spmi: pmic_arb: add support for hw version 2
Subsystems (3)
PATCH 1/1 spmi: pmic_arb: add support for hw version 2
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
- New observer register space. All rx trafic uses this space.
- Diffrent command format for spmi command registers.
--- .../bindings/spmi/qcom,spmi-pmic-arb.txt | 11 +- drivers/spmi/spmi-pmic-arb.c | 295 ++++++++++++++++++--- 2 files changed, 263 insertions(+), 43 deletions(-)
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +8 -3
@@ -1,11 +1,11 @@ Qualcomm SPMI Controller (PMIC Arbiter) -The SPMI PMIC Arbiter is found on the Snapdragon 800 Series. It is an SPMI +The SPMI PMIC Arbiter is found on Snapdragon chipsets. It is an SPMI controller with wrapping arbitration logic to allow for multiple on-chip devices to control a single SPMI master. -The PMIC Arbiter can also act as an interrupt controller, providing interrupts -to slave devices. +The PMIC Arbiter is also an interrupt controller, interrupting the Snapdragon +on dtection of a sequence initiated by a request-capable-slave to the master. See spmi.txt for the generic SPMI controller binding requirements for child nodes.
@@ -38,6 +38,11 @@ Required properties: cell 4: interrupt flags indicating level-sense information, as defined in dt-bindings/interrupt-controller/irq.h +Optional properties: +- reg-names : may have "chnls", "obsrvr" + "chnls" - tx-channel per virtual slave registers. + "obsrvr" - rx-channel (called observer) per virtual slave registers. + Example: spmi {
drivers/spmi/spmi-pmic-arb.c +255 -40
450 lines not on this page · raw message
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
Hi Gilad,
Just few comments.
Quoted 11 lines · Gilad Avidov
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
- New observer register space. All rx trafic uses this space.
- Diffrent command format for spmi command registers.Signed-off-by Gilad Avidov <[email protected]>Acked-by Sagar Dharia <[email protected]>---
<snip>
drivers/spmi/spmi-pmic-arb.c enum pmic_arb_cmd_op_code { · line 106
Quoted 3 lines
drivers/spmi/spmi-pmic-arb.c +2
-0+struct pmic_arb_ver ; +
Is there a better name for this structure. Ok it contain version information, but.
Quoted 6 lines
/**
*spmi_pmic_arb_dev- SPMI PMIC Arbiter object
*- * @base: address of the PMIC Arbiter core registers. + * @rd_base: on v1 "core", on v2 "observer" register base off DT. + * @rd_base: on v1 "core", on v2 "chnls" register base off DT.
s/rd/wr/
Quoted 66 lines
* @intr: address of the SPMI interrupt control registers. * @cnfg: address of the PMIC Arbiter configuration registers. * @lock: lock to synchronize accesses. - * @channel: which channel to use for accesses. + * @channel: which ee channel to use for accesses. * @irq: PMIC ARB interrupt.* @ee: the current Execution Environment
* @min_apid: minimum APID (used for bounding IRQ search)drivers/spmi/spmi-pmic-arb.c +41
-1@@ -114,9 +121,11 @@ enum pmic_arb_cmd_op_code {* @domain: irq domain object for PMIC IRQ domain * @spmic: SPMI controller object * @apid_to_ppid: cached mapping from APID to PPID +* @ppid_to_chan cached mapping from APID to channel number, v2 only. */ struct spmi_pmic_arb_dev { -void __iomem *base; +void __iomem *rd_base; +void __iomem *wr_base;void __iomem *intr;void __iomem *cnfg;raw_spinlock_tlock ;@@ -129,17 +138,61 @@ struct spmi_pmic_arb_dev {struct irq_domain *domain;struct spmi_controller *spmic;u16 apid_to_ppid[256 ]; +const struct pmic_arb_ver *ver; +u8 *ppid_to_chan; +}; + +/** +* pmic_arb_ver: version dependent functionality and values. +* +* @non_data_cmd: on v1 issues an spmi non-data command. +* on v2 no HW support, returns -EOPNOTSUPP. +* @offset: on v1 offset of per-ee channel. +* on v2 offset of per-ee and per-ppid channel. +* @fmt_cmd: formats a GENI/SPMI command. +* @owner_acc_status: on v1 offset of PMIC_ARB_SPMI_PIC_OWNERm_ACC_STATUSn +* on v2 offset of SPMI_PIC_OWNERm_ACC_STATUSn. +* @acc_enable: on v1 offset of PMIC_ARB_SPMI_PIC_ACC_ENABLEn +* on v2 offset of SPMI_PIC_ACC_ENABLEn. +* @irq_status: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_STATUSn +* on v2 offset of SPMI_PIC_IRQ_STATUSn. +* @irq_clear: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_CLEARn +* on v2 offset of SPMI_PIC_IRQ_CLEARn. +* @geni_ctrl: PMIC_ARB_GENI_CTRL offset. +* @geni_status: PMIC_ARB_GENI_STATUS offset. +* @protocol_irq_status: PMIC_ARB_PROTOCOL_IRQ_STATUS offset. +*/ +struct pmic_arb_ver { +int (*non_data_cmd)(struct spmi_controller *ctrl,u8 opc,u8 sid); +/* SPMI commands (including data) related functionality */ +off_t (*offset)(struct spmi_pmic_arb_dev *dev,u8 sid,u16 addr); +u32 (*fmt_cmd)(u8 opc,u8 sid,u16 addr,u8 bc); +/* Interrupts related functionality (offset of PIC registers) */ +off_t (*owner_acc_status)(u8 m,u8 n); +off_t (*acc_enable)(u8 n); +off_t (*irq_status)(u8 n); +off_t (*irq_clear)(u8 n); +/* Register offsets */ +off_t geni_ctrl; +off_t geni_status; +off_t protocol_irq_status; };
<snip>
drivers/spmi/spmi-pmic-arb.c static int pmic_arb_wait_for_done(struct spmi_controller *ctrl) · line 299
Quoted 8 lines
drivers/spmi/spmi-pmic-arb.c +7
-0+ +/* Non-data command */ +static int pmic_arb_cmd (struct spmi_controller *ctrl,u8 opc,u8 sid) +{ +struct spmi_pmic_arb_dev *pmic_arb =spmi_controller_get_drvdata (ctrl); + +pr_debug ("op:0x%x sid:%d\n" , opc, sid);
Please use dev_dbg.
