| 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
| ||
|
Date: Mon, 5 Feb 2024 17:08:24 +0200
From: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>
To: Odelu Kukatla <quic_okukatla@...cinc.com>
Cc: djakov@...nel.org, rafael@...nel.org, corbet@....net,
linux-pm@...r.kernel.org, linux-doc@...r.kernel.org,
linux-kernel@...r.kernel.org, quic_viveka@...cinc.com, peterz@...radead.org,
quic_mdtipton@...cinc.com, linux-arm-msm@...r.kernel.org,
konrad.dybcio@...aro.org, krzysztof.kozlowski@...aro.org
Subject: Re: [PATCH v2 1/4] interconnect: qcom: icc-rpmh: Add QoS
configuration support
On Mon, 5 Feb 2024 at 16:57, Odelu Kukatla <quic_okukatla@...cinc.com> wrote:
>
> It adds QoS support for QNOC device and includes support for
> configuring priority, priority forward disable, urgency forwarding.
> This helps in priortizing the traffic originating from different
> interconnect masters at NoC(Network On Chip).
>
> Signed-off-by: Odelu Kukatla <quic_okukatla@...cinc.com>
> ---
> drivers/interconnect/qcom/icc-rpmh.c | 161 ++++++++++++++++++++++++++-
> drivers/interconnect/qcom/icc-rpmh.h | 32 ++++++
> 2 files changed, 189 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/interconnect/qcom/icc-rpmh.c b/drivers/interconnect/qcom/icc-rpmh.c
> index c1aa265c1f4e..67e600f1191d 100644
> --- a/drivers/interconnect/qcom/icc-rpmh.c
> +++ b/drivers/interconnect/qcom/icc-rpmh.c
> @@ -1,28 +1,69 @@
> // SPDX-License-Identifier: GPL-2.0
> /*
> * Copyright (c) 2020, The Linux Foundation. All rights reserved.
> + * Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved.
> */
>
> +#include <linux/clk.h>
> #include <linux/interconnect.h>
> #include <linux/interconnect-provider.h>
> #include <linux/module.h>
> #include <linux/of.h>
> #include <linux/of_platform.h>
> #include <linux/slab.h>
> +#include <linux/bits.h>
>
> #include "bcm-voter.h"
> #include "icc-common.h"
> #include "icc-rpmh.h"
>
> +/* QNOC QoS */
> +#define QOSGEN_MAINCTL_LO(p, qp) (0x8 + (p->port_offsets[qp]))
> +#define QOS_SLV_URG_MSG_EN_MASK BIT_MASK(3)
> +#define QOS_DFLT_PRIO_MASK GENMASK(6, 4)
> +#define QOS_DISABLE_MASK BIT_MASK(24)
> +
> +/**
> + * qcom_icc_set_qos - initialize static QoS configurations
> + * @qp: qcom icc provider to which @node belongs
> + * @node: qcom icc node to operate on
> + */
> +static void qcom_icc_set_qos(struct qcom_icc_provider *qp,
> + struct qcom_icc_node *node)
> +{
> + const struct qcom_icc_qosbox *qos = node->qosbox;
> + int port;
> +
> + if (!qp->regmap)
> + return;
> +
> + if (!qos)
> + return;
> +
> + for (port = 0; port < qos->num_ports; port++) {
> + regmap_update_bits(qp->regmap, QOSGEN_MAINCTL_LO(qos, port),
> + QOS_DISABLE_MASK,
> + FIELD_PREP(QOS_DISABLE_MASK, qos->prio_fwd_disable));
> +
> + regmap_update_bits(qp->regmap, QOSGEN_MAINCTL_LO(qos, port),
> + QOS_DFLT_PRIO_MASK,
> + FIELD_PREP(QOS_DFLT_PRIO_MASK, qos->prio));
> +
> + regmap_update_bits(qp->regmap, QOSGEN_MAINCTL_LO(qos, port),
> + QOS_SLV_URG_MSG_EN_MASK,
> + FIELD_PREP(QOS_SLV_URG_MSG_EN_MASK, qos->urg_fwd));
> + }
> +}
> +
> /**
> * qcom_icc_pre_aggregate - cleans up stale values from prior icc_set
> * @node: icc node to operate on
> */
> void qcom_icc_pre_aggregate(struct icc_node *node)
> {
> - size_t i;
> - struct qcom_icc_node *qn;
> struct qcom_icc_provider *qp;
> + struct qcom_icc_node *qn;
> + size_t i;
>
> qn = node->data;
> qp = to_qcom_provider(node->provider);
> @@ -49,8 +90,8 @@ EXPORT_SYMBOL_GPL(qcom_icc_pre_aggregate);
> int qcom_icc_aggregate(struct icc_node *node, u32 tag, u32 avg_bw,
> u32 peak_bw, u32 *agg_avg, u32 *agg_peak)
> {
> - size_t i;
> struct qcom_icc_node *qn;
> + size_t i;
Please abstain from making the cleanup changes in the same patch as
the functional changes. These changes either should be dropped or
moved to a separate patch.
