Hi Greg,
On Fri, Sep 7, 2018 at 8:14 AM, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> On Mon, Sep 03, 2018 at 10:39:04AM -0300, Fabio Estevam wrote:
>> From: Fabio Estevam <fabio.estevam@xxxxxxx>
>>
>> This reverts commit df06ca1f565077596512fa2ab91df502e48e4f00.
>>03e6275ae381 ("usb: chipidea: Fix ULPI on imx51")
>> This commit causes reboot to fail on imx6 wandboard, so let's
>> revert it.
>>
>> Cc: <stable@xxxxxxxxxxxxxxx> #4.17
>> Reported-by: Rasmus Villemoes <rasmus.villemoes@xxxxxxxxx>
>> Signed-off-by: Fabio Estevam <fabio.estevam@xxxxxxx>
>
> Why is this a stable-tree-only patch? What is wrong with those trees
> that having this in Linus's tree (and 4.18) is ok, but stable trees are
> not ok?
In Linus' tree we have a930d8bd94d8 ("usb: chipidea: Always build ULPI code "),
which removes CONFIG_USB_CHIPIDEA_ULPI and CONFIG_USB_ULPI_BUS.
Since it fixes 03e6275ae381 ("usb: chipidea: Fix ULPI on imx51") then
we would also need to have this one in the stable tree.
And its equivalent fix for imx53: be9cae2479f4 ("usb: chipidea: imx:
Fix ULPI on imx53")
So in order to fix the reboot regression in stable as reported by
Rasmus I see two options:
1. Backport these three commits to stable:
a930d8bd94d8 ("usb: chipidea: Always build ULPI code ")
03e6275ae381 ("usb: chipidea: Fix ULPI on imx51")
be9cae2479f4 ("usb: chipidea: imx: Fix ULPI on imx53")
2. Just revert the defconfig change as originally proposed.
If everyone agrees that option 1 is preferred I can prepare these
patches for 4.4/4.9/4.14/4.17.
Please let me know if option 1 is the way to go.
Thanks
