can: fsl,flexcan: add fsl,s32v234-flexcan and imx95 wakeup
Subsystems (5)
Frank Li
maintainerv2 cover can: fsl,flexcan: add fsl,s32v234-flexcan and imx95 wakeup
To: Marc Kleine-Budde <[email protected]>
To: Vincent Mailhol <[email protected]>
To: David S. Miller <[email protected]>
To: Eric Dumazet <[email protected]>
To: Jakub Kicinski <[email protected]>
To: Paolo Abeni <[email protected]>
To: Rob Herring <[email protected]>
To: Krzysztof Kozlowski <[email protected]>
To: Conor Dooley <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
---
Changes in v2:
- fixed typo an rework commit message for patch 4
- Add rob's review tag
- Add Vincent's review tag
- Add const for fsl_imx95_devtype_data
- Link to v1: https://lore.kernel.org/r/[email protected]
---
Chircu-Mare Bogdan-Petru (1):
can: flexcan: Add S32V234 support to FlexCAN driver
Frank Li (1):
dt-bindings: can: fsl,flexcan: add compatible string fsl,s32v234-flexcan
Haibo Chen (2):
bingdings: can: flexcan: move fsl,imx95-flexcan standalone
can: flexcan: add wakeup support for imx95
.../devicetree/bindings/net/can/fsl,flexcan.yaml | 5 +- drivers/net/can/flexcan/flexcan-core.c | 54 ++++++++++++++++++++-- drivers/net/can/flexcan/flexcan.h | 2 + 3 files changed, 53 insertions(+), 8 deletions(-)
---
base-commit: a4cf44c1ad6471737cf687b1f1644cf33641e738
change-id: 20240709-flexcan-c75cfef2acb9
Best regards,
---
Frank Li <[email protected]>
Frank Li
maintainerPATCH v2 1/4 dt-bindings: can: fsl,flexcan: add compatible string fsl,s32v234-flexcan
Add compatible string fsl,s32v234-flexcan for s32 chips.
--- Documentation/devicetree/bindings/net/can/fsl,flexcan.yaml | 1 + 1 file changed, 1 insertion(+)
Documentation/devicetree/bindings/net/can/fsl,flexcan.yaml +1 -0
@@ -27,6 +27,7 @@ properties: - fsl,vf610-flexcan - fsl,ls1021ar2-flexcan - fsl,lx2160ar1-flexcan + - fsl,s32v234-flexcan - items: - enum: - fsl,imx53-flexcan
Signature · 1 line
2.34.1
Frank Li
maintainerPATCH v2 2/4 can: flexcan: Add S32V234 support to FlexCAN driver
In reply to Frank LiFrom: Chircu-Mare Bogdan-Petru <[email protected]>
Add flexcan support for S32V234.
--- drivers/net/can/flexcan/flexcan-core.c | 5 +++++ 1 file changed, 5 insertions(+)
drivers/net/can/flexcan/flexcan-core.c +5 -0
@@ -378,6 +378,10 @@ static const struct flexcan_devtype_data fsl_lx2160a_r1_devtype_data = { FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR, }; +static struct flexcan_devtype_data fsl_s32v234_devtype_data = { + .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_DISABLE_MECR , +}; +static const struct can_bittiming_const flexcan_bittiming_const = { .name =DRV_NAME , .tseg1_min =4 ,
@@ -2018,6 +2022,7 @@ static const struct of_device_id flexcan_of_match[] = { { .compatible = "fsl,vf610-flexcan", .data = &fsl_vf610_devtype_data, }, { .compatible = "fsl,ls1021ar2-flexcan", .data = &fsl_ls1021a_r2_devtype_data, }, { .compatible = "fsl,lx2160ar1-flexcan", .data = &fsl_lx2160a_r1_devtype_data, }, + { .compatible = "fsl,s32v234-flexcan", .data = &fsl_s32v234_devtype_data, }, { /* sentinel */ }, }; MODULE_DEVICE_TABLE(of, flexcan_of_match);
Signature · 1 line
2.34.1
Marc Kleine-Budde
maintainerRe: [PATCH v2 2/4] can: flexcan: Add S32V234 support to FlexCAN driver
Quoted 20 lines · Frank Li
From: Chircu-Mare Bogdan-Petru <[email protected]>
Add flexcan support for S32V234.
