twl4030_usb porting advice

twl4030_usb porting advice

[Tomasz Mloduchowski]

Hi!

I'm trying to port the support patches for an OMAP3530 based development 
kit by Technexion.
Technexion provides patches for PSP kernel trees, I need to use a kernel 
with modern IR stack, so I'm developing against linux-omap-2.6.git


I'm mostly done with the initial port, however, I'm stuck with the 
following issue:

[    2.991973] ehci-omap ehci-omap.0: port 2 full speed --> companion
[    2.992034] ehci-omap ehci-omap.0: GetStatus port:2 status 003801 0 
ACK POWER OWNER sig=j CONNECT
[    2.992095] hub 1-0:1.0: port 2 not reset yet, waiting 50ms
[    3.055511] asoc: twl4030-hifi <-> omap-mcbsp-dai.1 mapping ok
[    3.061920] ehci-omap ehci-omap.0: GetStatus port:2 status 003002 0 
ACK POWER OWNER sig=se0 CSC
[    3.061981] hub 1-0:1.0: unable to enumerate USB device on port 2
[    3.068389] hub 1-0:1.0: state 7 ports 3 chg 0000 evt 0004


I would gladly welcome any pointers that could help me solve this issue.

Cheers,
Tomasz

Re: twl4030_usb porting advice

[Felipe Balbi]

On Wed, Dec 15, 2010 at 11:55:24AM +0000, Tomasz Mloduchowski wrote:
>Hi!
>
>I'm trying to port the support patches for an OMAP3530 based 
>development kit by Technexion.
>Technexion provides patches for PSP kernel trees, I need to use a 
>kernel with modern IR stack, so I'm developing against 
>linux-omap-2.6.git
>
>
>I'm mostly done with the initial port, however, I'm stuck with the 
>following issue:
>
>[    2.991973] ehci-omap ehci-omap.0: port 2 full speed --> companion
>[    2.992034] ehci-omap ehci-omap.0: GetStatus port:2 status 003801 0 
>ACK POWER OWNER sig=j CONNECT
>[    2.992095] hub 1-0:1.0: port 2 not reset yet, waiting 50ms
>[    3.055511] asoc: twl4030-hifi <-> omap-mcbsp-dai.1 mapping ok
>[    3.061920] ehci-omap ehci-omap.0: GetStatus port:2 status 003002 0 
>ACK POWER OWNER sig=se0 CSC
>[    3.061981] hub 1-0:1.0: unable to enumerate USB device on port 2
>[    3.068389] hub 1-0:1.0: state 7 ports 3 chg 0000 evt 0004
>
>
>I would gladly welcome any pointers that could help me solve this issue.

you're connecting a FS/LS device to EHCI port, that doesn't work. OMAP
can't handle port handoff to companion OHCI controller, then it fails
not being able to enumerate the device. Try remuxing your port to OHCI
controller and disabling EHCI from Kconfig.

You should be using ohci-omap3.c driver.

-- 
balbi

Re: twl4030_usb porting advice

[Tomasz Mloduchowski]

>> I would gladly welcome any pointers that could help me solve this issue.
>
> you're connecting a FS/LS device to EHCI port, that doesn't work. OMAP
> can't handle port handoff to companion OHCI controller, then it fails
> not being able to enumerate the device. Try remuxing your port to OHCI
> controller and disabling EHCI from Kconfig.

It actually is a HS device (it's a 4-port hub) - when booting the vendor 
supplied kernel (based off 2.6.32, i.e. pre -IR redesign), the device 
identifies as follows:

Bus 001 Device 002: ID 05e3:0608 Genesys Logic, Inc. USB-2.0 4-Port HUB

with the following boot messages:

usb 1-2: new high speed USB device using ehci-omap and address 2
usb 1-2: New USB device found, idVendor=05e3, idProduct=0608
usb 1-2: New USB device strings: Mfr=0, Product=1, SerialNumber=0
usb 1-2: Product: USB2.0 Hub
hub 1-2:1.0: USB hub found
hub 1-2:1.0: 4 ports detected


