ACK/cmnt: [Utopic] [PULL] UBUNTU: fix no console in case of cloud environment V1

[Ming Lei]

On Fri, Apr 3, 2015 at 8:42 AM, Ming Lei <ming.lei at canonical.com> wrote:
> On Thu, Apr 2, 2015 at 6:40 PM, Stefan Bader <stefan.bader at canonical.com> wrote:
>> On 02.04.2015 03:20, Ming Lei wrote:
>>> Hi Guys,
>>>
>>> In current arm64 utopic cloud image, the 'console=' parameter
>>> has been removed from grub config, and always resort to 'stdout'
>>> DT property which is setup from QEMU.
>>>
>>> Unfortunately parsing 'stdout' DT property is just merged to 3.19
>>> kernel, so there is no console usable any more for arm64 utopic
>>> VM when booting via uefi/grub.
>>>
>>> 5 of these 6 patches are cherry picked from 3.19 to utopic, and the
>>> other one is backported by just dropping the dt self-test code, and
>>> with these 6 patches, arm64 utopic VM can be booted with console
>>> usable via uefi/grub.
>>>
>>> BugLink: http://bugs.launchpad.net/bugs/1438585
>>>
>>> SRU Justification:
>>>     Impact: no console if arm64 utopic based VM is booted from the cloud image
>>>     Fix: backport stdout enablement patches from 3.19
>>>     Testcase: boot one arm64 utopic VM via uefi/grub
>>>
>>> So please pull the top 6 patches from below tree/branch:
>>>
>>>      git://kernel.ubuntu.com/git/ming/ubuntu-utopic.git  console-fix_stdout
>>>
>>> 5ba2ee8 of: support passing console options with stdout-path
>>> 5c64126 of: add optional options parameter to of_find_node_by_path()
>>> 7507839 of: Add bindings for chosen node, stdout-path
>>> 54e15d7 of: correct of_console_check()'s return value
>> Those above look consistent and ok.
>>
>>> 4f2b1f0 of: Enable console on serial ports specified by /chosen/stdout-path
>> This touches common code though I assume of_node will be NULL for anything not
>> belonging to open firmware.
>
> It is NULL only if stdout property isn't set, and it can be handled well too.
>
>>
>>> a699dfd of: Create of_console_check() for selecting a console
>>> specified in /chosen
>> This may carry some risk as it also removes an exported function
>> (of_device_is_stdout_path). Maybe that should be modified to add the new
>> function in parallel instead. Even though it is claimed the function is not used
>> by anything, there always may be someone doing so in external modules.
>>
>
> It must be my fault to miss that part when reediting the patch, and
> the original patch did not remove of_device_is_stdout_path().

No, No, it is really removed originally, and there aren't other users
of the removed API even in 3.19, so I think it shouldn't be out-of-tree
users too for utopic.

Sorry, I thought it was 27a30f64a13c051c0, which didn't remove the API
of of_find_node_by_path().

Thanks,
Ming Lei