Skip to contentSkip to conversation

Use Device Lifecycle managed functions in TI R5 Remoteproc driver

PATCH In review Beleswar Padhi 3 participants 9 messages Dec 19, 2024 – Jan 6, 2025 Remote processor (remoteproc)

lore
Share & tools
lore
#0 lore raw

v2 cover Use Device Lifecycle managed functions in TI R5 Remoteproc driver

This series uses various devm_ helpers to simplify device removal path in ti_k3_r5_remoteproc driver. This is the first series in the TI K3 Remoteproc refactoring as long planned since [0].

Testing Done:
1. Tested boot of R5F remoteprocs in MCU and MAIN voltage domain in both IPC-Only mode and Kernel remoteproc mode in all Jacinto K3 devices.
2. Tested Lockstep, Split and Single-CPU Mode configuration (wherever applicable) of R5F remoteprocs in all Jacinto K3 devices.
3. Tested shutdown of R5F remoteprocs from Linux userspace and also by executing modprobe -r ti_k3_r5_remoteproc.
4. Tested that each patch in this series generates no new warnings/errors.

v2: Changelog:
1. Re-ordered patches in the series to use devm functions starting from the last called function in remove(), to ease review. [Andrew]
2. Fixed a missing return after dev_err_probe() call in [PATCH v2 3/5] ("remoteproc: k3-r5: Use devm_ioremap_wc() helper"). [Andrew]
3. Removed redundant rproc_del() call in [PATCH v2 4/5] ("remoteproc:
k3-r5: Use devm_rproc_add() helper").

Link to v1:
https://lore.kernel.org/all/[email protected]/

[0]: https://lore.kernel.org/all/Zr4w8Vj0mVo5sBsJ@p14s/

Beleswar Padhi (5):
remoteproc: k3-r5: Add devm action to release reserved memory
remoteproc: k3-r5: Use devm_kcalloc() helper
remoteproc: k3-r5: Use devm_ioremap_wc() helper
remoteproc: k3-r5: Use devm_rproc_add() helper
remoteproc: k3-r5: Add devm action to release tsp

 drivers/remoteproc/ti_k3_r5_remoteproc.c | 88 ++++++++++--------------
 1 file changed, 35 insertions(+), 53 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 1/5 remoteproc: k3-r5: Add devm action to release reserved memory

Use a device lifecycle managed action to release reserved memory. This helps prevent mistakes like releasing out of order in cleanup functions and forgetting to release on error paths.

---
v2: Changelog:
1. None to this patch

Link to v1:
https://lore.kernel.org/all/[email protected]/

 drivers/remoteproc/ti_k3_r5_remoteproc.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/remoteproc/ti_k3_r5_remoteproc.c +13 -8
index 6560b7954027..d51f88a5abc1 100644
@@ -955,6 +955,13 @@ static int k3_r5_rproc_configure(struct k3_r5_rproc *kproc)
 	return ret;
 }
 
+static void k3_r5_mem_release(void *data)
+{
+	struct device *dev = data;
+
+	of_reserved_mem_device_release(dev);
+}
+
 static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 {
 	struct device *dev = kproc->dev;
@@ -985,12 +992,14 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 		return ret;
 	}
 
+	ret = devm_add_action_or_reset(dev, k3_r5_mem_release, dev);
+	if (ret)
+		return ret;
+
 	num_rmems--;
 	kproc->rmem = kcalloc(num_rmems, sizeof(*kproc->rmem), GFP_KERNEL);
-	if (!kproc->rmem) {
-		ret = -ENOMEM;
-		goto release_rmem;
-	}
+	if (!kproc->rmem)
+		return -ENOMEM;
 
 	/* use remaining reserved memory regions for static carveouts */
 	for (i = 0; i < num_rmems; i++) {
@@ -1041,8 +1050,6 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 	for (i--; i >= 0; i--)
 		iounmap(kproc->rmem[i].cpu_addr);
 	kfree(kproc->rmem);
-release_rmem:
-	of_reserved_mem_device_release(dev);
 	return ret;
 }
 
@@ -1053,8 +1060,6 @@ static void k3_r5_reserved_mem_exit(struct k3_r5_rproc *kproc)
 	for (i = 0; i < kproc->num_rmems; i++)
 		iounmap(kproc->rmem[i].cpu_addr);
 	kfree(kproc->rmem);
-
-	of_reserved_mem_device_release(kproc->dev);
 }
 
 /*
Signature · 1 line
2.34.1
#2 lore raw

PATCH v2 2/5 remoteproc: k3-r5: Use devm_kcalloc() helper

In reply to Beleswar Padhi

Use a device lifecycle managed action to free memory. This helps prevent mistakes like freeing out of order in cleanup functions and forgetting to free on error paths.