I've pasted the boot log for my kernel:
http://pastie.org/1379102

And the boot log for the original kernel:
http://pastie.org/1379100

Cheers,
Tomasz

Re: twl4030_usb porting advice

[Felipe Balbi]

On Wed, Dec 15, 2010 at 12:27:10PM +0000, Tomasz Mloduchowski wrote:
>>>I would gladly welcome any pointers that could help me solve this issue.
>>
>>you're connecting a FS/LS device to EHCI port, that doesn't work. OMAP
>>can't handle port handoff to companion OHCI controller, then it fails
>>not being able to enumerate the device. Try remuxing your port to OHCI
>>controller and disabling EHCI from Kconfig.
>
>It actually is a HS device (it's a 4-port hub) - when booting the 
>vendor supplied kernel (based off 2.6.32, i.e. pre -IR redesign), the 
>device identifies as follows:
>
>Bus 001 Device 002: ID 05e3:0608 Genesys Logic, Inc. USB-2.0 4-Port HUB
>
>with the following boot messages:
>
>usb 1-2: new high speed USB device using ehci-omap and address 2
>usb 1-2: New USB device found, idVendor=05e3, idProduct=0608
>usb 1-2: New USB device strings: Mfr=0, Product=1, SerialNumber=0
>usb 1-2: Product: USB2.0 Hub
>hub 1-2:1.0: USB hub found
>hub 1-2:1.0: 4 ports detected
>
>
>I've pasted the boot log for my kernel:
>http://pastie.org/1379102
>
>And the boot log for the original kernel:
>http://pastie.org/1379100

It could be a bug on the reset sequence on ehci-omap.c, do you have
anything connected to any of the hub ports ?

Please enable CONFIG_USB_DEBUG on your kernel config and grab a bit more
information.

Also, boot the machine without the device attached and after booting
run:

# dmesg -c

then attach the device, then grab the output of dmesg.

-- 
balbi

Re: twl4030_usb porting advice

[Tomasz Mloduchowski]

On 12/15/10 12:39, Felipe Balbi wrote:
> On Wed, Dec 15, 2010 at 12:27:10PM +0000, Tomasz Mloduchowski wrote:
>>>> I would gladly welcome any pointers that could help me solve this
>>>> issue.
>>>
>>> you're connecting a FS/LS device to EHCI port, that doesn't work. OMAP
>>> can't handle port handoff to companion OHCI controller, then it fails
>>> not being able to enumerate the device. Try remuxing your port to OHCI
>>> controller and disabling EHCI from Kconfig.
>>
>> It actually is a HS device (it's a 4-port hub) - when booting the
>> vendor supplied kernel (based off 2.6.32, i.e. pre -IR redesign), the
>> device identifies as follows:
>>
>> Bus 001 Device 002: ID 05e3:0608 Genesys Logic, Inc. USB-2.0 4-Port HUB
>>
>> with the following boot messages:
>>
>> usb 1-2: new high speed USB device using ehci-omap and address 2
>> usb 1-2: New USB device found, idVendor=05e3, idProduct=0608
>> usb 1-2: New USB device strings: Mfr=0, Product=1, SerialNumber=0
>> usb 1-2: Product: USB2.0 Hub
>> hub 1-2:1.0: USB hub found
>> hub 1-2:1.0: 4 ports detected
>>
>>
>> I've pasted the boot log for my kernel:
>> http://pastie.org/1379102
>>
>> And the boot log for the original kernel:
>> http://pastie.org/1379100
>
> It could be a bug on the reset sequence on ehci-omap.c, do you have
> anything connected to any of the hub ports ?
>
> Please enable CONFIG_USB_DEBUG on your kernel config and grab a bit more
> information.
>
> Also, boot the machine without the device attached and after booting
> run:
>
> # dmesg -c
>
> then attach the device, then grab the output of dmesg.
>

I don't have the ability to disconnect the hub - it's integrated on the 
board :(

I've also made a mistake describing the pastebins - the working, PSP 
kernel boot log is here:
http://pastie.org/1379102

while the non-working git kernel messages are here (with CONFIG_USB_DEBUG):
http://pastie.org/1379100

