Re: [PATCH V4] firmware: arm_scmi: Make scmi core independent of the transport type

[Cristian Marussi]

On 23/01/2020 11:27, Viresh Kumar wrote:
> On 23-01-20, 10:30, Sudeep Holla wrote:
>> On Wed, Jan 22, 2020 at 12:15:38PM +0000, Sudeep Holla wrote:
>>> On Wed, Jan 22, 2020 at 08:06:23AM +0530, Viresh Kumar wrote:
>>>
>>
>> [...]
>>
>>>> Can you please help me getting this tested, now that I have rebased it
>>>> as well :) ?
>>>>
>>>
>>> Sure, I will give it a go on my Juno. Thanks for the rebase, makes it
>>> simpler.
>>>
>>
>> Sorry for the delay. I gave this a spin on my Juno. I am seeing below
>> warning once on boot but it continues and everything seem to work fine.
>> Also the warning is not related to this change I believe and this patch
>> is just helping to hit some corner case with deferred probe and devres.
>> I need to spend some time to debug it.
>>
>> Regards,
>> Sudeep
>>
>> --->8
>>
>> WARNING: CPU: 1 PID: 187 at drivers/base/dd.c:519 really_probe+0x11c/0x418
>> Modules linked in:
>> CPU: 1 PID: 187 Comm: kworker/1:2 Not tainted 5.5.0-rc7-00026-gf7231cd3108d-dirty #20
>> Hardware name: ARM LTD ARM Juno Development Platform/ARM Juno Development Platform, BIOS EDK II Jan 16 2020
>> Workqueue: events deferred_probe_work_func
>> pstate: 80000005 (Nzcv daif -PAN -UAO)
>> pc : really_probe+0x11c/0x418
>> lr : really_probe+0x10c/0x418
>> Call trace:
>>  really_probe+0x11c/0x418
>>  driver_probe_device+0xe4/0x138
>>  __device_attach_driver+0x90/0x110
>>  bus_for_each_drv+0x80/0xd0
>>  __device_attach+0xdc/0x160
>>  device_initial_probe+0x18/0x20
>>  bus_probe_device+0x98/0xa0
>>  deferred_probe_work_func+0x90/0xe0
>>  process_one_work+0x1ec/0x4a8
>>  worker_thread+0x210/0x490
>>  kthread+0x110/0x118
>>  ret_from_fork+0x10/0x18
>> ---[ end trace 06f96d55ce6093a8 ]---
> 
> Still it looks strange that the warning comes only after my patch :)

In fact, got the same warning while testing your patch on JUNO at top of SCMI for-next.
But then everything worked fine as Sudeep said.

Thanks

Cristian

> 
> Should I send V5 (fixed few comments after reviews) now ?
>