Review
Quoted 2 lines
(quoted diff) +2
-0+static const struct pmic_arb_ver pmic_arb_v1 = { + .non_data_cmd= pmic_arb_non_data_cmd_v1,
Missing space before =
drivers/spmi/spmi-pmic-arb.c static int qpnpint_irq_domain_map(struct irq_domain *d, · line 768
Quoted 2 lines
drivers/spmi/spmi-pmic-arb.c +2
-0+ .offset = pmic_arb_offset_v1, + .fmt_cmd = pmic_arb_fmt_cmd_v1,
Bad indentation... and bellow
Quoted 26 lines
drivers/spmi/spmi-pmic-arb.c +22
-0+ .owner_acc_status= pmic_arb_owner_acc_status_v1, + .acc_enable= pmic_arb_acc_enable_v1, + .irq_status= pmic_arb_irq_status_v1, + .irq_clear= pmic_arb_irq_clear_v1, + .geni_ctrl= 0x24, + .geni_status= 0x28, + .protocol_irq_status= (0x700 + 0x820), +}; + +static const struct pmic_arb_ver pmic_arb_v2 = { + .non_data_cmd= pmic_arb_non_data_cmd_v2, + .offset = pmic_arb_offset_v2, + .fmt_cmd = pmic_arb_fmt_cmd_v2, + .owner_acc_status= pmic_arb_owner_acc_status_v2, + .acc_enable= pmic_arb_acc_enable_v2, + .irq_status= pmic_arb_irq_status_v2, + .irq_clear= pmic_arb_irq_clear_v2, + .geni_ctrl=0x28 , + .geni_status=0x2C , + .protocol_irq_status= (0x700 +0x900 ), +}; +static const struct irq_domain_ops pmic_arb_irq_domain_ops = { .map = qpnpint_irq_domain_map, .xlate = qpnpint_irq_domain_dt_translate,
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev) · line 805
Quoted 9 lines
drivers/spmi/spmi-pmic-arb.c +3
-1@@ -634,9 +798,11 @@ static int spmi_pmic_arb_probe(struct platform_device *pdev)struct spmi_pmic_arb_dev *pa;struct spmi_controller *ctrl;struct resource *res; -u32 channel, ee; +u32 channel, ee, hw_ver;int err, i; +pr_err ("PMIC Arbiter\n" ); +
leftover from debug?
Quoted 19 lines
drivers/spmi/spmi-pmic-arb.c +6
-3ctrl =spmi_controller_alloc (&pdev->dev,sizeof (*pa));if (!ctrl)return -ENOMEM ;@@ -645,12 +811,64 @@ static int spmi_pmic_arb_probe(struct platform_device *pdev) pa->spmic = ctrl; res =platform_get_resource_byname (pdev,IORESOURCE_MEM ,"core" ); - pa->base =devm_ioremap_resource (&ctrl->dev, res); -if (IS_ERR (pa->base)) { - err =PTR_ERR (pa->base); + pa->rd_base =devm_ioremap_resource (&ctrl->dev, res); +if (IS_ERR (pa->rd_base)) { + err =PTR_ERR (pa->rd_base); goto err_put_ctrl; } + hw_ver =readl_relaxed (pa->rd_base +PMIC_ARB_VERSION ); + +if (hw_ver <PMIC_ARB_VERSION_V2_MIN ) {
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev) · line 832
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +10
-0+ pa->ver= &pmic_arb_v1; +dev_dbg (&ctrl->dev,"PMIC Arb Version-1 (0x%x)\n" , hw_ver); + pa->wr_base = pa->rd_base; + }else { +u8 chan; +u16 ppid; +u32 regval; + + pa->ver = &pmic_arb_v2; +dev_dbg (&ctrl->dev,"PMIC Arb Version-2 (0x%x)\n" , hw_ver);
Do we really need two almost the same dev_dbg's?
Quoted 24 lines
drivers/spmi/spmi-pmic-arb.c +23
-0+ + pa->ppid_to_chan =devm_kzalloc (&ctrl->dev, +PPID_TO_CHAN_TABLE_SZ ,GFP_KERNEL ); +if (!pa->ppid_to_chan) { +dev_err (&ctrl->dev, +"faild allocation of ppid_to_chan table\n" ); + err = -ENOMEM ; + goto err_put_ctrl; + } +/* +* PMIC_ARB_REG_CHNL is a table in HW mapping channel to ppid. +* ppid_to_chan is an inverted cache of that table. +*/ +for (chan =0 ; chan <PMIC_ARB_MAX_CHNL ; ++chan) { + regval =readl_relaxed (pa->rd_base + +PMIC_ARB_REG_CHNL (chan)); +if (!regval) +continue ; + + ppid = (regval >>8 ) &0xFFF ; + pa->ppid_to_chan[ppid] = chan; + } +
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev)
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +10
-0+ res =platform_get_resource_byname (pdev,IORESOURCE_MEM , +"obsrvr" ); + pa->rd_base =devm_ioremap_resource (&ctrl->dev, res); +if (IS_ERR (pa->rd_base)) { + err =PTR_ERR (pa->rd_base); + goto err_put_ctrl; + } + + res =platform_get_resource_byname (pdev,IORESOURCE_MEM , +"chnls" );
it looks like
pa->wr_base = devm_ioremap_resource(&ctrl->dev, res);
is missing?
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev)
Quoted 7 lines
drivers/spmi/spmi-pmic-arb.c +6
-0+if (IS_ERR (pa->wr_base)) { + err =PTR_ERR (pa->wr_base); + goto err_put_ctrl; + } + } +
Thanks Ivan.
Hi Ivan,
Quoted 16 lines · Ivan T. Ivanov
Hi Gilad,
Just few comments.
Quoted 11 lines · Gilad Avidov
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
- New observer register space. All rx trafic uses this space.
- Diffrent command format for spmi command registers.Signed-off-by Gilad Avidov <[email protected]>Acked-by Sagar Dharia <[email protected]>---
Quoted 6 lines
<snip>
drivers/spmi/spmi-pmic-arb.c enum pmic_arb_cmd_op_code { · line 106
Quoted 2 lines
drivers/spmi/spmi-pmic-arb.c +2
-0+struct pmic_arb_ver ; +Is there a better name for this structure. Ok it contain version information, but.
will rename to pmic_arb_ver_ops.
Quoted 77 lines
Quoted 7 lines
/**
*spmi_pmic_arb_dev- SPMI PMIC Arbiter object
*- * @base: address of the PMIC Arbiter core registers. + * @rd_base: on v1 "core", on v2 "observer" register base off DT. + * @rd_base: on v1 "core", on v2 "chnls" register base off DT.s/rd/wr/
Quoted 66 lines
* @intr: address of the SPMI interrupt control registers. * @cnfg: address of the PMIC Arbiter configuration registers. * @lock: lock to synchronize accesses.- * @channel: which channel to use for accesses.
+ * @channel: which ee channel to use for accesses.
