lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 02 Apr 2021 22:54:40 +0530
From:   manafm@...eaurora.org
To:     AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...ainline.org>
Cc:     linux-arm-msm@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, Andy Gross <agross@...nel.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Rob Herring <robh+dt@...nel.org>,
        Ram Chandrasekar <rkumbako@...eaurora.org>,
        Taniya Das <tdas@...eaurora.org>,
        Akhil P Oommen <akhilpo@...eaurora.org>
Subject: Re: [RFC PATCH 2/2] drivers: soc: qcom: Add SDPM clock monitor driver

On 2021-04-02 15:49, AngeloGioacchino Del Regno wrote:
> Il 02/04/21 08:59, Manaf Meethalavalappu Pallikunhi ha scritto:
>> Add SDPM clock monitor driver, which will register for clock rate
>> change notification and write the clock rate into SDPM CSR register.
>> 
>> Signed-off-by: Ram Chandrasekar <rkumbako@...eaurora.org>
>> Signed-off-by: Manaf Meethalavalappu Pallikunhi 
>> <manafm@...eaurora.org>
>> ---
>>   drivers/soc/qcom/Kconfig            |   8 ++
>>   drivers/soc/qcom/Makefile           |   1 +
>>   drivers/soc/qcom/sdpm_clk_monitor.c | 217 
>> ++++++++++++++++++++++++++++++++++++
>>   3 files changed, 226 insertions(+)
>>   create mode 100644 drivers/soc/qcom/sdpm_clk_monitor.c
>> 
>> diff --git a/drivers/soc/qcom/Kconfig b/drivers/soc/qcom/Kconfig
>> index 9464ff4..1f04d69 100644
>> --- a/drivers/soc/qcom/Kconfig
>> +++ b/drivers/soc/qcom/Kconfig
>> @@ -237,4 +237,12 @@ config QCOM_APR
>>   	  used by audio driver to configure QDSP6
>>   	  ASM, ADM and AFE modules.
>>   +config QCOM_SDPM_CLOCK_MONITOR
>> +	tristate "Qualcomm SDPM Clock Monitor"
>> +	depends on COMMON_CLK
>> +	help
>> +	    This enables the Qualcomm SDPM Clock Monitor. This driver can 
>> register
>> +	    for different clock rate change notifications and write the 
>> clock
>> +	    rate into the SDPM CSR register. This driver will receive the 
>> clock
>> +	    list and the CSR details from devicetree.
>>   endmenu
>> diff --git a/drivers/soc/qcom/Makefile b/drivers/soc/qcom/Makefile
>> index d658a10..4eef767 100644
>> --- a/drivers/soc/qcom/Makefile
>> +++ b/drivers/soc/qcom/Makefile
>> @@ -29,3 +29,4 @@ obj-$(CONFIG_QCOM_RPMPD) += rpmpd.o
>>   obj-$(CONFIG_QCOM_KRYO_L2_ACCESSORS) +=	kryo-l2-accessors.o
>>   obj-$(CONFIG_QCOM_LLCC_PERFMON) += llcc_perfmon.o
>>   obj-$(CONFIG_QCOM_MEMORY_DUMP_V2) += memory_dump_v2.o
>> +obj-$(CONFIG_QCOM_SDPM_CLOCK_MONITOR) += sdpm_clk_monitor.o
>> diff --git a/drivers/soc/qcom/sdpm_clk_monitor.c 
>> b/drivers/soc/qcom/sdpm_clk_monitor.c
>> new file mode 100644
>> index 00000000..1aee119
>> --- /dev/null
>> +++ b/drivers/soc/qcom/sdpm_clk_monitor.c
>> @@ -0,0 +1,217 @@
>> +// SPDX-License-Identifier: GPL-2.0-only
>> +/*
>> + * Copyright (c) 2021, The Linux Foundation. All rights reserved.
>> + */
>> +
>> +#define pr_fmt(fmt) "%s " fmt, KBUILD_MODNAME
>> +
>> +#include <linux/clk.h>
>> +#include <linux/err.h>
>> +#include <linux/module.h>
>> +#include <linux/of.h>
>> +#include <linux/of_address.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/slab.h>
>> +
>> +#define SDPM_DRIVER		"sdpm-clk-monitor"
>> +#define CSR_MAX_VAL		7
>> +#define CSR_OFFSET		0xF00
>> +#define FREQ_HZ_TO_MHZ(f)	DIV_ROUND_UP((f), 1000000)
>> +#define SDPM_CSR_OFFSET(id)	(CSR_OFFSET + (id * 4))
>> +
>> +struct sdpm_clk_instance;
>> +struct sdpm_clk_data {
>> +	struct list_head		sdpm_node;
>> +	struct notifier_block		clk_rate_nb;
>> +	struct clk			*clk;
>> +	const char			*clock_name;
