Skip to contentSkip to conversation

remoteproc: imx_rproc: allow attaching to running core kicked by the bootloader

PATCH Superseded Hiago De Franco 4 participants 31 messages May 7, 2025 – May 30, 2025

Subsystems (4)

v2 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v2 cover remoteproc: imx_rproc: allow attaching to running core kicked by the bootloader

From: Hiago De Franco <[email protected]>

For the i.MX8X and i.MX8 family SoCs, currently when the remotecore is started by the bootloader and the M core and A core are in the same partition, the driver is not capable to detect the remote core and report the correct state of it.

This series of patches implement an API call to the SCU which will return the power mode of a given resource (M core in this case) and if it is already powered on, the driver will attach to it. This SCU API was already being used inside pmdomain/imx/scu-pd.c driver, therefore it has been moved to firmware/imx/misc.c so it can be accessed by imx_rproc driver.

Finally, the imx_rproc_clk_enable() function was also changed to make it return before dev_clk_get() is called, as it currently generates an SCU fault reset if the remote core is already running and the kernel tries to enable the clock again. These changes are a follow up from a v1 sent to imx_rproc [1] and from a reported regression [2].

[1] https://lore.kernel.org/lkml/[email protected]/
[2] https://lore.kernel.org/lkml/20250404141713.ac2ntcsjsf7epdfa@hiago-nb/

v2:
- Order of commits changed, now "firmware: imx: move get power mode function from scu-pd.c to misc.c" comes at first.
- Updated commit description of "remoteproc: imx_rproc: skip clock enable when M-core is managed by the SCU" as suggested by Peng.
- Removed "imx_rproc_is_on" function and dropped unnecessary include as suggested by Peng.
- Email corrected: [email protected] -> [email protected]

v1:
- https://lore.kernel.org/lkml/[email protected]/

Hiago De Franco (3):
firmware: imx: move get power mode function from scu-pd.c to misc.c
remoteproc: imx_rproc: skip clock enable when M-core is managed by the
SCU
remoteproc: imx_rproc: add power mode check for remote core attachment

 drivers/firmware/imx/misc.c           | 47 +++++++++++++++++++++++++++
 drivers/pmdomain/imx/scu-pd.c         | 29 ++++-------------
 drivers/remoteproc/imx_rproc.c        | 17 ++++++++--
 include/linux/firmware/imx/svc/misc.h |  8 +++++
 4 files changed, 77 insertions(+), 24 deletions(-)
Signature · 1 line
2.39.5
#1 lore raw

PATCH v2 1/3 firmware: imx: move get power mode function from scu-pd.c to misc.c

From: Hiago De Franco <[email protected]>

Move imx_sc_get_pd_power() from pmdomain/imx/scu-pd.c to firmware/imx/misc.c and rename it to imx_sc_pm_get_resource_power_mode() to maintain the same naming logic with other functions in misc.c.

This makes the API available for other use cases. For example, remoteproc/imx_rproc.c can now use this function to check the power mode of the remote core.

---
v2: moved this patch to be the first one

---
 drivers/firmware/imx/misc.c           | 47 +++++++++++++++++++++++++++
 drivers/pmdomain/imx/scu-pd.c         | 29 ++++-------------
 include/linux/firmware/imx/svc/misc.h |  8 +++++
 3 files changed, 62 insertions(+), 22 deletions(-)
drivers/firmware/imx/misc.c +47 -0
index d073cb3ce699..61fcb0066ec9 100644
@@ -37,6 +37,18 @@ struct imx_sc_msg_resp_misc_get_ctrl {
 	u32 val;
 } __packed __aligned(4);
 
+struct imx_sc_msg_req_get_resource_power_mode {
+	struct imx_sc_rpc_msg hdr;
+	union {
+		struct {
+			u16 resource;
+		} req;
+		struct {
+			u8 mode;
+		} resp;
+	} data;
+} __packed __aligned(4);
+
 /*
  * This function sets a miscellaneous control value.
  *
@@ -135,3 +147,38 @@ int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 	return imx_scu_call_rpc(ipc, &msg, true);
 }
 EXPORT_SYMBOL(imx_sc_pm_cpu_start);
+
+/*
+ * This function gets the power mode from a given @resource
+ *
+ * @param[in]     ipc         IPC handle
+ * @param[in]     resource    resource to check the power mode
+ *
+ * @return Returns < 0 for errors or the following for success:
+ * IMX_SC_PM_PW_MODE_OFF  0	Power off
+ * IMX_SC_PM_PW_MODE_STBY 1	Power in standby
+ * IMX_SC_PM_PW_MODE_LP   2	Power in low-power
+ * IMX_SC_PM_PW_MODE_ON   3	Power on
+ *
+ * These are defined under firmware/imx/svc/pm.h
+ */
+int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc, u32 resource)
+{
+	struct imx_sc_msg_req_get_resource_power_mode msg;
+	struct imx_sc_rpc_msg *hdr = &msg.hdr;
+	int ret;
+
+	hdr->ver = IMX_SC_RPC_VERSION;
+	hdr->svc = IMX_SC_RPC_SVC_PM;
+	hdr->func = IMX_SC_PM_FUNC_GET_RESOURCE_POWER_MODE;
+	hdr->size = 2;
+
+	msg.data.req.resource = resource;
+
+	ret = imx_scu_call_rpc(ipc, &msg, true);
+	if (ret)
+		return ret;
+
+	return msg.data.resp.mode;
+}
+EXPORT_SYMBOL(imx_sc_pm_get_resource_power_mode);
drivers/pmdomain/imx/scu-pd.c +7 -22
index 01d465d88f60..945f972e636f 100644
@@ -54,6 +54,7 @@
 #include <dt-bindings/firmware/imx/rsrc.h>
 #include <linux/console.h>
 #include <linux/firmware/imx/sci.h>
+#include <linux/firmware/imx/svc/misc.h>
 #include <linux/firmware/imx/svc/rm.h>
 #include <linux/io.h>
 #include <linux/module.h>
@@ -328,27 +329,6 @@ static void imx_sc_pd_get_console_rsrc(void)
 	imx_con_rsrc = specs.args[0];
 }
 
-static int imx_sc_get_pd_power(struct device *dev, u32 rsrc)
-{
-	struct imx_sc_msg_req_get_resource_power_mode msg;
-	struct imx_sc_rpc_msg *hdr = &msg.hdr;
-	int ret;
-
-	hdr->ver = IMX_SC_RPC_VERSION;
-	hdr->svc = IMX_SC_RPC_SVC_PM;
-	hdr->func = IMX_SC_PM_FUNC_GET_RESOURCE_POWER_MODE;
-	hdr->size = 2;
-
-	msg.data.req.resource = rsrc;
-
-	ret = imx_scu_call_rpc(pm_ipc_handle, &msg, true);
-	if (ret)
-		dev_err(dev, "failed to get power resource %d mode, ret %d\n",
-			rsrc, ret);
-
-	return msg.data.resp.mode;
-}
-
 static int imx_sc_pd_power(struct generic_pm_domain *domain, bool power_on)
 {
 	struct imx_sc_msg_req_set_resource_power_mode msg;
@@ -438,7 +418,12 @@ imx_scu_add_pm_domain(struct device *dev, int idx,
 	if (imx_con_rsrc == sc_pd->rsrc)
 		sc_pd->pd.flags = GENPD_FLAG_RPM_ALWAYS_ON;
 
-	mode = imx_sc_get_pd_power(dev, pd_ranges->rsrc + idx);
+	mode = imx_sc_pm_get_resource_power_mode(pm_ipc_handle,
+						 pd_ranges->rsrc + idx);
+	if (mode < 0)
+		dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+			pd_ranges->rsrc + idx, mode);
+
 	if (mode == IMX_SC_PM_PW_MODE_ON)
 		is_off = false;
 	else
include/linux/firmware/imx/svc/misc.h +8 -0
index 760db08a67fc..376c800a88d0 100644
@@ -55,6 +55,8 @@ int imx_sc_misc_get_control(struct imx_sc_ipc *ipc, u32 resource,
 
 int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 			bool enable, u64 phys_addr);
+
+int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc, u32 resource);
 #else
 static inline int imx_sc_misc_set_control(struct imx_sc_ipc *ipc,
 					  u32 resource, u8 ctrl, u32 val)
@@ -73,5 +75,11 @@ static inline int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 {
 	return -ENOTSUPP;
 }
+
+static inline int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc,
+						    u32 resource)
+{
+	return -ENOTSUPP;
+}
 #endif
 #endif /* _SC_MISC_API_H */
Signature · 1 line
2.39.5
#2 lore raw

Re: [PATCH v2 1/3] firmware: imx: move get power mode function from scu-pd.c to misc.c

Hi Hiago,

Quoted 122 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

Move imx_sc_get_pd_power() from pmdomain/imx/scu-pd.c to firmware/imx/misc.c and rename it to imx_sc_pm_get_resource_power_mode() to maintain the same naming logic with other functions in misc.c.

This makes the API available for other use cases. For example, remoteproc/imx_rproc.c can now use this function to check the power mode of the remote core.

---
v2: moved this patch to be the first one
---

drivers/firmware/imx/misc.c           | 47 +++++++++++++++++++++++++++
drivers/pmdomain/imx/scu-pd.c         | 29 ++++-------------
include/linux/firmware/imx/svc/misc.h |  8 +++++
3 files changed, 62 insertions(+), 22 deletions(-)
drivers/firmware/imx/misc.c +47 -0
index d073cb3ce699..61fcb0066ec9 100644
@@ -37,6 +37,18 @@ struct imx_sc_msg_resp_misc_get_ctrl {
 	u32 val;
 } __packed __aligned(4);
 
+struct imx_sc_msg_req_get_resource_power_mode {
+	struct imx_sc_rpc_msg hdr;
+	union {
+		struct {
+			u16 resource;
+		} req;
+		struct {
+			u8 mode;
+		} resp;
+	} data;
+} __packed __aligned(4);
+
 /*
 * This function sets a miscellaneous control value.
 *
@@ -135,3 +147,38 @@ int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 	return imx_scu_call_rpc(ipc, &msg, true);
 }
 EXPORT_SYMBOL(imx_sc_pm_cpu_start);
+
+/*
+ * This function gets the power mode from a given @resource
+ *
+ * @param[in]     ipc         IPC handle
+ * @param[in]     resource    resource to check the power mode
+ *
+ * @return Returns < 0 for errors or the following for success:
+ * IMX_SC_PM_PW_MODE_OFF  0	Power off
+ * IMX_SC_PM_PW_MODE_STBY 1	Power in standby
+ * IMX_SC_PM_PW_MODE_LP   2	Power in low-power
+ * IMX_SC_PM_PW_MODE_ON   3	Power on
+ *
+ * These are defined under firmware/imx/svc/pm.h
+ */
+int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc, u32 resource)
+{
+	struct imx_sc_msg_req_get_resource_power_mode msg;
+	struct imx_sc_rpc_msg *hdr = &msg.hdr;
+	int ret;
+
+	hdr->ver = IMX_SC_RPC_VERSION;
+	hdr->svc = IMX_SC_RPC_SVC_PM;
+	hdr->func = IMX_SC_PM_FUNC_GET_RESOURCE_POWER_MODE;
+	hdr->size = 2;
+
+	msg.data.req.resource = resource;
+
+	ret = imx_scu_call_rpc(ipc, &msg, true);
+	if (ret)
+		return ret;
+
+	return msg.data.resp.mode;
+}
+EXPORT_SYMBOL(imx_sc_pm_get_resource_power_mode);
drivers/pmdomain/imx/scu-pd.c +1 -21
index 01d465d88f60..945f972e636f 100644
@@ -54,6 +54,7 @@
 #include <dt-bindings/firmware/imx/rsrc.h>
 #include <linux/console.h>
 #include <linux/firmware/imx/sci.h>
+#include <linux/firmware/imx/svc/misc.h>
 #include <linux/firmware/imx/svc/rm.h>
 #include <linux/io.h>
 #include <linux/module.h>
@@ -328,27 +329,6 @@ static void imx_sc_pd_get_console_rsrc(void)
 	imx_con_rsrc = specs.args[0];
 }
 