Cheers,
Tomasz

Re: twl4030_usb porting advice

[Felipe Balbi]

On Wed, Dec 15, 2010 at 01:09:42PM +0000, Tomasz Mloduchowski wrote:
>On 12/15/10 12:39, Felipe Balbi wrote:
>>On Wed, Dec 15, 2010 at 12:27:10PM +0000, Tomasz Mloduchowski wrote:
>>>>>I would gladly welcome any pointers that could help me solve this
>>>>>issue.
>>>>
>>>>you're connecting a FS/LS device to EHCI port, that doesn't work. OMAP
>>>>can't handle port handoff to companion OHCI controller, then it fails
>>>>not being able to enumerate the device. Try remuxing your port to OHCI
>>>>controller and disabling EHCI from Kconfig.
>>>
>>>It actually is a HS device (it's a 4-port hub) - when booting the
>>>vendor supplied kernel (based off 2.6.32, i.e. pre -IR redesign), the
>>>device identifies as follows:
>>>
>>>Bus 001 Device 002: ID 05e3:0608 Genesys Logic, Inc. USB-2.0 4-Port HUB
>>>
>>>with the following boot messages:
>>>
>>>usb 1-2: new high speed USB device using ehci-omap and address 2
>>>usb 1-2: New USB device found, idVendor=05e3, idProduct=0608
>>>usb 1-2: New USB device strings: Mfr=0, Product=1, SerialNumber=0
>>>usb 1-2: Product: USB2.0 Hub
>>>hub 1-2:1.0: USB hub found
>>>hub 1-2:1.0: 4 ports detected
>>>
>>>
>>>I've pasted the boot log for my kernel:
>>>http://pastie.org/1379102
>>>
>>>And the boot log for the original kernel:
>>>http://pastie.org/1379100
>>
>>It could be a bug on the reset sequence on ehci-omap.c, do you have
>>anything connected to any of the hub ports ?
>>
>>Please enable CONFIG_USB_DEBUG on your kernel config and grab a bit more
>>information.
>>
>>Also, boot the machine without the device attached and after booting
>>run:
>>
>># dmesg -c
>>
>>then attach the device, then grab the output of dmesg.
>>
>
>I don't have the ability to disconnect the hub - it's integrated on 
>the board :(

I see.

>I've also made a mistake describing the pastebins - the working, PSP 
>kernel boot log is here:
>http://pastie.org/1379102
>
>while the non-working git kernel messages are here (with CONFIG_USB_DEBUG):
>http://pastie.org/1379100

I'm quoting a snippet below for commenting.

> [    1.639892] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
> [    1.646759] ehci_hcd: block sizes: qh 60 qtd 96 itd 160 sitd 96
> [    1.646881] ehci-omap ehci-omap.0: failed to get ehci port1 regulator

This is alarming. Please add this patch to get the error code:

diff --git a/drivers/usb/host/ehci-omap.c b/drivers/usb/host/ehci-omap.c
index 116ae28..5da3fa3 100644
--- a/drivers/usb/host/ehci-omap.c
+++ b/drivers/usb/host/ehci-omap.c
@@ -682,7 +682,8 @@ static int ehci_hcd_omap_probe(struct platform_device *pdev)
                 if (IS_ERR(omap->regulator[i])) {
                         omap->regulator[i] = NULL;
                         dev_dbg(&pdev->dev,
-                       "failed to get ehci port%d regulator\n", i);
+                       "failed to get ehci port%d regulator --> %d\n", i,
+                       PTR_ERR(omap->regulator[i]));
                 } else {
                         regulator_enable(omap->regulator[i]);
                 }

> [    1.646911] ehci-omap ehci-omap.0: starting TI EHCI USB Controller
> [    1.647064] ehci-omap ehci-omap.0: OMAP UHH_REVISION 0x10
> [    1.647094] ehci-omap ehci-omap.0: TLL RESET DONE
> [    1.647094] ehci-omap ehci-omap.0: OMAP3 ES version > ES2.1
> [    1.647125] ehci-omap ehci-omap.0: UHH setup done, uhh_hostconfig=21c
> [    2.649841] ehci-omap ehci-omap.0: phy reset operation timed out