>> +	uint32_t			csr_id;
>> +	struct sdpm_clk_instance	*sdpm_inst;
>> +};
>> +
>> +struct sdpm_clk_instance {
>> +	struct device			*dev;
>> +	void __iomem			*regmap;
>> +	struct list_head		sdpm_instances;
>> +};
>> +
>> +static void sdpm_csr_write(struct sdpm_clk_data *sdpm_data,
>> +				unsigned long clk_rate)
>> +{
>> +	struct sdpm_clk_instance *sdpm_inst = sdpm_data->sdpm_inst;
>> +
>> +	writel_relaxed(clk_rate, sdpm_inst->regmap +
>> +			SDPM_CSR_OFFSET(sdpm_data->csr_id));
>> +	dev_dbg(sdpm_inst->dev, "clock:%s offset:0x%x frequency:%u\n",
>> +			sdpm_data->clock_name,
>> +			SDPM_CSR_OFFSET(sdpm_data->csr_id),
>> +			clk_rate);
>> +}
>> +
>> +static int sdpm_clock_notifier(struct notifier_block *nb,
>> +					unsigned long event, void *data)
>> +{
>> +	struct clk_notifier_data *ndata = data;
>> +	struct sdpm_clk_data *sdpm_data = container_of(nb,
>> +				struct sdpm_clk_data, clk_rate_nb);
>> +
>> +	dev_dbg(sdpm_data->sdpm_inst->dev, "clock:%s event:%lu\n",
>> +			sdpm_data->clock_name, event);
>> +	switch (event) {
>> +	case PRE_RATE_CHANGE:
>> +		if (ndata->new_rate > ndata->old_rate)
>> +			sdpm_csr_write(sdpm_data,
>> +					FREQ_HZ_TO_MHZ(ndata->new_rate));
>> +		return NOTIFY_DONE;
>> +	case POST_RATE_CHANGE:
>> +		if (ndata->new_rate < ndata->old_rate)
>> +			sdpm_csr_write(sdpm_data,
>> +					FREQ_HZ_TO_MHZ(ndata->new_rate));
>> +		return NOTIFY_DONE;
>> +	case ABORT_RATE_CHANGE:
>> +		if (ndata->new_rate > ndata->old_rate)
>> +			sdpm_csr_write(sdpm_data,
>> +					FREQ_HZ_TO_MHZ(ndata->old_rate));
>> +		return NOTIFY_DONE;
>> +	default:
>> +		return NOTIFY_DONE;
>> +	}
>> +}
>> +
>> +static int sdpm_clk_device_remove(struct platform_device *pdev)
>> +{
>> +	struct sdpm_clk_instance *sdpm_inst =
>> +		(struct sdpm_clk_instance *)dev_get_drvdata(&pdev->dev);
>> +	struct sdpm_clk_data *sdpm_data = NULL, *next = NULL;
>> +
>> +	list_for_each_entry_safe(sdpm_data, next,
>> +			&sdpm_inst->sdpm_instances, sdpm_node) {
>> +		clk_notifier_unregister(sdpm_data->clk,
>> +					&sdpm_data->clk_rate_nb);
>> +		list_del(&sdpm_data->sdpm_node);
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static int sdpm_clk_device_probe(struct platform_device *pdev)
>> +{
>> +	struct device *dev = &pdev->dev;
>> +	struct device_node *dev_node = dev->of_node;
>> +	int ret = 0, idx = 0, clk_ct = 0, csr = 0, csr_ct = 0;
>> +	struct sdpm_clk_instance *sdpm_inst = NULL;
>> +	struct sdpm_clk_data *sdpm_data = NULL;
>> +	struct resource *res;
>> +
>> +	sdpm_inst = devm_kzalloc(dev, sizeof(*sdpm_inst), GFP_KERNEL);
>> +	if (!sdpm_inst)
>> +		return -ENOMEM;
>> +	sdpm_inst->dev = dev;
>> +
>> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	if (!res) {
>> +		dev_err(dev, "Couldn't get MEM resource\n");
>> +		return -EINVAL;
>> +	}
>> +	dev_dbg(dev, "sdpm@...x size:%d\n", res->start,
>> +			resource_size(res));
>> +	dev_set_drvdata(dev, sdpm_inst);
>> +
>> +	sdpm_inst->regmap = devm_ioremap_resource(dev, res);
>> +	if (!sdpm_inst->regmap) {
>> +		dev_err(dev, "Couldn't get regmap\n");
>> +		return -EINVAL;
>> +	}
>> +
>> +	ret = of_property_count_strings(dev_node, "clock-names");
>> +	if (ret <= 0) {
>> +		dev_err(dev, "Couldn't get clock names. %d\n", ret);
>> +		return ret;
>> +	}
>> +	clk_ct = ret;
>> +	ret = of_property_count_u32_elems(dev_node, "csr-id");
>> +	if (ret <= 0) {
>> +		dev_err(dev, "Couldn't get csr ID array. %d\n", ret);
>> +		return ret;
>> +	}
>> +	csr_ct = ret;
>> +
>> +	if (csr_ct != clk_ct) {
>> +		dev_err(dev, "Invalid csr:%d and clk:%d count.\n", csr_ct,