-static int imx_sc_get_pd_power(struct device *dev, u32 rsrc)
-{
-	struct imx_sc_msg_req_get_resource_power_mode msg;
-	struct imx_sc_rpc_msg *hdr = &msg.hdr;
-	int ret;
-
-	hdr->ver = IMX_SC_RPC_VERSION;
-	hdr->svc = IMX_SC_RPC_SVC_PM;
-	hdr->func = IMX_SC_PM_FUNC_GET_RESOURCE_POWER_MODE;
-	hdr->size = 2;
-
-	msg.data.req.resource = rsrc;
-
-	ret = imx_scu_call_rpc(pm_ipc_handle, &msg, true);
-	if (ret)
-		dev_err(dev, "failed to get power resource %d mode, ret %d\n",
-			rsrc, ret);
-
-	return msg.data.resp.mode;
-}
-
 static int imx_sc_pd_power(struct generic_pm_domain *domain, bool power_on)
 {
 	struct imx_sc_msg_req_set_resource_power_mode msg;

drivers/pmdomain/imx/scu-pd.c imx_scu_add_pm_domain(struct device *dev, int idx, · line 421

Quoted 5 lines
drivers/pmdomain/imx/scu-pd.c +1 -1
@@ -438,7 +418,12 @@ imx_scu_add_pm_domain(struct device *dev, int idx,
 	if (imx_con_rsrc == sc_pd->rsrc)
 		sc_pd->pd.flags = GENPD_FLAG_RPM_ALWAYS_ON;
 
-	mode = imx_sc_get_pd_power(dev, pd_ranges->rsrc + idx);
+	mode = imx_sc_pm_get_resource_power_mode(pm_ipc_handle,

I mean, not change scu-pd.c in this patchset. After imx_sc_pm_get_resource_power_mode landed, you could update scu-pd.c.

Otherwise you are touch three trees which are a bit hard to manage for maintainers:
scu-pd.c under linux-pm tree
misc.c under linux-imx tree, I think using power.c should be better. imx_rproc.c under linux-remoteproc tree

My suggestion was that
patch 1: firmware: imx: Introduce imx_sc_pm_get_resource_power_mode patch 2: remoteproc: imx_rproc: Detecting power mode of M4 for i.MX8QX Patch 3: pmdomain: scu-pd: migrate to use imx common pm API

Patch 3 could be separate out in future patchset to avoid handling three trees.

Regards,
Peng

Quoted 36 lines
drivers/pmdomain/imx/scu-pd.c +5 -0
+						 pd_ranges->rsrc + idx);
+	if (mode < 0)
+		dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+			pd_ranges->rsrc + idx, mode);
+
 	if (mode == IMX_SC_PM_PW_MODE_ON)
 		is_off = false;
 	else
include/linux/firmware/imx/svc/misc.h +8 -0
index 760db08a67fc..376c800a88d0 100644
@@ -55,6 +55,8 @@ int imx_sc_misc_get_control(struct imx_sc_ipc *ipc, u32 resource,
 
 int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 			bool enable, u64 phys_addr);
+
+int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc, u32 resource);
 #else
 static inline int imx_sc_misc_set_control(struct imx_sc_ipc *ipc,
 					  u32 resource, u8 ctrl, u32 val)
@@ -73,5 +75,11 @@ static inline int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 {
 	return -ENOTSUPP;
 }
+
+static inline int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc,
+						    u32 resource)
+{
+	return -ENOTSUPP;
+}
 #endif
 #endif /* _SC_MISC_API_H */
Signature · 1 line
2.39.5
#3 lore raw

Re: [PATCH v2 1/3] firmware: imx: move get power mode function from scu-pd.c to misc.c

Hi Peng,

Quoted 140 lines · Peng Fan

Hi Hiago,

Quoted 122 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

Move imx_sc_get_pd_power() from pmdomain/imx/scu-pd.c to firmware/imx/misc.c and rename it to imx_sc_pm_get_resource_power_mode() to maintain the same naming logic with other functions in misc.c.

This makes the API available for other use cases. For example, remoteproc/imx_rproc.c can now use this function to check the power mode of the remote core.

---
v2: moved this patch to be the first one
---

drivers/firmware/imx/misc.c           | 47 +++++++++++++++++++++++++++
drivers/pmdomain/imx/scu-pd.c         | 29 ++++-------------
include/linux/firmware/imx/svc/misc.h |  8 +++++
3 files changed, 62 insertions(+), 22 deletions(-)
drivers/firmware/imx/misc.c +47 -0
index d073cb3ce699..61fcb0066ec9 100644
@@ -37,6 +37,18 @@ struct imx_sc_msg_resp_misc_get_ctrl {
 	u32 val;
 } __packed __aligned(4);
 
+struct imx_sc_msg_req_get_resource_power_mode {
+	struct imx_sc_rpc_msg hdr;
+	union {
+		struct {
+			u16 resource;
+		} req;
+		struct {
+			u8 mode;
+		} resp;
+	} data;
+} __packed __aligned(4);
+
 /*
 * This function sets a miscellaneous control value.
 *
@@ -135,3 +147,38 @@ int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 	return imx_scu_call_rpc(ipc, &msg, true);
 }
 EXPORT_SYMBOL(imx_sc_pm_cpu_start);
+
+/*
+ * This function gets the power mode from a given @resource
+ *
+ * @param[in]     ipc         IPC handle
+ * @param[in]     resource    resource to check the power mode
+ *
+ * @return Returns < 0 for errors or the following for success:
+ * IMX_SC_PM_PW_MODE_OFF  0	Power off
+ * IMX_SC_PM_PW_MODE_STBY 1	Power in standby
+ * IMX_SC_PM_PW_MODE_LP   2	Power in low-power
+ * IMX_SC_PM_PW_MODE_ON   3	Power on
+ *
+ * These are defined under firmware/imx/svc/pm.h
+ */
+int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc, u32 resource)
+{
+	struct imx_sc_msg_req_get_resource_power_mode msg;
+	struct imx_sc_rpc_msg *hdr = &msg.hdr;
+	int ret;
+
+	hdr->ver = IMX_SC_RPC_VERSION;
+	hdr->svc = IMX_SC_RPC_SVC_PM;
+	hdr->func = IMX_SC_PM_FUNC_GET_RESOURCE_POWER_MODE;
+	hdr->size = 2;
+
+	msg.data.req.resource = resource;
+
+	ret = imx_scu_call_rpc(ipc, &msg, true);
+	if (ret)
+		return ret;
+
+	return msg.data.resp.mode;
+}
+EXPORT_SYMBOL(imx_sc_pm_get_resource_power_mode);
drivers/pmdomain/imx/scu-pd.c +1 -21
index 01d465d88f60..945f972e636f 100644
@@ -54,6 +54,7 @@
 #include <dt-bindings/firmware/imx/rsrc.h>
 #include <linux/console.h>
 #include <linux/firmware/imx/sci.h>
+#include <linux/firmware/imx/svc/misc.h>
 #include <linux/firmware/imx/svc/rm.h>
 #include <linux/io.h>
 #include <linux/module.h>
@@ -328,27 +329,6 @@ static void imx_sc_pd_get_console_rsrc(void)
 	imx_con_rsrc = specs.args[0];
 }
 
-static int imx_sc_get_pd_power(struct device *dev, u32 rsrc)
-{
-	struct imx_sc_msg_req_get_resource_power_mode msg;
-	struct imx_sc_rpc_msg *hdr = &msg.hdr;
-	int ret;
-
-	hdr->ver = IMX_SC_RPC_VERSION;
-	hdr->svc = IMX_SC_RPC_SVC_PM;
-	hdr->func = IMX_SC_PM_FUNC_GET_RESOURCE_POWER_MODE;
-	hdr->size = 2;
-
-	msg.data.req.resource = rsrc;
-
-	ret = imx_scu_call_rpc(pm_ipc_handle, &msg, true);
-	if (ret)
-		dev_err(dev, "failed to get power resource %d mode, ret %d\n",
-			rsrc, ret);
-
-	return msg.data.resp.mode;
-}
-
 static int imx_sc_pd_power(struct generic_pm_domain *domain, bool power_on)
 {
 	struct imx_sc_msg_req_set_resource_power_mode msg;

drivers/pmdomain/imx/scu-pd.c imx_scu_add_pm_domain(struct device *dev, int idx, · line 421

Quoted 5 lines
drivers/pmdomain/imx/scu-pd.c +1 -1
@@ -438,7 +418,12 @@ imx_scu_add_pm_domain(struct device *dev, int idx,
 	if (imx_con_rsrc == sc_pd->rsrc)
 		sc_pd->pd.flags = GENPD_FLAG_RPM_ALWAYS_ON;
 
-	mode = imx_sc_get_pd_power(dev, pd_ranges->rsrc + idx);
+	mode = imx_sc_pm_get_resource_power_mode(pm_ipc_handle,

I mean, not change scu-pd.c in this patchset. After imx_sc_pm_get_resource_power_mode landed, you could update scu-pd.c.

Otherwise you are touch three trees which are a bit hard to manage for maintainers:
scu-pd.c under linux-pm tree
misc.c under linux-imx tree, I think using power.c should be better. imx_rproc.c under linux-remoteproc tree

Quoted 6 lines

My suggestion was that
patch 1: firmware: imx: Introduce imx_sc_pm_get_resource_power_mode patch 2: remoteproc: imx_rproc: Detecting power mode of M4 for i.MX8QX Patch 3: pmdomain: scu-pd: migrate to use imx common pm API

Patch 3 could be separate out in future patchset to avoid handling three trees.

Ok, got it. I will prepare the v3 with your suggestions. Thanks.

Quoted 31 lines

Regards,
Peng

Quoted 28 lines
drivers/pmdomain/imx/scu-pd.c +5 -0
+						 pd_ranges->rsrc + idx);
+	if (mode < 0)
+		dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+			pd_ranges->rsrc + idx, mode);
+
 	if (mode == IMX_SC_PM_PW_MODE_ON)
 		is_off = false;
 	else
include/linux/firmware/imx/svc/misc.h +5 -0
index 760db08a67fc..376c800a88d0 100644
@@ -55,6 +55,8 @@ int imx_sc_misc_get_control(struct imx_sc_ipc *ipc, u32 resource,
 
 int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 			bool enable, u64 phys_addr);
+
+int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc, u32 resource);
 #else
 static inline int imx_sc_misc_set_control(struct imx_sc_ipc *ipc,
 					  u32 resource, u8 ctrl, u32 val)
@@ -73,5 +75,11 @@ static inline int imx_sc_pm_cpu_start(struct imx_sc_ipc *ipc, u32 resource,
 {
 	return -ENOTSUPP;
 }
+
+static inline int imx_sc_pm_get_resource_power_mode(struct imx_sc_ipc *ipc,
+						    u32 resource)
Quoted 8 lines
Quoted 8 lines
include/linux/firmware/imx/svc/misc.h +3 -0
+{
+	return -ENOTSUPP;
+}
 #endif
 #endif /* _SC_MISC_API_H */
Signature · 1 line
2.39.5

Best Regards,
Hiago.