Signed-off-by Chircu-Mare Bogdan-Petru <[email protected]>Signed-off-by Dan Nica <[email protected]>Signed-off-by Stefan-Gabriel Mirea <[email protected]>Reviewed-by Li Yang <[email protected]>Reviewed-by Joakim Zhang <[email protected]>Reviewed-by Leonard Crestez <[email protected]>Signed-off-by Frank Li <[email protected]>--- drivers/net/can/flexcan/flexcan-core.c | 5 +++++ 1 file changed, 5 insertions(+)drivers/net/can/flexcan/flexcan-core.c +0
-0
drivers/net/can/flexcan/flexcan-core.c static const struct flexcan_devtype_data fsl_lx2160a_r1_devtype_data = { · line 383
Quoted 6 lines
drivers/net/can/flexcan/flexcan-core.c +3
-0@@ -378,6 +378,10 @@ static const struct flexcan_devtype_data fsl_lx2160a_r1_devtype_data = { FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR, }; +static struct flexcan_devtype_data fsl_s32v234_devtype_data = { + .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_DISABLE_MECR , +};
- Does the core support CAN-FD?
- Does the core generate an error interrupt when going from error warning to error passive?
- Are you sure, you don't need to enable FLEXCAN_QUIRK_ENABLE_EACEN_RRS?
- You probably want to use the mailboxes instead of the FIFO for the RX-path.
regards,
Marc
Signature · 4 lines
Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung Nürnberg | Phone: +49-5121-206917-129 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
Frank Li
maintainerPATCH v2 3/4 bingdings: can: flexcan: move fsl,imx95-flexcan standalone
In reply to Frank LiFrom: Haibo Chen <[email protected]>
The flexcan in iMX95 is not compatible with imx93 because wakeup method is difference. Make fsl,imx95-flexcan not fallback to fsl,imx93-flexcan.
--- Documentation/devicetree/bindings/net/can/fsl,flexcan.yaml | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-)
Documentation/devicetree/bindings/net/can/fsl,flexcan.yaml +1 -3
@@ -17,6 +17,7 @@ properties: compatible: oneOf: - enum: + - fsl,imx95-flexcan - fsl,imx93-flexcan - fsl,imx8qm-flexcan - fsl,imx8mp-flexcan
@@ -39,9 +40,6 @@ properties: - fsl,imx6ul-flexcan - fsl,imx6sx-flexcan - const: fsl,imx6q-flexcan - - items: - - const: fsl,imx95-flexcan - - const: fsl,imx93-flexcan - items: - enum: - fsl,ls1028ar1-flexcan
Signature · 1 line
2.34.1
Frank Li
maintainerPATCH v2 4/4 can: flexcan: add wakeup support for imx95
In reply to Frank LiFrom: Haibo Chen <[email protected]>
iMX95 defines a bit in GPR that sets/unsets the IPG_STOP signal to the FlexCAN module, controlling its entry into STOP mode. Wakeup should work even if FlexCAN is in STOP mode.
Due to iMX95 architecture design, the A-Core cannot access GPR; only the system manager (SM) can configure GPR. To support the wakeup feature, follow these steps:
- For suspend:
1) During Linux suspend, when CAN suspends, do nothing for GPR and keep CAN-related clocks on.
2) In ATF, check whether CAN needs to support wakeup; if yes, send a request to SM through the SCMI protocol.
3) In SM, configure the GPR and unset IPG_STOP.
4) A-Core suspends.
- For wakeup and resume:
1) A-Core wakeup event arrives.
2) In SM, deassert IPG_STOP.
3) Linux resumes.
Add a new fsl_imx95_devtype_data and FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI to reflect this.
--- drivers/net/can/flexcan/flexcan-core.c | 49 ++++++++++++++++++++++++++++++---- drivers/net/can/flexcan/flexcan.h | 2 ++ 2 files changed, 46 insertions(+), 5 deletions(-)
drivers/net/can/flexcan/flexcan-core.c +44 -5
@@ -354,6 +354,14 @@ static struct flexcan_devtype_data fsl_imx93_devtype_data = { FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR, }; +static const struct flexcan_devtype_data fsl_imx95_devtype_data = { + .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_ENABLE_EACEN_RRS | +FLEXCAN_QUIRK_DISABLE_MECR |FLEXCAN_QUIRK_USE_RX_MAILBOX | +FLEXCAN_QUIRK_BROKEN_PERR_STATE |FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI | +FLEXCAN_QUIRK_SUPPORT_FD |FLEXCAN_QUIRK_SUPPORT_ECC | +FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX | +FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR , +};static const struct flexcan_devtype_data fsl_vf610_devtype_data = { .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_ENABLE_EACEN_RRS | FLEXCAN_QUIRK_DISABLE_MECR | FLEXCAN_QUIRK_USE_RX_MAILBOX |
