usb clock not found on imx27 when using dt

usb clock not found on imx27 when using dt

[Philippe Reynes]

Hi all,

I try to add the support of usb on my armadeus apf27,
but there is a problem when try to reclaim the clock.
imx_usb mxc-ehci.0: Failed to get clock, err=-2

To avoid adding too many attachment to this mail,
I've put all the information on a website. You can
found them here:? http://ryu.zarb.org/~trem/armadeus/


The initial issue is in the directory : 20121110

boot-3.7-rc4.log : boot log with the issue
linux-3.7-rc4.config : the configuration file for the kernel
investifation.txt : the description of my investigation.


I've tried to fix the clock issue, but another issue occurs.
The information about this "fix" is in : 20121111
boot-3.7-rc4.log : boot log with the new issue
issue.txt : the description of the change done on the kernel

I've got severals questions :
- does my dts correct with usb change ?
- does someone has already tested usb on imx27 with dt ?

Regards,
Philippe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20121111/3bbd3fdc/attachment-0001.html>

usb clock not found on imx27 when using dt

[s.hauer at pengutronix.de]

On Sun, Nov 11, 2012 at 08:34:34PM +0000, Philippe Reynes wrote:
> Hi all,
> 
> I try to add the support of usb on my armadeus apf27,
> but there is a problem when try to reclaim the clock.
> imx_usb mxc-ehci.0: Failed to get clock, err=-2
> 
> To avoid adding too many attachment to this mail,
> I've put all the information on a website. You can
> found them here:? http://ryu.zarb.org/~trem/armadeus/
> 
> 
> The initial issue is in the directory : 20121110
> 
> boot-3.7-rc4.log : boot log with the issue
> linux-3.7-rc4.config : the configuration file for the kernel
> investifation.txt : the description of my investigation.

You should really post the relevant bits on the mailing list. Nobody
visits websites for investigating your issue. Besides, it will not be
in the mailing list archives so people will not find it later.

> 
> 
> I've tried to fix the clock issue, but another issue occurs.
> The information about this "fix" is in : 20121111
> boot-3.7-rc4.log : boot log with the new issue
> issue.txt : the description of the change done on the kernel
> 
> I've got severals questions :
> - does my dts correct with usb change ?
> - does someone has already tested usb on imx27 with dt ?

Please investigate in using the chipidea driver. This will be the way
forward for i.MX.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

usb clock not found on imx27 when using dt

[trem]

On 12/11/12 08:54, s.hauer at pengutronix.de wrote:
> On Sun, Nov 11, 2012 at 08:34:34PM +0000, Philippe Reynes wrote:
>> Hi all,
>>
>> I try to add the support of usb on my armadeus apf27,
>> but there is a problem when try to reclaim the clock.
>> imx_usb mxc-ehci.0: Failed to get clock, err=-2
>>
>> To avoid adding too many attachment to this mail,
>> I've put all the information on a website. You can
>> found them here:  http://ryu.zarb.org/~trem/armadeus/
>>
>>
>> The initial issue is in the directory : 20121110
>>
>> boot-3.7-rc4.log : boot log with the issue
>> linux-3.7-rc4.config : the configuration file for the kernel
>> investifation.txt : the description of my investigation.
>
> You should really post the relevant bits on the mailing list. Nobody
> visits websites for investigating your issue. Besides, it will not be
> in the mailing list archives so people will not find it later.

You're right, I add an archive with all log as attachment.

>>
>>
>> I've tried to fix the clock issue, but another issue occurs.
>> The information about this "fix" is in : 20121111
>> boot-3.7-rc4.log : boot log with the new issue
>> issue.txt : the description of the change done on the kernel
>>
>> I've got severals questions :
>> - does my dts correct with usb change ?
>> - does someone has already tested usb on imx27 with dt ?
>
> Please investigate in using the chipidea driver. This will be the way
> forward for i.MX.

I use the chipidea driver for the usb.

The main issue is that the chipidea driver don't found the clock :

[    0.761693] imx_usb mxc-ehci.0: Failed to get clock, err=-2
[    0.767620] imx_usb: probe of mxc-ehci.0 failed with error -2

The driver ci13xxx_imx request the clock in the probe function
in the following line:
data->clk = devm_clk_get(&pdev->dev, NULL);
calling: clk = clk_get(dev, id);
calling: clk_get_sys(dev_id, con_id);
calling: clk_find(dev_id, con_id);
with dev_id = mxc-ehci.0 and con_id = NULL
This last function fail to found a clock.

In the function mx27_clocks_init, several mxc-ehci.0
clocks are registered, but all with a con_id.

So the function clk_find fails.