#4 lore raw

PATCH v2 2/3 remoteproc: imx_rproc: skip clock enable when M-core is managed by the SCU

In reply to Hiago De Franco

From: Hiago De Franco <[email protected]>

For the i.MX8X and i.MX8 family SoCs, when the M-core is powered up by the bootloader, M-core and Linux are in same SCFW (System Controller Firmware) partition, so linux has permission to control M-core.

But when M-core is started, the SCFW will automatically enable the clock and configure the rate, and any users that want to enable the clock will get error 'LOCKED' from SCFW. So current imx_rproc.c probe function fails because clk_prepare_enable also fails. With that, the M-core power domain is powered off when it is still running, causing a SCU (System Controller Unit) fault reset, and the system restarts.

To address the issue, ignore handling the clk for i.MX8X and i.MX8 M-core, because SCFW will automatically enable and configure the clock.

---
v2: commit description updated, as suggested. Fixed Peng Fan email.

---
 drivers/remoteproc/imx_rproc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/remoteproc/imx_rproc.c +2 -2
index 74299af1d7f1..627e57a88db2 100644
@@ -1029,8 +1029,8 @@ static int imx_rproc_clk_enable(struct imx_rproc *priv)
 	struct device *dev = priv->dev;
 	int ret;
 
-	/* Remote core is not under control of Linux */
-	if (dcfg->method == IMX_RPROC_NONE)
+	/* Remote core is not under control of Linux or it is managed by SCU API */
+	if (dcfg->method == IMX_RPROC_NONE || dcfg->method == IMX_RPROC_SCU_API)
 		return 0;
 
 	priv->clk = devm_clk_get(dev, NULL);
Signature · 1 line
2.39.5
#5 lore raw

PATCH v2 3/3 remoteproc: imx_rproc: add power mode check for remote core attachment

In reply to Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +13 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
 			if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
 				return -EINVAL;
 
+			/*
+			 * If remote core is already running (e.g. kicked by
+			 * the bootloader), attach to it.
+			 */
+			ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+								priv->rsrc_id);
+			if (ret < 0)
+				dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+					priv->rsrc_id, ret);
+
+			if (ret == IMX_SC_PM_PW_MODE_ON)
+				priv->rproc->state = RPROC_DETACHED;
+
 			return imx_rproc_attach_pd(priv);
 		}
 
Signature · 1 line
2.39.5
#6 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines
drivers/remoteproc/imx_rproc.c +0 -1
                }
 
--

2.39.5

Kind regards
Uffe

#7 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

Quoted 9 lines
Quoted 6 lines
drivers/remoteproc/imx_rproc.c +0 -1
                }
 
--

2.39.5

Kind regards
Uffe

Best Regards,
Hiago.

#8 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

If you think it can work, I can help to cook a formal patch for the above helper that you can fold into your series. Let me know.

Quoted 5 lines

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

I see, thanks for clarifying!

Kind regards
Uffe

#9 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

Quoted 13 lines

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

If you think it can work, I can help to cook a formal patch for the above helper that you can fold into your series. Let me know.

Quoted 5 lines

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

Quoted 4 lines

I see, thanks for clarifying!

Kind regards
Uffe

Thank you!
Hiago.

#10 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 103 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON
Quoted 7 lines

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

Ulf's new API dev_pm_genpd_is_on needs to run after power domain attached.

But if run after power domain attached, there is no API to know whether M4 is kicked by bootloader or now.

Even imx_rproc_attach_pd has a check for single power domain, but I just give a look again on current i.MX8QM/QX, all are using two power domain entries.

Quoted 4 lines
Quoted 4 lines

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

I think still need to export firmware API. My idea is
1. introduce a new firmware API and put under firmware/imx/power.c
2. Use this new firmware API in imx_rproc.c
3. Replace scu-pd.c to use this new firmware API.

Or
1. Export the API in scu-pd.c
2. Use the API in imx_rproc.c

With approach two, you need to handle them in three trees in one patchset:
imx/pd/rproc.

With approach one, you need to handle two trees in one patchset: imx/rproc tree, then after done, pd tree

Regards,
Peng

Quoted 17 lines
Quoted 10 lines

If you think it can work, I can help to cook a formal patch for the above helper that you can fold into your series. Let me know.

Quoted 5 lines

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

Quoted 4 lines

I see, thanks for clarifying!

Kind regards
Uffe

Thank you!
Hiago.

#11 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Hi Peng,

Quoted 25 lines · Peng Fan

Ulf's new API dev_pm_genpd_is_on needs to run after power domain attached.

But if run after power domain attached, there is no API to know whether M4 is kicked by bootloader or now.

Even imx_rproc_attach_pd has a check for single power domain, but I just give a look again on current i.MX8QM/QX, all are using two power domain entries.

Quoted 4 lines
Quoted 4 lines

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

I think still need to export firmware API. My idea is
1. introduce a new firmware API and put under firmware/imx/power.c
2. Use this new firmware API in imx_rproc.c
3. Replace scu-pd.c to use this new firmware API.

Or
1. Export the API in scu-pd.c
2. Use the API in imx_rproc.c

Quoted 5 lines

With approach two, you need to handle them in three trees in one patchset:
imx/pd/rproc.

With approach one, you need to handle two trees in one patchset: imx/rproc tree, then after done, pd tree

This patch series is already implementing approach one, as I understand, right?

The difference is I moved this API from drivers/pmdomain/imx/scu-pd.c to firmware/imx/misc.c, and exported it there. But you think I should create this new file firmware/imx/power.c and move the API from scu-pd.c to power.c, is my understanding correct?

Quoted 3 lines

Regards,
Peng

Best Regards,
Hiago.

#12 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

In reply to Peng Fan
Quoted 107 lines · Peng Fan
Quoted 105 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

Quoted 7 lines
Quoted 5 lines
[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

Ulf's new API dev_pm_genpd_is_on needs to run after power domain attached.

Correct, but you need to provide the correct "dev" to it. See my other reply to Hiago.

Quoted 3 lines

But if run after power domain attached, there is no API to know whether M4 is kicked by bootloader or now.

As long as you have multiple PM domains attached for a device, genpd will *not* power on the PM domain(s).

Genpd does a power-on in the single PM domain case (for legacy reasons), but that should not be a problem here, right?

So what am I missing?

Quoted 20 lines

Even imx_rproc_attach_pd has a check for single power domain, but I just give a look again on current i.MX8QM/QX, all are using two power domain entries.

Quoted 4 lines
Quoted 4 lines

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

I think still need to export firmware API. My idea is
1. introduce a new firmware API and put under firmware/imx/power.c
2. Use this new firmware API in imx_rproc.c
3. Replace scu-pd.c to use this new firmware API.

Or
1. Export the API in scu-pd.c
2. Use the API in imx_rproc.c

Quoted 25 lines

With approach two, you need to handle them in three trees in one patchset:
imx/pd/rproc.

With approach one, you need to handle two trees in one patchset: imx/rproc tree, then after done, pd tree

Regards,
Peng

Quoted 17 lines
Quoted 10 lines

If you think it can work, I can help to cook a formal patch for the above helper that you can fold into your series. Let me know.

Quoted 5 lines

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

Quoted 4 lines

I see, thanks for clarifying!

Kind regards
Uffe

Thank you!
Hiago.

Kind regards
Uffe

#13 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 124 lines · Ulf Hansson
Quoted 107 lines · Peng Fan
Quoted 105 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

Quoted 7 lines
Quoted 5 lines
[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

Ulf's new API dev_pm_genpd_is_on needs to run after power domain attached.

Correct, but you need to provide the correct "dev" to it. See my other reply to Hiago.

Quoted 3 lines

But if run after power domain attached, there is no API to know whether M4 is kicked by bootloader or now.

Quoted 7 lines

As long as you have multiple PM domains attached for a device, genpd will *not* power on the PM domain(s).

Genpd does a power-on in the single PM domain case (for legacy reasons), but that should not be a problem here, right?

So what am I missing?

Aha, PD_FLAG_DEV_LINK_ON is being used when you attach to the PM domains. I would re-work things in the driver (if needed) so you can avoid using this flag, then the PM domains should stay power-off until there is a call to pm_runtime_get_sync().

[...]

Kind regards
Uffe

#14 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

In reply to Hiago De Franco
Quoted 104 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON
Quoted 7 lines

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

As your device has multiple PM domains and those gets attached with dev_pm_domain_attach_list(), the device(s) that you should use with dev_pm_genpd_is_on() are in imx_rproc->pd_list->pd_devs[n].

Quoted 14 lines
Quoted 14 lines

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

If you think it can work, I can help to cook a formal patch for the above helper that you can fold into your series. Let me know.

Quoted 5 lines

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

Quoted 7 lines
Quoted 4 lines

I see, thanks for clarifying!

Kind regards
Uffe

Thank you!
Hiago.

Kind regards
Uffe

#15 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Hi Ulf,

Quoted 109 lines · Ulf Hansson
Quoted 108 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF
Quoted 7 lines
Quoted 3 lines

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

As your device has multiple PM domains and those gets attached with dev_pm_domain_attach_list(), the device(s) that you should use with dev_pm_genpd_is_on() are in imx_rproc->pd_list->pd_devs[n].

Ok got it, thanks for sharing.

I just send the v3 with the changes Peng proposed (here
https://lore.kernel.org/lkml/[email protected]/T/#t),
but I am a bit confused which path we should take, the initial approach proposed or using these PD functions. Maybe we can discuss this in the new v3 patch series?

Quoted 17 lines
Quoted 17 lines
Quoted 17 lines

In this way we don't need to export unnecessary firmware functions from firmware/imx/misc.c, as patch1/3 does.

If you think it can work, I can help to cook a formal patch for the above helper that you can fold into your series. Let me know.

Quoted 5 lines

With that we can correcly attach to the remote core running, which was not possible before, where the function returns at "return imx_rproc_attach_pd(priv);" with the RPROC_OFFLINE state to rproc_validate().

I see, thanks for clarifying!

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Kind regards
Uffe

Thank you!
Hiago.

Kind regards
Uffe

Thanks,
Hiago.

#16 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 121 lines · Hiago De Franco

Hi Ulf,

Quoted 109 lines · Ulf Hansson
Quoted 108 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF
Quoted 7 lines
Quoted 3 lines

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

As your device has multiple PM domains and those gets attached with dev_pm_domain_attach_list(), the device(s) that you should use with dev_pm_genpd_is_on() are in imx_rproc->pd_list->pd_devs[n].

Quoted 7 lines

Ok got it, thanks for sharing.

I just send the v3 with the changes Peng proposed (here
https://lore.kernel.org/lkml/[email protected]/T/#t),
but I am a bit confused which path we should take, the initial approach proposed or using these PD functions. Maybe we can discuss this in the new v3 patch series?

I think it would be better if we can avoid sharing low-level firmware functions for PM domains. I am worried that they may become abused for other future use-cases.

So, if possible, I would rather make us try to use dev_pm_genpd_is_on() (or something along those lines), but let's see what Peng thinks about it before we make the decision.

[...]

Kind regards
Uffe

#17 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Hi Ulf,

Quoted 130 lines · Ulf Hansson
Quoted 121 lines · Hiago De Franco

Hi Ulf,

Quoted 109 lines · Ulf Hansson
Quoted 108 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF
Quoted 7 lines
Quoted 3 lines

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

As your device has multiple PM domains and those gets attached with dev_pm_domain_attach_list(), the device(s) that you should use with dev_pm_genpd_is_on() are in imx_rproc->pd_list->pd_devs[n].

