From 0d6526f82c3cdefcca47f73f5fc08dc6f335eac6 Mon Sep 17 00:00:00 2001 From: Tim Chen Date: Mon, 21 Sep 2026 17:37:22 -0700 Subject: sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling bug alb_break_llc() decides whether to break LLC preference during active load balance. It does so by testing that every runnable fair task on the source rq prefers its LLC: env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable But the two counters cover different sets. nr_pref_llc_running is updated in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued, so it follows queued tasks. h_nr_runnable is updated in set_delayed()/ clear_delayed() and drops delay-dequeued tasks. So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks, alb_break_llc() returns false, and active balance is free to pull a task off its preferred LLC. Active balance only moves runnable tasks, and this is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB skips the per-task test in can_migrate_task(). The runnable set is the one we want. Fix it on the counter side. A task should be counted in nr_pref_llc_running exactly while it is both queued on its preferred LLC (pref_llc_queued) and runnable (!sched_delayed). Define that membership once in task_pref_llc_runnable(), and adjust the counter only through pref_llc_running_inc()/pref_llc_running_dec() from the four sites that change either input: account_llc_enqueue(), account_llc_dequeue(), set_delayed() and clear_delayed(). Gating every update on the same predicate keeps the delay, wake and dequeue paths from double-counting or underflowing; see the comments at those sites for the ordering. nr_llc_running and sd->llc_counts are not touched and stay on queued semantics. Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC") Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/ Reported-by: Zhan Xusheng Suggested-by: Chen Yu Signed-off-by: Tim Chen Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Reviewed-by: Kayra Cizmeci Cc: # v7.2.x Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com --- drivers/rtc/rtc-xgene.c | 278 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 278 insertions(+) create mode 100644 drivers/rtc/rtc-xgene.c (limited to 'drivers/rtc/rtc-xgene.c') diff --git a/drivers/rtc/rtc-xgene.c b/drivers/rtc/rtc-xgene.c new file mode 100644 index 000000000..6660b664e --- /dev/null +++ b/drivers/rtc/rtc-xgene.c @@ -0,0 +1,278 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * APM X-Gene SoC Real Time Clock Driver + * + * Copyright (c) 2014, Applied Micro Circuits Corporation + * Author: Rameshwar Prasad Sahu + * Loc Ho + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* RTC CSR Registers */ +#define RTC_CCVR 0x00 +#define RTC_CMR 0x04 +#define RTC_CLR 0x08 +#define RTC_CCR 0x0C +#define RTC_CCR_IE BIT(0) +#define RTC_CCR_MASK BIT(1) +#define RTC_CCR_EN BIT(2) +#define RTC_CCR_WEN BIT(3) +#define RTC_STAT 0x10 +#define RTC_STAT_BIT BIT(0) +#define RTC_RSTAT 0x14 +#define RTC_EOI 0x18 +#define RTC_VER 0x1C + +struct xgene_rtc_dev { + struct rtc_device *rtc; + void __iomem *csr_base; + struct clk *clk; + unsigned int irq_wake; + unsigned int irq_enabled; +}; + +static int xgene_rtc_read_time(struct device *dev, struct rtc_time *tm) +{ + struct xgene_rtc_dev *pdata = dev_get_drvdata(dev); + + rtc_time64_to_tm(readl(pdata->csr_base + RTC_CCVR), tm); + return 0; +} + +static int xgene_rtc_set_time(struct device *dev, struct rtc_time *tm) +{ + struct xgene_rtc_dev *pdata = dev_get_drvdata(dev); + + /* + * NOTE: After the following write, the RTC_CCVR is only reflected + * after the update cycle of 1 seconds. + */ + writel((u32)rtc_tm_to_time64(tm), pdata->csr_base + RTC_CLR); + readl(pdata->csr_base + RTC_CLR); /* Force a barrier */ + + return 0; +} + +static int xgene_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm) +{ + struct xgene_rtc_dev *pdata = dev_get_drvdata(dev); + + /* If possible, CMR should be read here */ + rtc_time64_to_tm(0, &alrm->time); + alrm->enabled = readl(pdata->csr_base + RTC_CCR) & RTC_CCR_IE; + + return 0; +} + +static int xgene_rtc_alarm_irq_enable(struct device *dev, u32 enabled) +{ + struct xgene_rtc_dev *pdata = dev_get_drvdata(dev); + u32 ccr; + + ccr = readl(pdata->csr_base + RTC_CCR); + if (enabled) { + ccr &= ~RTC_CCR_MASK; + ccr |= RTC_CCR_IE; + } else { + ccr &= ~RTC_CCR_IE; + ccr |= RTC_CCR_MASK; + } + writel(ccr, pdata->csr_base + RTC_CCR); + + return 0; +} + +static int xgene_rtc_alarm_irq_enabled(struct device *dev) +{ + struct xgene_rtc_dev *pdata = dev_get_drvdata(dev); + + return readl(pdata->csr_base + RTC_CCR) & RTC_CCR_IE ? 