@@ -548,6 +556,13 @@ static inline int flexcan_enter_stop_mode(struct flexcan_priv *priv) }else if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_GPR ) {regmap_update_bits (priv->stm.gpr, priv->stm.req_gpr,1 << priv->stm.req_bit,1 << priv->stm.req_bit); + }else if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI ) { +/* For the SCMI mode, driver do nothing, ATF will send request to +* SM(system manager, M33 core) through SCMI protocol after linux +* suspend. Once SM get this request, it will send IPG_STOP signal +* to Flex_CAN, let CAN in STOP mode. +*/ +return 0 ; }return flexcan_low_power_enter_ack (priv);
@@ -559,7 +574,11 @@ static inline int flexcan_exit_stop_mode(struct flexcan_priv *priv)u32 reg_mcr;int ret; -/* remove stop request */ +/* Remove stop request, for FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, +* do nothing here, because ATF already send request to SM before +* linux resume. Once SM get this request, it will deassert the +* IPG_STOP signal to Flex_CAN. +*/ if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCFW ) { ret =flexcan_stop_mode_enable_scfw (priv, false);if (ret <0 )
@@ -1987,6 +2006,9 @@ static int flexcan_setup_stop_mode(struct platform_device *pdev) ret =flexcan_setup_stop_mode_scfw (pdev); else if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_GPR) ret =flexcan_setup_stop_mode_gpr (pdev); + else if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI) +/* ATF will handle all STOP_IPG related work */ + ret =0 ; else/* return 0 directly if doesn't support stop mode feature */ return 0;
@@ -2013,6 +2035,7 @@ static const struct of_device_id flexcan_of_match[] = { { .compatible = "fsl,imx8qm-flexcan", .data = &fsl_imx8qm_devtype_data, }, { .compatible = "fsl,imx8mp-flexcan", .data = &fsl_imx8mp_devtype_data, }, { .compatible = "fsl,imx93-flexcan", .data = &fsl_imx93_devtype_data, }, + { .compatible = "fsl,imx95-flexcan", .data = &fsl_imx95_devtype_data, }, { .compatible = "fsl,imx6q-flexcan", .data = &fsl_imx6q_devtype_data, }, { .compatible = "fsl,imx28-flexcan", .data = &fsl_imx28_devtype_data, }, { .compatible = "fsl,imx53-flexcan", .data = &fsl_imx25_devtype_data, },
@@ -2311,9 +2334,22 @@ static int __maybe_unused flexcan_noirq_suspend(struct device *device)if (netif_running (dev)) {int err; -if (device_may_wakeup (device)) +if (device_may_wakeup (device)) {flexcan_enable_wakeup_irq (priv, true); +/* For FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, it need +* ATF to send request to SM through SCMI protocol, +* SM will assert the IPG_STOP signal. But all this +* works need the CAN clocks keep on. +* After the CAN module get the IPG_STOP mode, and +* switch to STOP mode, whether still keep the CAN +* clocks on or gate them off depend on the Hardware +* design. +*/ +if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI ) +return 0 ; + } + err =pm_runtime_force_suspend (device);if (err)return err;
@@ -2330,9 +2366,12 @@ static int __maybe_unused flexcan_noirq_resume(struct device *device)if (netif_running (dev)) {int err; - err =pm_runtime_force_resume (device); -if (err) -return err; +if (!(device_may_wakeup (device) && + priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI )) { + err =pm_runtime_force_resume (device); +if (err) +return err; + }if (device_may_wakeup (device))flexcan_enable_wakeup_irq (priv, false);
drivers/net/can/flexcan/flexcan.h +2 -0
@@ -68,6 +68,8 @@#define FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR BIT(15)/* Device supports RX via FIFO */ #define FLEXCAN_QUIRK_SUPPORT_RX_FIFO BIT(16) +/* Setup stop mode with ATF SCMI protocol to support wakeup */ +#define FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI BIT(17)struct flexcan_devtype_data {u32 quirks;/* quirks needed for different IP cores */
Signature · 1 line
2.34.1
Marc Kleine-Budde
maintainerRe: [PATCH v2 4/4] can: flexcan: add wakeup support for imx95
Quoted 40 lines · Frank Li
From: Haibo Chen <[email protected]>
iMX95 defines a bit in GPR that sets/unsets the
IPG_STOPsignal to the FlexCAN module, controlling its entry into STOP mode. Wakeup should work even if FlexCAN is in STOP mode.Due to iMX95 architecture design, the A-Core cannot access GPR; only the system manager (SM) can configure GPR. To support the wakeup feature, follow these steps:
- For suspend:
1) During Linux suspend, when CAN suspends, do nothing for GPR and keep CAN-related clocks on.
2) In ATF, check whether CAN needs to support wakeup; if yes, send a request to SM through the SCMI protocol.
3) In SM, configure the GPR and unsetIPG_STOP.
4) A-Core suspends.- For wakeup and resume:
1) A-Core wakeup event arrives.
2) In SM, deassertIPG_STOP.