Quoted 7 lines

Ok got it, thanks for sharing.

I just send the v3 with the changes Peng proposed (here
https://lore.kernel.org/lkml/[email protected]/T/#t),
but I am a bit confused which path we should take, the initial approach proposed or using these PD functions. Maybe we can discuss this in the new v3 patch series?

Quoted 7 lines

I think it would be better if we can avoid sharing low-level firmware functions for PM domains. I am worried that they may become abused for other future use-cases.

So, if possible, I would rather make us try to use dev_pm_genpd_is_on() (or something along those lines), but let's see what Peng thinks about it before we make the decision.

There are two power domains for this m4:
power-domains = <&pd IMX_SC_R_M4_0_PID0>, <&pd IMX_SC_R_M4_0_MU_1A>;

So before attach the pd, dev_pm_genpd_is_on should also return false per my understanding. If run dev_pm_genpd_is_on after attaching the pd, the pd will be powered on. So we are not able to know whether M4 is started by bootloader or not.

Hiago's case needs the real power status before attaching the pd to set remoteproc as DETACHED(M4 kicked by bootloader) or OFFLINE( M4 not kicked by bootloader) state.

Seems there is no other SCFW API to check whether M4 is started by bootloader.

I not have good idea as of now except directly checking the real power status to indicate M4 started by bootloader or not. Or using a device tree property runtime added by bootloader,
saying "fsl,rproc-started"?

Thanks,
Peng

Quoted 5 lines

[...]

Kind regards
Uffe

#18 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 141 lines · Peng Fan

Hi Ulf,

Quoted 130 lines · Ulf Hansson
Quoted 121 lines · Hiago De Franco

Hi Ulf,

Quoted 109 lines · Ulf Hansson
Quoted 108 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF
Quoted 7 lines
Quoted 3 lines

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

As your device has multiple PM domains and those gets attached with dev_pm_domain_attach_list(), the device(s) that you should use with dev_pm_genpd_is_on() are in imx_rproc->pd_list->pd_devs[n].

Quoted 7 lines

Ok got it, thanks for sharing.

I just send the v3 with the changes Peng proposed (here
https://lore.kernel.org/lkml/[email protected]/T/#t),
but I am a bit confused which path we should take, the initial approach proposed or using these PD functions. Maybe we can discuss this in the new v3 patch series?

Quoted 7 lines

I think it would be better if we can avoid sharing low-level firmware functions for PM domains. I am worried that they may become abused for other future use-cases.

So, if possible, I would rather make us try to use dev_pm_genpd_is_on() (or something along those lines), but let's see what Peng thinks about it before we make the decision.

Quoted 7 lines

There are two power domains for this m4:
power-domains = <&pd IMX_SC_R_M4_0_PID0>, <&pd IMX_SC_R_M4_0_MU_1A>;

So before attach the pd, dev_pm_genpd_is_on should also return false per my understanding. If run dev_pm_genpd_is_on after attaching the pd, the pd will be powered on. So we are not able to know whether M4 is started by bootloader or not.

Could we use PD_FLAG_NO_DEV_LINK when attach the PD, then use dev_pm_genpd_is_on to detect the status of genpd?

we set is_off as true when pm_genpd_init if the PD is physical ON.

Thanks,
Peng

Quoted 23 lines

Hiago's case needs the real power status before attaching the pd to set remoteproc as DETACHED(M4 kicked by bootloader) or OFFLINE( M4 not kicked by bootloader) state.

Seems there is no other SCFW API to check whether M4 is started by bootloader.

I not have good idea as of now except directly checking the real power status to indicate M4 started by bootloader or not. Or using a device tree property runtime added by bootloader,
saying "fsl,rproc-started"?

Thanks,
Peng

Quoted 5 lines

[...]

Kind regards
Uffe

#19 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 152 lines · Peng Fan
Quoted 141 lines · Peng Fan

Hi Ulf,

Quoted 130 lines · Ulf Hansson
Quoted 121 lines · Hiago De Franco

Hi Ulf,

Quoted 109 lines · Ulf Hansson
Quoted 108 lines · Hiago De Franco
Quoted 89 lines · Ulf Hansson
Quoted 61 lines · Hiago De Franco

Hello,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 7 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines

I am not sure if I understood correcly, let me know if I missed something. From my understanding in this case it does not matter, since the RPROC_DETACHED will only be a flag to trigger the attach callback from rproc_validate(), when rproc_add() is called inside remoteproc_core.c.

Okay, I see.

To me, it sounds like we should introduce a new genpd helper function instead. Something along the lines of this (drivers/pmdomain/core.c)

bool dev_pm_genpd_is_on(struct device *dev)

{
        struct generic_pm_domain *genpd;
        bool is_on;

        genpd = dev_to_genpd_safe(dev);
        if (!genpd)
                return false;

        genpd_lock(genpd);
        is_on = genpd_status_on(genpd);
        genpd_unlock(genpd);

        return is_on;
}
Quoted 5 lines

After imx_rproc_attach_pd() has run, we have the devices that correspond to the genpd(s). Those can then be passed as in-parameters to the above function to get the power-state of their PM domains (genpds). Based on that, we can decide if priv->rproc->state should be to RPROC_DETACHED or not. Right?

Got your idea, I think it should work yes, I am not so sure how. From what I can see these power domains are managed by drivers/pmdomain/imx/scu-pd.c and by enabling the debug messages I can see the power mode is correct when the remote core is powered on:

[    0.317369] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_ON

and powered off:

[    0.314953] imx-scu-pd system-controller:power-controller: cm40-pid0 : IMX_SC_PM_PW_MODE_OFF
Quoted 7 lines
Quoted 3 lines

But I cannot see how to integrate this into the dev_pm_genpd_is_on() you proposed. For a quick check, I added this function and it always return NULL at dev_to_genpd_safe(). Can you help me to understand this part?

As your device has multiple PM domains and those gets attached with dev_pm_domain_attach_list(), the device(s) that you should use with dev_pm_genpd_is_on() are in imx_rproc->pd_list->pd_devs[n].

Quoted 7 lines

Ok got it, thanks for sharing.

I just send the v3 with the changes Peng proposed (here
https://lore.kernel.org/lkml/[email protected]/T/#t),
but I am a bit confused which path we should take, the initial approach proposed or using these PD functions. Maybe we can discuss this in the new v3 patch series?

Quoted 7 lines

I think it would be better if we can avoid sharing low-level firmware functions for PM domains. I am worried that they may become abused for other future use-cases.

So, if possible, I would rather make us try to use dev_pm_genpd_is_on() (or something along those lines), but let's see what Peng thinks about it before we make the decision.

Quoted 7 lines

There are two power domains for this m4:
power-domains = <&pd IMX_SC_R_M4_0_PID0>, <&pd IMX_SC_R_M4_0_MU_1A>;

So before attach the pd, dev_pm_genpd_is_on should also return false per my understanding. If run dev_pm_genpd_is_on after attaching the pd, the pd will be powered on. So we are not able to know whether M4 is started by bootloader or not.

Quoted 5 lines

Could we use PD_FLAG_NO_DEV_LINK when attach the PD, then use dev_pm_genpd_is_on to detect the status of genpd?

we set is_off as true when pm_genpd_init if the PD is physical ON.

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

[...]

Kind regards
Uffe

#20 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Hi Ulf,

Quoted 20 lines · Ulf Hansson

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

Quoted 4 lines

[...]

Kind regards
Uffe

I did some tests here and I might be missing something. I used the dev_pm_genpd_is_on() inside imx_rproc.c with the following changes:

@@ -902,7 +902,12 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)

        if (dev->pm_domain)
                return 0;

        ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);

+ printk("hfranco: returned pd devs is %d", ret);
+ for (int i = 0; i < ret; i++) {
+ test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+ printk("hfranco: returned value is %d", test);
+ }

        return ret < 0 ? ret : 0;

}

This was a quick test to check the returned value, and it always return 1 for both pds, even if I did not boot the remote core.

So I was wondering if it was because of PD_FLAG_DEV_LINK_ON, I removed it and passed NULL to dev_pm_domain_attach_list(). Booting the kernel now it correctly reports 0 for both pds, however when I start the remote core with a hello world firmware and boot the kernel, the CPU resets with a fault reset ("Reset cause: SCFW fault reset").

I added both pm functions to probe, just to test:

@@ -1152,6 +1158,9 @@ static int imx_rproc_probe(struct platform_device *pdev)

                goto err_put_clk;
        }

+ pm_runtime_enable(dev);
+ pm_runtime_get_sync(dev);
+
return 0

Now the kernel boot with the remote core running, but it still returns 0 from dev_pm_genpd_is_on(). So basically now it always returns 0, with or without the remote core running.

I tried to move pm_runtime_get_sync() to .prepare function but it make the kernel not boot anymore (with the SCU fault reset).

Do you have any suggestions? Am I doing something wrong with these PDs?

Best regards,
Hiago.

#21 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 48 lines · Hiago De Franco

Hi Ulf,

Quoted 20 lines · Ulf Hansson

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

Quoted 4 lines

[...]

Kind regards
Uffe

I did some tests here and I might be missing something. I used the dev_pm_genpd_is_on() inside imx_rproc.c with the following changes:

(quoted diff) +5 -0
@@ -902,7 +902,12 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
        if (dev->pm_domain)
                return 0;
 
        ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+       printk("hfranco: returned pd devs is %d", ret);
+       for (int i = 0; i < ret; i++) {
+               test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+               printk("hfranco: returned value is %d", test);
+       }
        return ret < 0 ? ret : 0;
 }
 

This was a quick test to check the returned value, and it always return 1 for both pds, even if I did not boot the remote core.

Quoted 2 lines

So I was wondering if it was because of PD_FLAG_DEV_LINK_ON, I removed it and passed NULL to dev_pm_domain_attach_list().

Right, that's exactly what we should be doing.

Quoted 8 lines

Booting the kernel
now it correctly reports 0 for both pds, however when I start the remote core with a hello world firmware and boot the kernel, the CPU resets with a fault reset ("Reset cause: SCFW fault reset").

I added both pm functions to probe, just to test:

Review static int imx_rproc_probe(struct platform_device *pdev) · line 1163

Quoted 6 lines
(quoted diff) +3 -0
@@ -1152,6 +1158,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
                goto err_put_clk;
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);
+

Indeed, calling pm_runtime_enable() and then pm_runtime_get_sync() should turn on the PM domains for the device, which I assume is needed at some point.

Although, I wonder if this may be a bit too late, I would expect that you at least need to call these *before* the call to rproc_add(), as I assume the rproc-core may start using the device/driver beyond that point.

Quoted 5 lines
(quoted diff) +0 -0
        return 0
 

Now the kernel boot with the remote core running, but it still returns 0 from dev_pm_genpd_is_on(). So basically now it always returns 0, with or without the remote core running.

dev_pm_genpd_is_on() is returning the current status of the PM domain (genpd) for the device.

Could it be that the genpd provider doesn't register its PM domains with the state that the HW is really in? pm_genpd_init() is the call that allows the genpd provider to specify the initial state.

I think we need Peng's help here to understand what goes on.

Quoted 3 lines

I tried to move pm_runtime_get_sync() to .prepare function but it make the kernel not boot anymore (with the SCU fault reset).

Try move pm_runtime_enable() before rproc_add().

Quoted 5 lines

Do you have any suggestions? Am I doing something wrong with these PDs?

