User:Joerg rw/hostmode/summary

N900 (rx51) hostmode notes

There were several (semi-) successful encounters of using the
hostmode, some via tricking the OTG statemachine, some via
implementing the FORCE_HOST testmode properly, we'll concentrate on
the latter

* Hardware
The detailed overview is at

** There're leaked schematics

** There're datasheets for SoC (spurf98d is the best version) and

** Comparision with N8x0
*** Same USB core
*** N8x0 uses a discrete component TUSB6010 that integrates musb core and phy
while N900 uses SoC-integrated musb with ISP1707 PHY via ULPI link
*** ID pin of the USB core on N8x0 can be forcibly grounded
programmatically, thus activating the host mode in the core, not
possible with ISP1707
*** ID pin is connected to the micro-usb receptable on N8x0, but some
stupid decided to not route it to ISP1707 on N900 (he routed it to
TWL4030 instead)

** The battery charger chip is controlled via i2c and can
supply 5V/200mA on usb vbus when asked to.
** twl4030 has connection to the ID pin and can sense it
OTG-like operation (at least automatic switching to/from hostmode)

* Software
** Driver
*** Shitty
*** There's a considerably improved version at linux-omap
(nobody's tried to use it on N900 so far)
*** Needs implementation of forced hostmode via one of the testmodes
See the patchsets in the repository, there're no incremental patches
accompanied by the meaningful logs (yet); when it starts working,
implement it in a separate file as a "platform" function for ``echo
host >/sys/devices/platform/musb_hdrc/mode'' to start working.
*** Needs (why?) disabling of some current limiting features
diff --git a/drivers/usb/core/generic.c b/drivers/usb/core/generic.c
index 7e912f2..b51e5ea 100644
--- a/drivers/usb/core/generic.c
+++ b/drivers/usb/core/generic.c
@@ -97,10 +97,10 @@ int usb_choose_configuration(struct usb_device *udev)
                /* Rule out configs that draw too much bus current */
-               if (c->desc.bMaxPower * 2 > udev->bus_mA) {
+               /*if (c->desc.bMaxPower * 2 > udev->bus_mA) {
-               }
+               }*/
                /* When the first config's first interface is one of Microsoft's
                 * pet nonstandard Ethernet-over-USB protocols, ignore it unless
@@ -132,10 +132,10 @@ int usb_choose_configuration(struct usb_device *udev)
                        best = c;
-       if (insufficient_power > 0)
+       /*if (insufficient_power > 0)
                dev_info(&udev->dev, "rejected %d configuration%s "
                        "due to insufficient available bus power\n",
-                       insufficient_power, plural(insufficient_power));
+                       insufficient_power, plural(insufficient_power));*/
        if (best) {
                i = best->desc.bConfigurationValue;

*** Needs (why?) disabling of some suspend features
diff --git a/drivers/usb/otg/twl4030-usb.c b/drivers/usb/otg/twl4030-usb.c
index ecffb2a..1577e7a 100644
--- a/drivers/usb/otg/twl4030-usb.c
+++ b/drivers/usb/otg/twl4030-usb.c
@@ -480,7 +480,7 @@ extern void musb_restore_ctx_and_resume(struct usb_gadget *gadget);
 static void twl4030_phy_suspend(struct twl4030_usb *twl, int controller_off)
-       if (twl->asleep)
+       if (!twl->asleep)
        twl4030_phy_power(twl, 0);

*** Needs (why?) manual step-through to get device recognised, enumerated, etc
echo H >/proc/driver/musb_hdrc
echo host >/sys/devices/platform/musb_hdrc/mode
echo e > /proc/driver/musb_hdrc
echo F > /proc/driver/musb_hdrc
echo host >/sys/devices/platform/musb_hdrc/mode
echo H >/proc/driver/musb_hdrc
echo I > /proc/driver/musb_hdrc
echo host >/sys/devices/platform/musb_hdrc/mode
echo H >/proc/driver/musb_hdrc
echo "Done echo'ing"
#echo "Connect the flash drive in 3 seconds!"
stop bme; sleep 3; i2cset -y -m 0x07 2 0x6b 0x01 0x05; while true; do sleep 28; i2cset -y -m 0x80 2 0x6b 0x00 0x80; done

*** How can it switch out of "waiting Vbus to rise" state?
*** No way to distinguish between full-speed and high-speed modes
by just looking at the D+, D- states (via debug register of isp1704),
needs special care

** BME
*** Messes with charging so should be disabled
stop bme
*** Incapable of providing VBUS to external device
One can use ``i2cset -y -m 0x07 2 0x6b 0x01 0x05'' instead to enable
the reverse boost converter and ``i2cset -y -m 0x80 2 0x6b 0x00 0x80''
to disable

** twl4030-usb.c
*** Gets in the way by switching states behind your back
Switches state to b_idle if the ID pin is not grounded), so should be
(at least partially) disabled.
*** Is registered as an OTG controller
Should we provide a minimal OTG driver for isp1407 instead? Or should
we rather get the whole stuff somehow working and then try to port the
upstream driver (it's supposed to be saner)?

** musb
*** SESSREQ (Session request) interrupt preconditions?
*** CONNECT interrupt preconditions?