3) Linux resumes.Add a new
fsl_imx95_devtype_dataandFLEXCAN_QUIRK_SETUP_STOP_MODE_SCMIto reflect this.Reviewed-by Han Xu <[email protected]>Signed-off-by Haibo Chen <[email protected]>Reviewed-by Vincent Mailhol <[email protected]>Signed-off-by Frank Li <[email protected]>--- drivers/net/can/flexcan/flexcan-core.c | 49 ++++++++++++++++++++++++++++++---- drivers/net/can/flexcan/flexcan.h | 2 ++ 2 files changed, 46 insertions(+), 5 deletions(-)drivers/net/can/flexcan/flexcan-core.c +0
-0
drivers/net/can/flexcan/flexcan-core.c static struct flexcan_devtype_data fsl_imx93_devtype_data = { · line 363
Quoted 10 lines
drivers/net/can/flexcan/flexcan-core.c +7
-0@@ -354,6 +354,14 @@ static struct flexcan_devtype_data fsl_imx93_devtype_data = { FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR, }; +static const struct flexcan_devtype_data fsl_imx95_devtype_data = { + .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_ENABLE_EACEN_RRS | +FLEXCAN_QUIRK_DISABLE_MECR |FLEXCAN_QUIRK_USE_RX_MAILBOX | +FLEXCAN_QUIRK_BROKEN_PERR_STATE |FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI | +FLEXCAN_QUIRK_SUPPORT_FD |FLEXCAN_QUIRK_SUPPORT_ECC | +FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX | +FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR ,
Please keep the flags sorted by their value.
drivers/net/can/flexcan/flexcan-core.c static struct flexcan_devtype_data fsl_imx93_devtype_data = { · line 364
Quoted 1 line
drivers/net/can/flexcan/flexcan-core.c +1
-0+};
Please add a newline here.
Quoted 49 lines
drivers/net/can/flexcan/flexcan-core.c +16
-1static const struct flexcan_devtype_data fsl_vf610_devtype_data = { .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_ENABLE_EACEN_RRS | FLEXCAN_QUIRK_DISABLE_MECR | FLEXCAN_QUIRK_USE_RX_MAILBOX |@@ -548,6 +556,13 @@ static inline int flexcan_enter_stop_mode(struct flexcan_priv *priv) }else if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_GPR ) {regmap_update_bits (priv->stm.gpr, priv->stm.req_gpr,1 << priv->stm.req_bit,1 << priv->stm.req_bit); + }else if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI ) { +/* For the SCMI mode, driver do nothing, ATF will send request to +* SM(system manager, M33 core) through SCMI protocol after linux +* suspend. Once SM get this request, it will send IPG_STOP signal +* to Flex_CAN, let CAN in STOP mode. +*/ +return 0 ; }return flexcan_low_power_enter_ack (priv);@@ -559,7 +574,11 @@ static inline int flexcan_exit_stop_mode(struct flexcan_priv *priv)u32 reg_mcr;int ret; -/* remove stop request */ +/* Remove stop request, for FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, +* do nothing here, because ATF already send request to SM before +* linux resume. Once SM get this request, it will deassert the +* IPG_STOP signal to Flex_CAN. +*/ if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCFW ) { ret =flexcan_stop_mode_enable_scfw (priv, false);if (ret <0 )@@ -1987,6 +2006,9 @@ static int flexcan_setup_stop_mode(struct platform_device *pdev) ret =flexcan_setup_stop_mode_scfw (pdev); else if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_GPR) ret =flexcan_setup_stop_mode_gpr (pdev); + else if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI) +/* ATF will handle all STOP_IPG related work */ + ret =0 ; else/* return 0 directly if doesn't support stop mode feature */ return 0;@@ -2013,6 +2035,7 @@ static const struct of_device_id flexcan_of_match[] = { { .compatible = "fsl,imx8qm-flexcan", .data = &fsl_imx8qm_devtype_data, }, { .compatible = "fsl,imx8mp-flexcan", .data = &fsl_imx8mp_devtype_data, }, { .compatible = "fsl,imx93-flexcan", .data = &fsl_imx93_devtype_data, }, + { .compatible = "fsl,imx95-flexcan", .data = &fsl_imx95_devtype_data, }, { .compatible = "fsl,imx6q-flexcan", .data = &fsl_imx6q_devtype_data, }, { .compatible = "fsl,imx28-flexcan", .data = &fsl_imx28_devtype_data, }, { .compatible = "fsl,imx53-flexcan", .data = &fsl_imx25_devtype_data, },
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_suspend(struct device *device) · line 2340
Quoted 8 lines
drivers/net/can/flexcan/flexcan-core.c +2
-1@@ -2311,9 +2334,22 @@ static int __maybe_unused flexcan_noirq_suspend(struct device *device)if (netif_running (dev)) {int err; -if (device_may_wakeup (device)) +if (device_may_wakeup (device)) {flexcan_enable_wakeup_irq (priv, true); +/* For FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, it need
needs
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_suspend(struct device *device) · line 2344
Quoted 4 lines
drivers/net/can/flexcan/flexcan-core.c +4
-0+* ATF to send request to SM through SCMI protocol, +* SM will assert the IPG_STOP signal. But all this +* works need the CAN clocks keep on. +* After the CAN module get the IPG_STOP mode, and
gets
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_suspend(struct device *device) · line 2345
Quoted 1 line
drivers/net/can/flexcan/flexcan-core.c +1
-0+* switch to STOP mode, whether still keep the CAN
switches
Quoted 11 lines
drivers/net/can/flexcan/flexcan-core.c +7
-0+* clocks on or gate them off depend on the Hardware +* design. +*/ +if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI ) +return 0 ; + } + err =pm_runtime_force_suspend (device);if (err)return err;
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_resume(struct device *device) · line 2369
Quoted 7 lines
drivers/net/can/flexcan/flexcan-core.c +1
-3@@ -2330,9 +2366,12 @@ static int __maybe_unused flexcan_noirq_resume(struct device *device)if (netif_running (dev)) {int err; - err =pm_runtime_force_resume (device); -if (err) -return err; + if (!(device_may_wakeup(device) &&
^^^^^^^^^^^^^^^^^^^^^^^^
Where does this come from?