Best regards,
Hiago.

Kind regards
Uffe

#22 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 91 lines · Ulf Hansson
Quoted 48 lines · Hiago De Franco

Hi Ulf,

Quoted 20 lines · Ulf Hansson

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

Quoted 4 lines

[...]

Kind regards
Uffe

I did some tests here and I might be missing something. I used the dev_pm_genpd_is_on() inside imx_rproc.c with the following changes:

(quoted diff) +5 -0
@@ -902,7 +902,12 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
        if (dev->pm_domain)
                return 0;
 
        ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+       printk("hfranco: returned pd devs is %d", ret);
+       for (int i = 0; i < ret; i++) {
+               test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+               printk("hfranco: returned value is %d", test);
+       }
        return ret < 0 ? ret : 0;
 }
 

This was a quick test to check the returned value, and it always return 1 for both pds, even if I did not boot the remote core.

Quoted 2 lines

So I was wondering if it was because of PD_FLAG_DEV_LINK_ON, I removed it and passed NULL to dev_pm_domain_attach_list().

Right, that's exactly what we should be doing.

Quoted 8 lines

Booting the kernel
now it correctly reports 0 for both pds, however when I start the remote core with a hello world firmware and boot the kernel, the CPU resets with a fault reset ("Reset cause: SCFW fault reset").

I added both pm functions to probe, just to test:

Review static int imx_rproc_probe(struct platform_device *pdev) · line 1163

Quoted 6 lines
(quoted diff) +3 -0
@@ -1152,6 +1158,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
                goto err_put_clk;
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);
+

Indeed, calling pm_runtime_enable() and then pm_runtime_get_sync() should turn on the PM domains for the device, which I assume is needed at some point.

Although, I wonder if this may be a bit too late, I would expect that you at least need to call these *before* the call to rproc_add(), as I assume the rproc-core may start using the device/driver beyond that point.

Quoted 5 lines
(quoted diff) +0 -0
        return 0
 

Now the kernel boot with the remote core running, but it still returns 0 from dev_pm_genpd_is_on(). So basically now it always returns 0, with or without the remote core running.

dev_pm_genpd_is_on() is returning the current status of the PM domain (genpd) for the device.

Could it be that the genpd provider doesn't register its PM domains with the state that the HW is really in? pm_genpd_init() is the call that allows the genpd provider to specify the initial state.

Quoted 7 lines

I think we need Peng's help here to understand what goes on.

Quoted 3 lines

I tried to move pm_runtime_get_sync() to .prepare function but it make the kernel not boot anymore (with the SCU fault reset).

Try move pm_runtime_enable() before rproc_add().

Thanks Ulf, that indeed made it work, at least now the kernel does not reset anymore with the SCU fault reset. However I am still only getting 0 from dev_pm_genpd_is_on(), no matter what the state of the remote core. Maybe I am missing something in between?

Peng, do you know what could be the issue here?

Quoted 9 lines
Quoted 6 lines

Do you have any suggestions? Am I doing something wrong with these PDs?

Best regards,
Hiago.

Kind regards
Uffe

Best regards,
Hiago

#23 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 98 lines · Hiago De Franco
Quoted 97 lines · Ulf Hansson
Quoted 48 lines · Hiago De Franco

Hi Ulf,

Quoted 20 lines · Ulf Hansson

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

Quoted 4 lines

[...]

Kind regards
Uffe

I did some tests here and I might be missing something. I used the dev_pm_genpd_is_on() inside imx_rproc.c with the following changes:

(quoted diff) +5 -0
@@ -902,7 +902,12 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
        if (dev->pm_domain)
                return 0;
 
        ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+       printk("hfranco: returned pd devs is %d", ret);
+       for (int i = 0; i < ret; i++) {
+               test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+               printk("hfranco: returned value is %d", test);
+       }
        return ret < 0 ? ret : 0;
 }
 

This was a quick test to check the returned value, and it always return 1 for both pds, even if I did not boot the remote core.

Quoted 2 lines

So I was wondering if it was because of PD_FLAG_DEV_LINK_ON, I removed it and passed NULL to dev_pm_domain_attach_list().

Right, that's exactly what we should be doing.

Quoted 8 lines

Booting the kernel
now it correctly reports 0 for both pds, however when I start the remote core with a hello world firmware and boot the kernel, the CPU resets with a fault reset ("Reset cause: SCFW fault reset").

I added both pm functions to probe, just to test:

Review static int imx_rproc_probe(struct platform_device *pdev) · line 1163

Quoted 6 lines
(quoted diff) +3 -0
@@ -1152,6 +1158,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
                goto err_put_clk;
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);
+

Indeed, calling pm_runtime_enable() and then pm_runtime_get_sync() should turn on the PM domains for the device, which I assume is needed at some point.

Although, I wonder if this may be a bit too late, I would expect that you at least need to call these *before* the call to rproc_add(), as I assume the rproc-core may start using the device/driver beyond that point.

Quoted 5 lines
(quoted diff) +0 -0
        return 0
 

Now the kernel boot with the remote core running, but it still returns 0 from dev_pm_genpd_is_on(). So basically now it always returns 0, with or without the remote core running.

dev_pm_genpd_is_on() is returning the current status of the PM domain (genpd) for the device.

Could it be that the genpd provider doesn't register its PM domains with the state that the HW is really in? pm_genpd_init() is the call that allows the genpd provider to specify the initial state.

I think we need Peng's help here to understand what goes on.

Quoted 3 lines

I tried to move pm_runtime_get_sync() to .prepare function but it make the kernel not boot anymore (with the SCU fault reset).

Quoted 8 lines
Quoted 1 line

Try move pm_runtime_enable() before rproc_add().

Thanks Ulf, that indeed made it work, at least now the kernel does not reset anymore with the SCU fault reset. However I am still only getting 0 from dev_pm_genpd_is_on(), no matter what the state of the remote core. Maybe I am missing something in between?

Peng, do you know what could be the issue here?

imx_rproc_attach_pd
->dev_pm_domain_attach_list

      ->genpd_dev_pm_attach_by_id
              ->genpd_queue_power_off_work
	         ->cm40_pid0 is powered off because the genpd is set with is_off=false

So dev_pm_genpd_is_on will return false after attach.

This means that with U-Boot kick M4, cm40 might be powered off when attaching the pd even with LINK_ON set, because genpd is set with is_off=false.

The reason we set genpd to match real hardware status is to avoid RPC call and to save power. But seems it could not work well with U-boot kicking M4.

I not have good idea on how to address this issue. The current driver could work with linux kick M4, M4 packed in flash.bin and M4 in a standalone partition.

Regards,
Peng

Quoted 13 lines
Quoted 10 lines
Quoted 7 lines

Do you have any suggestions? Am I doing something wrong with these PDs?

Best regards,
Hiago.

Kind regards
Uffe

Best regards,
Hiago

#24 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 120 lines · Peng Fan
Quoted 98 lines · Hiago De Franco
Quoted 97 lines · Ulf Hansson
Quoted 48 lines · Hiago De Franco

Hi Ulf,

Quoted 20 lines · Ulf Hansson

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

Quoted 4 lines

[...]

Kind regards
Uffe

I did some tests here and I might be missing something. I used the dev_pm_genpd_is_on() inside imx_rproc.c with the following changes:

(quoted diff) +5 -0
@@ -902,7 +902,12 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
        if (dev->pm_domain)
                return 0;
 
        ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+       printk("hfranco: returned pd devs is %d", ret);
+       for (int i = 0; i < ret; i++) {
+               test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+               printk("hfranco: returned value is %d", test);
+       }
        return ret < 0 ? ret : 0;
 }
 

This was a quick test to check the returned value, and it always return 1 for both pds, even if I did not boot the remote core.

Quoted 2 lines

So I was wondering if it was because of PD_FLAG_DEV_LINK_ON, I removed it and passed NULL to dev_pm_domain_attach_list().

Right, that's exactly what we should be doing.

Quoted 8 lines

Booting the kernel
now it correctly reports 0 for both pds, however when I start the remote core with a hello world firmware and boot the kernel, the CPU resets with a fault reset ("Reset cause: SCFW fault reset").

I added both pm functions to probe, just to test:

Review static int imx_rproc_probe(struct platform_device *pdev) · line 1163

Quoted 6 lines
(quoted diff) +3 -0
@@ -1152,6 +1158,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
                goto err_put_clk;
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);
+

Indeed, calling pm_runtime_enable() and then pm_runtime_get_sync() should turn on the PM domains for the device, which I assume is needed at some point.

Although, I wonder if this may be a bit too late, I would expect that you at least need to call these *before* the call to rproc_add(), as I assume the rproc-core may start using the device/driver beyond that point.

Quoted 5 lines
(quoted diff) +0 -0
        return 0
 

Now the kernel boot with the remote core running, but it still returns 0 from dev_pm_genpd_is_on(). So basically now it always returns 0, with or without the remote core running.

dev_pm_genpd_is_on() is returning the current status of the PM domain (genpd) for the device.

Could it be that the genpd provider doesn't register its PM domains with the state that the HW is really in? pm_genpd_init() is the call that allows the genpd provider to specify the initial state.

I think we need Peng's help here to understand what goes on.

Quoted 3 lines

I tried to move pm_runtime_get_sync() to .prepare function but it make the kernel not boot anymore (with the SCU fault reset).

Quoted 8 lines
Quoted 1 line

Try move pm_runtime_enable() before rproc_add().

Thanks Ulf, that indeed made it work, at least now the kernel does not reset anymore with the SCU fault reset. However I am still only getting 0 from dev_pm_genpd_is_on(), no matter what the state of the remote core. Maybe I am missing something in between?

Peng, do you know what could be the issue here?

imx_rproc_attach_pd
->dev_pm_domain_attach_list

      ->genpd_dev_pm_attach_by_id
              ->genpd_queue_power_off_work
                 ->cm40_pid0 is powered off because the genpd is set with is_off=false

So dev_pm_genpd_is_on will return false after attach.

This means that with U-Boot kick M4, cm40 might be powered off when attaching the pd even with LINK_ON set, because genpd is set with is_off=false.

Quoted 6 lines

The reason we set genpd to match real hardware status is to avoid RPC call and to save power. But seems it could not work well with U-boot kicking M4.

I not have good idea on how to address this issue. The current driver could work with linux kick M4, M4 packed in flash.bin and M4 in a standalone partition.

Thanks for the detailed analysis!

This is a very similar issue as many other genpd providers are suffering from - and something that I have been working on recently to fix.

A few days ago I posted a new version of a series [1], which is based upon using the fw_devlink and ->sync_state() support. In principle, we need to prevent genpd from power-off a PM domain if it was powered-on during boot , until all the consumer-drivers of a PM domain have been probed.

I had a look at the DT description of how imx describes power-domain providers/consumers, along with the corresponding genpd provider implementation in drivers/pmdomain/imx/scu-pd.c. Unless I missed something, I think [1] should do the trick for you, without any further changes. Can you please give it a try and see if that solves this problem?

[...]

Kind regards
Uffe

[1]
https://lore.kernel.org/all/[email protected]/

#25 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 140 lines · Ulf Hansson
Quoted 120 lines · Peng Fan
Quoted 98 lines · Hiago De Franco
Quoted 97 lines · Ulf Hansson
Quoted 48 lines · Hiago De Franco

Hi Ulf,

Quoted 20 lines · Ulf Hansson

You should not provide any flag (or attach_data to dev_pm_domain_attach_list()) at all. In other words just call dev_pm_domain_attach_list(dev, NULL, &priv->pd_list), similar to how drivers/remoteproc/imx_dsp_rproc.c does it.

In this way, the device_link is created by making the platform->dev the consumer and by keeping the supplier-devices (corresponding to the genpds) in RPM_SUSPENDED state.

The PM domains (genpds) are then left in their current state, which should allow us to call dev_pm_genpd_is_on() for the corresponding supplier-devices, to figure out whether the bootloader turned them on or not, I think.

Moreover, to make sure the genpds are turned on when needed, we also need to call pm_runtime_enable(platform->dev) and pm_runtime_get_sync(platform->dev). The easiest approach is probably to do that during ->probe() - and then as an improvement on top you may want to implement more fine-grained support for runtime PM.

Quoted 4 lines

[...]

Kind regards
Uffe

I did some tests here and I might be missing something. I used the dev_pm_genpd_is_on() inside imx_rproc.c with the following changes:

(quoted diff) +5 -0
@@ -902,7 +902,12 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
        if (dev->pm_domain)
                return 0;
 
        ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+       printk("hfranco: returned pd devs is %d", ret);
+       for (int i = 0; i < ret; i++) {
+               test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+               printk("hfranco: returned value is %d", test);
+       }
        return ret < 0 ? ret : 0;
 }
 

This was a quick test to check the returned value, and it always return 1 for both pds, even if I did not boot the remote core.

Quoted 2 lines

So I was wondering if it was because of PD_FLAG_DEV_LINK_ON, I removed it and passed NULL to dev_pm_domain_attach_list().

Right, that's exactly what we should be doing.

Quoted 8 lines

Booting the kernel
now it correctly reports 0 for both pds, however when I start the remote core with a hello world firmware and boot the kernel, the CPU resets with a fault reset ("Reset cause: SCFW fault reset").

I added both pm functions to probe, just to test:

Review static int imx_rproc_probe(struct platform_device *pdev) · line 1163

Quoted 6 lines
(quoted diff) +3 -0
@@ -1152,6 +1158,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
                goto err_put_clk;
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);
+

Indeed, calling pm_runtime_enable() and then pm_runtime_get_sync() should turn on the PM domains for the device, which I assume is needed at some point.

Although, I wonder if this may be a bit too late, I would expect that you at least need to call these *before* the call to rproc_add(), as I assume the rproc-core may start using the device/driver beyond that point.

Quoted 5 lines
(quoted diff) +0 -0
        return 0
 

Now the kernel boot with the remote core running, but it still returns 0 from dev_pm_genpd_is_on(). So basically now it always returns 0, with or without the remote core running.

dev_pm_genpd_is_on() is returning the current status of the PM domain (genpd) for the device.

Could it be that the genpd provider doesn't register its PM domains with the state that the HW is really in? pm_genpd_init() is the call that allows the genpd provider to specify the initial state.

I think we need Peng's help here to understand what goes on.

Quoted 3 lines

I tried to move pm_runtime_get_sync() to .prepare function but it make the kernel not boot anymore (with the SCU fault reset).

Quoted 8 lines
Quoted 1 line

Try move pm_runtime_enable() before rproc_add().

Thanks Ulf, that indeed made it work, at least now the kernel does not reset anymore with the SCU fault reset. However I am still only getting 0 from dev_pm_genpd_is_on(), no matter what the state of the remote core. Maybe I am missing something in between?

Peng, do you know what could be the issue here?

imx_rproc_attach_pd
->dev_pm_domain_attach_list

      ->genpd_dev_pm_attach_by_id
              ->genpd_queue_power_off_work
                 ->cm40_pid0 is powered off because the genpd is set with is_off=false

So dev_pm_genpd_is_on will return false after attach.

This means that with U-Boot kick M4, cm40 might be powered off when attaching the pd even with LINK_ON set, because genpd is set with is_off=false.

Quoted 6 lines

The reason we set genpd to match real hardware status is to avoid RPC call and to save power. But seems it could not work well with U-boot kicking M4.

I not have good idea on how to address this issue. The current driver could work with linux kick M4, M4 packed in flash.bin and M4 in a standalone partition.

Thanks for the detailed analysis!

This is a very similar issue as many other genpd providers are suffering from - and something that I have been working on recently to fix.

A few days ago I posted a new version of a series [1], which is based upon using the fw_devlink and ->sync_state() support. In principle, we need to prevent genpd from power-off a PM domain if it was powered-on during boot , until all the consumer-drivers of a PM domain have been probed.

Quoted 6 lines

I had a look at the DT description of how imx describes power-domain providers/consumers, along with the corresponding genpd provider implementation in drivers/pmdomain/imx/scu-pd.c. Unless I missed something, I think [1] should do the trick for you, without any further changes. Can you please give it a try and see if that solves this problem?

Cool! I can give a try and provide an answer soon. Thanks!

Quoted 8 lines

[...]

Kind regards
Uffe

[1]
https://lore.kernel.org/all/[email protected]/

Best regards,
Hiago

#26 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

On Tue, May 27, 2025 at 10:45:25AM -0300, Hiago De Franco wrote:

[...]

Quoted 13 lines
Quoted 13 lines

Thanks for the detailed analysis!

This is a very similar issue as many other genpd providers are suffering from - and something that I have been working on recently to fix.

A few days ago I posted a new version of a series [1], which is based upon using the fw_devlink and ->sync_state() support. In principle, we need to prevent genpd from power-off a PM domain if it was powered-on during boot , until all the consumer-drivers of a PM domain have been probed.

Quoted 8 lines
Quoted 6 lines

I had a look at the DT description of how imx describes power-domain providers/consumers, along with the corresponding genpd provider implementation in drivers/pmdomain/imx/scu-pd.c. Unless I missed something, I think [1] should do the trick for you, without any further changes. Can you please give it a try and see if that solves this problem?

Cool! I can give a try and provide an answer soon. Thanks!

We are making progress ;-)

With the patches you shared Ulf (I added them on top of the current master branch), it works as expected, dev_pm_genpd_is_on() returns 0 when I boot the kernel without M4 running and it returns 1 when I boot the kernel with M4 running with a hello-world demo.

However now I tried to, if dev_pm_genpd_is_on() returns 1, put the DETACHED state, something as

if (dev_pm_genpd_is_on(priv->pd_list->pd_devs[0]))

	priv->rproc->state = RPROC_DETACHED;

In this case I used 0 because I understand this is the IMX_SC_R_M4_0_PID0 defined in my device tree overlay:

		power-domains = <&pd IMX_SC_R_M4_0_PID0>,
				<&pd IMX_SC_R_M4_0_MU_1A>;

But in this case, the kernel does not boot anymore, I see the "Starting kernel..." and nothing else.

I am using the pm_runtime functions before rproc_add():

@@ -1146,6 +1154,9 @@ static int imx_rproc_probe(struct platform_device *pdev)

                }
        }

+ pm_runtime_enable(dev);
+ pm_runtime_get_sync(dev);
+

        ret = rproc_add(rproc);
        if (ret) {
                dev_err(dev, "rproc_add failed\n");

and calling dev_pm_genpd_is_on() after dev_pm_domain_attach_list() has been called.

With kernel not starting I cannot even debug it. Do you have any suggestion on why this might be happening?

Quoted 4 lines
Quoted 4 lines

[...]

Quoted 8 lines
Quoted 5 lines

Kind regards
Uffe

[1]
https://lore.kernel.org/all/[email protected]/

Best regards,
Hiago

Best regards,
Hiago

#27 lore raw

RE: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 46 lines

Subject: Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

On Tue, May 27, 2025 at 10:45:25AM -0300, Hiago De Franco wrote:

[...]

Quoted 5 lines
Quoted 5 lines

Thanks for the detailed analysis!

This is a very similar issue as many other genpd providers are suffering from - and something that I have been working on

recently

Quoted 5 lines
Quoted 5 lines

to fix.

A few days ago I posted a new version of a series [1], which is based upon using the fw_devlink and ->sync_state() support. In principle, we need to prevent genpd from power-off a PM domain if

it

Quoted 1 line
Quoted 1 line

was powered-on during boot , until all the consumer-drivers of a

PM

Quoted 3 lines
Quoted 3 lines

domain have been probed.

I had a look at the DT description of how imx describes power-

domain

Quoted 7 lines
Quoted 5 lines

providers/consumers, along with the corresponding genpd provider implementation in drivers/pmdomain/imx/scu-pd.c. Unless I missed something, I think [1] should do the trick for you, without any further changes. Can you please give it a try and see if that solves this problem?

Cool! I can give a try and provide an answer soon. Thanks!

We are making progress ;-)

With the patches you shared Ulf (I added them on top of the current master branch), it works as expected, dev_pm_genpd_is_on() returns 0 when I boot the kernel without M4 running and it returns 1 when I boot the kernel with M4 running with a hello-world demo.

However now I tried to, if dev_pm_genpd_is_on() returns 1, put the DETACHED state, something as

if (dev_pm_genpd_is_on(priv->pd_list->pd_devs[0]))

	priv->rproc->state = RPROC_DETACHED;
Quoted 8 lines

In this case I used 0 because I understand this is the IMX_SC_R_M4_0_PID0 defined in my device tree overlay:

		power-domains = <&pd IMX_SC_R_M4_0_PID0>,
				<&pd IMX_SC_R_M4_0_MU_1A>;

But in this case, the kernel does not boot anymore, I see the "Starting kernel..." and nothing else.

Please add "earlycon" in bootargs to see where it hangs.

Review static int imx_rproc_probe(struct · line 1159

Quoted 10 lines

I am using the pm_runtime functions before rproc_add():

(quoted diff) +2 -0
@@ -1146,6 +1154,9 @@ static int imx_rproc_probe(struct
 platform_device *pdev)
                }
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);

I think only make this apply for i.MX8QX/8QM/DX, then no impact to other patforms.

Regards,
Peng

#28 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

On Thu, May 29, 2025 at 03:54:47AM +0000, Peng Fan wrote:

[...]

Quoted 16 lines
Quoted 16 lines

We are making progress ;-)

With the patches you shared Ulf (I added them on top of the current master branch), it works as expected, dev_pm_genpd_is_on() returns 0 when I boot the kernel without M4 running and it returns 1 when I boot the kernel with M4 running with a hello-world demo.

However now I tried to, if dev_pm_genpd_is_on() returns 1, put the DETACHED state, something as

if (dev_pm_genpd_is_on(priv->pd_list->pd_devs[0]))

	priv->rproc->state = RPROC_DETACHED;

In this case I used 0 because I understand this is the IMX_SC_R_M4_0_PID0 defined in my device tree overlay:

Quoted 7 lines
Quoted 5 lines
		power-domains = <&pd IMX_SC_R_M4_0_PID0>,
				<&pd IMX_SC_R_M4_0_MU_1A>;

But in this case, the kernel does not boot anymore, I see the "Starting kernel..." and nothing else.

Please add "earlycon" in bootargs to see where it hangs.

Thanks Peng! I was able to catch the kernel panic yesterday, however I must say that today I was doing the tests again and the issue is gone. Sorry, I might have done something wrong yesterday with the tests. Anyway, here is the log:

[    1.271163] remoteproc remoteproc0: imx-rproc is available
[    1.280296] remoteproc remoteproc0: attaching to imx-rproc
[    1.285756] Unable to handle kernel paging request at virtual address ffff80005ae3dd79
[    1.293624] Mem abort info:
[    1.294655] mmc0: SDHCI controller on 5b010000.mmc [5b010000.mmc] using ADMA
[    1.296386]   ESR = 0x0000000096000005
[    1.307194]   EC = 0x25: DABT (current EL), IL = 32 bits
[    1.312473]   SET = 0, FnV = 0
[    1.315566]   EA = 0, S1PTW = 0
[    1.318649]   FSC = 0x05: level 1 translation fault
[    1.323510] Data abort info:
[    1.326370]   ISV = 0, ISS = 0x00000005, ISS2 = 0x00000000
[    1.331846]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
[    1.336882]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
[    1.342182] swapper pgtable: 4k pages, 48-bit VAs, pgdp=0000000096bc1000
[    1.348870] [ffff80005ae3dd79] pgd=0000000000000000, p4d=1000000097054003, pud=0000000000000000
[    1.357565] Internal error: Oops: 0000000096000005 [#1]  SMP
[    1.363198] Modules linked in:
[    1.366236] CPU: 2 UID: 0 PID: 47 Comm: kworker/u16:3 Not tainted 6.15.0-03667-g3f5f09105c40-dirty #826 PREEMPT
[    1.376405] Hardware name: Toradex Colibri iMX8QXP on Colibri Evaluation Board V3 (DT)
[    1.384313] Workqueue: events_unbound deferred_probe_work_func
[    1.390128] pstate: 00000005 (nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[    1.397076] pc : rproc_handle_resources.constprop.0+0x78/0x1d0
[    1.402896] lr : rproc_boot+0x368/0x56c
[    1.406717] sp : ffff8000819c3990
[    1.410017] x29: ffff8000819c3990 x28: ffff80005ae3dd7d x27: 0000000000000000
[    1.417145] x26: 0000000000000000 x25: ffff0000015ec038 x24: ffff800080f0c0a8
[    1.424268] x23: ffff8000813a6110 x22: 00000000d999ad79 x21: ffff0000015ec000
[    1.431392] x20: 0000000026665683 x19: ffff80005ae3dd79 x18: 0000000000000006
[    1.438516] x17: ffff000001799400 x16: ffff000001798e00 x15: 4addd15cca11c529
[    1.445639] x14: 53ebce6d5564d787 x13: 4addd15cca11c529 x12: 53ebce6d5564d787
[    1.452763] x11: 95a1e33b6b190674 x10: 9e3c9abdb41ca345 x9 : ab17b4eaffd6fd1c
[    1.459887] x8 : d5da055de4cfbb87 x7 : dfd7fa31596acbbc x6 : 9946d97107d0dcca
[    1.467011] x5 : ffff0000010c7800 x4 : 00000000000003fc x3 : ffff0000010c7780
[    1.474134] x2 : fffffffffffffff0 x1 : ffff8000814a3000 x0 : ffff8000814a3000
[    1.481261] Call trace:
[    1.483690]  rproc_handle_resources.constprop.0+0x78/0x1d0 (P)
[    1.487705] mmc0: new HS400 MMC card at address 0001
[    1.489502]  rproc_boot+0x368/0x56c
[    1.495349] mmcblk0: mmc0:0001 Q2J55L 7.09 GiB
[    1.497929]  rproc_add+0x184/0x190
[    1.504356]  mmcblk0: p1 p2
[    1.505747]  imx_rproc_probe+0x458/0x528
[    1.509238] mmcblk0boot0: mmc0:0001 Q2J55L 16.0 MiB
[    1.512437]  platform_probe+0x68/0xc0
[    1.512452]  really_probe+0xc0/0x38c
[    1.520584] mmcblk0boot1: mmc0:0001 Q2J55L 16.0 MiB
[    1.520951]  __driver_probe_device+0x7c/0x15c
[    1.527522] mmcblk0rpmb: mmc0:0001 Q2J55L 4.00 MiB, chardev (242:0)
[    1.529377]  driver_probe_device+0x3c/0x10c
[    1.544263]  __device_attach_driver+0xbc/0x158
[    1.548586]  bus_for_each_drv+0x84/0xe0
[    1.552407]  __device_attach+0x9c/0x1ac
[    1.556231]  device_initial_probe+0x14/0x20
[    1.560401]  bus_probe_device+0xac/0xb0
[    1.564221]  deferred_probe_work_func+0x9c/0xec
[    1.568741]  process_one_work+0x14c/0x28c
[    1.572735]  worker_thread+0x2cc/0x3d4
[    1.576473]  kthread+0x12c/0x208
[    1.579687]  ret_from_fork+0x10/0x20
[    1.583253] Code: 8b36c033 9100127c 54000924 d503201f (b9400261)
[    1.589337] ---[ end trace 0000000000000000 ]---

But again, the issue is not happening anymore ;-) I will keep testing to see if the issue happens again, but for now is working fine, I can now attach to the remote processor.

This is the git diff on top of Ulf patches I have been testing:

drivers/pmdomain/core.c +17 -0
index 6da25e2c81d2..661a6aad40a8 100644
@@ -599,6 +599,23 @@ int dev_pm_genpd_set_performance_state(struct device *dev, unsigned int state)
 }
 EXPORT_SYMBOL_GPL(dev_pm_genpd_set_performance_state);
 
+bool dev_pm_genpd_is_on(struct device *dev)
+{
+        struct generic_pm_domain *genpd;
+        bool is_on;
+
+        genpd = dev_to_genpd_safe(dev);
+        if (!genpd)
+                return false;
+
+        genpd_lock(genpd);
+        is_on = genpd_status_on(genpd);
+        genpd_unlock(genpd);
+
+        return is_on;
+}
+EXPORT_SYMBOL_GPL(dev_pm_genpd_is_on);
+
 /**
  * dev_pm_genpd_set_next_wakeup - Notify PM framework of an impending wakeup.
  *
drivers/remoteproc/imx_rproc.c +15 -4
index 627e57a88db2..9688370f9bb5 100644
@@ -18,6 +18,7 @@
 #include <linux/of_reserved_mem.h>
 #include <linux/platform_device.h>
 #include <linux/pm_domain.h>
+#include <linux/pm_runtime.h>
 #include <linux/reboot.h>
 #include <linux/regmap.h>
 #include <linux/remoteproc.h>
@@ -891,9 +892,7 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
 {
 	struct device *dev = priv->dev;
 	int ret;
-	struct dev_pm_domain_attach_data pd_data = {
-		.pd_flags = PD_FLAG_DEV_LINK_ON,
-	};
+	bool test;
 
 	/*
 	 * If there is only one power-domain entry, the platform driver framework
@@ -902,7 +901,16 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
 	if (dev->pm_domain)
 		return 0;
 
-	ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+	ret = dev_pm_domain_attach_list(dev, NULL, &priv->pd_list);
+	printk("hfranco: returned pd devs is %d", ret);
+	for (int i = 0; i < ret; i++) {
+		test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+		printk("hfranco: returned value is %d", test);
+		if (test) {
+			priv->rproc->state = RPROC_DETACHED;
+			break;
+		}
+	}
 	return ret < 0 ? ret : 0;
 }
 
@@ -1146,6 +1154,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
 		}
 	}
 
+	pm_runtime_enable(dev);
+	pm_runtime_get_sync(dev);
+
 	ret = rproc_add(rproc);
 	if (ret) {
 		dev_err(dev, "rproc_add failed\n");
include/linux/pm_runtime.h +1 -0
index 756b842dcd30..16d1fca2a8c5 100644
@@ -95,6 +95,7 @@ extern void pm_runtime_put_suppliers(struct device *dev);
 extern void pm_runtime_new_link(struct device *dev);
 extern void pm_runtime_drop_link(struct device_link *link);
 extern void pm_runtime_release_supplier(struct device_link *link);
+bool dev_pm_genpd_is_on(struct device *dev);
 
 int devm_pm_runtime_set_active_enabled(struct device *dev);
 extern int devm_pm_runtime_enable(struct device *dev);

This is the rproc output when bootaux is used:

root@colibri-imx8x-07308754:~# dmesg | grep hfranco
[    0.478475] hfranco: returned pd devs is 2
[    0.478496] hfranco: returned value is 1
root@colibri-imx8x-07308754:~# dmesg | grep rproc
[    0.478797] remoteproc remoteproc0: imx-rproc is available
[    0.478878] remoteproc remoteproc0: attaching to imx-rproc
[    0.478961] remoteproc remoteproc0: remote processor imx-rproc is now attached

I will cleanup everything and try to come up with a patch. Ulf, in this case, as your patches have not yet been merged, should I wait for them?

Thanks for all the help guys.

Quoted 9 lines
Quoted 9 lines

I am using the pm_runtime functions before rproc_add():

(quoted diff) +0 -0
@@ -1146,6 +1154,9 @@ static int imx_rproc_probe(struct
 platform_device *pdev)
                }
        }
 
Quoted 5 lines

Review static int imx_rproc_probe(struct · line 1159

Quoted 2 lines
(quoted diff) +2 -0
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);

I think only make this apply for i.MX8QX/8QM/DX, then no impact to other patforms.

All right, I will check how to do this. Thanks.

Quoted 3 lines

Regards,
Peng

Best regards,
Hiago.

#29 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

Quoted 205 lines · Hiago De Franco

On Thu, May 29, 2025 at 03:54:47AM +0000, Peng Fan wrote:

[...]

Quoted 16 lines
Quoted 16 lines

We are making progress ;-)

With the patches you shared Ulf (I added them on top of the current master branch), it works as expected, dev_pm_genpd_is_on() returns 0 when I boot the kernel without M4 running and it returns 1 when I boot the kernel with M4 running with a hello-world demo.

However now I tried to, if dev_pm_genpd_is_on() returns 1, put the DETACHED state, something as

if (dev_pm_genpd_is_on(priv->pd_list->pd_devs[0]))

    priv->rproc->state = RPROC_DETACHED;

In this case I used 0 because I understand this is the IMX_SC_R_M4_0_PID0 defined in my device tree overlay:

Quoted 7 lines
Quoted 5 lines
            power-domains = <&pd IMX_SC_R_M4_0_PID0>,
                            <&pd IMX_SC_R_M4_0_MU_1A>;

But in this case, the kernel does not boot anymore, I see the "Starting kernel..." and nothing else.

Please add "earlycon" in bootargs to see where it hangs.

Thanks Peng! I was able to catch the kernel panic yesterday, however I must say that today I was doing the tests again and the issue is gone. Sorry, I might have done something wrong yesterday with the tests. Anyway, here is the log:

[    1.271163] remoteproc remoteproc0: imx-rproc is available
[    1.280296] remoteproc remoteproc0: attaching to imx-rproc
[    1.285756] Unable to handle kernel paging request at virtual address ffff80005ae3dd79
[    1.293624] Mem abort info:
[    1.294655] mmc0: SDHCI controller on 5b010000.mmc [5b010000.mmc] using ADMA
[    1.296386]   ESR = 0x0000000096000005
[    1.307194]   EC = 0x25: DABT (current EL), IL = 32 bits
[    1.312473]   SET = 0, FnV = 0
[    1.315566]   EA = 0, S1PTW = 0
[    1.318649]   FSC = 0x05: level 1 translation fault
[    1.323510] Data abort info:
[    1.326370]   ISV = 0, ISS = 0x00000005, ISS2 = 0x00000000
[    1.331846]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
[    1.336882]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
[    1.342182] swapper pgtable: 4k pages, 48-bit VAs, pgdp=0000000096bc1000
[    1.348870] [ffff80005ae3dd79] pgd=0000000000000000, p4d=1000000097054003, pud=0000000000000000
[    1.357565] Internal error: Oops: 0000000096000005 [#1]  SMP
[    1.363198] Modules linked in:
[    1.366236] CPU: 2 UID: 0 PID: 47 Comm: kworker/u16:3 Not tainted 6.15.0-03667-g3f5f09105c40-dirty #826 PREEMPT
[    1.376405] Hardware name: Toradex Colibri iMX8QXP on Colibri Evaluation Board V3 (DT)
[    1.384313] Workqueue: events_unbound deferred_probe_work_func
[    1.390128] pstate: 00000005 (nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[    1.397076] pc : rproc_handle_resources.constprop.0+0x78/0x1d0
[    1.402896] lr : rproc_boot+0x368/0x56c
[    1.406717] sp : ffff8000819c3990
[    1.410017] x29: ffff8000819c3990 x28: ffff80005ae3dd7d x27: 0000000000000000
[    1.417145] x26: 0000000000000000 x25: ffff0000015ec038 x24: ffff800080f0c0a8
[    1.424268] x23: ffff8000813a6110 x22: 00000000d999ad79 x21: ffff0000015ec000
[    1.431392] x20: 0000000026665683 x19: ffff80005ae3dd79 x18: 0000000000000006
[    1.438516] x17: ffff000001799400 x16: ffff000001798e00 x15: 4addd15cca11c529
[    1.445639] x14: 53ebce6d5564d787 x13: 4addd15cca11c529 x12: 53ebce6d5564d787
[    1.452763] x11: 95a1e33b6b190674 x10: 9e3c9abdb41ca345 x9 : ab17b4eaffd6fd1c
[    1.459887] x8 : d5da055de4cfbb87 x7 : dfd7fa31596acbbc x6 : 9946d97107d0dcca
[    1.467011] x5 : ffff0000010c7800 x4 : 00000000000003fc x3 : ffff0000010c7780
[    1.474134] x2 : fffffffffffffff0 x1 : ffff8000814a3000 x0 : ffff8000814a3000
[    1.481261] Call trace:
[    1.483690]  rproc_handle_resources.constprop.0+0x78/0x1d0 (P)
[    1.487705] mmc0: new HS400 MMC card at address 0001
[    1.489502]  rproc_boot+0x368/0x56c
[    1.495349] mmcblk0: mmc0:0001 Q2J55L 7.09 GiB
[    1.497929]  rproc_add+0x184/0x190
[    1.504356]  mmcblk0: p1 p2
[    1.505747]  imx_rproc_probe+0x458/0x528
[    1.509238] mmcblk0boot0: mmc0:0001 Q2J55L 16.0 MiB
[    1.512437]  platform_probe+0x68/0xc0
[    1.512452]  really_probe+0xc0/0x38c
[    1.520584] mmcblk0boot1: mmc0:0001 Q2J55L 16.0 MiB
[    1.520951]  __driver_probe_device+0x7c/0x15c
[    1.527522] mmcblk0rpmb: mmc0:0001 Q2J55L 4.00 MiB, chardev (242:0)
[    1.529377]  driver_probe_device+0x3c/0x10c
[    1.544263]  __device_attach_driver+0xbc/0x158
[    1.548586]  bus_for_each_drv+0x84/0xe0
[    1.552407]  __device_attach+0x9c/0x1ac
[    1.556231]  device_initial_probe+0x14/0x20
[    1.560401]  bus_probe_device+0xac/0xb0
[    1.564221]  deferred_probe_work_func+0x9c/0xec
[    1.568741]  process_one_work+0x14c/0x28c
[    1.572735]  worker_thread+0x2cc/0x3d4
[    1.576473]  kthread+0x12c/0x208
[    1.579687]  ret_from_fork+0x10/0x20
[    1.583253] Code: 8b36c033 9100127c 54000924 d503201f (b9400261)
[    1.589337] ---[ end trace 0000000000000000 ]---

But again, the issue is not happening anymore ;-) I will keep testing to see if the issue happens again, but for now is working fine, I can now attach to the remote processor.

This is the git diff on top of Ulf patches I have been testing:

drivers/pmdomain/core.c +17 -0
index 6da25e2c81d2..661a6aad40a8 100644
@@ -599,6 +599,23 @@ int dev_pm_genpd_set_performance_state(struct device *dev, unsigned int state)
 }
 EXPORT_SYMBOL_GPL(dev_pm_genpd_set_performance_state);
 
+bool dev_pm_genpd_is_on(struct device *dev)
+{
+        struct generic_pm_domain *genpd;
+        bool is_on;
+
+        genpd = dev_to_genpd_safe(dev);
+        if (!genpd)
+                return false;
+
+        genpd_lock(genpd);
+        is_on = genpd_status_on(genpd);
+        genpd_unlock(genpd);
+
+        return is_on;
+}
+EXPORT_SYMBOL_GPL(dev_pm_genpd_is_on);
+
 /**
  * dev_pm_genpd_set_next_wakeup - Notify PM framework of an impending wakeup.
  *
drivers/remoteproc/imx_rproc.c +15 -4
index 627e57a88db2..9688370f9bb5 100644
@@ -18,6 +18,7 @@
 #include <linux/of_reserved_mem.h>
 #include <linux/platform_device.h>
 #include <linux/pm_domain.h>
+#include <linux/pm_runtime.h>
 #include <linux/reboot.h>
 #include <linux/regmap.h>
 #include <linux/remoteproc.h>
@@ -891,9 +892,7 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
 {
        struct device *dev = priv->dev;
        int ret;
-       struct dev_pm_domain_attach_data pd_data = {
-               .pd_flags = PD_FLAG_DEV_LINK_ON,
-       };
+       bool test;
 
        /*
         * If there is only one power-domain entry, the platform driver framework
@@ -902,7 +901,16 @@ static int imx_rproc_attach_pd(struct imx_rproc *priv)
        if (dev->pm_domain)
                return 0;
 
-       ret = dev_pm_domain_attach_list(dev, &pd_data, &priv->pd_list);
+       ret = dev_pm_domain_attach_list(dev, NULL, &priv->pd_list);
+       printk("hfranco: returned pd devs is %d", ret);
+       for (int i = 0; i < ret; i++) {
+               test = dev_pm_genpd_is_on(priv->pd_list->pd_devs[i]);
+               printk("hfranco: returned value is %d", test);
+               if (test) {
+                       priv->rproc->state = RPROC_DETACHED;
+                       break;
+               }
+       }
        return ret < 0 ? ret : 0;
 }
 
@@ -1146,6 +1154,9 @@ static int imx_rproc_probe(struct platform_device *pdev)
                }
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);
+
        ret = rproc_add(rproc);
        if (ret) {
                dev_err(dev, "rproc_add failed\n");
include/linux/pm_runtime.h +1 -0
index 756b842dcd30..16d1fca2a8c5 100644
@@ -95,6 +95,7 @@ extern void pm_runtime_put_suppliers(struct device *dev);
 extern void pm_runtime_new_link(struct device *dev);
 extern void pm_runtime_drop_link(struct device_link *link);
 extern void pm_runtime_release_supplier(struct device_link *link);
+bool dev_pm_genpd_is_on(struct device *dev);
 
 int devm_pm_runtime_set_active_enabled(struct device *dev);
 extern int devm_pm_runtime_enable(struct device *dev);

This is the rproc output when bootaux is used:

root@colibri-imx8x-07308754:~# dmesg | grep hfranco
[    0.478475] hfranco: returned pd devs is 2
[    0.478496] hfranco: returned value is 1
root@colibri-imx8x-07308754:~# dmesg | grep rproc
[    0.478797] remoteproc remoteproc0: imx-rproc is available
[    0.478878] remoteproc remoteproc0: attaching to imx-rproc
[    0.478961] remoteproc remoteproc0: remote processor imx-rproc is now attached
Quoted 2 lines

I will cleanup everything and try to come up with a patch. Ulf, in this case, as your patches have not yet been merged, should I wait for them?

I think you can state in the cover-letter that your series depends on mine, so please go ahead and submit them.

Quoted 17 lines

Thanks for all the help guys.

Quoted 14 lines

drivers/remoteproc/imx_rproc.c static int imx_rproc_probe(struct · line 1159

Quoted 11 lines

I am using the pm_runtime functions before rproc_add():

drivers/remoteproc/imx_rproc.c +2 -0
@@ -1146,6 +1154,9 @@ static int imx_rproc_probe(struct
 platform_device *pdev)
                }
        }
 
+       pm_runtime_enable(dev);
+       pm_runtime_get_sync(dev);

I think only make this apply for i.MX8QX/8QM/DX, then no impact to other patforms.

In general I think we should avoid such quirks in drivers, unless it's really needed. Just wanted to share my opinion, but it's totally up to you.

[...]

Kind regards
Uffe

#30 lore raw

Re: [PATCH v2 3/3] remoteproc: imx_rproc: add power mode check for remote core attachment

In reply to Ulf Hansson

Hi Ulf,

Quoted 49 lines · Ulf Hansson
Quoted 34 lines · Hiago De Franco

From: Hiago De Franco <[email protected]>

When the remote core is started before Linux boots (e.g., by the bootloader), the driver currently is not able to attach because it only checks for cores running in different partitions. If the core was kicked by the bootloader, it is in the same partition as Linux and it is already up and running.

This adds power mode verification through the SCU interface, enabling the driver to detect when the remote core is already running and properly attach to it.

---
v2: Dropped unecessary include. Removed the imx_rproc_is_on function, as suggested.

---
 drivers/remoteproc/imx_rproc.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/remoteproc/imx_rproc.c +4 -0
index 627e57a88db2..9b6e9e41b7fc 100644
@@ -949,6 +949,19 @@ static int imx_rproc_detect_mode(struct imx_rproc *priv)
                        if (of_property_read_u32(dev->of_node, "fsl,entry-address", &priv->entry))
                                return -EINVAL;
 
+                       /*
+                        * If remote core is already running (e.g. kicked by
+                        * the bootloader), attach to it.
+                        */

drivers/remoteproc/imx_rproc.c static int imx_rproc_detect_mode(struct imx_rproc *priv) · line 965

Quoted 10 lines
drivers/remoteproc/imx_rproc.c +9 -0
+                       ret = imx_sc_pm_get_resource_power_mode(priv->ipc_handle,
+                                                               priv->rsrc_id);
+                       if (ret < 0)
+                               dev_err(dev, "failed to get power resource %d mode, ret %d\n",
+                                       priv->rsrc_id, ret);
+
+                       if (ret == IMX_SC_PM_PW_MODE_ON)
+                               priv->rproc->state = RPROC_DETACHED;
+
                        return imx_rproc_attach_pd(priv);

Why is it important to potentially set "priv->rproc->state = RPROC_DETACHED" before calling imx_rproc_attach_pd()?

Quoted 3 lines

Would it be possible to do it the other way around? First calling imx_rproc_attach_pd() then get the power-mode to know if RPROC_DETACHED should be set or not?

If M4 is not powered up by bootloader, attach_pd is to power up the related power domains. And the rproc->state should be OFFLINE.

If M4 is powered up by bootloader, the rproc->state should be set as DETACHED, then attach_pd here will not touch the real pd, because scu-pd driver set is_off as false when doing pm_genpd_init. In this case, we still need attach_pd to avoid power shutdown when pd_ignore_unused and also need to support linux stop/start m4 even it is started by bootloader.

So we could not reverse the logic.

Regards,
Peng

Quoted 13 lines

The main reason why I ask, is because of how we handle the single PM domain case. In that case, the PM domain has already been attached (and powered-on) before we reach this point.

Quoted 5 lines
drivers/remoteproc/imx_rproc.c +0 -1
                }
 
--

2.39.5

Kind regards
Uffe