Hi, Sorry for the late response.
On 02/14/2018 04:19 PM, Marek Vasut wrote:
On 02/14/2018 03:14 PM, Faiz Abbas wrote:
Hi,
On Wednesday 14 February 2018 06:53 PM, Marek Vasut wrote:
On 02/14/2018 12:20 PM, Faiz Abbas wrote:
Hi,
On Wednesday 14 February 2018 03:46 PM, Marek Vasut wrote:
On 02/14/2018 11:10 AM, Faiz Abbas wrote:
board_usb_init()/_cleanup() should be in board files and don't have a place in the xhci-omap driver. Weak versions for board_usb_init()/_cleanup() already exist in common/usb.c (for host mode) and drivers/usb/gadget/g_dnl.c (for gadget mode).
Signed-off-by: Faiz Abbas faiz_abbas@ti.com
Reviewed-by: Marek Vasut marex@denx.de
I'd like some TBs from the people using those boards.
I have tested this for dra7xx, am43xx and am57xx.
So why is it marked RFC ?
Because I was unclear why Uri Mashiach did not do this in 1a9a5f7 ("fix double weak board_usb_init functions").
OK, submit it as a normal patch with my RB.
As mentioned in 1a9a5f7, the functions omap_xhci_board_usb_init() and omap_xhci_board_usb_cleanup() are called for the boards with the CONFIG_USB_XHCI_OMAP definition.
The weak implementation of the functions omap_xhci_board_usb_init() executed enable_usb_clocks(). The weak implementation of the function omap_xhci_board_usb_cleanup() executed the function disable_usb_clocks().
For the boards compulab/cl-som-am57x and compulab/cm_t43: * CONFIG_USB_XHCI_OMAP is defined * omap_xhci_board_usb_init is not implemented, relying on the weak implementation. * omap_xhci_board_usb_cleanup is not defined, relying on the weak implementation.
The fix is missing the implementation of board_usb_init and board_usb_cleanup in the compulab/cl-som-am57x and compulab/cm_t43. The implementation should include the content of the deleted weak functions omap_xhci_board_usb_init() and omap_xhci_board_usb_cleanup().