| Message ID | 20231201062053.1268492-1-anshuman.khandual@arm.com |
|---|---|
| Headers | show |
| Series | coresight: Move remaining AMBA ACPI devices into platform driver | expand |
On Fri, Dec 01, 2023 at 11:50:47AM +0530, Anshuman Khandual wrote: > Add support for the dynamic replicator device in the platform driver, which > can then be used on ACPI based platforms. This change would now allow > runtime power management for repliacator devices on ACPI based systems. > > The driver would try to enable the APB clock if available. Also, rename the > code to reflect the fact that it now handles both static and dynamic > replicators. > > Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> > Cc: Sudeep Holla <sudeep.holla@arm.com> Except the minor nit below which may apply also for few other patches in the series Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes Tested-by: Sudeep Holla <sudeep.holla@arm.com> # Boot and driver probe only [...] > diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c > index b6be73034996..64de0bee02ec 100644 > --- a/drivers/hwtracing/coresight/coresight-replicator.c > +++ b/drivers/hwtracing/coresight/coresight-replicator.c > @@ -38,6 +38,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); > struct replicator_drvdata { > void __iomem *base; > struct clk *atclk; > + struct clk *pclk; [minor nit] Perhaps can be documented as well ?
On Fri, Dec 01, 2023 at 11:50:49AM +0530, Anshuman Khandual wrote: > Add support for the catu devices in a new platform driver, which can then > be used on ACPI based platforms. This change would now allow runtime power > management for ACPI based systems. The driver would try to enable the APB > clock if available. > > Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> > Cc: Sudeep Holla <sudeep.holla@arm.com> Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes
On Fri, Dec 01, 2023 at 11:50:51AM +0530, Anshuman Khandual wrote: > Add support for the tmc devices in the platform driver, which can then be > used on ACPI based platforms. This change would now allow runtime power > management for ACPI based systems. The driver would try to enable the APB > clock if available. > > Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> > Cc: Sudeep Holla <sudeep.holla@arm.com> Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes Tested-by: Sudeep Holla <sudeep.holla@arm.com> # Boot and driver probe only
On Fri, Dec 01, 2023 at 11:50:53AM +0530, Anshuman Khandual wrote: > Add support for the cpu debug devices in a new platform driver, which can > then be used on ACPI based platforms. This change would now allow runtime > power management for ACPI based systems. The driver would try to enable > the APB clock if available. > > Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> > Cc: Sudeep Holla <sudeep.holla@arm.com> Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes
Hi Anshuman, On 01/12/2023 06:20, Anshuman Khandual wrote: > Add support for the catu devices in a new platform driver, which can then > be used on ACPI based platforms. This change would now allow runtime power > management for ACPI based systems. The driver would try to enable the APB > clock if available. This doesn't talk about the new helper. As such I would prefer that to be a separate preparatory patch. See below. > > Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> > Cc: Sudeep Holla <sudeep.holla@arm.com> > Cc: Suzuki K Poulose <suzuki.poulose@arm.com> > Cc: Mike Leach <mike.leach@linaro.org> > Cc: James Clark <james.clark@arm.com> > Cc: linux-acpi@vger.kernel.org > Cc: linux-arm-kernel@lists.infradead.org > Cc: linux-kernel@vger.kernel.org > Cc: coresight@lists.linaro.org > Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com> > --- > drivers/acpi/arm64/amba.c | 1 - > drivers/hwtracing/coresight/coresight-catu.c | 130 ++++++++++++++++--- > drivers/hwtracing/coresight/coresight-catu.h | 1 + > drivers/hwtracing/coresight/coresight-core.c | 29 +++++ > include/linux/coresight.h | 7 + > 5 files changed, 149 insertions(+), 19 deletions(-) > > diff --git a/drivers/acpi/arm64/amba.c b/drivers/acpi/arm64/amba.c > index afb6afb66967..587061b0fd2f 100644 > --- a/drivers/acpi/arm64/amba.c > +++ b/drivers/acpi/arm64/amba.c > @@ -27,7 +27,6 @@ static const struct acpi_device_id amba_id_list[] = { > {"ARMHC503", 0}, /* ARM CoreSight Debug */ > {"ARMHC979", 0}, /* ARM CoreSight TPIU */ > {"ARMHC97C", 0}, /* ARM CoreSight SoC-400 TMC, SoC-600 ETF/ETB */ > - {"ARMHC9CA", 0}, /* ARM CoreSight CATU */ > {"", 0}, > }; > > diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c > index 3949ded0d4fa..ba5ee7d158dd 100644 > --- a/drivers/hwtracing/coresight/coresight-catu.c > +++ b/drivers/hwtracing/coresight/coresight-catu.c > @@ -7,6 +7,8 @@ > * Author: Suzuki K Poulose <suzuki.poulose@arm.com> > */ > > +#include <linux/platform_device.h> > +#include <linux/acpi.h> > #include <linux/amba/bus.h> > #include <linux/device.h> > #include <linux/dma-mapping.h> > @@ -502,28 +504,20 @@ static const struct coresight_ops catu_ops = { > .helper_ops = &catu_helper_ops, > }; > > -static int catu_probe(struct amba_device *adev, const struct amba_id *id) > +static int __catu_probe(struct device *dev, struct resource *res) > { > int ret = 0; > u32 dma_mask; > - struct catu_drvdata *drvdata; > + struct catu_drvdata *drvdata = dev_get_drvdata(dev); > struct coresight_desc catu_desc; > struct coresight_platform_data *pdata = NULL; > - struct device *dev = &adev->dev; > void __iomem *base; > > catu_desc.name = coresight_alloc_device_name(&catu_devs, dev); > if (!catu_desc.name) > return -ENOMEM; > > - drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL); > - if (!drvdata) { > - ret = -ENOMEM; > - goto out; > - } > - > - dev_set_drvdata(dev, drvdata); > - base = devm_ioremap_resource(dev, &adev->res); > + base = devm_ioremap_resource(dev, res); > if (IS_ERR(base)) { > ret = PTR_ERR(base); > goto out; > @@ -568,18 +562,35 @@ static int catu_probe(struct amba_device *adev, const struct amba_id *id) > if (IS_ERR(drvdata->csdev)) > ret = PTR_ERR(drvdata->csdev); > else > - pm_runtime_put(&adev->dev); > + pm_runtime_put(dev); > out: > return ret; > } > > -static void catu_remove(struct amba_device *adev) > +static int catu_probe(struct amba_device *adev, const struct amba_id *id) > +{ > + struct catu_drvdata *drvdata; > + > + drvdata = devm_kzalloc(&adev->dev, sizeof(*drvdata), GFP_KERNEL); > + if (!drvdata) > + return -ENOMEM; > + > + amba_set_drvdata(adev, drvdata); > + return __catu_probe(&adev->dev, &adev->res); > +} > + > +static void __catu_remove(struct device *dev) > { > - struct catu_drvdata *drvdata = dev_get_drvdata(&adev->dev); > + struct catu_drvdata *drvdata = dev_get_drvdata(dev); > > coresight_unregister(drvdata->csdev); > } > > +static void catu_remove(struct amba_device *adev) > +{ > + __catu_remove(&adev->dev); > +} > + > static struct amba_id catu_ids[] = { > CS_AMBA_ID(0x000bb9ee), > {}, > @@ -598,13 +609,96 @@ static struct amba_driver catu_driver = { > .id_table = catu_ids, > }; > > +static int catu_platform_probe(struct platform_device *pdev) > +{ > + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + struct catu_drvdata *drvdata; > + int ret = 0; > + > + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); > + if (!drvdata) > + return -ENOMEM; > + > + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); > + if (IS_ERR(drvdata->pclk)) > + return -ENODEV; > + > + if (res) { > + drvdata->base = devm_ioremap_resource(&pdev->dev, res); > + if (IS_ERR(drvdata->base)) { > + clk_put(drvdata->pclk); > + return PTR_ERR(drvdata->base); > + } > + } > + > + pm_runtime_get_noresume(&pdev->dev); > + pm_runtime_set_active(&pdev->dev); > + pm_runtime_enable(&pdev->dev); > + > + dev_set_drvdata(&pdev->dev, drvdata); > + ret = __catu_probe(&pdev->dev, res); > + if (ret) { > + pm_runtime_put_noidle(&pdev->dev); > + pm_runtime_disable(&pdev->dev); > + } > + return ret; > +} > + > +static int catu_platform_remove(struct platform_device *pdev) > +{ > + __catu_remove(&pdev->dev); > + return 0; > +} > + > +#ifdef CONFIG_PM > +static int catu_runtime_suspend(struct device *dev) > +{ > + struct catu_drvdata *drvdata = dev_get_drvdata(dev); > + > + if (drvdata->pclk && !IS_ERR_OR_NULL(drvdata->pclk)) Only the second part is needed. IS_ERR_OR_NULL() already checks for NULL. > + clk_disable_unprepare(drvdata->pclk); > + return 0; > +} > + > +static int catu_runtime_resume(struct device *dev) > +{ > + struct catu_drvdata *drvdata = dev_get_drvdata(dev); > + > + if (drvdata->pclk && !IS_ERR_OR_NULL(drvdata->pclk)) Same here. > + clk_prepare_enable(drvdata->pclk); > + return 0; > +} > +#endif > + > +static const struct dev_pm_ops catu_dev_pm_ops = { > + SET_RUNTIME_PM_OPS(catu_runtime_suspend, catu_runtime_resume, NULL) > +}; > + > +#ifdef CONFIG_ACPI > +static const struct acpi_device_id catu_acpi_ids[] = { > + {"ARMHC9CA", 0}, /* ARM CoreSight CATU */ > + {}, > +}; > + > +MODULE_DEVICE_TABLE(acpi, catu_acpi_ids); > +#endif > + > +static struct platform_driver catu_platform_driver = { > + .probe = catu_platform_probe, > + .remove = catu_platform_remove, > + .driver = { > + .name = "coresight-catu-platform", > + .acpi_match_table = ACPI_PTR(catu_acpi_ids), > + .suppress_bind_attrs = true, > + .pm = &catu_dev_pm_ops, > + }, > +}; > + > static int __init catu_init(void) > { > int ret; > > - ret = amba_driver_register(&catu_driver); > - if (ret) > - pr_info("Error registering catu driver\n"); > + ret = coresight_init_driver("catu", &catu_driver, &catu_platform_driver); > tmc_etr_set_catu_ops(&etr_catu_buf_ops); > return ret; > } > @@ -612,7 +706,7 @@ static int __init catu_init(void) > static void __exit catu_exit(void) > { > tmc_etr_remove_catu_ops(); > - amba_driver_unregister(&catu_driver); > + coresight_remove_driver(&catu_driver, &catu_platform_driver); > } > > module_init(catu_init); > diff --git a/drivers/hwtracing/coresight/coresight-catu.h b/drivers/hwtracing/coresight/coresight-catu.h > index 442e034bbfba..141feac1c14b 100644 > --- a/drivers/hwtracing/coresight/coresight-catu.h > +++ b/drivers/hwtracing/coresight/coresight-catu.h > @@ -61,6 +61,7 @@ > #define CATU_IRQEN_OFF 0x0 > > struct catu_drvdata { > + struct clk *pclk; > void __iomem *base; > struct coresight_device *csdev; > int irq; > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c > index 9fabe00a40d6..ede9b0723f95 100644 > --- a/drivers/hwtracing/coresight/coresight-core.c > +++ b/drivers/hwtracing/coresight/coresight-core.c > @@ -1833,6 +1833,35 @@ static void __exit coresight_exit(void) > module_init(coresight_init); > module_exit(coresight_exit); > ---8>--- > +int coresight_init_driver(const char *drv, struct amba_driver *amba_drv, > + struct platform_driver *pdev_drv) > +{ > + int ret; > + > + ret = amba_driver_register(amba_drv); > + if (ret) { > + pr_err("%s: error registering AMBA driver\n", drv); > + return ret; > + } > + > + ret = platform_driver_register(pdev_drv); > + if (!ret) > + return 0; > + > + pr_err("%s: error registering platform driver\n", drv); > + amba_driver_unregister(amba_drv); > + return ret; > +} > +EXPORT_SYMBOL_GPL(coresight_init_driver); > + > +void coresight_remove_driver(struct amba_driver *amba_drv, > + struct platform_driver *pdev_drv) > +{ > + amba_driver_unregister(amba_drv); > + platform_driver_unregister(pdev_drv); > +} > +EXPORT_SYMBOL_GPL(coresight_remove_driver); Please could we split this into a separate patch itself ? Also, can we not use them for the other components ? funnel, replicator ? > + > MODULE_LICENSE("GPL v2"); > MODULE_AUTHOR("Pratik Patel <pratikp@codeaurora.org>"); > MODULE_AUTHOR("Mathieu Poirier <mathieu.poirier@linaro.org>"); > diff --git a/include/linux/coresight.h b/include/linux/coresight.h > index a269fffaf991..be7fe3793763 100644 > --- a/include/linux/coresight.h > +++ b/include/linux/coresight.h > @@ -12,6 +12,8 @@ > #include <linux/io.h> > #include <linux/perf_event.h> > #include <linux/sched.h> > +#include <linux/amba/bus.h> > +#include <linux/platform_device.h> > > /* Peripheral id registers (0xFD0-0xFEC) */ > #define CORESIGHT_PERIPHIDR4 0xfd0 > @@ -597,6 +599,11 @@ void coresight_relaxed_write64(struct coresight_device *csdev, > u64 val, u32 offset); > void coresight_write64(struct coresight_device *csdev, u64 val, u32 offset); > > +int coresight_init_driver(const char *drv, struct amba_driver *amba_drv, > + struct platform_driver *pdev_drv); > + > +void coresight_remove_driver(struct amba_driver *amba_drv, > + struct platform_driver *pdev_drv); > #else > static inline struct coresight_device * > coresight_register(struct coresight_desc *desc) { return NULL; } Suzuki
On 12/1/23 18:05, Sudeep Holla wrote: > On Fri, Dec 01, 2023 at 11:50:47AM +0530, Anshuman Khandual wrote: >> Add support for the dynamic replicator device in the platform driver, which >> can then be used on ACPI based platforms. This change would now allow >> runtime power management for repliacator devices on ACPI based systems. >> >> The driver would try to enable the APB clock if available. Also, rename the >> code to reflect the fact that it now handles both static and dynamic >> replicators. >> >> Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> >> Cc: Sudeep Holla <sudeep.holla@arm.com> > > Except the minor nit below which may apply also for few other patches > in the series > > Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes > Tested-by: Sudeep Holla <sudeep.holla@arm.com> # Boot and driver probe only > > [...] > >> diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c >> index b6be73034996..64de0bee02ec 100644 >> --- a/drivers/hwtracing/coresight/coresight-replicator.c >> +++ b/drivers/hwtracing/coresight/coresight-replicator.c >> @@ -38,6 +38,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); >> struct replicator_drvdata { >> void __iomem *base; >> struct clk *atclk; >> + struct clk *pclk; > > [minor nit] Perhaps can be documented as well ? Sure, will add the following comment above the structure. @pclk: optional clock for the core parts of the replicator.
On 12/4/23 09:56, Anshuman Khandual wrote: > > > On 12/1/23 18:05, Sudeep Holla wrote: >> On Fri, Dec 01, 2023 at 11:50:47AM +0530, Anshuman Khandual wrote: >>> Add support for the dynamic replicator device in the platform driver, which >>> can then be used on ACPI based platforms. This change would now allow >>> runtime power management for repliacator devices on ACPI based systems. >>> >>> The driver would try to enable the APB clock if available. Also, rename the >>> code to reflect the fact that it now handles both static and dynamic >>> replicators. >>> >>> Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> >>> Cc: Sudeep Holla <sudeep.holla@arm.com> >> >> Except the minor nit below which may apply also for few other patches >> in the series >> >> Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes >> Tested-by: Sudeep Holla <sudeep.holla@arm.com> # Boot and driver probe only >> >> [...] >> >>> diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c >>> index b6be73034996..64de0bee02ec 100644 >>> --- a/drivers/hwtracing/coresight/coresight-replicator.c >>> +++ b/drivers/hwtracing/coresight/coresight-replicator.c >>> @@ -38,6 +38,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); >>> struct replicator_drvdata { >>> void __iomem *base; >>> struct clk *atclk; >>> + struct clk *pclk; >> >> [minor nit] Perhaps can be documented as well ? > > Sure, will add the following comment above the structure. > > @pclk: optional clock for the core parts of the replicator. My bad, this will be the following instead. diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c index 64de0bee02ec..44b9a77ec8aa 100644 --- a/drivers/hwtracing/coresight/coresight-replicator.c +++ b/drivers/hwtracing/coresight/coresight-replicator.c @@ -31,6 +31,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); * @base: memory mapped base address for this component. Also indicates * whether this one is programmable or not. * @atclk: optional clock for the core parts of the replicator. + * @pclk: APB clock if present, otherwise NULL * @csdev: component vitals needed by the framework * @spinlock: serialize enable/disable operations. * @check_idfilter_val: check if the context is lost upon clock removal. I will update replicator, tpiu, tmc, and stm devices. struct catu_drvdata and debug_drvdata do not have the comment section itself.
