, as the VBUS is needed by PHY and musb-core to establish a USB session, and of course it's needed by your USB peripheral for the same reason, even when the peripheral has a dedicated power supply separate from USB. You can think of VBUS as a 1 bit data signal line that tells "start session now" - both host and peripheral are monitoring that VBUS line to sync their USB chat.

cheers
jOERG

Hi biketool,
no, there will never be "full OTG" on N900. It's not possible physically, as the needed lines are not connected to the USB receptacle. Anyway what is your special interest in "full OTG"? It only brings HNP (Host Negotiation Protocol - host and peripheral swapping roles on the fly [1]) to USB that I never have seen being used anywhere (SRP seems to work, and ADP.is not supported by N900 hardware anyway even with better circuitry). So what is it you're missing in H-E-N that you hope "full OTG" could deliver?

Yes, that's feasible if you use charge.sh script rather than bme, to keep your battery charged and N900 powered. One of the bytes sent to bq24150 via i2cset defines battery-full-voltage as 4200mV. You of course can change that to a lower voltage, e.g. 4000mV, and thus treat battery nicely.
* Get a modified charge.sh that uses this 4000mV theshold
* stop bme
* charge.sh&
now your battery will kept floating at 4.00V. Alas also hal-addon-bme breaks and thus no more proper info to whole system about battery state and mode (e.g. your battery icon in status applet will not show the correct state and charge, no bat-low warnings, probably no gentle shutdown when you unplug and run flat on battery, etc)
So before unplugging from USB you want to do:
* killall charge.sh; # or 'kill %1' in the shell you did 'charge-sh&", or simply '^C' if you didn't start charge.sh as a background& process
* start bme
and hope for all the rocks fit together again so after a while bme talks to hal-addon-bme and also reports somewhat correct charge state (bme is known to get severely confused on restarts, so don't get scared when it thinks your battery is empty or whatever, it usually will adjust after a few minutes).

If you're honestly interested in this then ping me, I might prepare the modified script, maybe even with a icon for desktop to click on to toggle the states

Anyway, my warning in charge.sh still applies: DO NOT MESS WITH THIS unless you know what you're doing! Even while bq24150 is inherently safe and you can't really make battery explode, you are still messing with I2C bus(ses) that also run to other chip you actually could kill with a "properly" designed I2C command. (I'll not elaborate on this here, for obvious reasons)

cheers
jOERG


[1] http://en.wikipedia.org/wiki/USB_On-The-Go#Protocols

Deep in the bowels of this topic there were once a few posts about using USB host to send the N900 's audio signal to an external DAC. Has anyone succeeded in doing this?

I had misread upthread that there was a possible hardware hack to connect the ID pin for brave soldering iron jocks. Having full OTG is more for the status, even now h-e-n it just as useful.

I guess that's a question for the "problems" thread, not for here. Anyway from here it looks like "plug in USB soundcard, enable hostmode, load drivers (which is the point where things get tricky, not before)"

/j

USB gurus, any comment on this?

http://talk.maemo.org/showpost.php?p...6&postcount=55
(Not the MHL part)

/Estel