>
> qn = node->data;
>
> @@ -159,13 +200,96 @@ int qcom_icc_bcm_init(struct qcom_icc_bcm *bcm, struct device *dev)
> }
> EXPORT_SYMBOL_GPL(qcom_icc_bcm_init);
>
> +static bool bcm_needs_qos_proxy(struct qcom_icc_bcm *bcm)
Although this part is not exported to the modules, I think it deserves
some documentation. Either in the commit message or in the source
file.
> +{
> + int i;
> +
> + for (i = 0; i < bcm->num_nodes; i++)
> + if (bcm->nodes[i]->qosbox)
> + return true;
> +
> + return false;
> +}
> +
> +static int enable_qos_deps(struct qcom_icc_provider *qp)
> +{
> + struct qcom_icc_bcm *bcm;
> + bool keepalive;
> + int ret, i;
> +
> + for (i = 0; i < qp->num_bcms; i++) {
> + bcm = qp->bcms[i];
> + if (bcm_needs_qos_proxy(bcm)) {
> + keepalive = bcm->keepalive;
> + bcm->keepalive = true;
> +
> + qcom_icc_bcm_voter_add(qp->voter, bcm);
> + ret = qcom_icc_bcm_voter_commit(qp->voter);
> +
> + bcm->keepalive = keepalive;
> +
> + if (ret) {
> + dev_err(qp->dev, "failed to vote BW to %s for QoS\n",
> + bcm->name);
> + return ret;
> + }
> + }
> + }
> +
> + ret = clk_bulk_prepare_enable(qp->num_clks, qp->clks);
> + if (ret)
> + dev_err(qp->dev, "failed to enable clocks for QoS\n");
> +
> + return ret;
> +}
> +
> +static void disable_qos_deps(struct qcom_icc_provider *qp)
> +{
> + struct qcom_icc_bcm *bcm;
> + int i;
> +
> + clk_bulk_disable_unprepare(qp->num_clks, qp->clks);
> +
> + for (i = 0; i < qp->num_bcms; i++) {
> + bcm = qp->bcms[i];
> + if (bcm_needs_qos_proxy(bcm)) {
> + qcom_icc_bcm_voter_add(qp->voter, bcm);
> + qcom_icc_bcm_voter_commit(qp->voter);
> + }
> + }
> +}
> +
> +static int qcom_icc_rpmh_configure_qos(struct qcom_icc_provider *qp)
> +{
> + struct qcom_icc_node *qnode;
> + size_t i;
> + int ret;
> +
> + ret = enable_qos_deps(qp);
> + if (ret)
> + return ret;
> +
> + for (i = 0; i < qp->num_nodes; i++) {
> + qnode = qp->nodes[i];
> + if (!qnode)
> + continue;
> +
> + if (qnode->qosbox)
> + qcom_icc_set_qos(qp, qnode);
> + }
> +
> + disable_qos_deps(qp);
> +
> + return ret;
> +}
> +
> int qcom_icc_rpmh_probe(struct platform_device *pdev)
> {
> + struct qcom_icc_node * const *qnodes, *qn;
> const struct qcom_icc_desc *desc;
> struct device *dev = &pdev->dev;
> struct icc_onecell_data *data;
> struct icc_provider *provider;
> - struct qcom_icc_node * const *qnodes, *qn;
> struct qcom_icc_provider *qp;
> struct icc_node *node;
> size_t num_nodes, i, j;
> @@ -199,12 +323,35 @@ int qcom_icc_rpmh_probe(struct platform_device *pdev)
>
> qp->dev = dev;
> qp->bcms = desc->bcms;
> + qp->nodes = desc->nodes;
> qp->num_bcms = desc->num_bcms;
> + qp->num_nodes = desc->num_nodes;
>
> qp->voter = of_bcm_voter_get(qp->dev, NULL);
> if (IS_ERR(qp->voter))
> return PTR_ERR(qp->voter);
>
> + if (desc->config) {
> + struct resource *res;
> + void __iomem *base;
> +
> + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> + if (!res)
> + return -ENODEV;
Having a working ICC even without QoS changes is still important. I'd
suggest to warn the user here and continue probe, skipping QoS
settings.
> +
> + base = devm_ioremap(dev, res->start, resource_size(res));
> + if (IS_ERR(base))
> + return PTR_ERR(base);
> +
> + qp->regmap = devm_regmap_init_mmio(dev, base, desc->config);
> + if (IS_ERR(qp->regmap))
> + return PTR_ERR(qp->regmap);
> + }
> +
> + qp->num_clks = devm_clk_bulk_get_all(qp->dev, &qp->clks);
> + if (qp->num_clks < 0)
> + return qp->num_clks;
> +
> for (i = 0; i < qp->num_bcms; i++)
> qcom_icc_bcm_init(qp->bcms[i], dev);
>
> @@ -229,6 +376,10 @@ int qcom_icc_rpmh_probe(struct platform_device *pdev)
> data->nodes[i] = node;
> }
>
> + ret = qcom_icc_rpmh_configure_qos(qp);
> + if (ret)
> + goto err_remove_nodes;
> +
> ret = icc_provider_register(provider);
> if (ret)
> goto err_remove_nodes;
> @@ -247,6 +398,7 @@ int qcom_icc_rpmh_probe(struct platform_device *pdev)
> err_deregister_provider:
> icc_provider_deregister(provider);
> err_remove_nodes:
> + clk_bulk_put_all(qp->num_clks, qp->clks);
No!!! You should never manually free the resources that are handled by
the devm_ functions. In rare cases it is really necessary for some
reason, there exists a counterpart devm_free_something funciton.