Quoted 26 lines
drivers/net/can/flexcan/flexcan-core.c +5
-0+ priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI)) { + err =pm_runtime_force_resume (device); +if (err) +return err; + }if (device_may_wakeup (device))flexcan_enable_wakeup_irq (priv, false);drivers/net/can/flexcan/flexcan.h +2
-0@@ -68,6 +68,8 @@#define FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR BIT(15)/* Device supports RX via FIFO */ #define FLEXCAN_QUIRK_SUPPORT_RX_FIFO BIT(16) +/* Setup stop mode with ATF SCMI protocol to support wakeup */ +#define FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI BIT(17)struct flexcan_devtype_data {u32 quirks;/* quirks needed for different IP cores */ Signature · 2 lines
-- 2.34.1
regards,
Marc
Signature · 4 lines
Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung Nürnberg | Phone: +49-5121-206917-129 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
Frank Li
maintainerRe: [PATCH v2 4/4] can: flexcan: add wakeup support for imx95
Quoted 137 lines · Marc Kleine-Budde
Quoted 40 lines · Frank Li
From: Haibo Chen <[email protected]>
iMX95 defines a bit in GPR that sets/unsets the
IPG_STOPsignal to the FlexCAN module, controlling its entry into STOP mode. Wakeup should work even if FlexCAN is in STOP mode.Due to iMX95 architecture design, the A-Core cannot access GPR; only the system manager (SM) can configure GPR. To support the wakeup feature, follow these steps:
- For suspend:
1) During Linux suspend, when CAN suspends, do nothing for GPR and keep CAN-related clocks on.
2) In ATF, check whether CAN needs to support wakeup; if yes, send a request to SM through the SCMI protocol.
3) In SM, configure the GPR and unsetIPG_STOP.
4) A-Core suspends.- For wakeup and resume:
1) A-Core wakeup event arrives.
2) In SM, deassertIPG_STOP.
3) Linux resumes.Add a new
fsl_imx95_devtype_dataandFLEXCAN_QUIRK_SETUP_STOP_MODE_SCMIto reflect this.Reviewed-by Han Xu <[email protected]>Signed-off-by Haibo Chen <[email protected]>Reviewed-by Vincent Mailhol <[email protected]>Signed-off-by Frank Li <[email protected]>--- drivers/net/can/flexcan/flexcan-core.c | 49 ++++++++++++++++++++++++++++++---- drivers/net/can/flexcan/flexcan.h | 2 ++ 2 files changed, 46 insertions(+), 5 deletions(-)drivers/net/can/flexcan/flexcan-core.c +0
-0drivers/net/can/flexcan/flexcan-core.c static struct flexcan_devtype_data fsl_imx93_devtype_data = { · line 363
Quoted 10 lines
drivers/net/can/flexcan/flexcan-core.c +7
-0@@ -354,6 +354,14 @@ static struct flexcan_devtype_data fsl_imx93_devtype_data = { FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR, }; +static const struct flexcan_devtype_data fsl_imx95_devtype_data = { + .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_ENABLE_EACEN_RRS | +FLEXCAN_QUIRK_DISABLE_MECR |FLEXCAN_QUIRK_USE_RX_MAILBOX | +FLEXCAN_QUIRK_BROKEN_PERR_STATE |FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI | +FLEXCAN_QUIRK_SUPPORT_FD |FLEXCAN_QUIRK_SUPPORT_ECC | +FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX | +FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR ,Please keep the flags sorted by their value.
drivers/net/can/flexcan/flexcan-core.c static struct flexcan_devtype_data fsl_imx93_devtype_data = { · line 364
Quoted 1 line
drivers/net/can/flexcan/flexcan-core.c +1
-0+};Please add a newline here.
