Re: hostmode . the script
 

for speed detection a single sequence of i2c commands is enough ( first i2c command from booston )
the kernel don't need a circular timer for stable vbus for speedcheck
a single 32s pulse** for first hardware testing, and a second pulse for full - high selection ( ** turned off when not necesary )
then set a modified procfs. echo A(utosense) > /sys/.....?
a simple program from userspace can check this output then start with the corect speed.
i think is there is a great complication stopping bme and start a circular timer from kernel space. for sure the timer can do the job easy "from inside"

i will make a modification in the script with single i2c command for first speed ckeck.
this avoid ugly killall booston in first step
a simple daemon can watch ID* pin so triggering the host will make things simplier and natural

*with filled CA-157 cable or uAsoket

please provide feedback about the script (PM or .. your choice )

Re: [Announce] USB hostmode beta release
 

blue_led, wouldn't it be profitable, to write how correct order of doing things should look, using Your script? I mean executing it before/after connecting device physically, etc. I think it would save us plentora of false bug reports/"it doesn't work for me" posts.

As for script itself, I got strange results testing it with my low-speed USB keyboard, connected to high-speed hub. It result in connecting hub at full-speed (I'm 100% sure hub is high-speed), but I'm not rising alarm, until I get opportunity to check it with another female-female adaptor (loose connection may be the case here). In HEN, sometimes full/low-speed tolerates loose cables, while high doesn't...

I'm waiting eagerly for other reports, cause I've lost hope for new GUI released by MAG.

/Estel

// Edit
I'm using joerg_rw's improved booston.

Re: [Announce] USB hostmode beta release
 

I'm writing bq2415x charging kernel module, so something related should be done in this module too (as replacement for BME).
http://atrey.karlin.mff.cuni.cz/~pali/bq2415x/

Re: [Announce] USB hostmode beta release
 

yes. otg vbus controll and force boost sysfs :D

@estel
i always start with device conected using original h-e-n or my script. connect the device and tap host icon. that's all :) is irelevant the moment of the connection until vbus go on
without external power i have no problem with 1hub, 1 sd card reader+card, and a mouse all connected together copying files and working with the mouse !
if your n900 loose vbus ... no hope. any vbus drop below 4.4 V will lost the host
you should have an icon pointing to the script. the shortcut can be added as usual on any desktop .
if icon is a blue square then reboot.
why you need gui. it is an programmer tool !! i say with full respect to Mohammad work

addendum
@estel
i confirm there is a bug.
with some hubs with integrated cardreaders the script will fail.( timeout problems )
an updated version will come

Re: [Announce] USB hostmode beta release
 

Thank you. With this file it has worked well.

Re: [Announce] USB hostmode beta release
 

The script works with several pendrive I've tried.

What if I'm using an external power source, the N900 starts loading?

Re: [Announce] USB hostmode beta release
 

@blue_led

Thanks for answer. Another small suggestion - could You consider including improved booston in Your script set? I've copied it manually - to check if it works - but, normally, it contain vanilla version of booston straight from H-E-N.

/Estel

//Edit
Not rellevant to script set, but I'm quite surprised, that You can use H-E-N with USB stuff connected before vbooston. I was never able to do that - in my case, correct order is to plug device shortly after booston, before enumerating (obviously). Otherwise, I get bunch of errors like "device not accepting address xyz".

// Edit 2
As for whitewolf suggestion, please *don't* enable charging with external power detected, at least by default. There are use cases - at least, in my situation - hen I don't want N900 to charge, at the same time powering USB devices (connected to it) externally via Y-cable/powered HUB.

In other cases, executing boostoff && charge.sh is simple enough.

Re: [Announce] USB hostmode beta release
 

// Edit 2
As for whitewolf suggestion, please *don't* enable charging with external power detected, at least by default. There are use cases - at least, in my situation - hen I don't want N900 to charge, at the same time powering USB devices (connected to it) externally via Y-cable/powered HUB.


why not? if you're using HEN it's going to run out of battery alot sooner then if you wasn't using HEN.

Maybe you shouldn't even be using the Y adapter, instead just use plain powered hub, which wont return charge to the phone.

So for me, I would rather it charge on external power.
Besides, I think the charge.sh script only charges when the battery isn't full.

Re: [Announce] USB hostmode beta release
 

First of all, some powered hubs does provide power to uplink. Yea, I know they shouldn't, but...

Second, it's not healthy for battery (life) to keep it at max voltage. LiPo's (LiIon's, in particular) like best to be keep between 20-80%. For people working for hours with keyboard/mouse connected to N900, keeping it auto-charged is extremely unhealthy for battery.

Not mentioning plentora of other situations, like calibrating battery chip, etc.

As for "charging only to 100%", what else do You want? Battery exploding due to overcharge? ;)

In any case, invoking charge.sh when someone want charging, is easy enough. You won't get other charging method during hostmode, anyway.

/Estel

Re: [Announce] USB hostmode beta release
 

While I am aware. My English is not very good.

¿I can or not to load while I have connected a USB device? Should I not do? I refer to your script, Estel

If yes, what steps should I follow.

I have a cable ready with diodes to prevent the arrival of current to the N900.

As a Y cable like this.
http://t2.gstatic.com/images?q=tbn:A...DG3SpQeL_ImYWv
The plug and cable with thinner section for additional power.