[meta-freescale] USB host issue on 2.6.35

[Otavio Salvador]

Hello Fabio, +Mahesh,

On Thu, Jul 25, 2013 at 9:29 PM, Fabio Estevam <festevam at gmail.com> wrote:
> Today I have been debugging an issue with the USB Host1 on mx53qsb
> with kernel 3.11-rc2 and the good news is that it is working now.
>
> However, when I tried it with the latest 2.6.35 maintain branch I get
> the following crash whenever I insert a USB thumb:
>
> usb wakeup is here
> irq 14: nobody cared (try booting with the "irqpoll" option)
> [<80036b20>] (unwind_backtrace+0x0/0xe0) from [<8041f7c0>]
> (__report_bad_irq+0x34/0x8c)
> [<8041f7c0>] (__report_bad_irq+0x34/0x8c) from [<80096f78>]
> (note_interrupt+0x148/0x1d8)
> [<80096f78>] (note_interrupt+0x148/0x1d8) from [<800975d8>]
> (handle_level_irq+0xf8/0x184)
> [<800975d8>] (handle_level_irq+0xf8/0x184) from [<80032068>]
> (asm_do_IRQ+0x68/0x88)
> [<80032068>] (asm_do_IRQ+0x68/0x88) from [<8042620c>] (__irq_svc+0x4c/0xcc)
> Exception stack(0xdf0b7ec0 to 0xdf0b7f08)
> 7ec0: 80849d78 00000001 00001f08 00000002 8042bb5c df0aacd0 df0b6000 80866a08
> 7ee0: df11ba1c 80849d78 df11b880 df0b7f34 00000001 df0b7f08 80423ec4 80423fac
> 7f00: 00000013 ffffffff
> [<8042620c>] (__irq_svc+0x4c/0xcc) from [<80423fac>] (schedule+0x24c/0x3cc)
> [<80423fac>] (schedule+0x24c/0x3cc) from [<8042484c>]
> (schedule_timeout+0x18/0x1d8)
> [<8042484c>] (schedule_timeout+0x18/0x1d8) from [<804243c8>]
> (wait_for_common+0xc8/0x1ac)
> [<804243c8>] (wait_for_common+0xc8/0x1ac) from [<804244ec>]
> (wait_for_completion_interruptible+0x14/0x20)
> [<804244ec>] (wait_for_completion_interruptible+0x14/0x20) from
> [<8004d5e0>] (wakeup_event_thread+0x3c/0x16c)
> [<8004d5e0>] (wakeup_event_thread+0x3c/0x16c) from [<800774d8>]
> (kthread+0x78/0x84)
> [<800774d8>] (kthread+0x78/0x84) from [<80033294>] (kernel_thread_exit+0x0/0x8)
> handlers:
> [<8004d4cc>] (usb_wakeup_handler+0x0/0xd8)
> [<802c4d3c>] (ehci_fsl_pre_irq+0x0/0x60)
> [<802b1014>] (usb_hcd_irq+0x0/0x80)
> Disabling IRQ #14
>
> I don't have the time to fix this for 2.6.35, but doing the following
> reverts I no longer get the crashes:
>
> commit 3de0c5d3b991c68cc523b46379cb3aa6276d2f3a
> Author: Fabio Estevam <fabio.estevam at freescale.com>
> Date:   Thu Jul 25 21:22:39 2013 -0300
>
>     Revert "ENGR00238974-2: mx28-usb: Fix kinds of usb wakeup problems"
>
>     This reverts commit 08c82b59a39b41dfb2afe5dfa8249fe9642bf20d.
>
> commit 68ab0caabc9463882debd40a7acbc264adf9f717
> Author: Fabio Estevam <fabio.estevam at freescale.com>
> Date:   Thu Jul 25 21:22:18 2013 -0300
>
>     Revert "ENGR00238974-3: mx28-usb: Fix the bug of incorrect data
> type of wakeup_event"
>
>     This reverts commit cb5eabef73b4f362332f5bd4a11b28b92fd18e47.
>
> commit 40e5e30aed7224f4fe57fe630c2d2d231bad4a66
> Author: Fabio Estevam <fabio.estevam at freescale.com>
> Date:   Thu Jul 25 21:20:37 2013 -0300
>
>     Revert "ENGR00239205: mx28-usb: disable usb wakeup before module unload"
>
>     This reverts commit a074aa4a9d95b84fa8709a4d3d4aad5ac2ed38a4.
>
> ,so I would suggest that someone creates a patch for meta-fsl-arm
> reverting the three patches above.

Any news on this patches? Are anyone looking at fixing the regression
in the maintain branch?

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750