Quoted 49 lines
drivers/net/can/flexcan/flexcan-core.c +16
-1static const struct flexcan_devtype_data fsl_vf610_devtype_data = { .quirks =FLEXCAN_QUIRK_DISABLE_RXFG |FLEXCAN_QUIRK_ENABLE_EACEN_RRS | FLEXCAN_QUIRK_DISABLE_MECR | FLEXCAN_QUIRK_USE_RX_MAILBOX |@@ -548,6 +556,13 @@ static inline int flexcan_enter_stop_mode(struct flexcan_priv *priv) }else if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_GPR ) {regmap_update_bits (priv->stm.gpr, priv->stm.req_gpr,1 << priv->stm.req_bit,1 << priv->stm.req_bit); + }else if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI ) { +/* For the SCMI mode, driver do nothing, ATF will send request to +* SM(system manager, M33 core) through SCMI protocol after linux +* suspend. Once SM get this request, it will send IPG_STOP signal +* to Flex_CAN, let CAN in STOP mode. +*/ +return 0 ; }return flexcan_low_power_enter_ack (priv);@@ -559,7 +574,11 @@ static inline int flexcan_exit_stop_mode(struct flexcan_priv *priv)u32 reg_mcr;int ret; -/* remove stop request */ +/* Remove stop request, for FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, +* do nothing here, because ATF already send request to SM before +* linux resume. Once SM get this request, it will deassert the +* IPG_STOP signal to Flex_CAN. +*/ if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCFW ) { ret =flexcan_stop_mode_enable_scfw (priv, false);if (ret <0 )@@ -1987,6 +2006,9 @@ static int flexcan_setup_stop_mode(struct platform_device *pdev) ret =flexcan_setup_stop_mode_scfw (pdev); else if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_GPR) ret =flexcan_setup_stop_mode_gpr (pdev); + else if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI) +/* ATF will handle all STOP_IPG related work */ + ret =0 ; else/* return 0 directly if doesn't support stop mode feature */ return 0;@@ -2013,6 +2035,7 @@ static const struct of_device_id flexcan_of_match[] = { { .compatible = "fsl,imx8qm-flexcan", .data = &fsl_imx8qm_devtype_data, }, { .compatible = "fsl,imx8mp-flexcan", .data = &fsl_imx8mp_devtype_data, }, { .compatible = "fsl,imx93-flexcan", .data = &fsl_imx93_devtype_data, }, + { .compatible = "fsl,imx95-flexcan", .data = &fsl_imx95_devtype_data, }, { .compatible = "fsl,imx6q-flexcan", .data = &fsl_imx6q_devtype_data, }, { .compatible = "fsl,imx28-flexcan", .data = &fsl_imx28_devtype_data, }, { .compatible = "fsl,imx53-flexcan", .data = &fsl_imx25_devtype_data, },drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_suspend(struct device *device) · line 2340
Quoted 8 lines
drivers/net/can/flexcan/flexcan-core.c +2
-1@@ -2311,9 +2334,22 @@ static int __maybe_unused flexcan_noirq_suspend(struct device *device)if (netif_running (dev)) {int err; -if (device_may_wakeup (device)) +if (device_may_wakeup (device)) {flexcan_enable_wakeup_irq (priv, true); +/* For FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, it need needs
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_suspend(struct device *device) · line 2344
Quoted 4 lines
drivers/net/can/flexcan/flexcan-core.c +4
-0+* ATF to send request to SM through SCMI protocol, +* SM will assert the IPG_STOP signal. But all this +* works need the CAN clocks keep on. +* After the CAN module get the IPG_STOP mode, and gets
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_suspend(struct device *device) · line 2345
Quoted 1 line
drivers/net/can/flexcan/flexcan-core.c +1
-0+* switch to STOP mode, whether still keep the CAN switches
Quoted 14 lines
drivers/net/can/flexcan/flexcan-core.c +7
-0+* clocks on or gate them off depend on the Hardware +* design. +*/ +if (priv->devtype_data.quirks &FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI ) +return 0 ; + } + err =pm_runtime_force_suspend (device);if (err)return err;@@ -2330,9 +2366,12 @@ static int __maybe_unused flexcan_noirq_resume(struct device *device)if (netif_running (dev)) {int err;
Quoted 7 lines
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_resume(struct device *device) · line 2369
Quoted 4 lines
drivers/net/can/flexcan/flexcan-core.c +1
-3- err =pm_runtime_force_resume (device); -if (err) -return err; + if (!(device_may_wakeup(device) &&^^^^^^^^^^^^^^^^^^^^^^^^Where does this come from?