On 12/1/23 19:11, Suzuki K Poulose wrote: > Hi Anshuman, > > On 01/12/2023 06:20, Anshuman Khandual wrote: >> Add support for the catu devices in a new platform driver, which can then >> be used on ACPI based platforms. This change would now allow runtime power >> management for ACPI based systems. The driver would try to enable the APB >> clock if available. > > This doesn't talk about the new helper. As such I would prefer that to be a separate preparatory patch. See below. Makes sense. > >> >> Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> >> Cc: Sudeep Holla <sudeep.holla@arm.com> >> Cc: Suzuki K Poulose <suzuki.poulose@arm.com> >> Cc: Mike Leach <mike.leach@linaro.org> >> Cc: James Clark <james.clark@arm.com> >> Cc: linux-acpi@vger.kernel.org >> Cc: linux-arm-kernel@lists.infradead.org >> Cc: linux-kernel@vger.kernel.org >> Cc: coresight@lists.linaro.org >> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com> >> --- >> drivers/acpi/arm64/amba.c | 1 - >> drivers/hwtracing/coresight/coresight-catu.c | 130 ++++++++++++++++--- >> drivers/hwtracing/coresight/coresight-catu.h | 1 + >> drivers/hwtracing/coresight/coresight-core.c | 29 +++++ >> include/linux/coresight.h | 7 + >> 5 files changed, 149 insertions(+), 19 deletions(-) >> >> diff --git a/drivers/acpi/arm64/amba.c b/drivers/acpi/arm64/amba.c >> index afb6afb66967..587061b0fd2f 100644 >> --- a/drivers/acpi/arm64/amba.c >> +++ b/drivers/acpi/arm64/amba.c >> @@ -27,7 +27,6 @@ static const struct acpi_device_id amba_id_list[] = { >> {"ARMHC503", 0}, /* ARM CoreSight Debug */ >> {"ARMHC979", 0}, /* ARM CoreSight TPIU */ >> {"ARMHC97C", 0}, /* ARM CoreSight SoC-400 TMC, SoC-600 ETF/ETB */ >> - {"ARMHC9CA", 0}, /* ARM CoreSight CATU */ >> {"", 0}, >> }; >> diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c >> index 3949ded0d4fa..ba5ee7d158dd 100644 >> --- a/drivers/hwtracing/coresight/coresight-catu.c >> +++ b/drivers/hwtracing/coresight/coresight-catu.c >> @@ -7,6 +7,8 @@ >> * Author: Suzuki K Poulose <suzuki.poulose@arm.com> >> */ >> +#include <linux/platform_device.h> >> +#include <linux/acpi.h> >> #include <linux/amba/bus.h> >> #include <linux/device.h> >> #include <linux/dma-mapping.h> >> @@ -502,28 +504,20 @@ static const struct coresight_ops catu_ops = { >> .helper_ops = &catu_helper_ops, >> }; >> -static int catu_probe(struct amba_device *adev, const struct amba_id *id) >> +static int __catu_probe(struct device *dev, struct resource *res) >> { >> int ret = 0; >> u32 dma_mask; >> - struct catu_drvdata *drvdata; >> + struct catu_drvdata *drvdata = dev_get_drvdata(dev); >> struct coresight_desc catu_desc; >> struct coresight_platform_data *pdata = NULL; >> - struct device *dev = &adev->dev; >> void __iomem *base; >> catu_desc.name = coresight_alloc_device_name(&catu_devs, dev); >> if (!catu_desc.name) >> return -ENOMEM; >> - drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL); >> - if (!drvdata) { >> - ret = -ENOMEM; >> - goto out; >> - } >> - >> - dev_set_drvdata(dev, drvdata); >> - base = devm_ioremap_resource(dev, &adev->res); >> + base = devm_ioremap_resource(dev, res); >> if (IS_ERR(base)) { >> ret = PTR_ERR(base); >> goto out; >> @@ -568,18 +562,35 @@ static int catu_probe(struct amba_device *adev, const struct amba_id *id) >> if (IS_ERR(drvdata->csdev)) >> ret = PTR_ERR(drvdata->csdev); >> else >> - pm_runtime_put(&adev->dev); >> + pm_runtime_put(dev); >> out: >> return ret; >> } >> -static void catu_remove(struct amba_device *adev) >> +static int catu_probe(struct amba_device *adev, const struct amba_id *id) >> +{ >> + struct catu_drvdata *drvdata; >> + >> + drvdata = devm_kzalloc(&adev->dev, sizeof(*drvdata), GFP_KERNEL); >> + if (!drvdata) >> + return -ENOMEM; >> + >> + amba_set_drvdata(adev, drvdata); >> + return __catu_probe(&adev->dev, &adev->res); >> +} >> + >> +static void __catu_remove(struct device *dev) >> { >> - struct catu_drvdata *drvdata = dev_get_drvdata(&adev->dev); >> + struct catu_drvdata *drvdata = dev_get_drvdata(dev); >> coresight_unregister(drvdata->csdev); >> } >> +static void catu_remove(struct amba_device *adev) >> +{ >> + __catu_remove(&adev->dev); >> +} >> + >> static struct amba_id catu_ids[] = { >> CS_AMBA_ID(0x000bb9ee), >> {}, >> @@ -598,13 +609,96 @@ static struct amba_driver catu_driver = { >> .id_table = catu_ids, >> }; >> +static int catu_platform_probe(struct platform_device *pdev) >> +{ >> + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); >> + struct catu_drvdata *drvdata; >> + int ret = 0; >> + >> + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); >> + if (!drvdata) >> + return -ENOMEM; >> + >> + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); >> + if (IS_ERR(drvdata->pclk)) >> + return -ENODEV; >> + >> + if (res) { >> + drvdata->base = devm_ioremap_resource(&pdev->dev, res); >> + if (IS_ERR(drvdata->base)) { >> + clk_put(drvdata->pclk); >> + return PTR_ERR(drvdata->base); >> + } >> + } >> + >> + pm_runtime_get_noresume(&pdev->dev); >> + pm_runtime_set_active(&pdev->dev); >> + pm_runtime_enable(&pdev->dev); >> + >> + dev_set_drvdata(&pdev->dev, drvdata); >> + ret = __catu_probe(&pdev->dev, res); >> + if (ret) { >> + pm_runtime_put_noidle(&pdev->dev); >> + pm_runtime_disable(&pdev->dev); >> + } >> + return ret; >> +} >> + >> +static int catu_platform_remove(struct platform_device *pdev) >> +{ >> + __catu_remove(&pdev->dev); >> + return 0; >> +} >> + >> +#ifdef