System register ETMs and ETE are unlikely to be preserved on CPU power down. The ETE DT binding also never documented "arm,coresight-loses-context-with-cpu" so nobody would have legitimately been able to use that binding to fix it and ACPI has no such binding at all.
Fix it by hard coding the setting for sysreg ETMs (ETE is always sysreg) or ACPI boots. Use a local variable when setting up save_state so that it's immune to concurrent probing when devices have different configurations which is an issue with modifying the global.
This fixes the following error when using Coresight with ACPI on the FVP which supports CPU PM:
coresight ete0: External agent took claim tag WARNING: drivers/hwtracing/coresight/coresight-core.c:248 at coresight_disclaim_device_unlocked+0xe0/0xe8, CPU#0: perf/117
Fixes: 35e1c9163e02 ("coresight: ete: Add support for ETE tracing") Signed-off-by: James Clark james.clark@linaro.org --- Fix PM save on ETE, which is an issue that showed up on the FVP when booted with ACPI and the newly enabled idle states. --- Changes in v3: - Setup using a local variable and then check drvdata->save_state at runtime so it's immune to concurrent probing with different configs. - Link to v2: https://lore.kernel.org/r/20260505-james-cs-ete-pm_save_enable-v2-1-d39c1f1e...
Changes in v2: - Expand rule to also save for ACPI and sysreg ETMs. Now module param is only read for MMIO ETMs and DT. - Drop change to save the result in drvdata as the single global is simpler. - Link to v1: https://lore.kernel.org/r/20260428-james-cs-ete-pm_save_enable-v1-0-c7a90ca6... --- drivers/hwtracing/coresight/coresight-etm4x-core.c | 48 +++++++++++++++------- 1 file changed, 34 insertions(+), 14 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c index d565a73f0042..591dfe0bc635 100644 --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c @@ -56,10 +56,14 @@ MODULE_PARM_DESC(boot_enable, "Enable tracing on boot"); #define PARAM_PM_SAVE_NEVER 1 /* never save any state */ #define PARAM_PM_SAVE_SELF_HOSTED 2 /* save self-hosted state only */
+/* + * Save option for ETM4. ETE, sysreg ETM4s and ACPI boots ignore this option and + * will always save. + */ static int pm_save_enable = PARAM_PM_SAVE_FIRMWARE; module_param(pm_save_enable, int, 0444); MODULE_PARM_DESC(pm_save_enable, - "Save/restore state on power down: 1 = never, 2 = self-hosted"); + "Save/restore state on power down: 1 = never, 2 = self-hosted. MMIO and DT only.");
static struct etmv4_drvdata *etmdrvdata[NR_CPUS]; static void etm4_set_default_config(struct etmv4_config *config); @@ -2012,7 +2016,7 @@ static int etm4_cpu_save(struct etmv4_drvdata *drvdata) { int ret = 0;
- if (pm_save_enable != PARAM_PM_SAVE_SELF_HOSTED) + if (!drvdata->save_state) return 0;
/* @@ -2127,7 +2131,7 @@ static void __etm4_cpu_restore(struct etmv4_drvdata *drvdata)
static void etm4_cpu_restore(struct etmv4_drvdata *drvdata) { - if (pm_save_enable != PARAM_PM_SAVE_SELF_HOSTED) + if (!drvdata->save_state) return;
if (coresight_get_mode(drvdata->csdev)) @@ -2212,6 +2216,17 @@ static void etm4_pm_clear(void) } }
+static bool etm4x_always_pm_save(struct device *dev, struct csdev_access *csa) +{ + /* + * Only IO mem ETM devices will benefit from skipping PM save and only + * DT has the option to control it, not ACPI. Otherwise system register + * based ETMs and ETEs will always lose context on CPU power down, so + * always save. + */ + return !csa->io_mem || is_acpi_device_node(dev_fwnode(dev)); +} + static int etm4_add_coresight_dev(struct etm4_init_arg *init_arg) { int ret; @@ -2221,6 +2236,7 @@ static int etm4_add_coresight_dev(struct etm4_init_arg *init_arg) struct coresight_desc desc = { 0 }; u8 major, minor; char *type_name; + bool pm_save;
if (!drvdata) return -EINVAL; @@ -2248,6 +2264,21 @@ static int etm4_add_coresight_dev(struct etm4_init_arg *init_arg)
etm4_set_default(&drvdata->config);
+ if (etm4x_always_pm_save(dev, init_arg->csa)) + pm_save = true; + else if (pm_save_enable == PARAM_PM_SAVE_FIRMWARE) + pm_save = coresight_loses_context_with_cpu(dev); + else + pm_save = pm_save_enable != PARAM_PM_SAVE_NEVER; + + if (pm_save) { + drvdata->save_state = devm_kmalloc(dev, + sizeof(struct etmv4_save_state), + GFP_KERNEL); + if (!drvdata->save_state) + return -ENOMEM; + } + pdata = coresight_get_platform_data(dev); if (IS_ERR(pdata)) return PTR_ERR(pdata); @@ -2305,17 +2336,6 @@ static int etm4_probe(struct device *dev) if (ret) return ret;
- if (pm_save_enable == PARAM_PM_SAVE_FIRMWARE) - pm_save_enable = coresight_loses_context_with_cpu(dev) ? - PARAM_PM_SAVE_SELF_HOSTED : PARAM_PM_SAVE_NEVER; - - if (pm_save_enable != PARAM_PM_SAVE_NEVER) { - drvdata->save_state = devm_kmalloc(dev, - sizeof(struct etmv4_save_state), GFP_KERNEL); - if (!drvdata->save_state) - return -ENOMEM; - } - raw_spin_lock_init(&drvdata->spinlock);
drvdata->cpu = coresight_get_cpu(dev);
--- base-commit: 971f3474f8898ae8bbab19a9b547819a5e6fbcf1 change-id: 20260420-james-cs-ete-pm_save_enable-4e994e35cdac
Best regards,
On Tue, May 05, 2026 at 05:51:25PM +0100, James Clark wrote:
System register ETMs and ETE are unlikely to be preserved on CPU power down. The ETE DT binding also never documented "arm,coresight-loses-context-with-cpu" so nobody would have legitimately been able to use that binding to fix it and ACPI has no such binding at all.
Fix it by hard coding the setting for sysreg ETMs (ETE is always sysreg) or ACPI boots. Use a local variable when setting up save_state so that it's immune to concurrent probing when devices have different configurations which is an issue with modifying the global.
This fixes the following error when using Coresight with ACPI on the FVP which supports CPU PM:
coresight ete0: External agent took claim tag WARNING: drivers/hwtracing/coresight/coresight-core.c:248 at coresight_disclaim_device_unlocked+0xe0/0xe8, CPU#0: perf/117
Fixes: 35e1c9163e02 ("coresight: ete: Add support for ETE tracing") Signed-off-by: James Clark james.clark@linaro.org
Reviewed-by: Leo Yan leo.yan@arm.com
On Tue, 05 May 2026 17:51:25 +0100, James Clark wrote:
System register ETMs and ETE are unlikely to be preserved on CPU power down. The ETE DT binding also never documented "arm,coresight-loses-context-with-cpu" so nobody would have legitimately been able to use that binding to fix it and ACPI has no such binding at all.
Fix it by hard coding the setting for sysreg ETMs (ETE is always sysreg) or ACPI boots. Use a local variable when setting up save_state so that it's immune to concurrent probing when devices have different configurations which is an issue with modifying the global.
[...]
Applied, thanks!
[1/1] coresight: ete: Always save state on power down https://git.kernel.org/coresight/c/2ab4645fe420
Best regards,