---
v2: Changelog:
1. None to this patch

Link to v1:
https://lore.kernel.org/all/[email protected]/

 drivers/remoteproc/ti_k3_r5_remoteproc.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/remoteproc/ti_k3_r5_remoteproc.c +1 -3
index d51f88a5abc1..0753a5c35c7e 100644
@@ -997,7 +997,7 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 		return ret;
 
 	num_rmems--;
-	kproc->rmem = kcalloc(num_rmems, sizeof(*kproc->rmem), GFP_KERNEL);
+	kproc->rmem = devm_kcalloc(dev, num_rmems, sizeof(*kproc->rmem), GFP_KERNEL);
 	if (!kproc->rmem)
 		return -ENOMEM;
 
@@ -1049,7 +1049,6 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 unmap_rmem:
 	for (i--; i >= 0; i--)
 		iounmap(kproc->rmem[i].cpu_addr);
-	kfree(kproc->rmem);
 	return ret;
 }
 
@@ -1059,7 +1058,6 @@ static void k3_r5_reserved_mem_exit(struct k3_r5_rproc *kproc)
 
 	for (i = 0; i < kproc->num_rmems; i++)
 		iounmap(kproc->rmem[i].cpu_addr);
-	kfree(kproc->rmem);
 }
 
 /*
Signature · 1 line
2.34.1
#3 lore raw

PATCH v2 3/5 remoteproc: k3-r5: Use devm_ioremap_wc() helper

In reply to Beleswar Padhi

Use a device lifecycle managed ioremap helper function. This helps prevent mistakes like unmapping out of order in cleanup functions and forgetting to unmap on all error paths.

---
v2: Changelog:
1. Re-ordered [PATCH 4/5] from v1 to [PATCH v2 3/5] in v2. [Andrew]
2. Returned after dev_err_probe() call by jumping to 'out' label in k3_r5_cluster_rproc_init(). [Andrew]

Link to v1:
https://lore.kernel.org/all/[email protected]/

 drivers/remoteproc/ti_k3_r5_remoteproc.c | 38 +++++-------------------
 1 file changed, 8 insertions(+), 30 deletions(-)
drivers/remoteproc/ti_k3_r5_remoteproc.c +8 -30
index 0753a5c35c7e..3493bae95fdf 100644
@@ -1004,17 +1004,13 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 	/* use remaining reserved memory regions for static carveouts */
 	for (i = 0; i < num_rmems; i++) {
 		rmem_np = of_parse_phandle(np, "memory-region", i + 1);
-		if (!rmem_np) {
-			ret = -EINVAL;
-			goto unmap_rmem;
-		}
+		if (!rmem_np)
+			return -EINVAL;
 
 		rmem = of_reserved_mem_lookup(rmem_np);
 		of_node_put(rmem_np);
-		if (!rmem) {
-			ret = -EINVAL;
-			goto unmap_rmem;
-		}
+		if (!rmem)
+			return -EINVAL;
 
 		kproc->rmem[i].bus_addr = rmem->base;
 		/*
@@ -1029,12 +1025,11 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 		 */
 		kproc->rmem[i].dev_addr = (u32)rmem->base;
 		kproc->rmem[i].size = rmem->size;
-		kproc->rmem[i].cpu_addr = ioremap_wc(rmem->base, rmem->size);
+		kproc->rmem[i].cpu_addr = devm_ioremap_wc(dev, rmem->base, rmem->size);
 		if (!kproc->rmem[i].cpu_addr) {
 			dev_err(dev, "failed to map reserved memory#%d at %pa of size %pa\n",
 				i + 1, &rmem->base, &rmem->size);
-			ret = -ENOMEM;
-			goto unmap_rmem;
+			return -ENOMEM;
 		}
 
 		dev_dbg(dev, "reserved memory%d: bus addr %pa size 0x%zx va %pK da 0x%x\n",
@@ -1045,19 +1040,6 @@ static int k3_r5_reserved_mem_init(struct k3_r5_rproc *kproc)
 	kproc->num_rmems = num_rmems;
 
 	return 0;
-
-unmap_rmem:
-	for (i--; i >= 0; i--)
-		iounmap(kproc->rmem[i].cpu_addr);
-	return ret;
-}
-
-static void k3_r5_reserved_mem_exit(struct k3_r5_rproc *kproc)
-{
-	int i;
-
-	for (i = 0; i < kproc->num_rmems; i++)
-		iounmap(kproc->rmem[i].cpu_addr);
 }
 