CONFIG_PM >> +static int catu_runtime_suspend(struct device *dev) >> +{ >> + struct catu_drvdata *drvdata = dev_get_drvdata(dev); >> + >> + if (drvdata->pclk && !IS_ERR_OR_NULL(drvdata->pclk)) > > Only the second part is needed. IS_ERR_OR_NULL() already checks for NULL. Actually, the first check here should be for valid drvdata instead, ensuring that subsequent de-referencing for drvdata->pclk does not crash. Will do this replacement for all affected patches. > >> + clk_disable_unprepare(drvdata->pclk); >> + return 0; >> +} >> + >> +static int catu_runtime_resume(struct device *dev) >> +{ >> + struct catu_drvdata *drvdata = dev_get_drvdata(dev); >> + >> + if (drvdata->pclk && !IS_ERR_OR_NULL(drvdata->pclk)) > > Same here. > >> + clk_prepare_enable(drvdata->pclk); >> + return 0; >> +} >> +#endif >> + >> +static const struct dev_pm_ops catu_dev_pm_ops = { >> + SET_RUNTIME_PM_OPS(catu_runtime_suspend, catu_runtime_resume, NULL) >> +}; >> + >> +#ifdef CONFIG_ACPI >> +static const struct acpi_device_id catu_acpi_ids[] = { >> + {"ARMHC9CA", 0}, /* ARM CoreSight CATU */ >> + {}, >> +}; >> + >> +MODULE_DEVICE_TABLE(acpi, catu_acpi_ids); >> +#endif >> + >> +static struct platform_driver catu_platform_driver = { >> + .probe = catu_platform_probe, >> + .remove = catu_platform_remove, >> + .driver = { >> + .name = "coresight-catu-platform", >> + .acpi_match_table = ACPI_PTR(catu_acpi_ids), >> + .suppress_bind_attrs = true, >> + .pm = &catu_dev_pm_ops, >> + }, >> +}; >> + >> static int __init catu_init(void) >> { >> int ret; >> - ret = amba_driver_register(&catu_driver); >> - if (ret) >> - pr_info("Error registering catu driver\n"); >> + ret = coresight_init_driver("catu", &catu_driver, &catu_platform_driver); >> tmc_etr_set_catu_ops(&etr_catu_buf_ops); >> return ret; >> } >> @@ -612,7 +706,7 @@ static int __init catu_init(void) >> static void __exit catu_exit(void) >> { >> tmc_etr_remove_catu_ops(); >> - amba_driver_unregister(&catu_driver); >> + coresight_remove_driver(&catu_driver, &catu_platform_driver); >> } >> module_init(catu_init); >> diff --git a/drivers/hwtracing/coresight/coresight-catu.h b/drivers/hwtracing/coresight/coresight-catu.h >> index 442e034bbfba..141feac1c14b 100644 >> --- a/drivers/hwtracing/coresight/coresight-catu.h >> +++ b/drivers/hwtracing/coresight/coresight-catu.h >> @@ -61,6 +61,7 @@ >> #define CATU_IRQEN_OFF 0x0 >> struct catu_drvdata { >> + struct clk *pclk; >> void __iomem *base; >> struct coresight_device *csdev; >> int irq; >> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c >> index 9fabe00a40d6..ede9b0723f95 100644 >> --- a/drivers/hwtracing/coresight/coresight-core.c >> +++ b/drivers/hwtracing/coresight/coresight-core.c >> @@ -1833,6 +1833,35 @@ static void __exit coresight_exit(void) >> module_init(coresight_init); >> module_exit(coresight_exit); >> > > ---8>--- > >> +int coresight_init_driver(const char *drv, struct amba_driver *amba_drv, >> + struct platform_driver *pdev_drv) >> +{ >> + int ret; >> + >> + ret = amba_driver_register(amba_drv); >> + if (ret) { >> + pr_err("%s: error registering AMBA driver\n", drv); >> + return ret; >> + } >> + >> + ret = platform_driver_register(pdev_drv); >> + if (!ret) >> + return 0; >> + >> + pr_err("%s: error registering platform driver\n", drv); >> + amba_driver_unregister(amba_drv); >> + return ret; >> +} >> +EXPORT_SYMBOL_GPL(coresight_init_driver); >> + >> +void coresight_remove_driver(struct amba_driver *amba_drv, >> + struct platform_driver *pdev_drv) >> +{ >> + amba_driver_unregister(amba_drv); >> + platform_driver_unregister(pdev_drv); >> +} >> +EXPORT_SYMBOL_GPL(coresight_remove_driver); > > Please could we split this into a separate patch itself ? Also, can we not use them for the other components ? funnel, replicator ? Sure, will split the helpers addition along with their header changes into a separate patch at the beginning of this series, and then use them for funnel and replicator devices in subsequent patches. > > >> + >> MODULE_LICENSE("GPL v2"); >> MODULE_AUTHOR("Pratik Patel <pratikp@codeaurora.org>"); >> MODULE_AUTHOR("Mathieu Poirier <mathieu.poirier@linaro.org>"); >> diff --git a/include/linux/coresight.h b/include/linux/coresight.h >> index a269fffaf991..be7fe3793763 100644 >> --- a/include/linux/coresight.h >> +++ b/include/linux/coresight.h >> @@ -12,6 +12,8 @@ >> #include <linux/io.h> >> #include <linux/perf_event.h> >> #include <linux/sched.h> >> +#include <linux/amba/bus.h> >> +#include <linux/platform_device.h> >> /* Peripheral id registers (0xFD0-0xFEC) */ >> #define CORESIGHT_PERIPHIDR4 0xfd0 >> @@ -597,6 +599,11 @@ void coresight_relaxed_write64(struct coresight_device *csdev, >> u64 val, u32 offset); >> void coresight_write64(struct coresight_device *csdev, u64 val, u32 offset); >> +int coresight_init_driver(const char *drv, struct amba_driver *amba_drv, >> + struct platform_driver *pdev_drv); >> + >> +void coresight_remove_driver(struct amba_driver *amba_drv, >> + struct platform_driver *pdev_drv); >> #else >> static inline struct coresight_device * >> coresight_register(struct coresight_desc *desc) { return NULL; } > > > Suzuki