1 : 0; +} + +static int xgene_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm) +{ + struct xgene_rtc_dev *pdata = dev_get_drvdata(dev); + + writel((u32)rtc_tm_to_time64(&alrm->time), pdata->csr_base + RTC_CMR); + + xgene_rtc_alarm_irq_enable(dev, alrm->enabled); + + return 0; +} + +static const struct rtc_class_ops xgene_rtc_ops = { + .read_time = xgene_rtc_read_time, + .set_time = xgene_rtc_set_time, + .read_alarm = xgene_rtc_read_alarm, + .set_alarm = xgene_rtc_set_alarm, + .alarm_irq_enable = xgene_rtc_alarm_irq_enable, +}; + +static irqreturn_t xgene_rtc_interrupt(int irq, void *id) +{ + struct xgene_rtc_dev *pdata = id; + + /* Check if interrupt asserted */ + if (!(readl(pdata->csr_base + RTC_STAT) & RTC_STAT_BIT)) + return IRQ_NONE; + + /* Clear interrupt */ + readl(pdata->csr_base + RTC_EOI); + + rtc_update_irq(pdata->rtc, 1, RTC_IRQF | RTC_AF); + + return IRQ_HANDLED; +} + +static int xgene_rtc_probe(struct platform_device *pdev) +{ + struct xgene_rtc_dev *pdata; + int ret; + int irq; + + pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL); + if (!pdata) + return -ENOMEM; + platform_set_drvdata(pdev, pdata); + + pdata->csr_base = devm_platform_ioremap_resource(pdev, 0); + if (IS_ERR(pdata->csr_base)) + return PTR_ERR(pdata->csr_base); + + pdata->rtc = devm_rtc_allocate_device(&pdev->dev); + if (IS_ERR(pdata->rtc)) + return PTR_ERR(pdata->rtc); + + irq = platform_get_irq(pdev, 0); + if (irq < 0) + return irq; + ret = devm_request_irq(&pdev->dev, irq, xgene_rtc_interrupt, 0, + dev_name(&pdev->dev), pdata); + if (ret) { + dev_err(&pdev->dev, "Could not request IRQ\n"); + return ret; + } + + pdata->clk = devm_clk_get(&pdev->dev, NULL); + if (IS_ERR(pdata->clk)) { + dev_err(&pdev->dev, "Couldn't get the clock for RTC\n"); + return -ENODEV; + } + ret = clk_prepare_enable(pdata->clk); + if (ret) + return ret; + + /* Turn on the clock and the crystal */ + writel(RTC_CCR_EN, pdata->csr_base + RTC_CCR); + + ret = device_init_wakeup(&pdev->dev, true); + if (ret) { + clk_disable_unprepare(pdata->clk); + return ret; + } + + pdata->rtc->ops = &xgene_rtc_ops; + pdata->rtc->range_max = U32_MAX; + + ret = devm_rtc_register_device(pdata->rtc); + if (ret) { + clk_disable_unprepare(pdata->clk); + return ret; + } + + return 0; +} + +static void xgene_rtc_remove(struct platform_device *pdev) +{ + struct xgene_rtc_dev *pdata = platform_get_drvdata(pdev); + + xgene_rtc_alarm_irq_enable(&pdev->dev, 0); + device_init_wakeup(&pdev->dev, false); + clk_disable_unprepare(pdata->clk); +} + +static int __maybe_unused xgene_rtc_suspend(struct device *dev) +{ + struct platform_device *pdev = to_platform_device(dev); + struct xgene_rtc_dev *pdata = platform_get_drvdata(pdev); + int irq; + + irq = platform_get_irq(pdev, 0); + + /* + * If this RTC alarm will be used for waking the system up, + * don't disable it of course. Else we just disable the alarm + * and await suspension. + */ + if (device_may_wakeup(&pdev->dev)) { + if (!enable_irq_wake(irq)) + pdata->irq_wake = 1; + } else { + pdata->irq_enabled = xgene_rtc_alarm_irq_enabled(dev); + xgene_rtc_alarm_irq_enable(dev, 0); + clk_disable_unprepare(pdata->clk); + } + return 0; +} + +static int __maybe_unused xgene_rtc_resume(struct device *dev) +{ + struct platform_device *pdev = to_platform_device(dev); + struct xgene_rtc_dev *pdata = platform_get_drvdata(pdev); + int irq; + int rc; + + irq = platform_get_irq(pdev, 0); + + if (device_may_wakeup(&pdev->dev)) { + if (pdata->irq_wake) { + disable_irq_wake(irq); + pdata->irq_wake = 0; + } + } else { + rc = clk_prepare_enable(pdata->clk); + if (rc) { + dev_err(dev, "Unable to enable clock error %d\n", rc); + return rc; + } + xgene_rtc_alarm_irq_enable(dev, pdata->irq_enabled); + } + + return 0; +} + +static SIMPLE_DEV_PM_OPS(xgene_rtc_pm_ops, xgene_rtc_suspend, xgene_rtc_resume); + +#ifdef CONFIG_OF +static const struct of_device_id xgene_rtc_of_match[] = { + {.compatible = "apm,xgene-rtc" }, + { } +}; +MODULE_DEVICE_TABLE(of, xgene_rtc_of_match); +#endif + +static struct platform_driver xgene_rtc_driver = { + .probe = xgene_rtc_probe, + .remove = xgene_rtc_remove, + .driver = { + .name = "xgene-rtc", + .pm = &xgene_rtc_pm_ops, + .of_match_table = of_match_ptr(xgene_rtc_of_match), + }, +}; + +module_platform_driver(xgene_rtc_driver); + +MODULE_DESCRIPTION("APM X-Gene SoC RTC driver"); +MODULE_AUTHOR("Rameshwar Sahu "); +MODULE_LICENSE("GPL"); -- cgit v1.3.1