Re: [PATCH V2 2/2] remoteproc: support attach recovery after rproc crash

[Bjorn Andersson <bjorn.andersson@linaro.org>]

On Tue 08 Mar 00:48 CST 2022, Peng Fan (OSS) wrote:

> From: Peng Fan <peng.fan@nxp.com>
> 
> Current logic only support main processor to stop/start the remote
> processor after rproc crash. However to SoC, such as i.MX8QM/QXP, the
> remote processor could do attach recovery after crash and trigger watchdog

Does it really do something called "attach recovery and trigger watchdog
reboot"? Doesn't it just reboot itself and Linux needs to detach and
reattach to get something (what?) reset?

> reboot. It does not need main processor to load image, or stop/start M4
> core.
> 
> Introduce two functions: rproc_attach_recovery, rproc_firmware_recovery
> for the two cases. Firmware recovery is as before, let main processor to
> help recovery, while attach recovery is recover itself withou help.
> To attach recovery, we only do detach and attach.
> 
> Signed-off-by: Peng Fan <peng.fan@nxp.com>
> ---
> 
> V2:
>  use rproc_has_feature in patch 1/2
> 
>  drivers/remoteproc/remoteproc_core.c | 67 ++++++++++++++++++++--------
>  1 file changed, 48 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c
> index 69f51acf235e..366fad475898 100644
> --- a/drivers/remoteproc/remoteproc_core.c
> +++ b/drivers/remoteproc/remoteproc_core.c
> @@ -1887,6 +1887,50 @@ static int __rproc_detach(struct rproc *rproc)
>  	return 0;
>  }
>  
> +static int rproc_attach_recovery(struct rproc *rproc)
> +{
> +	int ret;
> +
> +	mutex_unlock(&rproc->lock);
> +	ret = rproc_detach(rproc);
> +	mutex_lock(&rproc->lock);
> +	if (ret)
> +		return ret;
> +
> +	if (atomic_inc_return(&rproc->power) > 1)

In the stop/coredump/start path the code _will_ attempt to recover the
remote processor. With rproc_detach() and rproc_attach() fiddling with
the rproc->power refcount this might do something, or it might not do
something. And with the mutex_unlock() it's likely that you're opening
of up for various race conditions inbetween.


PS. Does anyone actually use this refcount, or are we just all holding
our breath for it never going beyond 1?

Regards,
Bjorn

> +		return 0;
> +
> +	return rproc_attach(rproc);
> +}
> +
> +static int rproc_firmware_recovery(struct rproc *rproc)
> +{
> +	const struct firmware *firmware_p;
> +	struct device *dev = &rproc->dev;
> +	int ret;
> +
> +	ret = rproc_stop(rproc, true);
> +	if (ret)
> +		return ret;
> +
> +	/* generate coredump */
> +	rproc->ops->coredump(rproc);
> +
> +	/* load firmware */
> +	ret = request_firmware(&firmware_p, rproc->firmware, dev);
> +	if (ret < 0) {
> +		dev_err(dev, "request_firmware failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/* boot the remote processor up again */
> +	ret = rproc_start(rproc, firmware_p);
> +
> +	release_firmware(firmware_p);
> +
> +	return ret;
> +}
> +
>  /**
>   * rproc_trigger_recovery() - recover a remoteproc
>   * @rproc: the remote processor
> @@ -1901,7 +1945,6 @@ static int __rproc_detach(struct rproc *rproc)
>   */
>  int rproc_trigger_recovery(struct rproc *rproc)
>  {
> -	const struct firmware *firmware_p;
>  	struct device *dev = &rproc->dev;
>  	int ret;
>  
> @@ -1915,24 +1958,10 @@ int rproc_trigger_recovery(struct rproc *rproc)
>  
>  	dev_err(dev, "recovering %s\n", rproc->name);
>  
> -	ret = rproc_stop(rproc, true);
> -	if (ret)
> -		goto unlock_mutex;
> -
> -	/* generate coredump */
> -	rproc->ops->coredump(rproc);
> -
> -	/* load firmware */
> -	ret = request_firmware(&firmware_p, rproc->firmware, dev);
> -	if (ret < 0) {
> -		dev_err(dev, "request_firmware failed: %d\n", ret);
> -		goto unlock_mutex;
> -	}
> -
> -	/* boot the remote processor up again */
> -	ret = rproc_start(rproc, firmware_p);
> -
> -	release_firmware(firmware_p);
> +	if (rproc_has_feature(rproc, RPROC_FEAT_ATTACH_RECOVERY))
> +		ret = rproc_attach_recovery(rproc);
> +	else
> +		ret = rproc_firmware_recovery(rproc);
>  
>  unlock_mutex:
>  	mutex_unlock(&rproc->lock);
> -- 
> 2.30.0
>