On 04/12/2023 04:48, Anshuman Khandual wrote: > > > On 12/4/23 09:56, Anshuman Khandual wrote: >> >> >> On 12/1/23 18:05, Sudeep Holla wrote: >>> On Fri, Dec 01, 2023 at 11:50:47AM +0530, Anshuman Khandual wrote: >>>> Add support for the dynamic replicator device in the platform driver, which >>>> can then be used on ACPI based platforms. This change would now allow >>>> runtime power management for repliacator devices on ACPI based systems. >>>> >>>> The driver would try to enable the APB clock if available. Also, rename the >>>> code to reflect the fact that it now handles both static and dynamic >>>> replicators. >>>> >>>> Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> >>>> Cc: Sudeep Holla <sudeep.holla@arm.com> >>> >>> Except the minor nit below which may apply also for few other patches >>> in the series >>> >>> Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes >>> Tested-by: Sudeep Holla <sudeep.holla@arm.com> # Boot and driver probe only >>> >>> [...] >>> >>>> diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c >>>> index b6be73034996..64de0bee02ec 100644 >>>> --- a/drivers/hwtracing/coresight/coresight-replicator.c >>>> +++ b/drivers/hwtracing/coresight/coresight-replicator.c >>>> @@ -38,6 +38,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); >>>> struct replicator_drvdata { >>>> void __iomem *base; >>>> struct clk *atclk; >>>> + struct clk *pclk; >>> >>> [minor nit] Perhaps can be documented as well ? >> >> Sure, will add the following comment above the structure. >> >> @pclk: optional clock for the core parts of the replicator. > > My bad, this will be the following instead. > > diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c > index 64de0bee02ec..44b9a77ec8aa 100644 > --- a/drivers/hwtracing/coresight/coresight-replicator.c > +++ b/drivers/hwtracing/coresight/coresight-replicator.c > @@ -31,6 +31,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); > * @base: memory mapped base address for this component. Also indicates > * whether this one is programmable or not. > * @atclk: optional clock for the core parts of the replicator. > + * @pclk: APB clock if present, otherwise NULL > * @csdev: component vitals needed by the framework > * @spinlock: serialize enable/disable operations. > * @check_idfilter_val: check if the context is lost upon clock removal. > > I will update replicator, tpiu, tmc, and stm devices. > funnel is missing it as well. If you build with W=1 it shows up. > struct catu_drvdata and debug_drvdata do not have the comment section itself. > _______________________________________________ > CoreSight mailing list -- coresight@lists.linaro.org > To unsubscribe send an email to coresight-leave@lists.linaro.org
On 01/12/2023 06:20, Anshuman Khandual wrote: > Add support for the tmc devices in the platform driver, which can then be > used on ACPI based platforms. This change would now allow runtime power > management for ACPI based systems. The driver would try to enable the APB > clock if available. > > Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> > Cc: Sudeep Holla <sudeep.holla@arm.com> > Cc: Suzuki K Poulose <suzuki.poulose@arm.com> > Cc: Mike Leach <mike.leach@linaro.org> > Cc: James Clark <james.clark@arm.com> > Cc: linux-acpi@vger.kernel.org > Cc: linux-arm-kernel@lists.infradead.org > Cc: linux-kernel@vger.kernel.org > Cc: coresight@lists.linaro.org > Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com> > --- > drivers/acpi/arm64/amba.c | 2 - > .../hwtracing/coresight/coresight-tmc-core.c | 130 +++++++++++++++--- > drivers/hwtracing/coresight/coresight-tmc.h | 1 + > 3 files changed, 115 insertions(+), 18 deletions(-) > [...] > -static int tmc_probe(struct amba_device *adev, const struct amba_id *id) > +static int __tmc_probe(struct device *dev, struct resource *res, void *dev_caps) > { > int ret = 0; > u32 devid; > void __iomem *base; > - struct device *dev = &adev->dev; > struct coresight_platform_data *pdata = NULL; > - struct tmc_drvdata *drvdata; > - struct resource *res = &adev->res; > + struct tmc_drvdata *drvdata = dev_get_drvdata(dev); > struct coresight_desc desc = { 0 }; > struct coresight_dev_list *dev_list = NULL; > > ret = -ENOMEM; > - drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL); > - if (!drvdata) > - goto out; > - > - dev_set_drvdata(dev, drvdata); > > /* Validity for the resource is already checked by the AMBA core */ > base = devm_ioremap_resource(dev, res); > @@ -487,8 +482,7 @@ static int tmc_probe(struct amba_device *adev, const struct amba_id *id) > desc.type = CORESIGHT_DEV_TYPE_SINK; > desc.subtype.sink_subtype = CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM; > desc.ops = &tmc_etr_cs_ops; > - ret = tmc_etr_setup_caps(dev, devid, > - coresight_get_uci_data(id)); > + ret = tmc_etr_setup_caps(dev, devid, dev_caps); > if (ret) > goto out; [...] > -module_amba_driver(tmc_driver); > +static int tmc_platform_probe(struct platform_device *pdev) > +{ > + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + struct tmc_drvdata *drvdata; > + int ret = 0; > + > + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); > + if (!drvdata) > + return -ENOMEM; > + > + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); > + if (IS_ERR(drvdata->pclk)) > + return -ENODEV; > + > + dev_set_drvdata(&pdev->dev, drvdata); > + pm_runtime_get_noresume(&pdev->dev); > + pm_runtime_set_active(&pdev->dev); > + pm_runtime_enable(&pdev->dev); > + > + ret = __tmc_probe(&pdev->dev, res, NULL); What effect does passing NULL here have? It seems like the dev_caps did something useful when it's set to CORESIGHT_SOC_600_ETR_CAPS. Maybe Suzuki knows as he added it?
