[PATCH 2/2] ARM: dts: omap5: Add dr_mode for dwc3

Dmitry Lifshitz lifshitz at compulab.co.il
Mon Sep 8 05:32:37 PDT 2014

Hi Roger,

Thank you for a quick reply. See, my reply below.

On 09/08/2014 02:39 PM, Roger Quadros wrote:
> On 09/08/2014 02:34 PM, Dmitry Lifshitz wrote:
>> Hi Felipe, Roger
>> On 04/16/2014 07:16 PM, Felipe Balbi wrote:
>>> On Fri, Oct 11, 2013 at 05:46:12PM +0300, Roger Quadros wrote:
>>>> Hi,
>>>> On 10/10/2013 01:49 PM, Kishon Vijay Abraham I wrote:
>>>>> From: George Cherian <george.cherian at ti.com>
>>>>> Added dr_mode property in dwc3 and set its default mode to device.
>>>> If there is a specific reason why this is not set to "otg", we need
>>>> to explain it here.
>>>> AFAIK the port is meant to be used as OTG port.
>>> we still don't have the workaround for dual role ports, I've been
>>> working on it but still need to retest and make sure it's working fine
>>> before merging upstream.
>> Does the OTG port suppose to work with OMAP5? I mean role switching mechanism?
>> I'm testing kernel v3.16 with OMAP5 based board (CM-T54).
>> If I set CONFIG_USB_DWC3_DUAL_ROLE, than nothing is going on when I plug some device/host cable into micro USB3.0 port (VBUS/ID detection with extcon events works fine).
>> CONFIG_USB_DWC3_HOST/CONFIG_USB_DWC3_GADGET bring port to appropriate mode and it works.
>> What is the purpose of setting dr_mode = "peripheral" in omap5.dtsi, since it is overridden by config options in dwc3_probe()?
> If dual role works it should be "otg" then. I'm not sure why it was set to peripheral. Maybe the role switching mechanism
> was not in place earlier.

This is what I have when setting dr_mode = "otg" (and compiling with 

when load gadget driver and plugging host PC cable:

root at cm-debian:~# modprobe g_ether
using random self ethernet address
using random host ethernet address
usb0: HOST MAC fa:db:52:7f:95:c5
usb0: MAC be:75:8d:71:6d:02
using random self ethernet address
using random host ethernet address
g_ether gadget: Ethernet Gadget, version: Memorial Day 2008
g_ether gadget: g_ether ready
root at cm-debian:~# ------------[ cut here ]------------
WARNING: CPU: 0 PID: 65 at 
44000000.ocp:L3 Standard Error: MASTER UN IDENTIFIED TARGET OCMRAM 
(Read): At Address: 0x00000000 : Data Access in User mode during 
Functional access
Modules linked in: usb_f_eem g_ether usb_f_rndis libcomposite u_ether 
configfs rfcomm ecb btmrvl_sdio btmrvl bluetooth mwifiex_sdio mwifiex 
dwc3 ahci_platform libahci_platform libahci libata snd_soc_cm_t54 
ads7846 dwc3_omap extcon_palmas rtc_palmas extcon_class at24 
snd_soc_wm8731 snd_soc_omap_mcbsp
CPU: 0 PID: 65 Comm: irq/39-palmas Tainted: G        W 
3.16.0-cm-t54+ #99
[<c00156d0>] (unwind_backtrace) from [<c00124ac>] (show_stack+0x10/0x14)
[<c00124ac>] (show_stack) from [<c05e29dc>] (dump_stack+0x74/0x8c)
[<c05e29dc>] (dump_stack) from [<c0037dbc>] (warn_slowpath_common+0x70/0x88)
[<c0037dbc>] (warn_slowpath_common) from [<c0037e68>] 
[<c0037e68>] (warn_slowpath_fmt) from [<c02d88b8>] 
[<c02d88b8>] (l3_handle_target.isra.2) from [<c02d89b0>] 
[<c02d89b0>] (l3_interrupt_handler) from [<c007ea00>] 
[<c007ea00>] (handle_irq_event_percpu) from [<c007ebf8>] 
[<c007ebf8>] (handle_irq_event) from [<c0081b78>] 
[<c0081b78>] (handle_fasteoi_irq) from [<c007e398>] 
[<c007e398>] (generic_handle_irq) from [<c000f4ac>] (handle_IRQ+0x40/0x90)
[<c000f4ac>] (handle_IRQ) from [<c0008620>] (gic_handle_irq+0x2c/0x60)
[<c0008620>] (gic_handle_irq) from [<c05eacc0>] (__irq_svc+0x40/0x50)
Exception stack(0xec939d80 to 0xec939dc8)
9d80: ecb729b0 20000013 00000000 00000001 ecb72800 eca1e000 00000000 
9da0: ec939ed0 20000013 ecb72818 ed005e14 ec939ec8 ec939dc8 bf015f54 
9dc0: 80000013 ffffffff
[<c05eacc0>] (__irq_svc) from [<c05ea548>] 
[<c05ea548>] (_raw_spin_unlock_irqrestore) from [<bf015f54>] 
(extcon_update_state+0x144/0x1c4 [extcon_class])
[<bf015f54>] (extcon_update_state [extcon_class]) from [<bf023288>] 
(palmas_vbus_irq_handler+0x64/0xa0 [extcon_palmas])
[<bf023288>] (palmas_vbus_irq_handler [extcon_palmas]) from [<c00815a8>] 
[<c00815a8>] (handle_nested_irq) from [<c03ab85c>] 
[<c03ab85c>] (regmap_irq_thread) from [<c007f8b8>] (irq_thread_fn+0x1c/0x40)
[<c007f8b8>] (irq_thread_fn) from [<c007f6a4>] (irq_thread+0x120/0x15c)
[<c007f6a4>] (irq_thread) from [<c0056434>] (kthread+0xdc/0xf0)
[<c0056434>] (kthread) from [<c000eb98>] (ret_from_fork+0x14/0x3c)
---[ end trace 3b365a734f7f3202 ]---
palmas-usb 48070000.i2c:palmas at 48:palmas_usb: USB cable is attached

when plugging storage device:

root at cm-debian:~# dmesg | tail -n20
omap-dwc3 4a020000.omap_dwc3: ID GND
palmas-usb 48070000.i2c:palmas at 48:palmas_usb: USB-HOST cable is attached
palmas-usb 48070000.i2c:palmas at 48:palmas_usb: USB cable is attached

and nothing is going on.

Once again, when compiling with predefined role 
(CONFIG_USB_DWC3_HOST/CONFIG_USB_DWC3_GADGET) the port is functional in 
both modes.

I'm wondering, if dwc3_probe() calls
dwc3_core_init_mode() -> dwc3_set_mode() than why it is not required in 

> George any objections for "otg"?
> cheers,
> -roger
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo at vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



More information about the linux-arm-kernel mailing list