From 899f44531fe6cac4b024710fec647ecc127724b8 Mon Sep 17 00:00:00 2001 From: Sibi Sankar Date: Wed, 30 Oct 2024 18:25:10 +0530 Subject: pmdomain: core: Add GENPD_FLAG_DEV_NAME_FW flag Introduce GENPD_FLAG_DEV_NAME_FW flag which instructs genpd to generate an unique device name using ida. It is aimed to be used by genpd providers which derive their names directly from FW making them susceptible to debugfs node creation failures. Reported-by: Johan Hovold Closes: https://lore.kernel.org/lkml/ZoQjAWse2YxwyRJv@hovoldconsulting.com/ Fixes: 718072ceb211 ("PM: domains: create debugfs nodes when adding power domains") Suggested-by: Ulf Hansson Suggested-by: Dmitry Baryshkov Signed-off-by: Sibi Sankar Cc: stable@vger.kernel.org Message-ID: <20241030125512.2884761-5-quic_sibis@quicinc.com> Signed-off-by: Ulf Hansson --- drivers/pmdomain/core.c | 49 ++++++++++++++++++++++++++++++++--------------- include/linux/pm_domain.h | 6 ++++++ 2 files changed, 40 insertions(+), 15 deletions(-) diff --git a/drivers/pmdomain/core.c b/drivers/pmdomain/core.c index 5ede0f7eda09..29ad510e881c 100644 --- a/drivers/pmdomain/core.c +++ b/drivers/pmdomain/core.c @@ -7,6 +7,7 @@ #define pr_fmt(fmt) "PM: " fmt #include +#include #include #include #include @@ -23,6 +24,9 @@ #include #include +/* Provides a unique ID for each genpd device */ +static DEFINE_IDA(genpd_ida); + #define GENPD_RETRY_MAX_MS 250 /* Approximate */ #define GENPD_DEV_CALLBACK(genpd, type, callback, dev) \ @@ -171,6 +175,7 @@ static const struct genpd_lock_ops genpd_raw_spin_ops = { #define genpd_is_cpu_domain(genpd) (genpd->flags & GENPD_FLAG_CPU_DOMAIN) #define genpd_is_rpm_always_on(genpd) (genpd->flags & GENPD_FLAG_RPM_ALWAYS_ON) #define genpd_is_opp_table_fw(genpd) (genpd->flags & GENPD_FLAG_OPP_TABLE_FW) +#define genpd_is_dev_name_fw(genpd) (genpd->flags & GENPD_FLAG_DEV_NAME_FW) static inline bool irq_safe_dev_in_sleep_domain(struct device *dev, const struct generic_pm_domain *genpd) @@ -189,7 +194,7 @@ static inline bool irq_safe_dev_in_sleep_domain(struct device *dev, if (ret) dev_warn_once(dev, "PM domain %s will not be powered off\n", - genpd->name); + dev_name(&genpd->dev)); return ret; } @@ -274,7 +279,7 @@ static void genpd_debug_remove(struct generic_pm_domain *genpd) if (!genpd_debugfs_dir) return; - debugfs_lookup_and_remove(genpd->name, genpd_debugfs_dir); + debugfs_lookup_and_remove(dev_name(&genpd->dev), genpd_debugfs_dir); } static void genpd_update_accounting(struct generic_pm_domain *genpd) @@ -731,7 +736,7 @@ static int _genpd_power_on(struct generic_pm_domain *genpd, bool timed) genpd->states[state_idx].power_on_latency_ns = elapsed_ns; genpd->gd->max_off_time_changed = true; pr_debug("%s: Power-%s latency exceeded, new value %lld ns\n", - genpd->name, "on", elapsed_ns); + dev_name(&genpd->dev), "on", elapsed_ns); out: raw_notifier_call_chain(&genpd->power_notifiers, GENPD_NOTIFY_ON, NULL); @@ -782,7 +787,7 @@ static int _genpd_power_off(struct generic_pm_domain *genpd, bool timed) genpd->states[state_idx].power_off_latency_ns = elapsed_ns; genpd->gd->max_off_time_changed = true; pr_debug("%s: Power-%s latency exceeded, new value %lld ns\n", - genpd->name, "off", elapsed_ns); + dev_name(&genpd->dev), "off", elapsed_ns); out: raw_notifier_call_chain(&genpd->power_notifiers, GENPD_NOTIFY_OFF, @@ -1940,7 +1945,7 @@ int dev_pm_genpd_add_notifier(struct device *dev, struct notifier_block *nb) if (ret) { dev_warn(dev, "failed to add notifier for PM domain %s\n", - genpd->name); + dev_name(&genpd->dev)); return ret; } @@ -1987,7 +1992,7 @@ int dev_pm_genpd_remove_notifier(struct device *dev) if (ret) { dev_warn(dev, "failed to remove notifier for PM domain %s\n", - genpd->name); + dev_name(&genpd->dev)); return ret; } @@ -2013,7 +2018,7 @@ static int genpd_add_subdomain(struct generic_pm_domain *genpd, */ if (!genpd_is_irq_safe(genpd) && genpd_is_irq_safe(subdomain)) { WARN(1, "Parent %s of subdomain %s must be IRQ safe\n", - genpd->name, subdomain->name); + dev_name(&genpd->dev), subdomain->name); return -EINVAL; } @@ -2088,7 +2093,7 @@ int pm_genpd_remove_subdomain(struct generic_pm_domain *genpd, if (!list_empty(&subdomain->parent_links) || subdomain->device_count) { pr_warn("%s: unable to remove subdomain %s\n", - genpd->name, subdomain->name); + dev_name(&genpd->dev), subdomain->name); ret = -EBUSY; goto out; } @@ -2225,6 +2230,7 @@ int pm_genpd_init(struct generic_pm_domain *genpd, genpd->status = is_off ? GENPD_STATE_OFF : GENPD_STATE_ON; genpd->device_count = 0; genpd->provider = NULL; + genpd->device_id = -ENXIO; genpd->has_provider = false; genpd->accounting_time = ktime_get_mono_fast_ns(); genpd->domain.ops.runtime_suspend = genpd_runtime_suspend; @@ -2265,7 +2271,18 @@ int pm_genpd_init(struct generic_pm_domain *genpd, return ret; device_initialize(&genpd->dev); - dev_set_name(&genpd->dev, "%s", genpd->name); + + if (!genpd_is_dev_name_fw(genpd)) { + dev_set_name(&genpd->dev, "%s", genpd->name); + } else { + ret = ida_alloc(&genpd_ida, GFP_KERNEL); + if (ret < 0) { + put_device(&genpd->dev); + return ret; + } + genpd->device_id = ret; + dev_set_name(&genpd->dev, "%s_%u", genpd->name, genpd->device_id); + } mutex_lock(&gpd_list_lock); list_add(&genpd->gpd_list_node, &gpd_list); @@ -2287,13 +2304,13 @@ static int genpd_remove(struct generic_pm_domain *genpd) if (genpd->has_provider) { genpd_unlock(genpd); - pr_err("Provider present, unable to remove %s\n", genpd->name); + pr_err("Provider present, unable to remove %s\n", dev_name(&genpd->dev)); return -EBUSY; } if (!list_empty(&genpd->parent_links) || genpd->device_count) { genpd_unlock(genpd); - pr_err("%s: unable to remove %s\n", __func__, genpd->name); + pr_err("%s: unable to remove %s\n", __func__, dev_name(&genpd->dev)); return -EBUSY; } @@ -2307,9 +2324,11 @@ static int genpd_remove(struct generic_pm_domain *genpd) genpd_unlock(genpd); genpd_debug_remove(genpd); cancel_work_sync(&genpd->power_off_work); + if (genpd->device_id != -ENXIO) + ida_free(&genpd_ida, genpd->device_id); genpd_free_data(genpd); - pr_debug("%s: removed %s\n", __func__, genpd->name); + pr_debug("%s: removed %s\n", __func__, dev_name(&genpd->dev)); return 0; } @@ -3272,12 +3291,12 @@ static int