On 04/12/2023 10:42, James Clark wrote: > > > On 01/12/2023 06:20, Anshuman Khandual wrote: >> Add support for the tmc devices in the platform driver, which can then be >> used on ACPI based platforms. This change would now allow runtime power >> management for ACPI based systems. The driver would try to enable the APB >> clock if available. >> >> Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> >> Cc: Sudeep Holla <sudeep.holla@arm.com> >> Cc: Suzuki K Poulose <suzuki.poulose@arm.com> >> Cc: Mike Leach <mike.leach@linaro.org> >> Cc: James Clark <james.clark@arm.com> >> Cc: linux-acpi@vger.kernel.org >> Cc: linux-arm-kernel@lists.infradead.org >> Cc: linux-kernel@vger.kernel.org >> Cc: coresight@lists.linaro.org >> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com> >> --- >> drivers/acpi/arm64/amba.c | 2 - >> .../hwtracing/coresight/coresight-tmc-core.c | 130 +++++++++++++++--- >> drivers/hwtracing/coresight/coresight-tmc.h | 1 + >> 3 files changed, 115 insertions(+), 18 deletions(-) >> > [...] >> -static int tmc_probe(struct amba_device *adev, const struct amba_id *id) >> +static int __tmc_probe(struct device *dev, struct resource *res, void *dev_caps) >> { >> int ret = 0; >> u32 devid; >> void __iomem *base; >> - struct device *dev = &adev->dev; >> struct coresight_platform_data *pdata = NULL; >> - struct tmc_drvdata *drvdata; >> - struct resource *res = &adev->res; >> + struct tmc_drvdata *drvdata = dev_get_drvdata(dev); >> struct coresight_desc desc = { 0 }; >> struct coresight_dev_list *dev_list = NULL; >> >> ret = -ENOMEM; >> - drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL); >> - if (!drvdata) >> - goto out; >> - >> - dev_set_drvdata(dev, drvdata); >> >> /* Validity for the resource is already checked by the AMBA core */ >> base = devm_ioremap_resource(dev, res); >> @@ -487,8 +482,7 @@ static int tmc_probe(struct amba_device *adev, const struct amba_id *id) >> desc.type = CORESIGHT_DEV_TYPE_SINK; >> desc.subtype.sink_subtype = CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM; >> desc.ops = &tmc_etr_cs_ops; >> - ret = tmc_etr_setup_caps(dev, devid, >> - coresight_get_uci_data(id)); >> + ret = tmc_etr_setup_caps(dev, devid, dev_caps); >> if (ret) >> goto out; > [...] >> -module_amba_driver(tmc_driver); >> +static int tmc_platform_probe(struct platform_device *pdev) >> +{ >> + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); >> + struct tmc_drvdata *drvdata; >> + int ret = 0; >> + >> + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); >> + if (!drvdata) >> + return -ENOMEM; >> + >> + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); >> + if (IS_ERR(drvdata->pclk)) >> + return -ENODEV; >> + >> + dev_set_drvdata(&pdev->dev, drvdata); >> + pm_runtime_get_noresume(&pdev->dev); >> + pm_runtime_set_active(&pdev->dev); >> + pm_runtime_enable(&pdev->dev); >> + >> + ret = __tmc_probe(&pdev->dev, res, NULL); > > What effect does passing NULL here have? It seems like the dev_caps did > something useful when it's set to CORESIGHT_SOC_600_ETR_CAPS. Maybe > Suzuki knows as he added it? That was supposed to initialise the capabilities of the TMC-ETR, which are not discoverable by ID registers. (e.g., SAVE restore of RRP/RWP). I guess, we could get rid of that cap and use the "PID" register match to figure out the capabilities. Suzuki
On 01/12/2023 06:20, Anshuman Khandual wrote: > Add support for the tmc devices in the platform driver, which can then be > used on ACPI based platforms. This change would now allow runtime power > management for ACPI based systems. The driver would try to enable the APB > clock if available. > [...] > -module_amba_driver(tmc_driver); > +static int tmc_platform_probe(struct platform_device *pdev) > +{ > + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + struct tmc_drvdata *drvdata; > + int ret = 0; > + > + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); > + if (!drvdata) > + return -ENOMEM; > + > + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); > + if (IS_ERR(drvdata->pclk)) > + return -ENODEV; > + > + dev_set_drvdata(&pdev->dev, drvdata); > + pm_runtime_get_noresume(&pdev->dev); > + pm_runtime_set_active(&pdev->dev); > + pm_runtime_enable(&pdev->dev); > + > + ret = __tmc_probe(&pdev->dev, res, NULL); > + if (ret) { > + pm_runtime_put_noidle(&pdev->dev); > + pm_runtime_disable(&pdev->dev); > + } I'm not sure if these pm_runtime()s are right because there is already a put inside of __tmc_probe() if it fails. If you unload and then reload all the coresight modules with these patches you get these errors which are new: coresight-tpiu-platform ARMHC979:00: Unbalanced pm_runtime_enable! CSCFG registered etm0 coresight etm0: CPU0: etm v4.2 initialized CSCFG registered etm1 coresight etm1: CPU1: etm v4.2 initialized CSCFG registered etm2 coresight etm2: CPU2: etm v4.2 initialized CSCFG registered etm3 coresight etm3: CPU3: etm v4.2 initialized coresight-tmc-platform ARMHC97C:00: Unbalanced pm_runtime_enable! coresight-tmc-platform ARMHC97C:01: Unbalanced pm_runtime_enable! coresight-tmc-platform ARMHC97C:02: Unbalanced pm_runtime_enable! coresight-tmc-platform ARMHC97C:03: Unbalanced pm_runtime_enable! It might be worth testing all of these pm_runtime()s, including the error case ones, because loading and unloading the modules doesn't even include the error scenarios, so there are probably more bad ones in there too.
On 12/4/23 15:27, James Clark wrote: > > On 04/12/2023 04:48, Anshuman Khandual wrote: >> >> On 12/4/23 09:56, Anshuman Khandual wrote: >>> >>> On 12/1/23 18:05, Sudeep Holla wrote: >>>> On Fri, Dec 01, 2023 at 11:50:47AM +0530, Anshuman Khandual wrote: >>>>> Add support for the dynamic replicator device in the platform driver, which >>>>> can then be used on ACPI based platforms. This change would now allow >>>>> runtime power management for repliacator devices on ACPI based systems. >>>>> >>>>> The driver would try to enable the APB clock if available. Also, rename the >>>>> code to reflect the fact that it now handles both static and dynamic >>>>> replicators. >>>>> >>>>> Cc: Lorenzo Pieralisi <lpieralisi@kernel.org> >>>>> Cc: Sudeep Holla <sudeep.holla@arm.com> >>>> Except the minor nit below which may apply also for few other patches >>>> in the series >>>> >>>> Acked-by: Sudeep Holla <sudeep.holla@arm.com> # For ACPI related changes >>>> Tested-by: Sudeep Holla <sudeep.holla@arm.com> # Boot and driver probe only >>>> >>>> [...] >>>> >>>>> diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c >>>>> index b6be73034996..64de0bee02ec 100644 >>>>> --- a/drivers/hwtracing/coresight/coresight-replicator.c >>>>> +++ b/drivers/hwtracing/coresight/coresight-replicator.c >>>>> @@ -38,6 +38,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); >>>>> struct replicator_drvdata { >>>>> void __iomem *base; >>>>> struct clk *atclk; >>>>> + struct clk *pclk; >>>> [minor nit] Perhaps can be documented as well ? >>> Sure, will add the following comment above the structure. >>> >>> @pclk: optional clock for the core parts of the replicator. >> My bad, this will be the following instead. >> >> diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c >> index 64de0bee02ec..44b9a77ec8aa 100644 >> --- a/drivers/hwtracing/coresight/coresight-replicator.c >> +++ b/drivers/hwtracing/coresight/coresight-replicator.c >> @@ -31,6 +31,7 @@ DEFINE_CORESIGHT_DEVLIST(replicator_devs, "replicator"); >> * @base: memory mapped base address for this component. Also indicates >> * whether this one is programmable or not. >> * @atclk: optional clock for the core parts of the replicator. >> + * @pclk: APB clock if present, otherwise NULL >> * @csdev: component vitals needed by the framework >> * @spinlock: serialize enable/disable operations. >> * @check_idfilter_val: check if the context is lost upon clock removal. >> >> I will update replicator, tpiu, tmc, and stm devices. >> > funnel is missing it as well. If you build with W=1 it shows up. Updated, done.