>> +				clk_ct);
>> +		return -EINVAL;
>> +	}
>> +
>> +	INIT_LIST_HEAD(&sdpm_inst->sdpm_instances);
>> +
>> +	for (idx = 0; idx < clk_ct; idx++) {
>> +
>> +		sdpm_data = devm_kzalloc(dev, sizeof(*sdpm_data), GFP_KERNEL);
>> +		if (!sdpm_data) {
>> +			ret = -ENOMEM;
>> +			goto clk_err_exit;
>> +		}
>> +
>> +		ret = of_property_read_string_index(dev_node, "clock-names",
>> +				idx, &sdpm_data->clock_name);
>> +		if (ret < 0) {
>> +			dev_err(dev, "Couldn't get clk name index:%d. %d\n",
>> +					idx, ret);
>> +			goto clk_err_exit;
>> +		}
>> +
>> +		sdpm_data->clk = devm_clk_get(dev, sdpm_data->clock_name);
>> +		if (IS_ERR(sdpm_data->clk)) {
>> +			ret = PTR_ERR(sdpm_data->clk);
>> +			goto clk_err_exit;
>> +		}
>> +
>> +		ret = of_property_read_u32_index(dev_node, "csr-id", idx, &csr);
>> +		if (ret < 0) {
>> +			dev_err(dev, "Couldn't get CSR for index:%d. %d\n",
>> +					idx, ret);
>> +			goto clk_err_exit;
>> +		}
>> +
>> +		if (ret > CSR_MAX_VAL) {
>> +			dev_err(dev, "Invalid CSR %d\n", csr);
>> +			ret = -EINVAL;
>> +			goto clk_err_exit;
>> +		}
>> +
>> +		dev_dbg(dev, "SDPM clock:%s csr:%d initialized\n",
>> +				sdpm_data->clock_name, csr);
>> +		sdpm_data->csr_id = csr;
>> +		sdpm_data->sdpm_inst = sdpm_inst;
>> +		sdpm_data->clk_rate_nb.notifier_call = sdpm_clock_notifier;
>> +		sdpm_csr_write(sdpm_data,
>> +			FREQ_HZ_TO_MHZ(clk_get_rate(sdpm_data->clk)));
>> +		clk_notifier_register(sdpm_data->clk, &sdpm_data->clk_rate_nb);
>> +		list_add(&sdpm_data->sdpm_node, &sdpm_inst->sdpm_instances);
>> +	}
>> +
>> +	return 0;
>> +
>> +clk_err_exit:
>> +	sdpm_clk_device_remove(pdev);
>> +
>> +	return ret;
>> +}
>> +
>> +static const struct of_device_id sdpm_clk_device_match[] = {
>> +	{.compatible = "qcom,sdpm"},
>> +	{}
>> +};
>> +
>> +static struct platform_driver sdpm_clk_device_driver = {
>> +	.probe          = sdpm_clk_device_probe,
>> +	.remove         = sdpm_clk_device_remove,
>> +	.driver         = {
>> +		.name   = SDPM_DRIVER,
>> +		.of_match_table = sdpm_clk_device_match,
>> +	},
>> +};
>> +module_platform_driver(sdpm_clk_device_driver);
>> +
>> +MODULE_DESCRIPTION("Qualcomm Technologies, Inc. SDPM Clock Monitor 
>> Driver");
>> +MODULE_LICENSE("GPL v2");
>> 
> Hello!
> 
> Thanks for the patch. A few things to point out:
> 
> Before getting into a big lack of time situation, I have written a
> functional driver for the Qualcomm System Performance Dynamic
> Monitoring (SPDM), which has to yet be cleaned up before sending.
> This driver appears to be "SDPM" instead, but the finalities of this
> new block seem to be very similar - if not the same - as the SPDM.
> 
> My SPDM implementation was done as an interconnect driver, as that is
> actually controlling the bandwidth of the CCI on the SoC.
> 
> 
> Is this block doing the same and/or anything similar?
> In case it does, I suggest to look at my *still a bit dirty* (sorry!)
> driver for the SPDM-TZ.
> 
> Please, look at the last three commits from this branch:
> https://github.com/SoMainline/linux/commits/angelo/v5.12-rc4-feat-spdmtz
> 
> Yours,
> - Angelo

Hi Angelo,

I have gone through above commit and description. I don't see any 
relation with
this SDPM (Simple Digital power meter) clock monitor driver. The SPDM 
(System
Performance Dynamic Monitoring) is trying to control CCI bandwidth.
But this driver monitors different IPs operating clock those are handled
by linux clock framework and informs the same to Qualcomm RDPM hardware.

Thank You,
Manaf

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