genpd_summary_one(struct seq_file *s, else snprintf(state, sizeof(state), "%s", status_lookup[genpd->status]); - seq_printf(s, "%-30s %-30s %u", genpd->name, state, genpd->performance_state); + seq_printf(s, "%-30s %-30s %u", dev_name(&genpd->dev), state, genpd->performance_state); /* * Modifications on the list require holding locks on both * parent and child, so we are safe. - * Also genpd->name is immutable. + * Also the device name is immutable. */ list_for_each_entry(link, &genpd->parent_links, parent_node) { if (list_is_first(&link->parent_node, &genpd->parent_links)) @@ -3502,7 +3521,7 @@ static void genpd_debug_add(struct generic_pm_domain *genpd) if (!genpd_debugfs_dir) return; - d = debugfs_create_dir(genpd->name, genpd_debugfs_dir); + d = debugfs_create_dir(dev_name(&genpd->dev), genpd_debugfs_dir); debugfs_create_file("current_state", 0444, d, genpd, &status_fops); diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h index b637ec14025f..cf4b11be3709 100644 --- a/include/linux/pm_domain.h +++ b/include/linux/pm_domain.h @@ -92,6 +92,10 @@ struct dev_pm_domain_list { * GENPD_FLAG_OPP_TABLE_FW: The genpd provider supports performance states, * but its corresponding OPP tables are not * described in DT, but are given directly by FW. + * + * GENPD_FLAG_DEV_NAME_FW: Instructs genpd to generate an unique device name + * using ida. It is used by genpd providers which + * get their genpd-names directly from FW. */ #define GENPD_FLAG_PM_CLK (1U << 0) #define GENPD_FLAG_IRQ_SAFE (1U << 1) @@ -101,6 +105,7 @@ struct dev_pm_domain_list { #define GENPD_FLAG_RPM_ALWAYS_ON (1U << 5) #define GENPD_FLAG_MIN_RESIDENCY (1U << 6) #define GENPD_FLAG_OPP_TABLE_FW (1U << 7) +#define GENPD_FLAG_DEV_NAME_FW (1U << 8) enum gpd_status { GENPD_STATE_ON = 0, /* PM domain is on */ @@ -163,6 +168,7 @@ struct generic_pm_domain { atomic_t sd_count; /* Number of subdomains with power "on" */ enum gpd_status status; /* Current state of the domain */ unsigned int device_count; /* Number of devices */ + unsigned int device_id; /* unique device id */ unsigned int suspended_count; /* System suspend device counter */ unsigned int prepared_count; /* Suspend counter of prepared devices */ unsigned int performance_state; /* Aggregated max performance state */ -- cgit v1.2.3-73-gaa49b From 0bf020344204a2c1067b7562b6a247e6c689e28b Mon Sep 17 00:00:00 2001 From: Sibi Sankar Date: Wed, 30 Oct 2024 18:25:11 +0530 Subject: pmdomain: arm: Use FLAG_DEV_NAME_FW to ensure unique names The domain attributes returned by the perf protocol can end up reporting identical names across domains, resulting in debugfs node creation failure. Use the GENPD_FLAG_DEV_NAME_FW to ensure the genpd providers end up with an unique name. Logs: [X1E reports 'NCC' for all its scmi perf domains] debugfs: Directory 'NCC' with parent 'pm_genpd' already present! debugfs: Directory 'NCC' with parent 'pm_genpd' already present! Reported-by: Johan Hovold Closes: https://lore.kernel.org/lkml/ZoQjAWse2YxwyRJv@hovoldconsulting.com/ Suggested-by: Ulf Hansson Suggested-by: Dmitry Baryshkov Signed-off-by: Sibi Sankar Cc: stable@vger.kernel.org Message-ID: <20241030125512.2884761-6-quic_sibis@quicinc.com> Signed-off-by: Ulf Hansson --- drivers/pmdomain/arm/scmi_perf_domain.