As I'm not really sure of the dts, I don't know if the problem
is in the dts or in the driver (maybe both).


Regards,
Philippe

-------------- next part --------------
A non-text attachment was scrubbed...
Name: usb_imx27_dt.tar.bz2
Type: application/x-bzip
Size: 24021 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20121112/2d224f44/attachment-0001.bin>

usb clock not found on imx27 when using dt

[Marek Vasut]

Dear trem,

> On 12/11/12 08:54, s.hauer at pengutronix.de wrote:
> > On Sun, Nov 11, 2012 at 08:34:34PM +0000, Philippe Reynes wrote:
> >> Hi all,
> >> 
> >> I try to add the support of usb on my armadeus apf27,
> >> but there is a problem when try to reclaim the clock.
> >> imx_usb mxc-ehci.0: Failed to get clock, err=-2
> >> 
> >> To avoid adding too many attachment to this mail,
> >> I've put all the information on a website. You can
> >> found them here:  http://ryu.zarb.org/~trem/armadeus/
> >> 
> >> 
> >> The initial issue is in the directory : 20121110
> >> 
> >> boot-3.7-rc4.log : boot log with the issue
> >> linux-3.7-rc4.config : the configuration file for the kernel
> >> investifation.txt : the description of my investigation.
> > 
> > You should really post the relevant bits on the mailing list. Nobody
> > visits websites for investigating your issue. Besides, it will not be
> > in the mailing list archives so people will not find it later.
> 
> You're right, I add an archive with all log as attachment.
> 
> >> I've tried to fix the clock issue, but another issue occurs.
> >> The information about this "fix" is in : 20121111
> >> boot-3.7-rc4.log : boot log with the new issue
> >> issue.txt : the description of the change done on the kernel
> >> 
> >> I've got severals questions :
> >> - does my dts correct with usb change ?
> >> - does someone has already tested usb on imx27 with dt ?
> > 
> > Please investigate in using the chipidea driver. This will be the way
> > forward for i.MX.
> 
> I use the chipidea driver for the usb.
> 
> The main issue is that the chipidea driver don't found the clock :
> 
> [    0.761693] imx_usb mxc-ehci.0: Failed to get clock, err=-2
> [    0.767620] imx_usb: probe of mxc-ehci.0 failed with error -2
> 
> The driver ci13xxx_imx request the clock in the probe function
> in the following line:
> data->clk = devm_clk_get(&pdev->dev, NULL);
> calling: clk = clk_get(dev, id);
> calling: clk_get_sys(dev_id, con_id);
> calling: clk_find(dev_id, con_id);
> with dev_id = mxc-ehci.0 and con_id = NULL
> This last function fail to found a clock.
> 
> In the function mx27_clocks_init, several mxc-ehci.0
> clocks are registered, but all with a con_id.
> 
> So the function clk_find fails.
> 
> As I'm not really sure of the dts, I don't know if the problem
> is in the dts or in the driver (maybe both).

Please pull the relevant parts from the attachment and post them inline.

Best regards,
Marek Vasut

usb clock not found on imx27 when using dt

[s.hauer at pengutronix.de]

On Mon, Nov 12, 2012 at 09:17:42PM +0100, trem wrote:
> On 12/11/12 08:54, s.hauer at pengutronix.de wrote:
> >On Sun, Nov 11, 2012 at 08:34:34PM +0000, Philippe Reynes wrote:
> >>Hi all,
> >>
> >>I try to add the support of usb on my armadeus apf27,
> >>but there is a problem when try to reclaim the clock.
> >>imx_usb mxc-ehci.0: Failed to get clock, err=-2
> >>
> >>To avoid adding too many attachment to this mail,
> >>I've put all the information on a website. You can
> >>found them here:  http://ryu.zarb.org/~trem/armadeus/
> >>
> >>
> >>The initial issue is in the directory : 20121110
> >>
> >>boot-3.7-rc4.log : boot log with the issue
> >>linux-3.7-rc4.config : the configuration file for the kernel
> >>investifation.txt : the description of my investigation.
> >
> >You should really post the relevant bits on the mailing list. Nobody
> >visits websites for investigating your issue. Besides, it will not be
> >in the mailing list archives so people will not find it later.
> 
> You're right, I add an archive with all log as attachment.
> 
> >>
> >>
> >>I've tried to fix the clock issue, but another issue occurs.
> >>The information about this "fix" is in : 20121111
> >>boot-3.7-rc4.log : boot log with the new issue
> >>issue.txt : the description of the change done on the kernel
> >>
> >>I've got severals questions :
> >>- does my dts correct with usb change ?
> >>- does someone has already tested usb on imx27 with dt ?
> >
> >Please investigate in using the chipidea driver. This will be the way
> >forward for i.MX.
> 
> I use the chipidea driver for the usb.
> 
> The main issue is that the chipidea driver don't found the clock :
> 
> [    0.761693] imx_usb mxc-ehci.0: Failed to get clock, err=-2
> [    0.767620] imx_usb: probe of mxc-ehci.0 failed with error -2
> 
> The driver ci13xxx_imx request the clock in the probe function
> in the following line:
> data->clk = devm_clk_get(&pdev->dev, NULL);
> calling: clk = clk_get(dev, id);
> calling: clk_get_sys(dev_id, con_id);
> calling: clk_find(dev_id, con_id);
> with dev_id = mxc-ehci.0 and con_id = NULL
> This last function fail to found a clock.
> 
> In the function mx27_clocks_init, several mxc-ehci.0
> clocks are registered, but all with a con_id.
> 
> So the function clk_find fails.
> 
> As I'm not really sure of the dts, I don't know if the problem
> is in the dts or in the driver (maybe both).