* @irq: PMIC ARB interrupt.
* @ee: the current Execution Environment
* @min_apid: minimum APID (used for bounding IRQ search)drivers/spmi/spmi-pmic-arb.c +41
-1@@ -114,9 +121,11 @@ enum pmic_arb_cmd_op_code {* @domain: irq domain object for PMIC IRQ domain * @spmic: SPMI controller object * @apid_to_ppid: cached mapping from APID to PPID +* @ppid_to_chan cached mapping from APID to channel number, v2 only. */ struct spmi_pmic_arb_dev { -void __iomem *base; +void __iomem *rd_base; +void __iomem *wr_base;void __iomem *intr;void __iomem *cnfg;raw_spinlock_tlock ;@@ -129,17 +138,61 @@ struct spmi_pmic_arb_dev {struct irq_domain *domain;struct spmi_controller *spmic;u16 apid_to_ppid[256 ]; +const struct pmic_arb_ver *ver; +u8 *ppid_to_chan; +}; + +/** +* pmic_arb_ver: version dependent functionality and values. +* +* @non_data_cmd: on v1 issues an spmi non-data command. +* on v2 no HW support, returns -EOPNOTSUPP. +* @offset: on v1 offset of per-ee channel. +* on v2 offset of per-ee and per-ppid channel. +* @fmt_cmd: formats a GENI/SPMI command. +* @owner_acc_status: on v1 offset of PMIC_ARB_SPMI_PIC_OWNERm_ACC_STATUSn +* on v2 offset of SPMI_PIC_OWNERm_ACC_STATUSn. +* @acc_enable: on v1 offset of PMIC_ARB_SPMI_PIC_ACC_ENABLEn +* on v2 offset of SPMI_PIC_ACC_ENABLEn. +* @irq_status: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_STATUSn +* on v2 offset of SPMI_PIC_IRQ_STATUSn. +* @irq_clear: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_CLEARn +* on v2 offset of SPMI_PIC_IRQ_CLEARn. +* @geni_ctrl: PMIC_ARB_GENI_CTRL offset. +* @geni_status: PMIC_ARB_GENI_STATUS offset. +* @protocol_irq_status: PMIC_ARB_PROTOCOL_IRQ_STATUS offset. +*/ +struct pmic_arb_ver { +int (*non_data_cmd)(struct spmi_controller *ctrl,u8 opc,u8 sid); +/* SPMI commands (including data) related functionality */ +off_t (*offset)(struct spmi_pmic_arb_dev *dev,u8 sid,u16 addr); +u32 (*fmt_cmd)(u8 opc,u8 sid,u16 addr,u8 bc); +/* Interrupts related functionality (offset of PIC registers) */ +off_t (*owner_acc_status)(u8 m,u8 n); +off_t (*acc_enable)(u8 n); +off_t (*irq_status)(u8 n); +off_t (*irq_clear)(u8 n); +/* Register offsets */ +off_t geni_ctrl; +off_t geni_status; +off_t protocol_irq_status; };<snip>
Quoted 9 lines
drivers/spmi/spmi-pmic-arb.c static int pmic_arb_wait_for_done(struct spmi_controller *ctrl) · line 299
Quoted 8 lines
drivers/spmi/spmi-pmic-arb.c +7
-0+ +/* Non-data command */ +static int pmic_arb_cmd (struct spmi_controller *ctrl,u8 opc,u8 sid) +{ +struct spmi_pmic_arb_dev *pmic_arb =spmi_controller_get_drvdata (ctrl); + +pr_debug ("op:0x%x sid:%d\n" , opc, sid);Please use
dev_dbg.
agree
Quoted 5 lines
Review
Quoted 4 lines
(quoted diff) +2
-0+static const struct pmic_arb_ver pmic_arb_v1 = { + .non_data_cmd= pmic_arb_non_data_cmd_v1,Missing space before =
will fix
Quoted 38 lines
drivers/spmi/spmi-pmic-arb.c static int qpnpint_irq_domain_map(struct irq_domain *d, · line 768
Quoted 3 lines
drivers/spmi/spmi-pmic-arb.c +2
-0+ .offset = pmic_arb_offset_v1, + .fmt_cmd = pmic_arb_fmt_cmd_v1,Bad indentation... and bellow
Quoted 33 lines
drivers/spmi/spmi-pmic-arb.c +23
-1+ .owner_acc_status= pmic_arb_owner_acc_status_v1, + .acc_enable= pmic_arb_acc_enable_v1, + .irq_status= pmic_arb_irq_status_v1, + .irq_clear= pmic_arb_irq_clear_v1, + .geni_ctrl= 0x24, + .geni_status= 0x28, + .protocol_irq_status= (0x700 + 0x820), +}; + +static const struct pmic_arb_ver pmic_arb_v2 = { + .non_data_cmd= pmic_arb_non_data_cmd_v2, + .offset = pmic_arb_offset_v2, + .fmt_cmd = pmic_arb_fmt_cmd_v2, + .owner_acc_status= pmic_arb_owner_acc_status_v2, + .acc_enable= pmic_arb_acc_enable_v2, + .irq_status= pmic_arb_irq_status_v2, + .irq_clear= pmic_arb_irq_clear_v2, + .geni_ctrl=0x28 , + .geni_status=0x2C , + .protocol_irq_status= (0x700 +0x900 ), +}; +static const struct irq_domain_ops pmic_arb_irq_domain_ops = { .map = qpnpint_irq_domain_map, .xlate = qpnpint_irq_domain_dt_translate,@@ -634,9 +798,11 @@ static int spmi_pmic_arb_probe(struct platform_device *pdev)struct spmi_pmic_arb_dev *pa;struct spmi_controller *ctrl;struct resource *res; -u32 channel, ee; +u32 channel, ee, hw_ver;int err, i;
Quoted 3 lines
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev) · line 805
Quoted 2 lines
drivers/spmi/spmi-pmic-arb.c +2
-0+pr_err ("PMIC Arbiter\n" ); +leftover from debug?
will remove.