This is also alarming... Let's increase the timeout just for a debugging
purpose:

diff --git a/drivers/usb/host/ehci-omap.c b/drivers/usb/host/ehci-omap.c
index 116ae28..fa8bff0 100644
--- a/drivers/usb/host/ehci-omap.c
+++ b/drivers/usb/host/ehci-omap.c
@@ -239,7 +239,7 @@ static void omap_usb_utmi_init(struct ehci_hcd_omap *omap, u8 tll_channel_mask)
  
  static void omap_ehci_soft_phy_reset(struct ehci_hcd_omap *omap, u8 port)
  {
-       unsigned long timeout = jiffies + msecs_to_jiffies(1000);
+       unsigned long timeout = jiffies + msecs_to_jiffies(10000);
         unsigned reg = 0;
  
         reg = ULPI_FUNC_CTRL_RESET

beware, this might slow your boot time quite a lot :-p

The rest is just continuing the enumeration and there's not much
interesting there besides the "unable to enumerate" message, which could
be caused by the two problems above.

-- 
balbi

Re: twl4030_usb porting advice

[Tomasz Mloduchowski]

On 12/16/10 09:09, Felipe Balbi wrote:
> On Wed, Dec 15, 2010 at 01:09:42PM +0000, Tomasz Mloduchowski wrote:
>> On 12/15/10 12:39, Felipe Balbi wrote:
>>> On Wed, Dec 15, 2010 at 12:27:10PM +0000, Tomasz Mloduchowski wrote:
>>>>>> I would gladly welcome any pointers that could help me solve this
>>>>>> issue.
>>>>>
>>>>> you're connecting a FS/LS device to EHCI port, that doesn't work. OMAP
>>>>> can't handle port handoff to companion OHCI controller, then it fails
>>>>> not being able to enumerate the device. Try remuxing your port to OHCI
>>>>> controller and disabling EHCI from Kconfig.
>>>>
>>>> It actually is a HS device (it's a 4-port hub) - when booting the
>>>> vendor supplied kernel (based off 2.6.32, i.e. pre -IR redesign), the
>>>> device identifies as follows:
>> I've also made a mistake describing the pastebins - the working, PSP
>> kernel boot log is here:
>> http://pastie.org/1379102
>>
>> while the non-working git kernel messages are here (with
>> CONFIG_USB_DEBUG):
>> http://pastie.org/1379100
>
> I'm quoting a snippet below for commenting.
>
>> [ 1.639892] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
>> [ 1.646759] ehci_hcd: block sizes: qh 60 qtd 96 itd 160 sitd 96
>> [ 1.646881] ehci-omap ehci-omap.0: failed to get ehci port1 regulator
>
> This is alarming. Please add this patch to get the error code:
 >
> This is also alarming... Let's increase the timeout just for a debugging
> purpose:
I've applied both patches. The new kernel log in its entirety resides here:
http://pastie.org/1384896

And the line of interest is
[    2.085449] ehci-omap ehci-omap.0: failed to get ehci port1 regulator 
--> 0

Unfortunately, I don't understand what an error code of '0' might mean 
here, but it surely does something.

Cheers,
Tomasz

Re: twl4030_usb porting advice

[Felipe Balbi]

Hi,

On Fri, Dec 17, 2010 at 07:23:43AM +0000, Tomasz Mloduchowski wrote:
>>I'm quoting a snippet below for commenting.
>>
>>>[ 1.639892] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
>>>[ 1.646759] ehci_hcd: block sizes: qh 60 qtd 96 itd 160 sitd 96
>>>[ 1.646881] ehci-omap ehci-omap.0: failed to get ehci port1 regulator
>>
>>This is alarming. Please add this patch to get the error code:
>>
>>This is also alarming... Let's increase the timeout just for a debugging
>>purpose:
>I've applied both patches. The new kernel log in its entirety resides here:
>http://pastie.org/1384896

Unfortunately seems like pastie.org is down:

503 Service Unavailable
No server is available to handle this request.

>And the line of interest is
>[    2.085449] ehci-omap ehci-omap.0: failed to get ehci port1 
>regulator --> 0
>
>Unfortunately, I don't understand what an error code of '0' might mean 
>here, but it surely does something.

