Re: [PATCH 0/3] remoteproc: xlnx: remote crash recovery

[Tanmay Shah <tanmay.shah@amd.com>]

On 11/11/25 1:12 AM, Peng Fan wrote:
> Hi Mathieu, Tanmay
> 
>> Subject: Re: [PATCH 0/3] remoteproc: xlnx: remote crash recovery
> ....
>>>
>>
>> So what is happening here - Peng, do you plan on providing more
>> debugging information? Tanmay - are you planning on sending a
>> second revision?
>>
> 
> Sorry for delay.  I gave a hack with below changes(at end) and it works.
> The below change is just hack code to let me verify the rproc->power value.
> The issue with current patchset is that after rproc_attach(),
> the rproc->power will be 0. So recovery only works for the 1st time,
> when it is triggered again, rproc_detach() will abort early in the if check:
> if (!atomic_dec_and_test(&rproc->power)) {
> ret = 0;
> return ret;
> }
> 
> ---------------
> diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c
> index a92c6cd6df67b..2b69304084d11 100644
> --- a/drivers/remoteproc/remoteproc_core.c
> +++ b/drivers/remoteproc/remoteproc_core.c
> @@ -1786,7 +1786,9 @@ static int rproc_attach_recovery(struct rproc *rproc)
>          if (ret)
>                  return ret;
>   
> -       return rproc_attach(rproc);
> +       ret = rproc_attach(rproc);
> +       atomic_set(&rproc->power, 1);
> +       return ret;
>   }
> 
> Thanks,
> Peng.
> 

Hi Peng,

This issue should be fixed after using rproc_boot instead of 
rproc_attach. rproc_boot makes sure that rproc_attach is atomic 
operation and it also increases power count as you mentioned above.

Thanks,
Tanmay

>>> Thanks,
>>> Peng
>>>
>>>>>
>>>>> Thanks,
>>>>> Tanmay
>>>>>
>>>>>> Thanks,
>>>>>> Peng
>>>>>
>>>>
>