> icc_nodes_remove(provider);
>
> return ret;
> @@ -258,6 +410,7 @@ void qcom_icc_rpmh_remove(struct platform_device *pdev)
> struct qcom_icc_provider *qp = platform_get_drvdata(pdev);
>
> icc_provider_deregister(&qp->provider);
> + clk_bulk_put_all(qp->num_clks, qp->clks);
And here again.
> icc_nodes_remove(&qp->provider);
> }
> EXPORT_SYMBOL_GPL(qcom_icc_rpmh_remove);
> diff --git a/drivers/interconnect/qcom/icc-rpmh.h b/drivers/interconnect/qcom/icc-rpmh.h
> index 2de29460e808..65cef156f212 100644
> --- a/drivers/interconnect/qcom/icc-rpmh.h
> +++ b/drivers/interconnect/qcom/icc-rpmh.h
> @@ -1,12 +1,14 @@
> /* SPDX-License-Identifier: GPL-2.0 */
> /*
> * Copyright (c) 2020, The Linux Foundation. All rights reserved.
> + * Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved.
> */
>
> #ifndef __DRIVERS_INTERCONNECT_QCOM_ICC_RPMH_H__
> #define __DRIVERS_INTERCONNECT_QCOM_ICC_RPMH_H__
>
> #include <dt-bindings/interconnect/qcom,icc.h>
> +#include <linux/regmap.h>
>
> #define to_qcom_provider(_provider) \
> container_of(_provider, struct qcom_icc_provider, provider)
> @@ -18,6 +20,11 @@
> * @bcms: list of bcms that maps to the provider
> * @num_bcms: number of @bcms
> * @voter: bcm voter targeted by this provider
> + * @nodes: list of icc nodes that maps to the provider
> + * @num_nodes: number of @nodes
> + * @regmap: used for QoS, register access
> + * @clks : clks required for register access
> + * @num_clks: number of @clks
> */
> struct qcom_icc_provider {
> struct icc_provider provider;
> @@ -25,6 +32,11 @@ struct qcom_icc_provider {
> struct qcom_icc_bcm * const *bcms;
> size_t num_bcms;
> struct bcm_voter *voter;
> + struct qcom_icc_node * const *nodes;
> + size_t num_nodes;
> + struct regmap *regmap;
> + struct clk_bulk_data *clks;
> + int num_clks;
Is it really necessary to store this in the global data? If I
understand correctly, QoS is only programmed during the boot and is
not to be reprogrammed later. Is my understanding correct?
> };
>
> /**
> @@ -41,6 +53,23 @@ struct bcm_db {
> u8 reserved;
> };
>
> +/**
> + * struct qcom_icc_qosbox - Qualcomm specific QoS config
Why is it qosbox?
> + * @prio: priority value assigned to requests on the node
> + * @urg_fwd: if set, master priority is used for requests.
> + * @prio_fwd_disable: if set, master priority is ignored and NoC's default priority is used.
> + * @num_ports: number of @ports
> + * @port_offsets: qos register offsets
> + */
> +
> +struct qcom_icc_qosbox {
> + const u32 prio;
> + const bool urg_fwd;
> + const bool prio_fwd_disable;
> + const u32 num_ports;
> + const u32 port_offsets[] __counted_by(num_ports);
> +};
> +
> #define MAX_LINKS 128
> #define MAX_BCMS 64
> #define MAX_BCM_PER_NODE 3
> @@ -58,6 +87,7 @@ struct bcm_db {
> * @max_peak: current max aggregate value of all peak bw requests
> * @bcms: list of bcms associated with this logical node
> * @num_bcms: num of @bcms
> + * @qosbox: qos config data associated with node
> */
> struct qcom_icc_node {
> const char *name;
> @@ -70,6 +100,7 @@ struct qcom_icc_node {
> u64 max_peak[QCOM_ICC_NUM_BUCKETS];
> struct qcom_icc_bcm *bcms[MAX_BCM_PER_NODE];
> size_t num_bcms;
> + const struct qcom_icc_qosbox *qosbox;
> };
>
> /**
> @@ -114,6 +145,7 @@ struct qcom_icc_fabric {
> };
>
> struct qcom_icc_desc {
> + const struct regmap_config *config;
> struct qcom_icc_node * const *nodes;
> size_t num_nodes;
> struct qcom_icc_bcm * const *bcms;
> --
> 2.17.1
>
>
--
With best wishes
Dmitry
Powered by blists - more mailing lists