[PATCH v4] remoteproc: Add device awake calls in rproc boot and shutdown path

[PATCH v4] remoteproc: Add device awake calls in rproc boot and shutdown path

[Souradeep Chowdhury]

Add device awake calls in case of rproc boot and rproc shutdown path.
Currently, device awake call is only present in the recovery path
of remoteproc. If a user stops and starts rproc by using the sysfs
interface, then on pm suspension the firmware loading fails as the
request_firmware call under adsp_load returns failure. Add device
awake calls to fix this.

Signed-off-by: Souradeep Chowdhury <quic_schowdhu@quicinc.com>
---
Changes in v4

*Remove stability from mailing list
*Remove the extra tab in v3
*Change the commit description

 drivers/remoteproc/remoteproc_core.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c
index c2cf0d277729..5d6c4e694b4c 100644
--- a/drivers/remoteproc/remoteproc_core.c
+++ b/drivers/remoteproc/remoteproc_core.c
@@ -1917,6 +1917,7 @@ int rproc_boot(struct rproc *rproc)
 		return -EINVAL;
 	}
 
+	pm_stay_awake(rproc->dev.parent);
 	dev = &rproc->dev;
 
 	ret = mutex_lock_interruptible(&rproc->lock);
@@ -1961,6 +1962,7 @@ int rproc_boot(struct rproc *rproc)
 		atomic_dec(&rproc->power);
 unlock_mutex:
 	mutex_unlock(&rproc->lock);
+	pm_relax(rproc->dev.parent);
 	return ret;
 }
 EXPORT_SYMBOL(rproc_boot);
@@ -1991,6 +1993,7 @@ int rproc_shutdown(struct rproc *rproc)
 	struct device *dev = &rproc->dev;
 	int ret = 0;
 
+	pm_stay_awake(rproc->dev.parent);
 	ret = mutex_lock_interruptible(&rproc->lock);
 	if (ret) {
 		dev_err(dev, "can't lock rproc %s: %d\n", rproc->name, ret);
@@ -2027,6 +2030,7 @@ int rproc_shutdown(struct rproc *rproc)
 	rproc->table_ptr = NULL;
 out:
 	mutex_unlock(&rproc->lock);
+	pm_relax(rproc->dev.parent);
 	return ret;
 }
 EXPORT_SYMBOL(rproc_shutdown);
-- 
2.34.1

Re: [PATCH v4] remoteproc: Add device awake calls in rproc boot and shutdown path

[Mukesh Ojha]

On Tue, Apr 01, 2025 at 01:29:21PM +0530, Souradeep Chowdhury wrote:
> Add device awake calls in case of rproc boot and rproc shutdown path.
> Currently, device awake call is only present in the recovery path
> of remoteproc. If a user stops and starts rproc by using the sysfs
> interface, then on pm suspension the firmware loading fails as the

"firmware loading which relies on userspace to load the firmware may
fail as during suspend freezing of processes happens and due to which
usermodehelper gets disable."

> request_firmware call under adsp_load returns failure. Add device
> awake calls to fix this.
> 
> Signed-off-by: Souradeep Chowdhury <quic_schowdhu@quicinc.com>
> ---
> Changes in v4
> 
> *Remove stability from mailing list
> *Remove the extra tab in v3
> *Change the commit description

nit: It would have been better if you have added detail what are all thing
happened between as v1-v3 here and link to respective lore links.

Changes in v3: 
..,

Changes in v2:
..,

> 
>  drivers/remoteproc/remoteproc_core.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c
> index c2cf0d277729..5d6c4e694b4c 100644
> --- a/drivers/remoteproc/remoteproc_core.c
> +++ b/drivers/remoteproc/remoteproc_core.c
> @@ -1917,6 +1917,7 @@ int rproc_boot(struct rproc *rproc)
>  		return -EINVAL;
>  	}
>  
> +	pm_stay_awake(rproc->dev.parent);
>  	dev = &rproc->dev;
>  
>  	ret = mutex_lock_interruptible(&rproc->lock);
> @@ -1961,6 +1962,7 @@ int rproc_boot(struct rproc *rproc)
>  		atomic_dec(&rproc->power);
>  unlock_mutex:
>  	mutex_unlock(&rproc->lock);
> +	pm_relax(rproc->dev.parent);
>  	return ret;
>  }
>  EXPORT_SYMBOL(rproc_boot);
> @@ -1991,6 +1993,7 @@ int rproc_shutdown(struct rproc *rproc)
>  	struct device *dev = &rproc->dev;
>  	int ret = 0;
>  
> +	pm_stay_awake(rproc->dev.parent);
>  	ret = mutex_lock_interruptible(&rproc->lock);
>  	if (ret) {
>  		dev_err(dev, "can't lock rproc %s: %d\n", rproc->name, ret);
> @@ -2027,6 +2030,7 @@ int rproc_shutdown(struct rproc *rproc)
>  	rproc->table_ptr = NULL;
>  out:
>  	mutex_unlock(&rproc->lock);
> +	pm_relax(rproc->dev.parent);
>  	return ret;
>  }
>  EXPORT_SYMBOL(rproc_shutdown);

Changes looks fine to me.
Feel free to add

Reviewed-by: Mukesh Ojha <mukesh.ojha@oss.qualcomm.com>

-Mukesh