 /*
@@ -1286,8 +1268,8 @@ static int k3_r5_cluster_rproc_init(struct platform_device *pdev)
 
 		ret = rproc_add(rproc);
 		if (ret) {
-			dev_err(dev, "rproc_add failed, ret = %d\n", ret);
-			goto err_add;
+			dev_err_probe(dev, ret, "rproc_add failed\n");
+			goto out;
 		}
 
 		/* create only one rproc in lockstep, single-cpu or
@@ -1333,8 +1315,6 @@ static int k3_r5_cluster_rproc_init(struct platform_device *pdev)
 
 err_powerup:
 	rproc_del(rproc);
-err_add:
-	k3_r5_reserved_mem_exit(kproc);
 out:
 	/* undo core0 upon any failures on core1 in split-mode */
 	if (cluster->mode == CLUSTER_MODE_SPLIT && core == core1) {
@@ -1379,8 +1359,6 @@ static void k3_r5_cluster_rproc_exit(void *data)
 		mbox_free_channel(kproc->mbox);
 
 		rproc_del(rproc);
-
-		k3_r5_reserved_mem_exit(kproc);
 	}
 }
 
Signature · 1 line
2.34.1
#4 lore raw

PATCH v2 4/5 remoteproc: k3-r5: Use devm_rproc_add() helper

In reply to Beleswar Padhi

Use device lifecycle managed devm_rproc_add() helper function. This helps prevent mistakes like deleting out of order in cleanup functions and forgetting to delete on all error paths.

---
v2: Changelog:
1. Re-ordered [PATCH 5/5] from v1 to [PATCH v2 4/5] in v2. [Andrew]
2. Removed redundant rproc_del() call in k3_r5_cluster_rproc_init() as devm_rproc_add is used.

Link to v1:
https://lore.kernel.org/all/[email protected]/

 drivers/remoteproc/ti_k3_r5_remoteproc.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)
drivers/remoteproc/ti_k3_r5_remoteproc.c +2 -6
index 3493bae95fdf..30b72bcb4d68 100644
@@ -1266,7 +1266,7 @@ static int k3_r5_cluster_rproc_init(struct platform_device *pdev)
 			goto out;
 		}
 
-		ret = rproc_add(rproc);
+		ret = devm_rproc_add(dev, rproc);
 		if (ret) {
 			dev_err_probe(dev, ret, "rproc_add failed\n");
 			goto out;
@@ -1297,7 +1297,7 @@ static int k3_r5_cluster_rproc_init(struct platform_device *pdev)
 			dev_err(dev,
 				"Timed out waiting for %s core to power up!\n",
 				rproc->name);
-			goto err_powerup;
+			goto out;
 		}
 	}
 
@@ -1313,8 +1313,6 @@ static int k3_r5_cluster_rproc_init(struct platform_device *pdev)
 		}
 	}
 
-err_powerup:
-	rproc_del(rproc);
 out:
 	/* undo core0 upon any failures on core1 in split-mode */
 	if (cluster->mode == CLUSTER_MODE_SPLIT && core == core1) {
@@ -1357,8 +1355,6 @@ static void k3_r5_cluster_rproc_exit(void *data)
 		}
 
 		mbox_free_channel(kproc->mbox);
-
-		rproc_del(rproc);
 	}
 }
 
Signature · 1 line
2.34.1
#5 lore raw

PATCH v2 5/5 remoteproc: k3-r5: Add devm action to release tsp

In reply to Beleswar Padhi

Use a device lifecycle managed action to release tsp ti_sci_proc handle. This helps prevent mistakes like releasing out of order in cleanup functions and forgetting to release on error paths.