On Mon, Dec 04, 2023 at 10:54:15AM +0000, James Clark wrote: > > > On 01/12/2023 06:20, Anshuman Khandual wrote: > > Add support for the tmc devices in the platform driver, which can then be > > used on ACPI based platforms. This change would now allow runtime power > > management for ACPI based systems. The driver would try to enable the APB > > clock if available. > > > [...] > > -module_amba_driver(tmc_driver); > > +static int tmc_platform_probe(struct platform_device *pdev) > > +{ > > + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > > + struct tmc_drvdata *drvdata; > > + int ret = 0; > > + > > + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); > > + if (!drvdata) > > + return -ENOMEM; > > + > > + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); > > + if (IS_ERR(drvdata->pclk)) > > + return -ENODEV; > > + > > + dev_set_drvdata(&pdev->dev, drvdata); > > + pm_runtime_get_noresume(&pdev->dev); > > + pm_runtime_set_active(&pdev->dev); > > + pm_runtime_enable(&pdev->dev); > > + > > + ret = __tmc_probe(&pdev->dev, res, NULL); > > + if (ret) { > > + pm_runtime_put_noidle(&pdev->dev); > > + pm_runtime_disable(&pdev->dev); > > + } > > I'm not sure if these pm_runtime()s are right because there is already a > put inside of __tmc_probe() if it fails. If you unload and then reload > all the coresight modules with these patches you get these errors which > are new: > Ah good find. I have just tested with ACPI and these modules as built-in to be more specific. -- Regards, Sudeep
On 12/4/23 16:24, James Clark wrote: > > > On 01/12/2023 06:20, Anshuman Khandual wrote: >> Add support for the tmc devices in the platform driver, which can then be >> used on ACPI based platforms. This change would now allow runtime power >> management for ACPI based systems. The driver would try to enable the APB >> clock if available. >> > [...] >> -module_amba_driver(tmc_driver); >> +static int tmc_platform_probe(struct platform_device *pdev) >> +{ >> + struct resource *res = platform_get_resource(pdev, IORESOURCE_MEM, 0); >> + struct tmc_drvdata *drvdata; >> + int ret = 0; >> + >> + drvdata = devm_kzalloc(&pdev->dev, sizeof(*drvdata), GFP_KERNEL); >> + if (!drvdata) >> + return -ENOMEM; >> + >> + drvdata->pclk = coresight_get_enable_apb_pclk(&pdev->dev); >> + if (IS_ERR(drvdata->pclk)) >> + return -ENODEV; >> + >> + dev_set_drvdata(&pdev->dev, drvdata); >> + pm_runtime_get_noresume(&pdev->dev); >> + pm_runtime_set_active(&pdev->dev); >> + pm_runtime_enable(&pdev->dev); >> + >> + ret = __tmc_probe(&pdev->dev, res, NULL); >> + if (ret) { >> + pm_runtime_put_noidle(&pdev->dev); >> + pm_runtime_disable(&pdev->dev); >> + } > > I'm not sure if these pm_runtime()s are right because there is already a > put inside of __tmc_probe() if it fails. If you unload and then reload Actually there is a pm_runtime_put() on the success path, not when it fails. So pm_runtime_put() gets called when __tmc_probe() returns 0. __tmc_probe() { .... ret = misc_register(&drvdata->miscdev); if (ret) coresight_unregister(drvdata->csdev); else pm_runtime_put(dev); out: return ret; } tmc_platform_probe() { .... pm_runtime_get_noresume(&pdev->dev); pm_runtime_set_active(&pdev->dev); pm_runtime_enable(&pdev->dev); ret = __tmc_probe(&pdev->dev, res, NULL); if (ret) { pm_runtime_put_noidle(&pdev->dev); pm_runtime_disable(&pdev->dev); } return ret; } tmc_probe() { .... return __tmc_probe(&adev->dev, &adev->res, coresight_get_uci_data(id)); } Currently pm_runtime_put() gets called - In success path both for AMBA and platform drivers - In error path only for platform driver Although the problem might be with pm_runtime_disable() instead - pm_runtime_disable() is not required in the platform driver probe() path - But might be required in tmc_platform_remove() along with a clk_put() > all the coresight modules with these patches you get these errors which > are new: > > coresight-tpiu-platform ARMHC979:00: Unbalanced pm_runtime_enable! The code is similar in TPIU platform driver as well. > CSCFG registered etm0 > coresight etm0: CPU0: etm v4.2 initialized > CSCFG registered etm1 > coresight etm1: CPU1: etm v4.2 initialized > CSCFG registered etm2 > coresight etm2: CPU2: etm v4.2 initialized > CSCFG registered etm3 > coresight etm3: CPU3: etm v4.2 initialized > coresight-tmc-platform ARMHC97C:00: Unbalanced pm_runtime_enable! > coresight-tmc-platform ARMHC97C:01: Unbalanced pm_runtime_enable! > coresight-tmc-platform ARMHC97C:02: Unbalanced pm_runtime_enable! > coresight-tmc-platform ARMHC97C:03: Unbalanced pm_runtime_enable! > > It might be worth testing all of these pm_runtime()s, including the > error case ones, because loading and unloading the modules doesn't even > include the error scenarios, so there are probably more bad ones in > there too. The code is very similar in CATU, STM as well but debug_platform_remove() seems to be doing this right. I am not very familiar with all the power management aspects in coresight, please do let me know if I missing something here.