include/linux/pm_wakeup.h
static inline bool device_may_wakeup(struct device *dev)
{
return dev->power.can_wakeup && !!dev->power.wakeup;
}Frank
Quoted 36 lines
Quoted 27 lines
drivers/net/can/flexcan/flexcan-core.c +5
-0+ priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI)) { + err =pm_runtime_force_resume (device); +if (err) +return err; + }if (device_may_wakeup (device))flexcan_enable_wakeup_irq (priv, false);drivers/net/can/flexcan/flexcan.h +2
-0@@ -68,6 +68,8 @@#define FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR BIT(15)/* Device supports RX via FIFO */ #define FLEXCAN_QUIRK_SUPPORT_RX_FIFO BIT(16) +/* Setup stop mode with ATF SCMI protocol to support wakeup */ +#define FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI BIT(17)struct flexcan_devtype_data {u32 quirks;/* quirks needed for different IP cores */ Signature · 2 lines
-- 2.34.1regards,
MarcSignature · 4 lines
Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung Nürnberg | Phone: +49-5121-206917-129 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
Marc Kleine-Budde
maintainerRe: [PATCH v2 4/4] can: flexcan: add wakeup support for imx95
Quoted 10 lines · Frank Li
Quoted 10 lines
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_resume(struct device *device) · line 2369
Quoted 8 lines
drivers/net/can/flexcan/flexcan-core.c +1
-3@@ -2330,9 +2366,12 @@ static int __maybe_unused flexcan_noirq_resume(struct device *device)if (netif_running (dev)) {int err; - err =pm_runtime_force_resume (device); -if (err) -return err; + if (!(device_may_wakeup(device) &&^^^^^^^^^^^^^^^^^^^^^^^^
Quoted 8 lines
Quoted 1 line
Where does this come from?
include/linux/pm_wakeup.hstatic inline bool
device_may_wakeup(struct device *dev){ return dev->power.can_wakeup && !!dev->power.wakeup; }
Sorry for the confusion. I wanted to point out, that the original driver doesn't have the check to device_may_wakeup(). Why was this added?
Quoted 20 lines
Quoted 20 lines
Quoted 20 lines
drivers/net/can/flexcan/flexcan-core.c +5
-0+ priv->devtype_data.quirks & FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI)) { + err =pm_runtime_force_resume (device); +if (err) +return err; + }if (device_may_wakeup (device))flexcan_enable_wakeup_irq (priv, false);drivers/net/can/flexcan/flexcan.h +2
-0@@ -68,6 +68,8 @@#define FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR BIT(15)/* Device supports RX via FIFO */ #define FLEXCAN_QUIRK_SUPPORT_RX_FIFO BIT(16) +/* Setup stop mode with ATF SCMI protocol to support wakeup */ +#define FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI BIT(17)
Quoted 2 lines
Quoted 2 lines
drivers/net/can/flexcan/flexcan.h line 75
Quoted 2 lines
drivers/net/can/flexcan/flexcan.h +0
-0struct flexcan_devtype_data {u32 quirks;/* quirks needed for different IP cores */
regards,
Marc
Signature · 4 lines
Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung Nürnberg | Phone: +49-5121-206917-129 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
RE: [PATCH v2 4/4] can: flexcan: add wakeup support for imx95
Quoted 30 lines
-----Original Message-----
From: Marc Kleine-Budde <[email protected]>
Sent: 2024年7月16日 22:46
To: Frank Li <[email protected]>
Cc: Vincent Mailhol <[email protected]>; David S. Miller <[email protected]>; Eric Dumazet <[email protected]>; Jakub Kicinski <[email protected]>; Paolo Abeni <[email protected]>; Rob Herring <[email protected]>; Krzysztof Kozlowski <[email protected]>; Conor Dooley <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; Bough Chen <[email protected]>; [email protected]; Han Xu <[email protected]>
Subject: Re: [PATCH v2 4/4] can: flexcan: add wakeup support for imx95Quoted 1 line · Frank Li
Quoted 1 line
Review static int __maybe_unused
Quoted 1 line
(quoted diff) +0
-0@@ -2330,9 +2366,12 @@ static int __maybe_unused
flexcan_noirq_resume(struct device *device)Quoted 13 lines
Quoted 10 lines
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_resume(struct device *device) · line 2369
Quoted 7 lines
drivers/net/can/flexcan/flexcan-core.c +1
-3if (netif_running (dev)) {int err; - err =pm_runtime_force_resume (device); -if (err) -return err; + if (!(device_may_wakeup(device) &&^^^^^^^^^^^^^^^^^^^^^^^^Where does this come from?
include/linux/pm_wakeup.h
Quoted 7 lines
Quoted 4 lines
static inline bool
device_may_wakeup(struct device *dev){ return dev->power.can_wakeup && !!dev->power.wakeup; }Sorry for the confusion. I wanted to point out, that the original driver doesn't have the check to
device_may_wakeup(). Why was this added?