---
v2: Changelog:
1. Re-ordered [PATCH 3/5] from v1 to [PATCH v2 5/5] in v2. [Andrew]

Link to v1:
https://lore.kernel.org/all/[email protected]/

 drivers/remoteproc/ti_k3_r5_remoteproc.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)
drivers/remoteproc/ti_k3_r5_remoteproc.c +11 -6
index 30b72bcb4d68..dbc513c5569c 100644
@@ -1487,6 +1487,13 @@ static int k3_r5_core_of_get_sram_memories(struct platform_device *pdev,
 	return 0;
 }
 
+static void k3_r5_release_tsp(void *data)
+{
+	struct ti_sci_proc *tsp = data;
+
+	ti_sci_proc_release(tsp);
+}
+
 static int k3_r5_core_of_init(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
@@ -1580,6 +1587,10 @@ static int k3_r5_core_of_init(struct platform_device *pdev)
 		goto err;
 	}
 
+	ret = devm_add_action_or_reset(dev, k3_r5_release_tsp, core->tsp);
+	if (ret)
+		goto err;
+
 	platform_set_drvdata(pdev, core);
 	devres_close_group(dev, k3_r5_core_of_init);
 
@@ -1596,13 +1607,7 @@ static int k3_r5_core_of_init(struct platform_device *pdev)
  */
 static void k3_r5_core_of_exit(struct platform_device *pdev)
 {
-	struct k3_r5_core *core = platform_get_drvdata(pdev);
 	struct device *dev = &pdev->dev;
-	int ret;
-
-	ret = ti_sci_proc_release(core->tsp);
-	if (ret)
-		dev_err(dev, "failed to release proc, ret = %d\n", ret);
 
 	platform_set_drvdata(pdev, NULL);
 	devres_release_group(dev, k3_r5_core_of_init);
Signature · 1 line
2.34.1
Reviewed-by on the v2 series
#6 lore raw
In reply to Beleswar Padhi
Quoted 11 lines · Beleswar Padhi

This series uses various devm_ helpers to simplify device removal path in ti_k3_r5_remoteproc driver. This is the first series in the TI K3 Remoteproc refactoring as long planned since [0].

Testing Done:
1. Tested boot of R5F remoteprocs in MCU and MAIN voltage domain in both IPC-Only mode and Kernel remoteproc mode in all Jacinto K3 devices.
2. Tested Lockstep, Split and Single-CPU Mode configuration (wherever applicable) of R5F remoteprocs in all Jacinto K3 devices.
3. Tested shutdown of R5F remoteprocs from Linux userspace and also by executing modprobe -r ti_k3_r5_remoteproc.

Did you also test that you could then start the cores back up?

I think that might need some firmware fixes we are working on, so might not work even before these patches, but just wanted to check if we have tried it yet.

Quoted 1 line

4. Tested that each patch in this series generates no new warnings/errors.

Was that with make W=1 C=1? Sparse checks will be done during -next testing so good to check for those too.

Otherwise patches all look good to me, for the series:

Quoted 24 lines

v2: Changelog:
1. Re-ordered patches in the series to use devm functions starting from the last called function in remove(), to ease review. [Andrew]
2. Fixed a missing return after dev_err_probe() call in [PATCH v2 3/5] ("remoteproc: k3-r5: Use devm_ioremap_wc() helper"). [Andrew]
3. Removed redundant rproc_del() call in [PATCH v2 4/5] ("remoteproc:
k3-r5: Use devm_rproc_add() helper").

Link to v1:
https://lore.kernel.org/all/[email protected]/

[0]: https://lore.kernel.org/all/Zr4w8Vj0mVo5sBsJ@p14s/

Beleswar Padhi (5):
remoteproc: k3-r5: Add devm action to release reserved memory
remoteproc: k3-r5: Use devm_kcalloc() helper
remoteproc: k3-r5: Use devm_ioremap_wc() helper
remoteproc: k3-r5: Use devm_rproc_add() helper
remoteproc: k3-r5: Add devm action to release tsp

  drivers/remoteproc/ti_k3_r5_remoteproc.c | 88 ++++++++++--------------
  1 file changed, 35 insertions(+), 53 deletions(-)
#7 lore raw
Quoted 13 lines · Andrew Davis
Quoted 11 lines · Beleswar Padhi

This series uses various devm_ helpers to simplify device removal path in ti_k3_r5_remoteproc driver. This is the first series in the TI K3 Remoteproc refactoring as long planned since [0].

Testing Done:
1. Tested boot of R5F remoteprocs in MCU and MAIN voltage domain in both IPC-Only mode and Kernel remoteproc mode in all Jacinto K3 devices.
2. Tested Lockstep, Split and Single-CPU Mode configuration (wherever applicable) of R5F remoteprocs in all Jacinto K3 devices.
3. Tested shutdown of R5F remoteprocs from Linux userspace and also by executing modprobe -r ti_k3_r5_remoteproc.

Quoted 5 lines

Did you also test that you could then start the cores back up?

I think that might need some firmware fixes we are working on, so might not work even before these patches, but just wanted to check if we have tried it yet.

Yes, the above graceful shutdown feature is part of firmware fixes. Also, some support needs to be added in the Linux driver, which sends a special "SHUTDOWN" mailbox message to the remotecore which signals the firmware to relinquish all resources and shutdown gracefully. So, support for turning the remoteprocs back on is not yet added, and not tested.

Quoted 5 lines
Quoted 2 lines

4. Tested that each patch in this series generates no new warnings/errors.

Was that with make W=1 C=1? Sparse checks will be done during -next testing so good to check for those too.

Yes, did that testing.

Quoted 4 lines

Otherwise patches all look good to me, for the series:

Thanks!

Quoted 25 lines
Quoted 25 lines

v2: Changelog:
1. Re-ordered patches in the series to use devm functions starting from the last called function in remove(), to ease review. [Andrew]
2. Fixed a missing return after dev_err_probe() call in [PATCH v2 3/5] ("remoteproc: k3-r5: Use devm_ioremap_wc() helper"). [Andrew]
3. Removed redundant rproc_del() call in [PATCH v2 4/5] ("remoteproc:
k3-r5: Use devm_rproc_add() helper").

Link to v1:
https://lore.kernel.org/all/[email protected]/

[0]: https://lore.kernel.org/all/Zr4w8Vj0mVo5sBsJ@p14s/

Beleswar Padhi (5):
   remoteproc: k3-r5: Add devm action to release reserved memory
   remoteproc: k3-r5: Use devm_kcalloc() helper
   remoteproc: k3-r5: Use devm_ioremap_wc() helper
   remoteproc: k3-r5: Use devm_rproc_add() helper
   remoteproc: k3-r5: Add devm action to release tsp

  drivers/remoteproc/ti_k3_r5_remoteproc.c | 88 ++++++++++--------------

  1 file changed, 35 insertions(+), 53 deletions(-)
maintainer
#8 lore raw
In reply to Beleswar Padhi
Quoted 33 lines · Beleswar Padhi

This series uses various devm_ helpers to simplify device removal path in ti_k3_r5_remoteproc driver. This is the first series in the TI K3 Remoteproc refactoring as long planned since [0].

Testing Done:
1. Tested boot of R5F remoteprocs in MCU and MAIN voltage domain in both IPC-Only mode and Kernel remoteproc mode in all Jacinto K3 devices.
2. Tested Lockstep, Split and Single-CPU Mode configuration (wherever applicable) of R5F remoteprocs in all Jacinto K3 devices.
3. Tested shutdown of R5F remoteprocs from Linux userspace and also by executing modprobe -r ti_k3_r5_remoteproc.
4. Tested that each patch in this series generates no new warnings/errors.

v2: Changelog:
1. Re-ordered patches in the series to use devm functions starting from the last called function in remove(), to ease review. [Andrew]
2. Fixed a missing return after dev_err_probe() call in [PATCH v2 3/5] ("remoteproc: k3-r5: Use devm_ioremap_wc() helper"). [Andrew]
3. Removed redundant rproc_del() call in [PATCH v2 4/5] ("remoteproc:
k3-r5: Use devm_rproc_add() helper").

Link to v1:
https://lore.kernel.org/all/[email protected]/

[0]: https://lore.kernel.org/all/Zr4w8Vj0mVo5sBsJ@p14s/

Beleswar Padhi (5):
remoteproc: k3-r5: Add devm action to release reserved memory
remoteproc: k3-r5: Use devm_kcalloc() helper
remoteproc: k3-r5: Use devm_ioremap_wc() helper
remoteproc: k3-r5: Use devm_rproc_add() helper
remoteproc: k3-r5: Add devm action to release tsp

Quoted 2 lines
 drivers/remoteproc/ti_k3_r5_remoteproc.c | 88 ++++++++++--------------
 1 file changed, 35 insertions(+), 53 deletions(-

I have applied this set.

Thanks,
Mathieu

Quoted 4 lines
Signature · 2 lines
-- 
2.34.1