Quoted 17 lines
Quoted 17 lines
drivers/spmi/spmi-pmic-arb.c +3
-3ctrl =spmi_controller_alloc (&pdev->dev,sizeof (*pa));if (!ctrl)return -ENOMEM ;@@ -645,12 +811,64 @@ static int spmi_pmic_arb_probe(struct platform_device *pdev) pa->spmic = ctrl; res =platform_get_resource_byname (pdev,IORESOURCE_MEM ,"core" ); - pa->base =devm_ioremap_resource (&ctrl->dev, res); -if (IS_ERR (pa->base)) { - err =PTR_ERR (pa->base); + pa->rd_base =devm_ioremap_resource (&ctrl->dev, res); +if (IS_ERR (pa->rd_base)) { + err =PTR_ERR (pa->rd_base); goto err_put_ctrl; }
Quoted 14 lines
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev) · line 832
Quoted 13 lines
drivers/spmi/spmi-pmic-arb.c +13
-0+ hw_ver =readl_relaxed (pa->rd_base +PMIC_ARB_VERSION ); + +if (hw_ver <PMIC_ARB_VERSION_V2_MIN ) { + pa->ver= &pmic_arb_v1; +dev_dbg (&ctrl->dev,"PMIC Arb Version-1 (0x%x)\n" , hw_ver); + pa->wr_base = pa->rd_base; + }else { +u8 chan; +u16 ppid; +u32 regval; + + pa->ver = &pmic_arb_v2; +dev_dbg (&ctrl->dev,"PMIC Arb Version-2 (0x%x)\n" , hw_ver);Do we really need two almost the same dev_dbg's?
I'll try to combine these two
Quoted 36 lines
Quoted 25 lines
drivers/spmi/spmi-pmic-arb.c +23
-0+ + pa->ppid_to_chan =devm_kzalloc (&ctrl->dev, +PPID_TO_CHAN_TABLE_SZ ,GFP_KERNEL ); +if (!pa->ppid_to_chan) { +dev_err (&ctrl->dev, +"faild allocation of ppid_to_chan table\n" ); + err = -ENOMEM ; + goto err_put_ctrl; + } +/* +* PMIC_ARB_REG_CHNL is a table in HW mapping channel to ppid. +* ppid_to_chan is an inverted cache of that table. +*/ +for (chan =0 ; chan <PMIC_ARB_MAX_CHNL ; ++chan) { + regval =readl_relaxed (pa->rd_base + +PMIC_ARB_REG_CHNL (chan)); +if (!regval) +continue ; + + ppid = (regval >>8 ) &0xFFF ; + pa->ppid_to_chan[ppid] = chan; + } +drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev)
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +10
-0+ res =platform_get_resource_byname (pdev,IORESOURCE_MEM , +"obsrvr" ); + pa->rd_base =devm_ioremap_resource (&ctrl->dev, res); +if (IS_ERR (pa->rd_base)) { + err =PTR_ERR (pa->rd_base); + goto err_put_ctrl; + } + + res =platform_get_resource_byname (pdev,IORESOURCE_MEM , +"chnls" );it looks like
Quoted 2 lines
pa->wr_base = devm_ioremap_resource(&ctrl->dev, res);is missing?
correct, will add.
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c static int spmi_pmic_arb_probe(struct platform_device *pdev)
Quoted 8 lines
drivers/spmi/spmi-pmic-arb.c +6
-0+if (IS_ERR (pa->wr_base)) { + err =PTR_ERR (pa->wr_base); + goto err_put_ctrl; + } + } +Thanks Ivan.
Thanks Gilad
Signature · 2 lines
Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project
Quoted 3 lines · Gilad Avidov
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
s/diffrent/different/
Quoted 2 lines
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
s/tarffic/traffic/
Quoted 1 line
- New observer register space. All rx trafic uses this space.
s/trafic/traffic/
Quoted 1 line
- Diffrent command format for spmi command registers.
s/Diffrent/Different/
Please run spell check.
Quoted 14 lines
Signed-off-by Gilad Avidov <[email protected]>Acked-by Sagar Dharia <[email protected]>--- .../bindings/spmi/qcom,spmi-pmic-arb.txt | 11 +- drivers/spmi/spmi-pmic-arb.c | 295 ++++++++++++++++++--- 2 files changed, 263 insertions(+), 43 deletions(-)Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +0
-0@@ -1,11 +1,11 @@ Qualcomm SPMI Controller (PMIC Arbiter)
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt line 8
Quoted 10 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +3
-3-The SPMI PMIC Arbiter is found on the Snapdragon 800 Series. It is an SPMI +The SPMI PMIC Arbiter is found on Snapdragon chipsets. It is an SPMI controller with wrapping arbitration logic to allow for multiple on-chip devices to control a single SPMI master. -The PMIC Arbiter can also act as an interrupt controller, providing interrupts -to slave devices. +The PMIC Arbiter is also an interrupt controller, interrupting the Snapdragon +on dtection of a sequence initiated by a request-capable-slave to the master.
s/dtection/detection/
Honestly I don't see why this part needs to change either. Please drop this hunk.
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt Required properties: · line 45
Quoted 12 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +5
-0See spmi.txt for the generic SPMI controller binding requirements for child nodes.@@ -38,6 +38,11 @@ Required properties: cell 4: interrupt flags indicating level-sense information, as defined in dt-bindings/interrupt-controller/irq.h +Optional properties: +- reg-names : may have "chnls", "obsrvr" + "chnls" - tx-channel per virtual slave registers. + "obsrvr" - rx-channel (called observer) per virtual slave registers. +
There's already a reg-names in this document and it's not optional. Please merge the two.
Quoted 24 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +0
-0Example: spmi {drivers/spmi/spmi-pmic-arb.c +3
-8@@ -25,16 +25,18 @@/* PMIC Arbiter configuration registers */ #define PMIC_ARB_VERSION 0x0000 +#define PMIC_ARB_VERSION_V2_MIN (0x20010000)#define PMIC_ARB_INT_EN 0x0004 -/* PMIC Arbiter channel registers */ -#define PMIC_ARB_CMD (N ) (0x0800 + (0x80 * (N))) -#define PMIC_ARB_CONFIG (N ) (0x0804 + (0x80 * (N))) -#define PMIC_ARB_STATUS (N ) (0x0808 + (0x80 * (N))) -#define PMIC_ARB_WDATA0 (N ) (0x0810 + (0x80 * (N))) -#define PMIC_ARB_WDATA1 (N ) (0x0814 + (0x80 * (N))) -#define PMIC_ARB_RDATA0 (N ) (0x0818 + (0x80 * (N))) -#define PMIC_ARB_RDATA1 (N ) (0x081C + (0x80 * (N))) +/* PMIC Arbiter channel registers offsets */ +#define PMIC_ARB_CMD (0x00)
drivers/spmi/spmi-pmic-arb.c line 42
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +7
-0+#define PMIC_ARB_CONFIG (0x04) +#define PMIC_ARB_STATUS (0x08) +#define PMIC_ARB_WDATA0 (0x10) +#define PMIC_ARB_WDATA1 (0x14) +#define PMIC_ARB_RDATA0 (0x18) +#define PMIC_ARB_RDATA1 (0x1C) +#define PMIC_ARB_REG_CHNL (N ) (0x800 + 0x4 * (N))/* Interrupt Controller */ #define SPMI_PIC_OWNER_ACC_STATUS (M ,N ) (0x0000 + ((32 * (M)) + (4 * (N))))
It looks like these macros would change too, but nothing has been done here. Interrupts haven't been tested?