Hi Marc,
Here add this to make sure for CAN with flag FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI and really used as wakeup source, do not need to call pm_runtime_force_resume(), keep it align with
what we do in flexcan_noirq_suspend.
As the comment in flexcan_noirq_suspend, CAN with flag FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, when used as wakeup source, need to keep CAN clock on when system suspend, let ATF part logic works, detail steps please refer to this patch commit log. Whether gate off the CAN clock or not depends on the Hardware design. So for this case, in flexcan_noirq_suspend, directly return0, do not call the pm_runtime_force_suspend(), then in flexcan_noirq_resume(), use the same logic to skip the pm_runtime_force_resume().
Best Regards
Haibo Chen
Quoted 34 lines
Quoted 3 lines
Quoted 3 lines
Review static int __maybe_unused · line 2370
Quoted 3 lines
(quoted diff) +1
-0+ priv->devtype_data.quirks &
FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI)) {Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
drivers/net/can/flexcan/flexcan-core.c +4
-0+ err =pm_runtime_force_resume (device); +if (err) +return err; + }if (device_may_wakeup (device)) flexcan_enable_wakeup_irq(priv, false); diff --git
a/drivers/net/can/flexcan/flexcan.hb/drivers/net/can/flexcan/flexcan.h
index025c3417031f4..4933d8c7439e6100644drivers/net/can/flexcan/flexcan.h +2
-0@@ -68,6 +68,8 @@#define FLEXCAN_QUIRK_SUPPORT_RX_MAILBOX_RTR BIT(15)/* Device supports RX via FIFO */ #define FLEXCAN_QUIRK_SUPPORT_RX_FIFO BIT(16) +/* Setup stop mode with ATF SCMI protocol to support wakeup */ +#define FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI BIT(17)Quoted 2 lines
Quoted 2 lines
drivers/net/can/flexcan/flexcan.h line 75
Quoted 2 lines
drivers/net/can/flexcan/flexcan.h +0
-0struct flexcan_devtype_data {u32 quirks;/* quirks needed for different IP cores */ regards,
Marc--
Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de |Vertretung Nürnberg | Phone: +49-5121-206917-129 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
Marc Kleine-Budde
maintainerRe: RE: [PATCH v2 4/4] can: flexcan: add wakeup support for imx95
Quoted 29 lines · Bough Chen
Quoted 16 lines · Frank Li
Quoted 1 line
Quoted 1 line
Review static int __maybe_unused
Quoted 1 line
(quoted diff) +0
-0@@ -2330,9 +2366,12 @@ static int __maybe_unused
flexcan_noirq_resume(struct device *device)Quoted 13 lines
Quoted 10 lines
drivers/net/can/flexcan/flexcan-core.c static int __maybe_unused flexcan_noirq_resume(struct device *device) · line 2369
Quoted 7 lines
drivers/net/can/flexcan/flexcan-core.c +1
-3if (netif_running (dev)) {int err; - err =pm_runtime_force_resume (device); -if (err) -return err; + if (!(device_may_wakeup(device) &&^^^^^^^^^^^^^^^^^^^^^^^^Where does this come from?
include/linux/pm_wakeup.hQuoted 7 lines
Quoted 4 lines
static inline bool
device_may_wakeup(struct device *dev){ return dev->power.can_wakeup && !!dev->power.wakeup; }Sorry for the confusion. I wanted to point out, that the original driver doesn't have the check to
device_may_wakeup(). Why was this added?Here add this to make sure for CAN with flag
FLEXCAN_QUIRK_SETUP_STOP_MODE_SCMIand really used as wakeup source, do not need to callpm_runtime_force_resume(), keep it align with what we do inflexcan_noirq_suspend.
Quoted 8 lines
As the comment in
flexcan_noirq_suspend, CAN with flagFLEXCAN_QUIRK_SETUP_STOP_MODE_SCMI, when used as wakeup source, need to keep CAN clock on when system suspend, let ATF part logic works, detail steps please refer to this patch commit log. Whether gate off the CAN clock or not depends on the Hardware design. So for this case, inflexcan_noirq_suspend, directly return0, do not call thepm_runtime_force_suspend(), then inflexcan_noirq_resume(), use the same logic to skip thepm_runtime_force_resume().
Please change the control flow, so that flexcan_noirq_suspend() and flexcan_noirq_resume() look symmetrical.
regards,
Marc
Signature · 4 lines
Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung Nürnberg | Phone: +49-5121-206917-129 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |