[PATCH v7 12/16] arm64: ras: Expose config abi through debugfs
Ruidong Tian
tianruidong at linux.alibaba.com
Tue Jun 2 00:15:35 PDT 2026
Expose per-node and per-record configuration through debugfs so
bring-up and validation can inspect and tweak driver state (CE
threshold, error counters, ...) at runtime.
Signed-off-by: Umang Chheda <umang.chheda at oss.qualcomm.com>
Signed-off-by: Ruidong Tian <tianruidong at linux.alibaba.com>
---
Documentation/ABI/testing/debugfs-arm64-ras | 50 +++++
MAINTAINERS | 1 +
drivers/ras/arm64/Makefile | 1 +
drivers/ras/arm64/ras-core.c | 31 +++
drivers/ras/arm64/ras-sysfs.c | 211 ++++++++++++++++++++
drivers/ras/arm64/ras.h | 17 ++
drivers/ras/debugfs.c | 3 +-
include/linux/ras.h | 2 +
8 files changed, 315 insertions(+), 1 deletion(-)
create mode 100644 Documentation/ABI/testing/debugfs-arm64-ras
create mode 100644 drivers/ras/arm64/ras-sysfs.c
diff --git a/Documentation/ABI/testing/debugfs-arm64-ras b/Documentation/ABI/testing/debugfs-arm64-ras
new file mode 100644
index 000000000000..d86bde83d0b9
--- /dev/null
+++ b/Documentation/ABI/testing/debugfs-arm64-ras
@@ -0,0 +1,50 @@
+What: /sys/kernel/debug/ras/arm64/<node_name>/
+Date: Dec 2025
+KernelVersion: 6.19
+Contact: Ruidong Tian <tianruidong at linux.alibaba.com>
+Description:
+ Directory representing a RAS node device, <name> means device
+ type, like:
+
+ - processor
+ - memory
+ - smmu
+ - ...
+
+What: /sys/kernel/debug/ras/arm64/<node_name>/record<index>/err_*
+Date: Dec 2025
+KernelVersion: 6.19
+Contact: Ruidong Tian <tianruidong at linux.alibaba.com>
+Description:
+ (RW) Read/Write err_* register.
+
+What: /sys/kernel/debug/ras/arm64/<node_name>/err_count
+Date: Dec 2025
+KernelVersion: 6.19
+Contact: Ruidong Tian <tianruidong at linux.alibaba.com>
+Description:
+ (RO) Outputs error statistics for all error records of this node.
+
+
+What: /sys/kernel/debug/ras/arm64/<node_name>/record<index>/err_count
+Date: Dec 2025
+KernelVersion: 6.19
+Contact: Ruidong Tian <tianruidong at linux.alibaba.com>
+Description:
+ (RO) Outputs error statistics for this record.
+
+What: /sys/kernel/debug/ras/arm64/<node_name>/ce_threshold
+Date: Dec 2025
+KernelVersion: 6.19
+Contact: Ruidong Tian <tianruidong at linux.alibaba.com>
+Description:
+ (WO) Write the CE threshold to all records of this node.
+ Returns error if input exceeded the maximum threshold.
+
+What: /sys/kernel/debug/ras/arm64/<node_name>/record<index>/ce_threshold
+Date: Dec 2025
+KernelVersion: 6.19
+Contact: Ruidong Tian <tianruidong at linux.alibaba.com>
+Description:
+ (RW) Read and write the CE threshold to this record.
+ Returns error if input exceeded the maximum threshold.