Quoted 23 lines
drivers/spmi/spmi-pmic-arb.c +4
-0@@ -52,6 +54,7 @@#define SPMI_MAPPING_TABLE_LEN 255#define SPMI_MAPPING_TABLE_TREE_DEPTH 16/* Maximum of 16-bits */ +#define PPID_TO_CHAN_TABLE_SZ BIT(12)/* PPID is 12bit chan is 1byte*/ /* Ownership Table */ #define SPMI_OWNERSHIP_TABLE_REG (N ) (0x0700 + (4 * (N)))@@ -88,6 +91,7 @@ enum pmic_arb_cmd_op_code {/* Maximum number of support PMIC peripherals */ #define PMIC_ARB_MAX_PERIPHS 256 +#define PMIC_ARB_MAX_CHNL 128#define PMIC_ARB_PERIPH_ID_VALID (1 << 15)#define PMIC_ARB_TIMEOUT_US 100#define PMIC_ARB_MAX_TRANS_BYTES (8)@@ -98,14 +102,17 @@ enum pmic_arb_cmd_op_code {/* interrupt enable bit */ #define SPMI_PIC_ACC_ENABLE_BIT BIT(0) +struct pmic_arb_ver ; +/**
drivers/spmi/spmi-pmic-arb.c enum pmic_arb_cmd_op_code { · line 115
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +3
-2* spmi_pmic_arb_dev - SPMI PMIC Arbiter object * -* @base: address of the PMIC Arbiter core registers. +* @rd_base: on v1 "core", on v2 "observer" register base off DT. +* @rd_base: on v1 "core", on v2 "chnls" register base off DT. * @intr: address of the SPMI interrupt control registers. * @cnfg: address of the PMIC Arbiter configuration registers. * @lock: lock to synchronize accesses. -* @channel: which channel to use for accesses. +* @channel: which ee channel to use for accesses.
Looks like an unnecessary change.
Quoted 38 lines
drivers/spmi/spmi-pmic-arb.c +21
-1* @irq: PMIC ARB interrupt. * @ee: the current Execution Environment * @min_apid: minimum APID (used for bounding IRQ search) @@ -114,9 +121,11 @@ enum pmic_arb_cmd_op_code {* @domain: irq domain object for PMIC IRQ domain * @spmic: SPMI controller object * @apid_to_ppid: cached mapping from APID to PPID +* @ppid_to_chan cached mapping from APID to channel number, v2 only. */ struct spmi_pmic_arb_dev { -void __iomem *base; +void __iomem *rd_base; +void __iomem *wr_base;void __iomem *intr;void __iomem *cnfg;raw_spinlock_t lock;@@ -129,17 +138,61 @@ struct spmi_pmic_arb_dev {struct irq_domain *domain;struct spmi_controller *spmic;u16 apid_to_ppid[256 ]; +const struct pmic_arb_ver *ver; +u8 *ppid_to_chan; +}; + +/** +* pmic_arb_ver: version dependent functionality and values. +* +* @non_data_cmd: on v1 issues an spmi non-data command. +* on v2 no HW support, returns -EOPNOTSUPP. +* @offset: on v1 offset of per-ee channel. +* on v2 offset of per-ee and per-ppid channel. +* @fmt_cmd: formats a GENI/SPMI command. +* @owner_acc_status: on v1 offset of PMIC_ARB_SPMI_PIC_OWNERm_ACC_STATUSn +* on v2 offset of SPMI_PIC_OWNERm_ACC_STATUSn. +* @acc_enable: on v1 offset of PMIC_ARB_SPMI_PIC_ACC_ENABLEn +* on v2 offset of SPMI_PIC_ACC_ENABLEn. +* @irq_status: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_STATUSn +* on v2 offset of SPMI_PIC_IRQ_STATUSn.
drivers/spmi/spmi-pmic-arb.c struct spmi_pmic_arb_dev { · line 168
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +10
-0+* @irq_clear: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_CLEARn +* on v2 offset of SPMI_PIC_IRQ_CLEARn. +* @geni_ctrl: PMIC_ARB_GENI_CTRL offset. +* @geni_status: PMIC_ARB_GENI_STATUS offset. +* @protocol_irq_status: PMIC_ARB_PROTOCOL_IRQ_STATUS offset. +*/ +struct pmic_arb_ver { +int (*non_data_cmd)(struct spmi_controller *ctrl,u8 opc,u8 sid); +/* SPMI commands (including data) related functionality */ +off_t (*offset)(struct spmi_pmic_arb_dev *dev,u8 sid,u16 addr);
Isn't off_t for file offsets? It doesn't seem right to use it for register offsets. Please use u32 or something similar.
Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project
Hi Stephen,
Quoted 5 lines · Stephen Boyd
Quoted 3 lines · Gilad Avidov
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
s/diffrent/different/
will correct the spelling. Thank you for pointing them out.
Quoted 34 lines
Quoted 5 lines
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
s/tarffic/traffic/
Quoted 1 line
- New observer register space. All rx trafic uses this space.
s/trafic/traffic/
Quoted 1 line
- Diffrent command format for spmi command registers.
s/Diffrent/Different/
Please run spell check.
Quoted 19 lines
Signed-off-by Gilad Avidov <[email protected]>Acked-by Sagar Dharia <[email protected]>--- .../bindings/spmi/qcom,spmi-pmic-arb.txt | 11 +- drivers/spmi/spmi-pmic-arb.c | 295 ++++++++++++++++++--- 2 files changed, 263 insertions(+), 43 deletions(-)Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +1
-1@@ -1,11 +1,11 @@ Qualcomm SPMI Controller (PMIC Arbiter) -The SPMI PMIC Arbiter is found on the Snapdragon 800 Series. It is an SPMI +The SPMI PMIC Arbiter is found on Snapdragon chipsets. It is an SPMI controller with wrapping arbitration logic to allow for multiple on-chip devices to control a single SPMI master.
Quoted 8 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt line 8
Quoted 4 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +2
-2-The PMIC Arbiter can also act as an interrupt controller, providing interrupts -to slave devices. +The PMIC Arbiter is also an interrupt controller, interrupting the Snapdragon +on dtection of a sequence initiated by a request-capable-slave to the master.s/dtection/detection/
Honestly I don't see why this part needs to change either. Please drop this hunk.
will drop this.
Quoted 8 lines
Quoted 8 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +0
-0See spmi.txt for the generic SPMI controller binding requirements for child nodes.@@ -38,6 +38,11 @@ Required properties: cell 4: interrupt flags indicating level-sense information, as defined in dt-bindings/interrupt-controller/irq.h
Quoted 7 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt Required properties: · line 45
Quoted 5 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +5
-0+Optional properties: +- reg-names : may have "chnls", "obsrvr" + "chnls" - tx-channel per virtual slave registers. + "obsrvr" - rx-channel (called observer) per virtual slave registers. +There's already a reg-names in this document and it's not optional. Please merge the two.
The existing regs are required for both v1 and v2, while the new ones are required for v2 only. Will combine with the existing reg-name as follows:
- reg-names : must contain:
"core" - core registers
...
v2 pmic arbiter registers:
"chnls" - tx-channel per virtual slave registers. "obsrvr" - rx-channel (called observer) per virtual slave registers.
Quoted 58 lines
Quoted 25 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +0
-0Example: spmi {drivers/spmi/spmi-pmic-arb.c +3
-8@@ -25,16 +25,18 @@/* PMIC Arbiter configuration registers */ #define PMIC_ARB_VERSION 0x0000 +#define PMIC_ARB_VERSION_V2_MIN (0x20010000)#define PMIC_ARB_INT_EN 0x0004 -/* PMIC Arbiter channel registers */ -#define PMIC_ARB_CMD (N ) (0x0800 + (0x80 * (N))) -#define PMIC_ARB_CONFIG (N ) (0x0804 + (0x80 * (N))) -#define PMIC_ARB_STATUS (N ) (0x0808 + (0x80 * (N))) -#define PMIC_ARB_WDATA0 (N ) (0x0810 + (0x80 * (N))) -#define PMIC_ARB_WDATA1 (N ) (0x0814 + (0x80 * (N))) -#define PMIC_ARB_RDATA0 (N ) (0x0818 + (0x80 * (N))) -#define PMIC_ARB_RDATA1 (N ) (0x081C + (0x80 * (N))) +/* PMIC Arbiter channel registers offsets */ +#define PMIC_ARB_CMD (0x00)drivers/spmi/spmi-pmic-arb.c line 42
Quoted 10 lines
drivers/spmi/spmi-pmic-arb.c +7
-0+#define PMIC_ARB_CONFIG (0x04) +#define PMIC_ARB_STATUS (0x08) +#define PMIC_ARB_WDATA0 (0x10) +#define PMIC_ARB_WDATA1 (0x14) +#define PMIC_ARB_RDATA0 (0x18) +#define PMIC_ARB_RDATA1 (0x1C) +#define PMIC_ARB_REG_CHNL (N ) (0x800 + 0x4 * (N))/* Interrupt Controller */ #define SPMI_PIC_OWNER_ACC_STATUS (M ,N ) (0x0000 + ((32 * (M)) + (4 * (N))))It looks like these macros would change too, but nothing has been done here. Interrupts haven't been tested?
Quoted 20 lines
drivers/spmi/spmi-pmic-arb.c +2
-0@@ -52,6 +54,7 @@#define SPMI_MAPPING_TABLE_LEN 255#define SPMI_MAPPING_TABLE_TREE_DEPTH 16/* Maximum of 16-bits */ +#define PPID_TO_CHAN_TABLE_SZ BIT(12)/* PPID is 12bit chan is 1byte*/ /* Ownership Table */ #define SPMI_OWNERSHIP_TABLE_REG (N ) (0x0700 + (4 * (N)))@@ -88,6 +91,7 @@ enum pmic_arb_cmd_op_code {/* Maximum number of support PMIC peripherals */ #define PMIC_ARB_MAX_PERIPHS 256 +#define PMIC_ARB_MAX_CHNL 128#define PMIC_ARB_PERIPH_ID_VALID (1 << 15)#define PMIC_ARB_TIMEOUT_US 100#define PMIC_ARB_MAX_TRANS_BYTES (8)@@ -98,14 +102,17 @@ enum pmic_arb_cmd_op_code {/* interrupt enable bit */ #define SPMI_PIC_ACC_ENABLE_BIT BIT(0)
Quoted 14 lines
drivers/spmi/spmi-pmic-arb.c enum pmic_arb_cmd_op_code { · line 115
Quoted 13 lines
drivers/spmi/spmi-pmic-arb.c +5
-2+struct pmic_arb_ver ; +/** * spmi_pmic_arb_dev - SPMI PMIC Arbiter object * -* @base: address of the PMIC Arbiter core registers. +* @rd_base: on v1 "core", on v2 "observer" register base off DT. +* @rd_base: on v1 "core", on v2 "chnls" register base off DT. * @intr: address of the SPMI interrupt control registers. * @cnfg: address of the PMIC Arbiter configuration registers. * @lock: lock to synchronize accesses. -* @channel: which channel to use for accesses. +* @channel: which ee channel to use for accesses. Looks like an unnecessary change.
Added the "ee" to the channel here, to differentiate it from the new channel registers. It should decrease confusion between the two unrelated concept named channel.
Quoted 43 lines
Quoted 43 lines
drivers/spmi/spmi-pmic-arb.c +25
-1* @irq: PMIC ARB interrupt. * @ee: the current Execution Environment * @min_apid: minimum APID (used for bounding IRQ search) @@ -114,9 +121,11 @@ enum pmic_arb_cmd_op_code {* @domain: irq domain object for PMIC IRQ domain * @spmic: SPMI controller object * @apid_to_ppid: cached mapping from APID to PPID +* @ppid_to_chan cached mapping from APID to channel number, v2 only. */ struct spmi_pmic_arb_dev { -void __iomem *base; +void __iomem *rd_base; +void __iomem *wr_base;void __iomem *intr;void __iomem *cnfg;raw_spinlock_t lock;@@ -129,17 +138,61 @@ struct spmi_pmic_arb_dev {struct irq_domain *domain;struct spmi_controller *spmic;u16 apid_to_ppid[256 ]; +const struct pmic_arb_ver *ver; +u8 *ppid_to_chan; +}; + +/** +* pmic_arb_ver: version dependent functionality and values. +* +* @non_data_cmd: on v1 issues an spmi non-data command. +* on v2 no HW support, returns -EOPNOTSUPP. +* @offset: on v1 offset of per-ee channel. +* on v2 offset of per-ee and per-ppid channel. +* @fmt_cmd: formats a GENI/SPMI command. +* @owner_acc_status: on v1 offset of PMIC_ARB_SPMI_PIC_OWNERm_ACC_STATUSn +* on v2 offset of SPMI_PIC_OWNERm_ACC_STATUSn. +* @acc_enable: on v1 offset of PMIC_ARB_SPMI_PIC_ACC_ENABLEn +* on v2 offset of SPMI_PIC_ACC_ENABLEn. +* @irq_status: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_STATUSn +* on v2 offset of SPMI_PIC_IRQ_STATUSn. +* @irq_clear: on v1 offset of PMIC_ARB_SPMI_PIC_IRQ_CLEARn +* on v2 offset of SPMI_PIC_IRQ_CLEARn. +* @geni_ctrl: PMIC_ARB_GENI_CTRL offset. +* @geni_status: PMIC_ARB_GENI_STATUS offset.
Quoted 9 lines
drivers/spmi/spmi-pmic-arb.c struct spmi_pmic_arb_dev { · line 168
Quoted 6 lines
drivers/spmi/spmi-pmic-arb.c +6
-0+* @protocol_irq_status: PMIC_ARB_PROTOCOL_IRQ_STATUS offset. +*/ +struct pmic_arb_ver { +int (*non_data_cmd)(struct spmi_controller *ctrl,u8 opc,u8 sid); +/* SPMI commands (including data) related functionality */ +off_t (*offset)(struct spmi_pmic_arb_dev *dev,u8 sid,u16 addr);Isn't
off_tfor file offsets? It doesn't seem right to use it for register offsets. Please use u32 or something similar.
Agree.
Thanks,
Gilad
Signature · 2 lines
Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project
Hi Gilad,
<snip>
Quoted 5 lines
Quoted 2 lines
/* Interrupt Controller */
#defineSPMI_PIC_OWNER_ACC_STATUS(M, N)(0x0000+ ((32 * (M)) + (4 * (N))))It looks like these macros would change too, but nothing has been done here. Interrupts haven't been tested?
Stephen is right, the irq related operations are not used in irq_chip methods. Do you plan to add it in next version?
<snip>
Quoted 20 lines
Quoted 20 lines
+/**
+ *pmic_arb_ver: version dependent functionality and values.
+ *
+ * @non_data_cmd: on v1 issues an spmi non-data command.
+ * on v2 no HW support, returns-EOPNOTSUPP.
+ * @offset: on v1 offset of per-ee channel.
+ * on v2 offset of per-ee and per-ppid channel.
+ * @fmt_cmd: formats a GENI/SPMI command.
+ * @owner_acc_status: on v1 offset ofPMIC_ARB_SPMI_PIC_OWNERm_ACC_STATUSn
+ * on v2 offset ofSPMI_PIC_OWNERm_ACC_STATUSn.
+ * @acc_enable: on v1 offset ofPMIC_ARB_SPMI_PIC_ACC_ENABLEn
+ * on v2 offset ofSPMI_PIC_ACC_ENABLEn.
+ * @irq_status: on v1 offset ofPMIC_ARB_SPMI_PIC_IRQ_STATUSn
+ * on v2 offset ofSPMI_PIC_IRQ_STATUSn.
+ * @irq_clear: on v1 offset ofPMIC_ARB_SPMI_PIC_IRQ_CLEARn
+ * on v2 offset ofSPMI_PIC_IRQ_CLEARn.
+ * @geni_ctrl:PMIC_ARB_GENI_CTRLoffset.
+ * @geni_status:PMIC_ARB_GENI_STATUSoffset.
+ * @protocol_irq_status:PMIC_ARB_PROTOCOL_IRQ_STATUSoffset.
+ */
<snip>
Signature · 2 lines
regards, Stan
Hi Gilad,
Quoted 32 lines · Gilad Avidov
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
- New observer register space. All rx trafic uses this space.
- Diffrent command format for spmi command registers.Signed-off-by Gilad Avidov <[email protected]>Acked-by Sagar Dharia <[email protected]>--- .../bindings/spmi/qcom,spmi-pmic-arb.txt | 11 +- drivers/spmi/spmi-pmic-arb.c | 295 ++++++++++++++++++--- 2 files changed, 263 insertions(+), 43 deletions(-)Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +3
-3@@ -1,11 +1,11 @@ Qualcomm SPMI Controller (PMIC Arbiter) -The SPMI PMIC Arbiter is found on the Snapdragon 800 Series. It is an SPMI +The SPMI PMIC Arbiter is found on Snapdragon chipsets. It is an SPMI controller with wrapping arbitration logic to allow for multiple on-chip devices to control a single SPMI master. -The PMIC Arbiter can also act as an interrupt controller, providing interrupts -to slave devices. +The PMIC Arbiter is also an interrupt controller, interrupting the Snapdragon +on dtection of a sequence initiated by a request-capable-slave to the master.
<snip>
Quoted 24 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +5
-8-/* Non-data command */ -static int pmic_arb_cmd(struct spmi_controller *ctrl, u8 opc, u8 sid) +static int +pmic_arb_non_data_cmd_v1(struct spmi_controller *ctrl, u8 opc, u8 sid) { struct spmi_pmic_arb_dev *pmic_arb = spmi_controller_get_drvdata(ctrl); unsigned long flags; u32 cmd; int rc; - - /* Check for valid non-data command */ - if (opc < SPMI_CMD_RESET || opc > SPMI_CMD_WAKEUP) - return -EINVAL; + u32 offset = pmic_arb->ver->offset(pmic_arb, sid, 0); cmd = ((opc | 0x40) << 27) | ((sid & 0xf) << 20); raw_spin_lock_irqsave(&pmic_arb->lock, flags); - pmic_arb_base_write(pmic_arb, PMIC_ARB_CMD(pmic_arb->channel), cmd); - rc = pmic_arb_wait_for_done(ctrl); + pmic_arb_base_write(pmic_arb, offset + PMIC_ARB_CMD, cmd); + rc = pmic_arb_wait_for_done(ctrl, pmic_arb->wr_base, sid, 0); raw_spin_unlock_irqrestore(&pmic_arb->lock, flags);
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt
Quoted 10 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +6
-0return rc; } +/* Unsupported by HW */ +static int +pmic_arb_non_data_cmd_v2(struct spmi_controller *ctrl, u8 opc, u8 sid) +{ + return -EOPNOTSUPP; +}
Does pmic arbiter v2 supports SPMI_CMD_WAKEUP and SPMI_CMD_SHUTDOWN commands? If so how we send those commands to the arbiter when the .non_data_cmd operation returns EOPNOTSUPP. If we returning EOPNOTSUPP the spmi bus .probe method will not call spmi driver .probe. See spmi.c spmi_drv_probe().
<snip>
Signature · 2 lines
regards, Stan
Quoted 65 lines · Stanimir Varbanov
Hi Gilad,
Quoted 32 lines · Gilad Avidov
Qualcomm PMIC Arbiter version-2 changes from version-1 are:
- Some diffrent register offsets.
- New channel register space, one per PMIC peripheral (ppid). All tx tarffic uses these channels.
- New observer register space. All rx trafic uses this space.
- Diffrent command format for spmi command registers.Signed-off-by Gilad Avidov <[email protected]>Acked-by Sagar Dharia <[email protected]>--- .../bindings/spmi/qcom,spmi-pmic-arb.txt | 11 +- drivers/spmi/spmi-pmic-arb.c | 295 ++++++++++++++++++--- 2 files changed, 263 insertions(+), 43 deletions(-)Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +3
-3@@ -1,11 +1,11 @@ Qualcomm SPMI Controller (PMIC Arbiter) -The SPMI PMIC Arbiter is found on the Snapdragon 800 Series. It is an SPMI +The SPMI PMIC Arbiter is found on Snapdragon chipsets. It is an SPMI controller with wrapping arbitration logic to allow for multiple on-chip devices to control a single SPMI master. -The PMIC Arbiter can also act as an interrupt controller, providing interrupts -to slave devices. +The PMIC Arbiter is also an interrupt controller, interrupting the Snapdragon +on dtection of a sequence initiated by a request-capable-slave to the master.<snip>
Quoted 28 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +5
-8-/* Non-data command */ -static int pmic_arb_cmd(struct spmi_controller *ctrl, u8 opc, u8 sid) +static int +pmic_arb_non_data_cmd_v1(struct spmi_controller *ctrl, u8 opc, u8 sid) { struct spmi_pmic_arb_dev *pmic_arb = spmi_controller_get_drvdata(ctrl); unsigned long flags; u32 cmd; int rc; - - /* Check for valid non-data command */ - if (opc < SPMI_CMD_RESET || opc > SPMI_CMD_WAKEUP) - return -EINVAL; + u32 offset = pmic_arb->ver->offset(pmic_arb, sid, 0); cmd = ((opc | 0x40) << 27) | ((sid & 0xf) << 20); raw_spin_lock_irqsave(&pmic_arb->lock, flags); - pmic_arb_base_write(pmic_arb, PMIC_ARB_CMD(pmic_arb->channel), cmd); - rc = pmic_arb_wait_for_done(ctrl); + pmic_arb_base_write(pmic_arb, offset + PMIC_ARB_CMD, cmd); + rc = pmic_arb_wait_for_done(ctrl, pmic_arb->wr_base, sid, 0); raw_spin_unlock_irqrestore(&pmic_arb->lock, flags); return rc; }
Quoted 6 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt
Quoted 6 lines
Documentation/devicetree/bindings/spmi/qcom,spmi-pmic-arb.txt +6
-0+/* Unsupported by HW */ +static int +pmic_arb_non_data_cmd_v2(struct spmi_controller *ctrl, u8 opc, u8 sid) +{ + return -EOPNOTSUPP; +}
Hi Stanimir,
Quoted 2 lines
Does pmic arbiter v2 supports
SPMI_CMD_WAKEUPandSPMI_CMD_SHUTDOWNcommands? If so how we send those commands to the arbiter when the
pmic-arbiter v2 does not support non-data commands including the two that you have mentioned above.
Quoted 3 lines
.non_data_cmd operation returns EOPNOTSUPP. If we returning EOPNOTSUPP the spmi bus .probe method will not call spmi driver .probe. See
spmi.cspmi_drv_probe().
Very keen observation!
The slaves that I'm working on do not need nor support the wakeup command.
I'll add a patch to add a new device tree boolean property to the framework, maybe "skip-wakeup", for similar slaves. Then change spmi_drv_probe() to skip the wakeup if the property is there.
Quoted 1 line
<snip>
Thanks,
Gilad
Signature · 2 lines
Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project
Hi Gilad,
<snip>
Quoted 42 lines
Quoted 24 lines
Quoted 24 lines
-/* Non-data command */
drivers/spmi/spmi-pmic-arb.c +2
-1-static int pmic_arb_cmd (struct spmi_controller *ctrl,u8 opc,u8 sid) +static int +pmic_arb_non_data_cmd_v1 (struct spmi_controller *ctrl,u8 opc,u8 sid) { struct spmi_pmic_arb_dev *pmic_arb =spmi_controller_get_drvdata(ctrl); unsigned long flags; u32 cmd; int rc;drivers/spmi/spmi-pmic-arb.c +1
-5- -/* Check for valid non-data command */ -if (opc <SPMI_CMD_RESET || opc >SPMI_CMD_WAKEUP ) -return -EINVAL ; +u32 offset = pmic_arb->ver->offset (pmic_arb, sid,0 ); cmd = ((opc |0x40 ) <<27 ) | ((sid &0xf ) <<20 );raw_spin_lock_irqsave (&pmic_arb->lock, flags); - pmic_arb_base_write(pmic_arb, PMIC_ARB_CMD(pmic_arb->channel),cmd);
drivers/spmi/spmi-pmic-arb.c +2
-1- rc =pmic_arb_wait_for_done (ctrl); +pmic_arb_base_write (pmic_arb, offset +PMIC_ARB_CMD , cmd); + rc =pmic_arb_wait_for_done (ctrl, pmic_arb->wr_base, sid,0 );raw_spin_unlock_irqrestore (&pmic_arb->lock, flags);Quoted 8 lines
drivers/spmi/spmi-pmic-arb.c static int pmic_arb_wait_for_done(struct spmi_controller *ctrl) · line 292
Quoted 8 lines
return rc;}
+/* Unsupported by HW */drivers/spmi/spmi-pmic-arb.c +5
-0+static int +pmic_arb_non_data_cmd_v2 (struct spmi_controller *ctrl,u8 opc,u8 sid) +{ +return -EOPNOTSUPP ; +}Hi Stanimir,
Quoted 2 lines
Does pmic arbiter v2 supports
SPMI_CMD_WAKEUPandSPMI_CMD_SHUTDOWNcommands? If so how we send those commands to the arbiter when thepmic-arbiter v2 does not support non-data commands including the two that you have mentioned above.
Quoted 3 lines
.non_data_cmd operation returns EOPNOTSUPP. If we returning EOPNOTSUPP the spmi bus .probe method will not call spmi driver .probe. See
spmi.cspmi_drv_probe().
Quoted 6 lines
Very keen observation!
The slaves that I'm working on do not need nor support the wakeup command. I'll add a patch to add a new device tree boolean property to the framework, maybe "skip-wakeup", for similar slaves.
Then changespmi_drv_probe()to skip the wakeup if the property is there.
No, instead of adding one more dt property just return zero from .non_data_cmd operation and keep the comment that this is not supported on v2. The other way is to fill the .non_data_cmd = NULL on v2 and add a check for .non_data_cmd != NULL in the callers.
Signature · 2 lines
regards, Stan