That's funny, 0 means "success", not sure why that branch is being
called.

I'll wait a bit until pastie is back, but if you could append to your
reply the output of dmesg | grep -i ehci I'd be glad.

-- 
balbi

Re: twl4030_usb porting advice

[Tomasz Mloduchowski]

On 12/17/10 10:11, Felipe Balbi wrote:
> Hi,
>
> On Fri, Dec 17, 2010 at 07:23:43AM +0000, Tomasz Mloduchowski wrote:
>>> I'm quoting a snippet below for commenting.
>>>
>>>> [ 1.639892] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
>>>> [ 1.646759] ehci_hcd: block sizes: qh 60 qtd 96 itd 160 sitd 96
>>>> [ 1.646881] ehci-omap ehci-omap.0: failed to get ehci port1 regulator
>>>
>>> This is alarming. Please add this patch to get the error code:
>>>
>>> This is also alarming... Let's increase the timeout just for a debugging
>>> purpose:
>> I've applied both patches. The new kernel log in its entirety resides
>> here:
>> http://pastie.org/1384896
>
> Unfortunately seems like pastie.org is down:
>
> 503 Service Unavailable
> No server is available to handle this request.
>
>> And the line of interest is
>> [ 2.085449] ehci-omap ehci-omap.0: failed to get ehci port1 regulator
>> --> 0
>>
>> Unfortunately, I don't understand what an error code of '0' might mean
>> here, but it surely does something.
>
> That's funny, 0 means "success", not sure why that branch is being
> called.
>
> I'll wait a bit until pastie is back, but if you could append to your
> reply the output of dmesg | grep -i ehci I'd be glad.
>
I really appreciate your help, this sort of a bug that really stunned me.

I've noticed that the variable storing the error code is set to NULL 
right before printing it out :(, so I corrected this mistake.

[    2.065216] ehci-omap ehci-omap.0: failed to get ehci port1 regulator 
--> -19

I also switched to pastebin, with the new, more meaningful dmesg:
http://pastebin.com/86kfLT2G

Cheers,
Tomasz

Re: twl4030_usb porting advice

[Felipe Balbi]

On Fri, Dec 17, 2010 at 11:14:50AM +0000, Tomasz Mloduchowski wrote:
>I really appreciate your help, this sort of a bug that really stunned me.
>
>I've noticed that the variable storing the error code is set to NULL 
>right before printing it out :(, so I corrected this mistake.
>
>[    2.065216] ehci-omap ehci-omap.0: failed to get ehci port1 
>regulator --> -19

Thanks for that :-)

Error -19 is -ENODEV, so it means there really isn't a regulator.

Maybe you can more debug to ehci_reset() function ?!?

>I also switched to pastebin, with the new, more meaningful dmesg:
>http://pastebin.com/86kfLT2G

I'll look at it and get back to you. But try to figure out exactly where
ehci decides to bounce the port to companion ohci and start adding
printks or dev_info to that. A good start is ehci_reset(), something
like below might help:

diff --git a/drivers/usb/host/ehci-dbg.c b/drivers/usb/host/ehci-dbg.c
index 6e25996..7fdaecf 100644
--- a/drivers/usb/host/ehci-dbg.c
+++ b/drivers/usb/host/ehci-dbg.c
@@ -330,7 +330,7 @@ dbg_port_buf (char *buf, unsigned len, const char *label, int port, u32 status)
  #define dbg_cmd(ehci, label, command) { \
         char _buf [80]; \
         dbg_command_buf (_buf, sizeof _buf, label, command); \
-       ehci_dbg (ehci, "%s\n", _buf); \
+       ehci_info (ehci, "%s\n", _buf); \
  }
  
  #define dbg_port(ehci, label, port, status) { \

a lot already.

-- 
balbi