\ No newline at end of file
diff --git a/MAINTAINERS b/MAINTAINERS
index 766d1240b465..007a5a69b6d9 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -349,6 +349,7 @@ M: Ruidong Tian <tianruidong at linux.alibaba.com>
L: linux-acpi at vger.kernel.org
L: linux-arm-kernel at lists.infradead.org
S: Supported
+F: Documentation/ABI/testing/debugfs-arm64-ras
F: arch/arm64/include/asm/ras.h
F: drivers/acpi/arm64/aest.c
F: drivers/ras/arm64/
diff --git a/drivers/ras/arm64/Makefile b/drivers/ras/arm64/Makefile
index c5387f05a067..e13e223107dd 100644
--- a/drivers/ras/arm64/Makefile
+++ b/drivers/ras/arm64/Makefile
@@ -3,3 +3,4 @@
obj-$(CONFIG_ARM64_RAS_DRIVER) += arm64_ras.o
arm64_ras-y := ras-core.o
+arm64_ras-y += ras-sysfs.o
diff --git a/drivers/ras/arm64/ras-core.c b/drivers/ras/arm64/ras-core.c
index 0b07b69545ad..c427c131a862 100644
--- a/drivers/ras/arm64/ras-core.c
+++ b/drivers/ras/arm64/ras-core.c
@@ -25,6 +25,8 @@ MODULE_PARM_DESC(aest_panic_on_ue,
static DEFINE_PER_CPU(struct ras_node, percpu_ras_node);
+struct dentry *arm64_ras_debugfs;
+
static const char *const ras_node_name[] = {
[ACPI_AEST_PROCESSOR_ERROR_NODE] = "processor",
[ACPI_AEST_MEMORY_ERROR_NODE] = "memory",
@@ -158,6 +160,27 @@ static void ras_do_proc(struct ras_record *record, struct ras_ext_regs *regs)
u64 status = regs->err_status, addr = regs->err_addr;
ras_print(record, regs);
+ if (regs->err_status & ERR_STATUS_CE)
+ record->count.ce++;
+ if (regs->err_status & ERR_STATUS_DE)
+ record->count.de++;
+ if (regs->err_status & ERR_STATUS_UE) {
+ switch (FIELD_GET(ERR_STATUS_UET, regs->err_status)) {
+ case ERR_STATUS_UET_UC:
+ record->count.uc++;
+ break;
+ case ERR_STATUS_UET_UEU:
+ record->count.ueu++;
+ break;
+ case ERR_STATUS_UET_UER:
+ record->count.uer++;
+ break;
+ case ERR_STATUS_UET_UEO:
+ record->count.ueo++;
+ break;
+ }
+ }
+
atomic_notifier_call_chain(&ras_decoder_chain, 0, record);
if (status & ERR_STATUS_CE)
@@ -887,6 +910,8 @@ static int arm64_ras_probe(struct platform_device *pdev)
platform_set_drvdata(pdev, node);
+ ras_node_init_debugfs(node);
+
return 0;
}
@@ -900,12 +925,18 @@ static struct platform_driver arm64_ras_driver = {
static int __init arm64_ras_init(void)
{
+#ifdef CONFIG_DEBUG_FS
+ arm64_ras_debugfs = debugfs_create_dir("arm64", ras_debugfs_dir);
+#endif
return platform_driver_register(&arm64_ras_driver);
}
module_init(arm64_ras_init);
static void __exit arm64_ras_exit(void)
{
+#ifdef CONFIG_DEBUG_FS
+ debugfs_remove_recursive(arm64_ras_debugfs);
+#endif
platform_driver_unregister(&arm64_ras_driver);
}
module_exit(arm64_ras_exit);
diff --git a/drivers/ras/arm64/ras-sysfs.c b/drivers/ras/arm64/ras-sysfs.c
new file mode 100644
index 000000000000..03cc00b820e2
--- /dev/null
+++ b/drivers/ras/arm64/ras-sysfs.c
@@ -0,0 +1,211 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * ARM Error Source Table Support
+ *
+ * Copyright (c) 2025, Alibaba Group.
+ */
+
+#include "ras.h"
+
+static int ras_store_threshold(struct ras_record *record, u64 threshold)
+{
+ struct ce_threshold *ce = &record->ce;
+ u64 err_misc0;
+
+ if (!ce->info)
+ return -EOPNOTSUPP;
+
+ if (threshold > ce->info->max_count)
+ return -EINVAL;
+
+ ce->threshold = threshold;
+ ce->count = ce->info->max_count - threshold + 1;
+
+ err_misc0 = record_read(record, ERXMISC0);
+ ce->reg_val = (err_misc0 & ~ce->info->mask) |
+ (ce->count << ce->info->shift);
+
+ record_write(record, ERXMISC0, ce->reg_val);
+ return 0;
+}
+
+static void ras_error_count(struct ras_record *record, struct record_count *count)
+{
+ count->ce += record->count.ce;
+ count->de += record->count.de;
+ count->uc += record->count.uc;
+ count->ueu += record->count.ueu;
+ count->uer += record->count.uer;
+ count->ueo += record->count.ueo;
+}
+
+/* Debugfs for RAS node */
+
+static int ras_node_err_count_show(struct seq_file *m, void *data)
+{
+ struct ras_node *node = m->private;
+ struct record_count count = { 0 };
+ int i;
+
+ for (i = 0; i < node->record_count; i++)
+ if (!test_bit(i, node->record_implemented))
+ ras_error_count(&node->records[i], &count);
+
+ seq_printf(m, "CE: %llu\n"
+ "DE: %llu\n"
+ "UC: %llu\n"
+ "UEU: %llu\n"
+ "UEO: %llu\n"
+ "UER: %llu\n",
+ count.ce, count.de, count.uc, count.ueu,
+ count.uer, count.ueo);
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(ras_node_err_count);
+
+/* Attribute for RAS record */
+
+#define DEFINE_RAS_DEBUGFS_ATTR(name, offset) \
+static int name##_get(void *data, u64 *val) \
+{ \
+ struct ras_record *record = data; \
+ *val = record_read(record, offset); \
+ return 0; \
+} \
+static int name##_set(void *data, u64 val) \
+{ \
+ struct ras_record *record = data; \
+ record_write(record, offset, val); \
+ return 0; \
+} \
+DEFINE_DEBUGFS_ATTRIBUTE(name##_ops, name##_get, name##_set, "%#llx\n")
+
+DEFINE_RAS_DEBUGFS_ATTR(err_fr, ERXFR);
+DEFINE_RAS_DEBUGFS_ATTR(err_ctrl, ERXCTLR);
+
+static int record_ce_threshold_get(void *data, u64 *val)
+{
+ struct ras_record *record = data;
+
+ *val = record->ce.threshold;
+ return 0;
+}
+
+static int record_ce_threshold_set(void *data, u64 val)
+{
+ struct ras_record *record = data;
+
+ return ras_store_threshold(record, val);
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(record_ce_threshold_ops, record_ce_threshold_get,
+ record_ce_threshold_set, "%llu\n");
+
+/* Node-level ce_threshold: write threshold to all records of this node */
+
+static int node_ce_threshold_set(void *data, u64 val)
+{
+ struct ras_node *node = data;
+ int i, ret, last_err = -EOPNOTSUPP;
+
+ for (i = 0; i < node->record_count; i++) {
+ ret = ras_store_threshold(&node->records[i], val);
+ if (ret == 0)
+ last_err = 0;
+ else if (ret == -EINVAL)
+ return ret;
+ }
+
+ return last_err;
+}
+
+DEFINE_DEBUGFS_ATTRIBUTE(node_ce_threshold_ops, NULL,
+ node_ce_threshold_set, "%llu\n");
+
+static int ras_record_err_count_show(struct seq_file *m, void *data)
+{
+ struct ras_record *record = m->private;
+ struct record_count count = { 0 };
+
+ ras_error_count(record, &count);
+
+ seq_printf(m, "CE: %llu\n"
+ "DE: %llu\n"
+ "UC: %llu\n"
+ "UEU: %llu\n"
+ "UEO: %llu\n"
+ "UER: %llu\n",
+ count.ce, count.de, count.uc, count.ueu,
+ count.uer, count.ueo);
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(ras_record_err_count);
+
+static void ras_record_init_debugfs(struct ras_record *record)
+{
+ debugfs_create_file("err_fr", 0600, record->debugfs,
+ record, &err_fr_ops);
+ debugfs_create_file("err_ctrl", 0600, record->debugfs,
+ record, &err_ctrl_ops);
+ debugfs_create_file("err_count", 0400, record->debugfs,
+ record, &ras_record_err_count_fops);
+ debugfs_create_file("ce_threshold", 0600, record->debugfs,
+ record, &record_ce_threshold_ops);
+}
+
+static void ras_init_records_debugfs(struct ras_node *node)
+{
+ struct ras_record *record;
+ int i;
+
+ for (i = 0; i < node->record_count; i++) {
+ record = &node->records[i];
+ if (!record->name || test_bit(i, node->record_implemented))
+ continue;
+ record->debugfs = debugfs_create_dir(record->name,
+ node->debugfs);
+
+ ras_record_init_debugfs(record);
+ }
+}
+
+static void ras_oncore_node_init_debugfs(struct ras_node *node)
+{
+ int cpu;
+ struct ras_node *percpu_node;
+ char name[16];
+
+ for_each_possible_cpu(cpu) {
+ percpu_node = per_cpu_ptr(node->oncore_node, cpu);
+
+ snprintf(name, sizeof(name), "processor%u", cpu);
+ percpu_node->debugfs = debugfs_create_dir(name, arm64_ras_debugfs);
+
+ debugfs_create_file("err_count", 0400, percpu_node->debugfs,
+ percpu_node, &ras_node_err_count_fops);
+ debugfs_create_file("ce_threshold", 0200, percpu_node->debugfs,
+ percpu_node, &node_ce_threshold_ops);
+ ras_init_records_debugfs(percpu_node);
+ }
+}
+
+void ras_node_init_debugfs(struct ras_node *node)
+{
+ if (!node->name)
+ return;
+
+ if (ras_node_is_oncore(node)) {
+ ras_oncore_node_init_debugfs(node);
+ return;
+ }
+
+ node->debugfs = debugfs_create_dir(node->name, arm64_ras_debugfs);
+ if (IS_ERR_OR_NULL(node->debugfs))
+ return;
+
+ debugfs_create_file("err_count", 0400, node->debugfs,
+ node, &ras_node_err_count_fops);
+ debugfs_create_file("ce_threshold", 0200, node->debugfs,
+ node, &node_ce_threshold_ops);
+ ras_init_records_debugfs(node);
+}
diff --git a/drivers/ras/arm64/ras.h b/drivers/ras/arm64/ras.h
index ac3876912495..92cbb975b4df 100644
--- a/drivers/ras/arm64/ras.h
+++ b/drivers/ras/arm64/ras.h
@@ -10,6 +10,7 @@
#include <linux/acpi_aest.h>
#include <asm/ras.h>
+#include <linux/debugfs.h>
#define DEFAULT_CE_THRESHOLD 1
@@ -62,6 +63,8 @@
#define GIC_ERRDEVARCH 0xFFBC
+extern struct dentry *arm64_ras_debugfs;
+
struct ras_access {
u64 (*read)(void __iomem *base, u32 offset);
void (*write)(void __iomem *base, u32 offset, u64 val);
@@ -80,14 +83,25 @@ struct ce_threshold {
u64 reg_val;
};
+struct record_count {
+ u64 ce;
+ u64 de;
+ u64 uc;
+ u64 uer;
+ u64 ueo;
+ u64 ueu;
+};
+
struct ras_record {
char *name;
void __iomem *regs_base;
struct ras_node *node;
const struct ras_access *access;
+ struct dentry *debugfs;
struct ce_threshold ce;
enum ras_ce_threshold threshold_type;
+ struct record_count count;
int index;
/*
@@ -116,6 +130,7 @@ struct ras_node {
struct device *dev;
const struct ras_group *group;
struct ras_node __percpu *oncore_node;
+ struct dentry *debugfs;
void __iomem *base;
void __iomem *errgsr;
@@ -286,4 +301,6 @@ static inline void ras_sync(struct ras_node *node)
isb();
}
+void ras_node_init_debugfs(struct ras_node *node);
+
#endif /* _DRIVERS_RAS_ARM64_RAS_H_ */
diff --git a/drivers/ras/debugfs.c b/drivers/ras/debugfs.c
index 42afd3de68b2..e4d9a5627e5f 100644
--- a/drivers/ras/debugfs.c
+++ b/drivers/ras/debugfs.c
@@ -3,7 +3,8 @@
#include <linux/ras.h>
#include "debugfs.h"
-static struct dentry *ras_debugfs_dir;
+struct dentry *ras_debugfs_dir;
+EXPORT_SYMBOL_GPL(ras_debugfs_dir);
static atomic_t trace_count = ATOMIC_INIT(0);
diff --git a/include/linux/ras.h b/include/linux/ras.h
index 11663150612f..976cd102f76c 100644
--- a/include/linux/ras.h
+++ b/include/linux/ras.h
@@ -71,4 +71,6 @@ static inline void ras_register_decode_chain(struct notifier_block *nb) {}
static inline void ras_unregister_decode_chain(struct notifier_block *nb) {}
#endif /* CONFIG_ARM64_RAS_DRIVER */
+extern struct dentry *ras_debugfs_dir;
+
#endif /* __RAS_H__ */
--
2.51.2.612.gdc70283dfc
More information about the linux-arm-kernel
mailing list