[U-Boot] [PATCH v3 08/11] dm: imx: Use gpio_request() to request GPIOs

[Simon Glass]

Hi Nikita,

On 1 October 2014 05:58, Nikita Kiryanov <nikita at compulab.co.il> wrote:
> Hi Simon,
>
> On 17/09/14 18:02, Simon Glass wrote:
>>
>> GPIOs should be requested before use. Without this, driver model will not
>> permit the GPIO to be used.
>>
>> Signed-off-by: Simon Glass <sjg at chromium.org>
>
>
> This patch introduces a bunch of errors (once the driver model stuff is
> turned on), all related to the gpios never being freed, but requested
> anew when reinitializing subsystems.
>
> The errors are:
>
> CM-FX6 # sata init
> Warning: iSSD setup failed!
> AHCI 0001.0300 32 slots 1 ports 3 Gbps 0x1 impl SATA mode
> flags: ncq stag pm led clo only pmp pio slum part
> SATA Device Info:
> S/N: 123900127157
> Product model number: SanDisk SSD i100 8GB
> Firmware version: 11.56.00
> Capacity: 15649200 sectors

Is it correct to init something twice? It has already been done when
U-Boot boots I think. If it is, then are you thinking of changing
cm_fx6_setup_issd() to cope with that?

>
> CM-FX6 # usb start
> (Re)start USB...
> USB0:   USB OTG pwr gpio request failed: -16
> USB OTG pwr gpio request failed: -16
> USB EHCI 1.00
> scanning bus 0 for devices... 2 USB Device(s) found
> USB1:   USB hub rst gpio request failed: -16
> USB hub rst gpio request failed: -16
> USB EHCI 1.00
> scanning bus 1 for devices... 6 USB Device(s) found
>        scanning usb for storage devices... max USB Storage Device reached: 5
> stopping
> 5 Storage Device(s) found
>
> CM-FX6 # sf probe
> mxc_spi: cannot setup gpio -16
> SF: Failed to set up slave
> Failed to initialize SPI flash at 0:0

I took at look at how this works for SPI. The approach of calling a
board function to find out the GPIO should go away with driver model -
we can use device tree, or platform data if that is not yet available.

Also if you change the board code to 'stash' the GPIO and not request
it a second time, you will need to do that in every board. It seems
better to me to make this change in the driver, at least for SPI.

board_spi_cs_gpio() was added very recently in commit 155fa9af9. If
you change it so that setup_cs_gpio() remembers the GPIO after calling
board_spi_cs_gpio() then we can avoid passing the problem on to
boards.

>
> CM-FX6 # saveenv
> Saving Environment to SPI Flash...
> mxc_spi: cannot setup gpio -16
> SF: Failed to set up slave
> *** Warning - spi_flash_probe() failed, using default environment

Same issue.

>
> I am going to submit a modified version of the cm_fx6 patches to address
> these problems.

I think those are already merged - I based my patches on your cm_fx6
patches and there were already in mainline.

>
> --
> Regards,
> Nikita Kiryanov

Regards,
Simon