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/pmdomain/arm/scmi_perf_domain.c b/drivers/pmdomain/arm/scmi_perf_domain.c index d7ef46ccd9b8..3693423459c9 100644 --- a/drivers/pmdomain/arm/scmi_perf_domain.c +++ b/drivers/pmdomain/arm/scmi_perf_domain.c @@ -125,7 +125,8 @@ static int scmi_perf_domain_probe(struct scmi_device *sdev) scmi_pd->ph = ph; scmi_pd->genpd.name = scmi_pd->info->name; scmi_pd->genpd.flags = GENPD_FLAG_ALWAYS_ON | - GENPD_FLAG_OPP_TABLE_FW; + GENPD_FLAG_OPP_TABLE_FW | + GENPD_FLAG_DEV_NAME_FW; scmi_pd->genpd.set_performance_state = scmi_pd_set_perf_state; scmi_pd->genpd.attach_dev = scmi_pd_attach_dev; scmi_pd->genpd.detach_dev = scmi_pd_detach_dev; -- cgit v1.2.3-73-gaa49b From f7c7c5aa556378a2c8da72c1f7f238b6648f95fb Mon Sep 17 00:00:00 2001 From: Peng Fan Date: Fri, 1 Nov 2024 18:12:51 +0800 Subject: pmdomain: imx93-blk-ctrl: correct remove path The check condition should be 'i < bc->onecell_data.num_domains', not 'bc->onecell_data.num_domains' which will make the look never finish and cause kernel panic. Also disable runtime to address "imx93-blk-ctrl 4ac10000.system-controller: Unbalanced pm_runtime_enable!" Fixes: e9aa77d413c9 ("soc: imx: add i.MX93 media blk ctrl driver") Signed-off-by: Peng Fan Reviewed-by: Stefan Wahren Cc: stable@vger.kernel.org Message-ID: <20241101101252.1448466-1-peng.fan@oss.nxp.com> Signed-off-by: Ulf Hansson --- drivers/pmdomain/imx/imx93-blk-ctrl.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/pmdomain/imx/imx93-blk-ctrl.c b/drivers/pmdomain/imx/imx93-blk-ctrl.c index 904ffa55b8f4..b10348ac10f0 100644 --- a/drivers/pmdomain/imx/imx93-blk-ctrl.c +++ b/drivers/pmdomain/imx/imx93-blk-ctrl.c @@ -313,7 +313,9 @@ static void imx93_blk_ctrl_remove(struct platform_device *pdev) of_genpd_del_provider(pdev->dev.of_node); - for (i = 0; bc->onecell_data.num_domains; i++) { + pm_runtime_disable(&pdev->dev); + + for (i = 0; i < bc->onecell_data.num_domains; i++) { struct imx93_blk_ctrl_domain *domain = &bc->domains[i]; pm_genpd_remove(&domain->genpd); -- cgit v1.2.3-73-gaa49b From 5d8a766226587d111620df520dd9239c009cb154 Mon Sep 17 00:00:00 2001 From: Cristian Marussi Date: Wed, 30 Oct 2024 18:25:08 +0530 Subject: firmware: arm_scmi: Skip opp duplicates Buggy firmware can reply with duplicated PERF opps descriptors. Ensure that the bad duplicates reported by the platform firmware doesn't get added to the opp-tables. Reported-by: Johan Hovold Closes: https://lore.kernel.org/lkml/ZoQjAWse2YxwyRJv@hovoldconsulting.com/ Signed-off-by: Cristian Marussi Tested-by: Johan Hovold Reviewed-by: Sudeep Holla Cc: stable@vger.kernel.org Message-ID: <20241030125512.2884761-3-quic_sibis@quicinc.com> Signed-off-by: Ulf Hansson --- drivers/firmware/arm_scmi/perf.c | 40 ++++++++++++++++++++++++++++++---------- 1 file changed, 30 insertions(+), 10 deletions(-) diff --git a/drivers/firmware/arm_scmi/perf.c b/drivers/firmware/arm_scmi/perf.c index 2d77b5f40ca7..32f9a9acd3e9 100644 --- a/drivers/firmware/arm_scmi/perf.c +++ b/drivers/firmware/arm_scmi/perf.c @@ -373,7 +373,7 @@ static int iter_perf_levels_update_state(struct scmi_iterator_state *st, return 0; } -static inline void +static inline int process_response_opp(struct device *dev, struct perf_dom_info *dom, struct scmi_opp *opp, unsigned int loop_idx, const struct scmi_msg_resp_perf_describe_levels *r) @@ -386,12 +386,16 @@ process_response_opp(struct device *dev, struct perf_dom_info *dom, le16_to_cpu(r->opp[loop_idx].transition_latency_us); ret = xa_insert(&dom->opps_by_lvl, opp->perf, opp, GFP_KERNEL); - if (ret) + if (ret) { dev_warn(dev, "Failed to add opps_by_lvl at %d for %s - ret:%d\n", opp->perf, dom->info.name, ret); + return ret; + } + + return 0; } -static inline void +static inline int process_response_opp_v4(struct device *dev, struct perf_dom_info *dom, struct scmi_opp *opp, unsigned int loop_idx, const struct scmi_msg_resp_perf_describe_levels_v4 *r) @@ -404,9 +408,11 @@ process_response_opp_v4(struct device *dev, struct perf_dom_info *dom, le16_to_cpu(r->opp[loop_idx].transition_latency_us); ret = xa_insert(&dom->opps_by_lvl, opp->perf, opp, GFP_KERNEL); - if (ret) + if (ret) { dev_warn(dev, "Failed to add opps_by_lvl at %d for %s - ret:%d\n", opp->perf, dom->info.name, ret); + return ret; + } /* Note that PERF v4 reports always five 32-bit words */ opp->indicative_freq = le32_to_cpu(r->opp[loop_idx].indicative_freq); @@ -415,13 +421,21 @@ process_response_opp_v4(struct device *dev, struct perf_dom_info *dom, ret = xa_insert(&dom->opps_by_idx, opp->level_index, opp, GFP_KERNEL); - if (ret) + if (ret) { dev_warn(dev, "Failed to add opps_by_idx at %d for %s - ret:%d\n", opp->level_index, dom->info.name, ret); + /* Cleanup by_lvl too */ + xa_erase(&dom->opps_by_lvl, opp->perf); + + return ret; + } + hash_add(dom->opps_by_freq, &opp->hash, opp->indicative_freq); } + + return 0; } static int @@ -429,16 +443,22 @@ iter_perf_levels_process_response(const struct scmi_protocol_handle *ph, const void *response, struct scmi_iterator_state *st, void *priv) { + int ret; struct scmi_opp *opp; struct scmi_perf_ipriv *p = priv; - opp = &p->perf_dom->opp[st->desc_index + st->loop_idx]; + opp = &p->perf_dom->opp[p->perf_dom->opp_count]; if (PROTOCOL_REV_MAJOR(p->version) <= 0x3) - process_response_opp(ph->dev, p->perf_dom, opp, st->loop_idx, - response); + ret = process_response_opp(ph->dev, p->perf_dom, opp, + st->loop_idx, response); else - process_response_opp_v4(ph->dev, p->perf_dom, opp, st->loop_idx, - response); + ret = process_response_opp_v4(ph->dev, p->perf_dom, opp, + st->loop_idx, response); + + /* Skip BAD duplicates received from firmware */ + if (ret) + return ret == -EBUSY ? 0 : ret; + p->perf_dom->opp_count++; dev_dbg(ph->dev, "Level %d Power %d Latency %dus Ifreq %d Index %d\n", -- cgit v1.2.3-73-gaa49b From e2261bb81e0db86c3c866734cf93232a58464ecd Mon Sep 17 00:00:00 2001 From: Sibi Sankar Date: Wed, 30 Oct 2024 18:25:09 +0530 Subject: firmware: arm_scmi: Report duplicate opps as firmware bugs Duplicate opps reported by buggy SCP firmware currently show up as warnings even though the only functional impact is that the level/index remain inaccessible. Make it less scary for the end user by using dev_info instead, along with FW_BUG tag. Suggested-by: Johan Hovold Signed-off-by: Sibi Sankar Reviewed-by: Cristian Marussi Reviewed-by: Florian Fainelli Reviewed-by: Sudeep Holla Cc: stable@vger.kernel.org Message-ID: <20241030125512.2884761-4-quic_sibis@quicinc.com> Signed-off-by: Ulf Hansson --- drivers/firmware/arm_scmi/perf.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/firmware/arm_scmi/perf.c b/drivers/firmware/arm_scmi/perf.c index 32f9a9acd3e9..c7e5a34b254b 100644 --- a/drivers/firmware/arm_scmi/perf.c +++ b/drivers/firmware/arm_scmi/perf.c @@ -387,7 +387,7 @@ process_response_opp(struct device *dev, struct perf_dom_info *dom, ret = xa_insert(&dom->opps_by_lvl, opp->perf, opp, GFP_KERNEL); if (ret) { - dev_warn(dev, "Failed to add opps_by_lvl at %d for %s - ret:%d\n", + dev_info(dev, FW_BUG "Failed to add opps_by_lvl at %d for %s - ret:%d\n", opp->perf, dom->info.name, ret); return ret; } @@ -409,7 +409,7 @@ process_response_opp_v4(struct device *dev, struct perf_dom_info *dom, ret = xa_insert(&dom->opps_by_lvl, opp->perf, opp, GFP_KERNEL); if (ret) { - dev_warn(dev, "Failed to add opps_by_lvl at %d for %s - ret:%d\n", + dev_info(dev, FW_BUG "Failed to add opps_by_lvl at %d for %s - ret:%d\n", opp->perf, dom->info.name, ret); return ret; } -- cgit v1.2.3-73-gaa49b From d2fab3fc27cbca7ba65c539a2c5fc7f941231983 Mon Sep 17 00:00:00 2001 From: Sibi Sankar Date: Wed, 30 Oct 2024 18:25:12 +0530 Subject: mailbox: qcom-cpucp: Mark the irq with IRQF_NO_SUSPEND flag The qcom-cpucp mailbox irq is expected to function during suspend-resume cycle particularly when the scmi cpufreq driver can query the current frequency using the get_level message after the cpus are brought up during resume. Hence mark the irq with IRQF_NO_SUSPEND flag to fix the do_xfer failures we see during resume. Err Logs: arm-scmi firmware:scmi: timed out in resp(caller:do_xfer+0x164/0x568) cpufreq: cpufreq_online: ->get() failed Reported-by: Johan Hovold Closes: https://lore.kernel.org/lkml/ZtgFj1y5ggipgEOS@hovoldconsulting.com/ Fixes: 0e2a9a03106c ("mailbox: Add support for QTI CPUCP mailbox controller") Signed-off-by: Sibi Sankar Reviewed-by: Konrad Dybcio Tested-by: Johan Hovold Cc: stable@vger.kernel.org Message-ID: <20241030125512.2884761-7-quic_sibis@quicinc.com> Signed-off-by: Ulf Hansson --- drivers/mailbox/qcom-cpucp-mbox.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/mailbox/qcom-cpucp-mbox.c b/drivers/mailbox/qcom-cpucp-mbox.c index e5437c294803..44f4ed15f818 100644 --- a/drivers/mailbox/qcom-cpucp-mbox.c +++ b/drivers/mailbox/qcom-cpucp-mbox.c @@ -138,7 +138,7 @@ static int qcom_cpucp_mbox_probe(struct platform_device *pdev) return irq; ret = devm_request_irq(dev, irq, qcom_cpucp_mbox_irq_fn, - IRQF_TRIGGER_HIGH, "apss_cpucp_mbox", cpucp); + IRQF_TRIGGER_HIGH | IRQF_NO_SUSPEND, "apss_cpucp_mbox", cpucp); if (ret < 0) return dev_err_probe(dev, ret, "Failed to register irq: %d\n", irq); -- cgit v1.2.3-73-gaa49b