You have to register the lookups for the chipidea driver with
clk_register_clkdev(). The lookups are only what the name suggests:
lookups. They are no clocks by themselves. This means that you have to
add another lookup which matches the device name.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

usb clock not found on imx27 when using dt

[trem]

On 13/11/12 08:38, s.hauer at pengutronix.de wrote:
>>> Please investigate in using the chipidea driver. This will be the way
>>> forward for i.MX.
>>
>> I use the chipidea driver for the usb.
>>
>> The main issue is that the chipidea driver don't found the clock :
>>
>> [    0.761693] imx_usb mxc-ehci.0: Failed to get clock, err=-2
>> [    0.767620] imx_usb: probe of mxc-ehci.0 failed with error -2
>>
>> The driver ci13xxx_imx request the clock in the probe function
>> in the following line:
>> data->clk = devm_clk_get(&pdev->dev, NULL);
>> calling: clk = clk_get(dev, id);
>> calling: clk_get_sys(dev_id, con_id);
>> calling: clk_find(dev_id, con_id);
>> with dev_id = mxc-ehci.0 and con_id = NULL
>> This last function fail to found a clock.
>>
>> In the function mx27_clocks_init, several mxc-ehci.0
>> clocks are registered, but all with a con_id.
>>
>> So the function clk_find fails.
>>
>> As I'm not really sure of the dts, I don't know if the problem
>> is in the dts or in the driver (maybe both).
>
> You have to register the lookups for the chipidea driver with
> clk_register_clkdev(). The lookups are only what the name suggests:
> lookups. They are no clocks by themselves. This means that you have to
> add another lookup which matches the device name.

Thanks for this help, I've added three "clocks" for the chipidea driver :
clk_register_clkdev(clk[usb_div], NULL, "mxc-ehci.0");

And I don't have the missing clock issue, but a kernel panic :

[    0.679565] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[    0.687807] usbcore: registered new interface driver cdc_acm
[    0.693503] cdc_acm: USB Abstract Control Model driver for USB modems and ISDN adapters
[    0.702506] usbcore: registered new interface driver usblp
[    0.708292] Initializing USB Mass Storage driver...
[    0.713885] usbcore: registered new interface driver usb-storage
[    0.720190] USB Mass Storage support registered.
[    0.726361] usbcore: registered new interface driver usbserial
[    0.732984] usbcore: registered new interface driver ftdi_sio
[    0.739510] usbserial: USB Serial support registered for FTDI USB Serial Device
[    0.747708] usbcore: registered new interface driver pl2303
[    0.754044] usbserial: USB Serial support registered for pl2303
[    0.762611] Unhandled fault: external abort on non-linefetch (0x008) at 0xf4424100
[    0.770227] Internal error: : 8 [#1] PREEMPT ARM
[    0.774861] Modules linked in:
[    0.777950] CPU: 0    Not tainted  (3.7.0-rc5-ge917a67 #2)
[    0.783469] PC is at ci_hdrc_probe+0xcc/0x430
[    0.787853] LR is at devm_kzalloc+0x4c/0x64
[    0.792058] pc : [<c03da658>]    lr : [<c023ffa8>]    psr: 60000013
[    0.792058] sp : d3833d98  ip : 40000013  fp : 00000000
[    0.803550] r10: c054db14  r9 : 00000000  r8 : f4424100
[    0.808788] r7 : d398c4e0  r6 : d39b1180  r5 : d39b1190  r4 : d3901010
[    0.815329] r3 : 00000100  r2 : 00000000  r1 : d3901010  r0 : d3901010
[    0.821871] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
[    0.829194] Control: 0005317f  Table: a0004000  DAC: 00000017
[    0.834954] Process swapper (pid: 1, stack limit = 0xd38321b8)
[    0.840799] Stack: (0xd3833d98 to 0xd3834000)
[    0.845174] 3d80:                                                       d39b1198 00000000
[    0.853384] 3da0: d39b1190 c055ea90 d39b1190 c0545010 c055ea6c c023e574 d39b1190 c023d288
[    0.861592] 3dc0: d39b1190 d3833dd8 c023d4a4 00000000 c055ea6c c023bb68 d380755c d39cdb74
[    0.869798] 3de0: d39cda50 d39b11c4 d39b1190 d39b1190 00000000 c023d588 d39b1190 c05367b0
[    0.878004] 3e00: d39b1190 c023c6a8 d39b1190 d38c99b0 d39b1198 c023b4d0 00000021 00000000
[    0.886210] 3e20: 00000002 c01e9aac d39b1180 d39b1190 d3832000 d39b1180 d39b1190 00000002
[    0.894416] 3e40: d38573a0 00000002 d3832000 c054db14 00000000 c023eb6c d39b1180 00000000
[    0.902622] 3e60: 00000000 d38573a0 00000002 c02b4cb0 ffffffff 00000000 d398c530 d38c99b0
[    0.910828] 3e80: d38c99a0 d398c530 00000000 00000000 ffffffff c03dac80 d38c99b0 c055ea90
[    0.919035] 3ea0: d38c99b0 c0545184 c050a32c 0000007a c051c8b8 c023e574 d38c99b0 c023d288
[    0.927241] 3ec0: d38c99b0 c0545184 d38c99e4 00000000 c050a32c c023d4a0 c0545184 d3833ee8
[    0.935448] 3ee0: c023d414 c023bbf4 d380754c d38c8d50 c0545184 c0545184 c05367b0 d39cda80
[    0.943654] 3f00: 00000000 c023c380 c04a6948 d3832000 c0545184 d3832000 c0545184 c054dc00
[    0.951860] 3f20: 00000000 c023da70 d3832000 00000007 c054dc00 00000000 c050a32c c00087ec
[    0.960065] 3f40: c04a70ec c04db800 00000006 00000006 00000000 c05147fc 00000006 c05147fc
[    0.968269] 3f60: 00000007 c054dc00 c0514808 c04f71ac 0000007a c03d7404 00000006 00000006
[    0.976475] 3f80: c04f71ac 01234567 d3832000 00000000 c03d7314 00000000 00000000 00000000
[    0.984678] 3fa0: 00000000 00000000 00000000 c0014730 00000000 00000000 00000000 00000000
[    0.992881] 3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    1.001084] 3fe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00008000
[    1.009335] [<c03da658>] (ci_hdrc_probe+0xcc/0x430) from [<c023e574>] (platform_drv_probe+0x14/0x18)
[    1.018523] [<c023e574>] (platform_drv_probe+0x14/0x18) from [<c023d288>] (driver_probe_device+0x80/0x20c)
[    1.028221] [<c023d288>] (driver_probe_device+0x80/0x20c) from [<c023bb68>] (bus_for_each_drv+0x64/0x8c)
[    1.037746] [<c023bb68>] (bus_for_each_drv+0x64/0x8c) from [<c023d588>] (device_attach+0x80/0x98)
[    1.046660] [<c023d588>] (device_attach+0x80/0x98) from [<c023c6a8>] (bus_probe_device+0x84/0xac)
[    1.055570] [<c023c6a8>] (bus_probe_device+0x84/0xac) from [<c023b4d0>] (device_add+0x534/0x604)
[    1.064392] [<c023b4d0>] (device_add+0x534/0x604) from [<c023eb6c>] (platform_device_add+0x134/0x20c)
[    1.073667] [<c023eb6c>] (platform_device_add+0x134/0x20c) from [<c02b4cb0>] (ci13xxx_add_device+0xd8/0x110)
[    1.083543] [<c02b4cb0>] (ci13xxx_add_device+0xd8/0x110) from [<c03dac80>] (ci13xxx_imx_probe+0x250/0x304)
[    1.093246] [<c03dac80>] (ci13xxx_imx_probe+0x250/0x304) from [<c023e574>] (platform_drv_probe+0x14/0x18)
[    1.102860] [<c023e574>] (platform_drv_probe+0x14/0x18) from [<c023d288>] (driver_probe_device+0x80/0x20c)
[    1.112557] [<c023d288>] (driver_probe_device+0x80/0x20c) from [<c023d4a0>] (__driver_attach+0x8c/0x90)
[    1.121991] [<c023d4a0>] (__driver_attach+0x8c/0x90) from [<c023bbf4>] (bus_for_each_dev+0x64/0x8c)
[    1.131077] [<c023bbf4>] (bus_for_each_dev+0x64/0x8c) from [<c023c380>] (bus_add_driver+0xa0/0x22c)
[    1.140167] [<c023c380>] (bus_add_driver+0xa0/0x22c) from [<c023da70>] (driver_register+0x78/0x140)
[    1.149253] [<c023da70>] (driver_register+0x78/0x140) from [<c00087ec>] (do_one_initcall+0x30/0x178)
[    1.158438] [<c00087ec>] (do_one_initcall+0x30/0x178) from [<c03d7404>] (kernel_init+0xf0/0x294)
[    1.167277] [<c03d7404>] (kernel_init+0xf0/0x294) from [<c0014730>] (ret_from_fork+0x14/0x24)
[    1.175833] Code: e584800c e5933004 e0888003 e5848010 (e5d83000)
[    1.182012] ---[ end trace f1f4ba6255ab3146 ]---
[    1.187015] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b


Regards,
Philippe

usb clock not found on imx27 when using dt

[s.hauer at pengutronix.de]

On Tue, Nov 13, 2012 at 09:43:20PM +0100, trem wrote:
> On 13/11/12 08:38, s.hauer at pengutronix.de wrote:
> >>>Please investigate in using the chipidea driver. This will be the way
> >>>forward for i.MX.
> >>
> >>I use the chipidea driver for the usb.
> >>
> >>The main issue is that the chipidea driver don't found the clock :
> >>
> >>[    0.761693] imx_usb mxc-ehci.0: Failed to get clock, err=-2
> >>[    0.767620] imx_usb: probe of mxc-ehci.0 failed with error -2
> >>
> >>The driver ci13xxx_imx request the clock in the probe function
> >>in the following line:
> >>data->clk = devm_clk_get(&pdev->dev, NULL);
> >>calling: clk = clk_get(dev, id);
> >>calling: clk_get_sys(dev_id, con_id);
> >>calling: clk_find(dev_id, con_id);
> >>with dev_id = mxc-ehci.0 and con_id = NULL
> >>This last function fail to found a clock.
> >>
> >>In the function mx27_clocks_init, several mxc-ehci.0
> >>clocks are registered, but all with a con_id.
> >>
> >>So the function clk_find fails.
> >>
> >>As I'm not really sure of the dts, I don't know if the problem
> >>is in the dts or in the driver (maybe both).
> >
> >You have to register the lookups for the chipidea driver with
> >clk_register_clkdev(). The lookups are only what the name suggests:
> >lookups. They are no clocks by themselves. This means that you have to
> >add another lookup which matches the device name.
> 
> Thanks for this help, I've added three "clocks" for the chipidea driver :
> clk_register_clkdev(clk[usb_div], NULL, "mxc-ehci.0");

That does not fit to the chipidea driver. The third argument has to
match the device name (which is the output of dev_name), so in your case
it must be something like 'imx_usb.0' (I don't know if that's correct,
please just add a printk("%s\n", dev_name(dev)); to the chipidea driver
to get the correct string)

Other than that I suggest that you read through drivers/clk/clkdev.c
to see how clock lookups are matched against devices, specifically:

/*
 * Find the correct struct clk for the device and connection ID.
 * We do slightly fuzzy matching here:
 *  An entry with a NULL ID is assumed to be a wildcard.
 *  If an entry has a device ID, it must match
 *  If an entry has a connection ID, it must match
 * Then we take the most specific entry - with the following
 * order of precedence: dev+con > dev only > con only.
 */

/**
 * clk_register_clkdev - register one clock lookup for a struct clk
 * @clk: struct clk to associate with all clk_lookups
 * @con_id: connection ID string on device
 * @dev_id: format string describing device name
 *
 * con_id or dev_id may be NULL as a wildcard, just as in the rest of
 * clkdev.
 */


> [    0.754044] usbserial: USB Serial support registered for pl2303
> [    0.762611] Unhandled fault: external abort on non-linefetch (0x008) at 0xf4424100
> [    0.770227] Internal error: : 8 [#1] PREEMPT ARM
> [    0.774861] Modules linked in:
> [    0.777950] CPU: 0    Not tainted  (3.7.0-rc5-ge917a67 #2)
> [    0.783469] PC is at ci_hdrc_probe+0xcc/0x430
> [    0.787853] LR is at devm_kzalloc+0x4c/0x64

This seems to be something else. Are you sure you are beyond the clk_get
point? I can't imagine this given the lookup you registered.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |