[PATCH v3 0/3] Add drm_ras netlink error event support

[PATCH v3 0/3] Add drm_ras netlink error event support

[Riana Tauro]

Define a new netlink event 'error-event' and a new multicast group
'error-report' in drm_ras. Each event contains device name, node and
error information to identify the error triggering the event.

Add drm_ras_nl_error_event() to trigger an event from the driver.
Wire this support to xe drm_ras to notify userspace whenever a GT or
SoC error occurs in PVC. Also add support for correctable errors in
CRI.

$ sudo ynl --family drm_ras --output-json --subscribe error-report

{
    "name": "error-event",
     "msg": {
         "device-name": "0000:03:00.0",
         "node-id": 1,
         "node-name": "uncorrectable-errors",
         "error-id": 1,
         "error-name": "core-compute",
         "error-value": 1
     }
}

Rev2: use ynl in document and commit message
      fix cosmetic review comments
      simplify caller 

Rev3: replace error-event with error-report
      had has_drm_ras check 
      add support for correctable errors in CRI

Riana Tauro (3):
  drm/drm_ras: Add drm_ras netlink error event
  drm/xe/xe_drm_ras: Add error-event support for PVC
  drm/xe/xe_ras: Add error-event support for Crescent Island

 Documentation/gpu/drm-ras.rst            | 21 ++++++
 Documentation/netlink/specs/drm_ras.yaml | 48 +++++++++++++
 drivers/gpu/drm/drm_ras.c                | 87 ++++++++++++++++++++++++
 drivers/gpu/drm/drm_ras_nl.c             |  6 ++
 drivers/gpu/drm/drm_ras_nl.h             |  4 ++
 drivers/gpu/drm/xe/xe_drm_ras.c          | 30 ++++++++
 drivers/gpu/drm/xe/xe_drm_ras.h          |  3 +
 drivers/gpu/drm/xe/xe_hw_error.c         |  5 +-
 drivers/gpu/drm/xe/xe_ras.c              | 53 +++++++++++++++
 include/drm/drm_ras.h                    |  5 ++
 include/uapi/drm/drm_ras.h               | 15 ++++
 11 files changed, 276 insertions(+), 1 deletion(-)

-- 
2.47.1

[PATCH v3 1/3] drm/drm_ras: Add drm_ras netlink error event

[Riana Tauro]

Define a new netlink event 'error-event' and a new multicast group
'error-report' in drm_ras. Each event contains device name, node and
error information to identify the error triggering the event.

Add drm_ras_nl_error_event() to trigger an event from the driver.
Userspace must subscribe to 'error-report' to receive 'error-event'
notifications.

Usage:

$ sudo ynl --family drm_ras --subscribe error-report

Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Zack McKevitt <zachary.mckevitt@oss.qualcomm.com>
Cc: Lijo Lazar <lijo.lazar@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Riana Tauro <riana.tauro@intel.com>
Reviewed-by: Raag Jadav <raag.jadav@intel.com>
---
v2: remove redundant initialization
    remove unnecessary space
    use ynl in commit message and doc (Raag)
    simplify doc for error-event attrs

v3: rename error-notify to error-report
    Replace notify with report across the file (Raag)
---
 Documentation/gpu/drm-ras.rst            | 21 ++++++
 Documentation/netlink/specs/drm_ras.yaml | 48 +++++++++++++
 drivers/gpu/drm/drm_ras.c                | 87 ++++++++++++++++++++++++
 drivers/gpu/drm/drm_ras_nl.c             |  6 ++
 drivers/gpu/drm/drm_ras_nl.h             |  4 ++
 include/drm/drm_ras.h                    |  5 ++
 include/uapi/drm/drm_ras.h               | 15 ++++
 7 files changed, 186 insertions(+)

diff --git a/Documentation/gpu/drm-ras.rst b/Documentation/gpu/drm-ras.rst
index 83c21853b74b..406e4c49bac1 100644
--- a/Documentation/gpu/drm-ras.rst
+++ b/Documentation/gpu/drm-ras.rst
@@ -56,6 +56,7 @@ User space tools can:
   ``node-id`` and ``error-id`` as parameters.
 * Clear specific error counters with the ``clear-error-counter`` command, using both
   ``node-id`` and ``error-id`` as parameters.
+* Subscribe to the ``error-report`` multicast group to receive ``error-event``.
 
 YAML-based Interface
 --------------------
@@ -111,3 +112,23 @@ Example: Clear an error counter for a given node
 
     sudo ynl --family drm_ras --do clear-error-counter --json '{"node-id":0, "error-id":1}'
     None
+
+Example: Subscribe to ``error-report`` multicast group
+
+.. code-block:: bash
+
+    sudo ynl --family drm_ras --output-json --subscribe error-report
+
+.. code-block:: json
+
+    {
+        "name": "error-event",
+        "msg": {
+            "device-name": "0000:03:00.0",
+            "node-id": 1,
+            "node-name": "uncorrectable-errors",
+            "error-id": 1,
+            "error-name": "error_name1",
+            "error-value": 1
+        }
+    }
diff --git a/Documentation/netlink/specs/drm_ras.yaml b/Documentation/netlink/specs/drm_ras.yaml
index e113056f8c01..8aed3d4515e5 100644
--- a/Documentation/netlink/specs/drm_ras.yaml
+++ b/Documentation/netlink/specs/drm_ras.yaml
@@ -69,6 +69,33 @@ attribute-sets:
         name: error-value
         type: u32
         doc: Current value of the requested error counter.
+  -
+    name: error-event-attrs
+    attributes:
+      -
+        name: device-name
+        type: string
+        doc: Device (PCI BDF, UUID) that reported the error.
+      -
+        name: node-id
+        type: u32
+        doc: ID of the node that reported the error.
+      -
+        name: node-name
+        type: string
+        doc: Name of the node that reported the error.
+      -
+        name: error-id
+        type: u32
+        doc: ID of the error counter.
+      -
+        name: error-name
+        type: string
+        doc: Name of the error.
+      -
+        name: error-value
+        type: u32
+        doc: Current value of the error counter.
 
 operations:
   list:
@@ -124,3 +151,24 @@ operations:
       do:
         request:
           attributes: *id-attrs
+    -
+      name: error-event
+      doc: >-
+           Report an error event to userspace.
+           The event includes the device, node and error information
+           of the error that triggered the event.
+      attribute-set: error-event-attrs
+      mcgrp: error-report
+      event:
+        attributes:
+          - device-name
+          - node-id
+          - node-name
+          - error-id
+          - error-name
+          - error-value
+
+mcast-groups:
+  list:
+    -
+      name: error-report
diff --git a/drivers/gpu/drm/drm_ras.c b/drivers/gpu/drm/drm_ras.c
index d6eab29a1394..77f912a4d101 100644
--- a/drivers/gpu/drm/drm_ras.c
+++ b/drivers/gpu/drm/drm_ras.c
@@ -41,6 +41,11 @@
  *    Userspace must provide Node ID, Error ID.
  *    Clears specific error counter of a node if supported.
  *
+ * 4. ERROR_REPORT: Subscribe to this multicast group to receive error events
+ *
+ * 5. ERROR_EVENT: Report an error event to userspace. The event contains device, node
+ *    and error information that triggered the event.
+ *
  * Node registration:
  *
  * - drm_ras_node_register(): Registers a new node and assigns
@@ -186,6 +191,34 @@ static int msg_reply_value(struct sk_buff *msg, u32 error_id,
 			   value);
 }
 
+static int msg_put_error_event_attrs(struct sk_buff *msg, struct drm_ras_node *node,
+				     u32 error_id, const char *error_name, u32 value)
+{
+	int ret;
+
+	ret = nla_put_string(msg, DRM_RAS_A_ERROR_EVENT_ATTRS_DEVICE_NAME, node->device_name);
+	if (ret)
+		return ret;
+
+	ret = nla_put_u32(msg, DRM_RAS_A_ERROR_EVENT_ATTRS_NODE_ID, node->id);
+	if (ret)
+		return ret;
+
+	ret = nla_put_string(msg, DRM_RAS_A_ERROR_EVENT_ATTRS_NODE_NAME, node->node_name);
+	if (ret)
+		return ret;
+
+	ret = nla_put_u32(msg, DRM_RAS_A_ERROR_EVENT_ATTRS_ERROR_ID, error_id);
+	if (ret)
+		return ret;
+
+	ret = nla_put_string(msg, DRM_RAS_A_ERROR_EVENT_ATTRS_ERROR_NAME, error_name);
+	if (ret)
+		return ret;
+
+	return nla_put_u32(msg, DRM_RAS_A_ERROR_EVENT_ATTRS_ERROR_VALUE, value);
+}
+
 static int doit_reply_value(struct genl_info *info, u32 node_id,
 			    u32 error_id)
 {
@@ -222,6 +255,60 @@ static int doit_reply_value(struct genl_info *info, u32 node_id,
 	return genlmsg_reply(msg, info);
 }
 
+/**
+ * drm_ras_nl_error_event() - Report an error event
+ * @node: Node structure
+ * @error_id: ID of the error
+ * @error_name: Name of the error
+ * @value: Value associated with the error
+ * @flags: GFP flags for memory allocation
+ *
+ * Report an error-event to userspace using the error-report multicast group.
+ *
+ * Return: 0 on success, or negative errno on failure.
+ */
+int drm_ras_nl_error_event(struct drm_ras_node *node, u32 error_id, const char *error_name,
+			   u32 value, gfp_t flags)
+{
+	struct genl_info info;
+	struct sk_buff *msg;
+	struct nlattr *hdr;
+	int ret;
+
+	if (!error_name)
+		return -EINVAL;
+
+	if (!genl_has_listeners(&drm_ras_nl_family, &init_net, DRM_RAS_NLGRP_ERROR_REPORT))
+		return 0;
+
+	genl_info_init_ntf(&info, &drm_ras_nl_family, DRM_RAS_CMD_ERROR_EVENT);
+
+	msg = genlmsg_new(NLMSG_GOODSIZE, flags);
+	if (!msg)
+		return -ENOMEM;
+
+	hdr = genlmsg_iput(msg, &info);
+	if (!hdr) {
+		ret = -EMSGSIZE;
+		goto free_msg;
+	}
+
+	ret = msg_put_error_event_attrs(msg, node, error_id, error_name, value);
+	if (ret)
+		goto cancel_msg;
+
+	genlmsg_end(msg, hdr);
+	genlmsg_multicast(&drm_ras_nl_family, msg, 0, DRM_RAS_NLGRP_ERROR_REPORT, flags);
+	return 0;
+
+cancel_msg:
+	genlmsg_cancel(msg, hdr);
+free_msg:
+	nlmsg_free(msg);
+	return ret;
+}
+EXPORT_SYMBOL(drm_ras_nl_error_event);
+
 /**
  * drm_ras_nl_get_error_counter_dumpit() - Dump all Error Counters
  * @skb: Netlink message buffer
diff --git a/drivers/gpu/drm/drm_ras_nl.c b/drivers/gpu/drm/drm_ras_nl.c
index dea1c1b2494e..9d3123cc9f9c 100644
--- a/drivers/gpu/drm/drm_ras_nl.c
+++ b/drivers/gpu/drm/drm_ras_nl.c
@@ -58,6 +58,10 @@ static const struct genl_split_ops drm_ras_nl_ops[] = {
 	},
 };
 
+static const struct genl_multicast_group drm_ras_nl_mcgrps[] = {
+	[DRM_RAS_NLGRP_ERROR_REPORT] = { "error-report", },
+};
+
 struct genl_family drm_ras_nl_family __ro_after_init = {
 	.name		= DRM_RAS_FAMILY_NAME,
 	.version	= DRM_RAS_FAMILY_VERSION,
@@ -66,4 +70,6 @@ struct genl_family drm_ras_nl_family __ro_after_init = {
 	.module		= THIS_MODULE,
 	.split_ops	= drm_ras_nl_ops,
 	.n_split_ops	= ARRAY_SIZE(drm_ras_nl_ops),
+	.mcgrps		= drm_ras_nl_mcgrps,
+	.n_mcgrps	= ARRAY_SIZE(drm_ras_nl_mcgrps),
 };
diff --git a/drivers/gpu/drm/drm_ras_nl.h b/drivers/gpu/drm/drm_ras_nl.h
index a398643572a5..03ec275aca92 100644
--- a/drivers/gpu/drm/drm_ras_nl.h
+++ b/drivers/gpu/drm/drm_ras_nl.h
@@ -21,6 +21,10 @@ int drm_ras_nl_get_error_counter_dumpit(struct sk_buff *skb,
 int drm_ras_nl_clear_error_counter_doit(struct sk_buff *skb,
 					struct genl_info *info);
 
+enum {
+	DRM_RAS_NLGRP_ERROR_REPORT,
+};
+
 extern struct genl_family drm_ras_nl_family;
 
 #endif /* _LINUX_DRM_RAS_GEN_H */
diff --git a/include/drm/drm_ras.h b/include/drm/drm_ras.h
index f2a787bc4f64..d4a275efdbb0 100644
--- a/include/drm/drm_ras.h
+++ b/include/drm/drm_ras.h
@@ -78,9 +78,14 @@ struct drm_device;
 #if IS_ENABLED(CONFIG_DRM_RAS)
 int drm_ras_node_register(struct drm_ras_node *node);
 void drm_ras_node_unregister(struct drm_ras_node *node);
+int drm_ras_nl_error_event(struct drm_ras_node *node, u32 error_id, const char *error_name,
+			   u32 value, gfp_t flags);
 #else
 static inline int drm_ras_node_register(struct drm_ras_node *node) { return 0; }
 static inline void drm_ras_node_unregister(struct drm_ras_node *node) { }
+static inline int drm_ras_nl_error_event(struct drm_ras_node *node, u32 error_id,
+					 const char *error_name, u32 value, gfp_t flags)
+{ return 0; }
 #endif
 
 #endif
diff --git a/include/uapi/drm/drm_ras.h b/include/uapi/drm/drm_ras.h
index 218a3ee86805..eab8231aa87c 100644
--- a/include/uapi/drm/drm_ras.h
+++ b/include/uapi/drm/drm_ras.h
@@ -38,13 +38,28 @@ enum {
 	DRM_RAS_A_ERROR_COUNTER_ATTRS_MAX = (__DRM_RAS_A_ERROR_COUNTER_ATTRS_MAX - 1)
 };
 
+enum {
+	DRM_RAS_A_ERROR_EVENT_ATTRS_DEVICE_NAME = 1,
+	DRM_RAS_A_ERROR_EVENT_ATTRS_NODE_ID,
+	DRM_RAS_A_ERROR_EVENT_ATTRS_NODE_NAME,
+	DRM_RAS_A_ERROR_EVENT_ATTRS_ERROR_ID,
+	DRM_RAS_A_ERROR_EVENT_ATTRS_ERROR_NAME,
+	DRM_RAS_A_ERROR_EVENT_ATTRS_ERROR_VALUE,
+
+	__DRM_RAS_A_ERROR_EVENT_ATTRS_MAX,
+	DRM_RAS_A_ERROR_EVENT_ATTRS_MAX = (__DRM_RAS_A_ERROR_EVENT_ATTRS_MAX - 1)
+};
+
 enum {
 	DRM_RAS_CMD_LIST_NODES = 1,
 	DRM_RAS_CMD_GET_ERROR_COUNTER,
 	DRM_RAS_CMD_CLEAR_ERROR_COUNTER,
+	DRM_RAS_CMD_ERROR_EVENT,
 
 	__DRM_RAS_CMD_MAX,
 	DRM_RAS_CMD_MAX = (__DRM_RAS_CMD_MAX - 1)
 };
 
+#define DRM_RAS_MCGRP_ERROR_REPORT	"error-report"
+
 #endif /* _UAPI_LINUX_DRM_RAS_H */
-- 
2.47.1

[PATCH v3 2/3] drm/xe/xe_drm_ras: Add error-event support for PVC

[Riana Tauro]

Report drm_ras error event to userspace when an error occurs.
Add support for core-compute and SoC errors in PVC.

$ sudo ynl --family drm_ras --output-json --subscribe error-report

{
    "name": "error-event",
     "msg": {
         "device-name": "0000:03:00.0",
         "node-id": 1,
         "node-name": "uncorrectable-errors",
         "error-id": 1,
         "error-name": "core-compute",
         "error-value": 1
     }
}

Signed-off-by: Riana Tauro <riana.tauro@intel.com>
Reviewed-by: Raag Jadav <raag.jadav@intel.com>
---
v2: use ynl (Raag)
    use value as function parameter
    move error event call to hw_error_source_handler 

v3: add has_drm_ras check
---
 drivers/gpu/drm/xe/xe_drm_ras.c  | 30 ++++++++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_drm_ras.h  |  3 +++
 drivers/gpu/drm/xe/xe_hw_error.c |  5 ++++-
 3 files changed, 37 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_drm_ras.c b/drivers/gpu/drm/xe/xe_drm_ras.c
index 7937d8ba0ed9..36afdfb5e412 100644
--- a/drivers/gpu/drm/xe/xe_drm_ras.c
+++ b/drivers/gpu/drm/xe/xe_drm_ras.c
@@ -185,6 +185,36 @@ static int register_nodes(struct xe_device *xe)
 	return ret;
 }
 
+/**
+ * xe_drm_ras_event() - Report drm-ras error event to userspace
+ * @xe: xe device structure
+ * @component: error component (see &enum drm_xe_ras_error_component)
+ * @severity: error severity (see &enum drm_xe_ras_error_severity)
+ * @value: value of error counter
+ * @flags: flags for allocation
+ *
+ * Report an error-event to userspace.
+ */
+void xe_drm_ras_event(struct xe_device *xe, u32 component, u32 severity, u32 value, gfp_t flags)
+{
+	struct xe_drm_ras *ras = &xe->ras;
+	struct xe_drm_ras_counter *info = ras->info[severity];
+	struct drm_ras_node *node = &ras->node[severity];
+	int ret;
+
+	/* Event is supported only if drm_ras is enabled */
+	if (!xe->info.has_drm_ras)
+		return;
+
+	if (!info || !info[component].name)
+		return;
+
+	ret = drm_ras_nl_error_event(node, component, info[component].name, value, flags);
+	if (ret)
+		drm_err(&xe->drm, "RAS error-event failed: %d for %s %s\n", ret,
+			info[component].name, error_severity[severity]);
+}
+
 /**
  * xe_drm_ras_init() - Initialize DRM RAS
  * @xe: xe device instance
diff --git a/drivers/gpu/drm/xe/xe_drm_ras.h b/drivers/gpu/drm/xe/xe_drm_ras.h
index 365c70e93e82..2a694bf69478 100644
--- a/drivers/gpu/drm/xe/xe_drm_ras.h
+++ b/drivers/gpu/drm/xe/xe_drm_ras.h
@@ -5,11 +5,14 @@
 #ifndef _XE_DRM_RAS_H_
 #define _XE_DRM_RAS_H_
 
+#include <linux/types.h>
+
 struct xe_device;
 
 #define for_each_error_severity(i)	\
 	for (i = 0; i < DRM_XE_RAS_ERR_SEV_MAX; i++)
 
 int xe_drm_ras_init(struct xe_device *xe);
+void xe_drm_ras_event(struct xe_device *xe, u32 component, u32 severity, u32 value, gfp_t flags);
 
 #endif
diff --git a/drivers/gpu/drm/xe/xe_hw_error.c b/drivers/gpu/drm/xe/xe_hw_error.c
index 4a4b363fc844..a833cecc74ec 100644
--- a/drivers/gpu/drm/xe/xe_hw_error.c
+++ b/drivers/gpu/drm/xe/xe_hw_error.c
@@ -432,7 +432,7 @@ static void hw_error_source_handler(struct xe_tile *tile, const enum hardware_er
 	struct xe_drm_ras *ras = &xe->ras;
 	struct xe_drm_ras_counter *info = ras->info[severity];
 	unsigned long flags, err_src;
-	u32 err_bit;
+	u32 err_bit, value;
 
 	if (!IS_DGFX(xe))
 		return;
@@ -495,6 +495,9 @@ static void hw_error_source_handler(struct xe_tile *tile, const enum hardware_er
 			gt_hw_error_handler(tile, hw_err, error_id);
 		if (err_bit == XE_SOC_ERROR)
 			soc_hw_error_handler(tile, hw_err, error_id);
+
+		value = atomic_read(&info[error_id].counter);
+		xe_drm_ras_event(xe, error_id, severity, value, GFP_ATOMIC);
 	}
 
 clear_reg:
-- 
2.47.1

[PATCH v3 3/3] drm/xe/xe_ras: Add error-event support for CRI

[Riana Tauro]

Add error-event support for Correctable errors in CRI.
error-event is reported to  userspace for all errors that crossed
threshold on receiving an interrupt for correctable errors.

Signed-off-by: Riana Tauro <riana.tauro@intel.com>
---
 drivers/gpu/drm/xe/xe_ras.c | 53 +++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_ras.c b/drivers/gpu/drm/xe/xe_ras.c
index 44f4e1a3455b..acf3207aa2fd 100644
--- a/drivers/gpu/drm/xe/xe_ras.c
+++ b/drivers/gpu/drm/xe/xe_ras.c
@@ -77,6 +77,18 @@ static u8 drm_to_xe_ras_severity(u8 severity)
 	}
 }
 
+static u8 xe_to_drm_ras_severity(u8 severity)
+{
+	switch (severity) {
+	case XE_RAS_SEV_CORRECTABLE:
+		return DRM_XE_RAS_ERR_SEV_CORRECTABLE;
+	case XE_RAS_SEV_UNCORRECTABLE:
+		return DRM_XE_RAS_ERR_SEV_UNCORRECTABLE;
+	default:
+		return DRM_XE_RAS_ERR_SEV_MAX;
+	}
+}
+
 static u8 drm_to_xe_ras_component(u8 component)
 {
 	switch (component) {
@@ -95,6 +107,24 @@ static u8 drm_to_xe_ras_component(u8 component)
 	}
 }
 
+static u8 xe_to_drm_ras_component(u8 component)
+{
+	switch (component) {
+	case XE_RAS_COMP_DEVICE_MEMORY:
+		return DRM_XE_RAS_ERR_COMP_DEVICE_MEMORY;
+	case XE_RAS_COMP_CORE_COMPUTE:
+		return DRM_XE_RAS_ERR_COMP_CORE_COMPUTE;
+	case XE_RAS_COMP_PCIE:
+		return DRM_XE_RAS_ERR_COMP_PCIE;
+	case XE_RAS_COMP_FABRIC:
+		return DRM_XE_RAS_ERR_COMP_FABRIC;
+	case XE_RAS_COMP_SOC_INTERNAL:
+		return DRM_XE_RAS_ERR_COMP_SOC_INTERNAL;
+	default:
+		return DRM_XE_RAS_ERR_COMP_MAX;
+	}
+}
+
 static int ras_status_to_errno(u32 status)
 {
 	switch (status) {
@@ -131,6 +161,27 @@ static inline const char *comp_to_str(u8 component)
 	return xe_ras_components[component];
 }
 
+static void ras_send_error_event(struct xe_device *xe, u8 severity, u8 component)
+{
+	u8 drm_severity, drm_component;
+	u32 value;
+	int ret;
+
+	drm_severity = xe_to_drm_ras_severity(severity);
+	if (drm_severity == DRM_XE_RAS_ERR_SEV_MAX)
+		return;
+
+	drm_component = xe_to_drm_ras_component(component);
+	if (drm_component == DRM_XE_RAS_ERR_COMP_MAX)
+		return;
+
+	ret = xe_ras_get_counter(xe, severity, component, &value);
+	if (ret)
+		return;
+
+	xe_drm_ras_event(xe, drm_component, drm_severity, value, GFP_KERNEL);
+}
+
 void xe_ras_counter_threshold_crossed(struct xe_device *xe,
 				      struct xe_sysctrl_event_response *response)
 {
@@ -152,6 +203,8 @@ void xe_ras_counter_threshold_crossed(struct xe_device *xe,
 		severity = errors[id].common.severity;
 		component = errors[id].common.component;
 
+		ras_send_error_event(xe, severity, component);
+
 		xe_warn(xe, "[RAS]: %s %s detected\n",
 			comp_to_str(component), sev_to_str(severity));
 	}
-- 
2.47.1

Re: [PATCH v3 1/3] drm/drm_ras: Add drm_ras netlink error event

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

New issues:
- [Medium] Multicast events are restricted to `init_net`, silently dropping events for listeners in other network namespaces despite `.netnsok = true`.

Pre-existing issues:
- [High] Lockless access to `struct drm_ras_node` pointers retrieved from xarray leads to Use-After-Free.
--

commit aab90b1a7792dc491d0dcac6947d349046f95876
Author: Riana Tauro <riana.tauro@intel.com>

drm/drm_ras: Add drm_ras netlink error event

This commit defines a new netlink event 'error-event' and a multicast
group 'error-report' for the DRM RAS subsystem. It provides a function
for drivers to report device and node error information to subscribed
userspace listeners.

Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Zack McKevitt <zachary.mckevitt@oss.qualcomm.com>
Cc: Lijo Lazar <lijo.lazar@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Riana Tauro <riana.tauro@intel.com>
Reviewed-by: Raag Jadav <raag.jadav@intel.com>

> diff --git a/drivers/gpu/drm/drm_ras.c b/drivers/gpu/drm/drm_ras.c
> index d6eab29a1394c..77f912a4d101f 100644
> --- a/drivers/gpu/drm/drm_ras.c
> +++ b/drivers/gpu/drm/drm_ras.c

[ ... ]

> @@ -222,6 +255,60 @@ static int doit_reply_value(struct genl_info *info, u32 node_id,
>  	return genlmsg_reply(msg, info);
>  }
>  
> +/**
> + * drm_ras_nl_error_event() - Report an error event
> + * @node: Node structure
> + * @error_id: ID of the error
> + * @error_name: Name of the error
> + * @value: Value associated with the error
> + * @flags: GFP flags for memory allocation
> + *
> + * Report an error-event to userspace using the error-report multicast group.
> + *
> + * Return: 0 on success, or negative errno on failure.
> + */
> +int drm_ras_nl_error_event(struct drm_ras_node *node, u32 error_id, const char *error_name,
> +			   u32 value, gfp_t flags)
> +{
> +	struct genl_info info;
> +	struct sk_buff *msg;
> +	struct nlattr *hdr;
> +	int ret;
> +
> +	if (!error_name)
> +		return -EINVAL;
> +
> +	if (!genl_has_listeners(&drm_ras_nl_family, &init_net, DRM_RAS_NLGRP_ERROR_REPORT))

[Severity: Medium]
Will this silently drop events for users subscribing from non-init network
namespaces? The drm_ras_nl_family sets .netnsok = true, but this check
is hardcoded to init_net.

> +		return 0;
> +
> +	genl_info_init_ntf(&info, &drm_ras_nl_family, DRM_RAS_CMD_ERROR_EVENT);
> +
> +	msg = genlmsg_new(NLMSG_GOODSIZE, flags);
> +	if (!msg)
> +		return -ENOMEM;
> +
> +	hdr = genlmsg_iput(msg, &info);
> +	if (!hdr) {
> +		ret = -EMSGSIZE;
> +		goto free_msg;
> +	}
> +
> +	ret = msg_put_error_event_attrs(msg, node, error_id, error_name, value);
> +	if (ret)
> +		goto cancel_msg;
> +
> +	genlmsg_end(msg, hdr);
> +	genlmsg_multicast(&drm_ras_nl_family, msg, 0, DRM_RAS_NLGRP_ERROR_REPORT, flags);

[Severity: Medium]
Similarly, genlmsg_multicast() implicitly uses init_net under the hood. Does
this prevent containerized listeners in other network namespaces from receiving
the multicast events?

> +	return 0;
> +
> +cancel_msg:
> +	genlmsg_cancel(msg, hdr);
> +free_msg:
> +	nlmsg_free(msg);
> +	return ret;
> +}
> +EXPORT_SYMBOL(drm_ras_nl_error_event);

[Severity: High]
This isn't a bug introduced by this patch, but does locklessly accessing
drm_ras_node pointers retrieved from drm_ras_xa lead to a use-after-free?

In drivers/gpu/drm/drm_ras.c, get_node_error_counter() retrieves the node:

drivers/gpu/drm/drm_ras.c:get_node_error_counter() {
    ...
    node = xa_load(&drm_ras_xa, node_id);
    if (!node)
        return -ENOENT;
    ...
    return node->query_error_counter(node, error_id, name, value);
}

A similar lockless pattern is present in drm_ras_nl_list_nodes_dumpit():

drivers/gpu/drm/drm_ras.c:drm_ras_nl_list_nodes_dumpit() {
    ...
    xa_for_each_start(&drm_ras_xa, id, node, ctx->restart) {
        ...
        ret = nla_put_string(skb, DRM_RAS_A_NODE_ATTRS_DEVICE_NAME, node->device_name);
    ...
}

If a concurrent device teardown or hot-unplug removes and frees the node
while these functions are executing, will accessing node->query_error_counter
or node->device_name dereference a freed pointer?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260622101716.3313496-5-riana.tauro@intel.com?part=1

Re: [PATCH v3 2/3] drm/xe/xe_drm_ras: Add error-event support for PVC

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
- [High] Potential console flood due to unratelimited `drm_err` in atomic context when netlink multicast fails.
- [Medium] Emitting RAS events with unchanged/zero counter values for uncounted errors.
- [Medium] Pointer arithmetic is performed on a potentially NULL pointer before checking if the DRM RAS feature is enabled.
--

commit 948141f8cea37215cdae8a5d0e089b0c6b181396
Author: Riana Tauro <riana.tauro@intel.com>

drm/xe/xe_drm_ras: Add error-event support for PVC

Report drm_ras error event to userspace when an error occurs.
Add support for core-compute and SoC errors in PVC.

> diff --git a/drivers/gpu/drm/xe/xe_drm_ras.c b/drivers/gpu/drm/xe/xe_drm_ras.c
> index 7937d8ba0ed9c..36afdfb5e4120 100644
> --- a/drivers/gpu/drm/xe/xe_drm_ras.c
> +++ b/drivers/gpu/drm/xe/xe_drm_ras.c

[ ... ]

> @@ -185,6 +185,36 @@ static int register_nodes(struct xe_device *xe)
>  	return ret;
>  }
>  
> +/**
> + * xe_drm_ras_event() - Report drm-ras error event to userspace
> + * @xe: xe device structure
> + * @component: error component (see &enum drm_xe_ras_error_component)
> + * @severity: error severity (see &enum drm_xe_ras_error_severity)
> + * @value: value of error counter
> + * @flags: flags for allocation
> + *
> + * Report an error-event to userspace.
> + */
> +void xe_drm_ras_event(struct xe_device *xe, u32 component, u32 severity, u32 value, gfp_t flags)
> +{
> +	struct xe_drm_ras *ras = &xe->ras;
> +	struct xe_drm_ras_counter *info = ras->info[severity];
> +	struct drm_ras_node *node = &ras->node[severity];

[Severity: Medium]
Is it safe to compute the node pointer here before checking if the DRM
RAS feature is enabled?

If xe->info.has_drm_ras is false, xe_drm_ras_init() might not have been
called, leaving ras->node as NULL. Evaluating &ras->node[severity] on a
NULL pointer results in undefined behavior which could be flagged by UBSAN
or compiler optimizations.

> +	int ret;
> +
> +	/* Event is supported only if drm_ras is enabled */
> +	if (!xe->info.has_drm_ras)
> +		return;
> +
> +	if (!info || !info[component].name)
> +		return;
> +
> +	ret = drm_ras_nl_error_event(node, component, info[component].name, value, flags);
> +	if (ret)
> +		drm_err(&xe->drm, "RAS error-event failed: %d for %s %s\n", ret,
> +			info[component].name, error_severity[severity]);

[Severity: High]
Could this unratelimited drm_err() cause a console flood?

Since xe_drm_ras_event() is called from hw_error_source_handler() in
hardware error interrupt context, an interrupt storm could occur.
If drm_ras_nl_error_event() fails (e.g., due to GFP_ATOMIC allocation
failures under memory pressure), this will rapidly print errors to
the console.

Should this use drm_err_ratelimited() similar to how the caller
handles its logging?

> +}
> +
>  /**
>   * xe_drm_ras_init() - Initialize DRM RAS
>   * @xe: xe device instance

> diff --git a/drivers/gpu/drm/xe/xe_hw_error.c b/drivers/gpu/drm/xe/xe_hw_error.c
> index 4a4b363fc844b..a833cecc74ec5 100644
> --- a/drivers/gpu/drm/xe/xe_hw_error.c
> +++ b/drivers/gpu/drm/xe/xe_hw_error.c

[ ... ]

> @@ -495,6 +495,9 @@ static void hw_error_source_handler(struct xe_tile *tile, const enum hardware_er
>  			gt_hw_error_handler(tile, hw_err, error_id);
>  		if (err_bit == XE_SOC_ERROR)
>  			soc_hw_error_handler(tile, hw_err, error_id);
> +
> +		value = atomic_read(&info[error_id].counter);
> +		xe_drm_ras_event(xe, error_id, severity, value, GFP_ATOMIC);

[Severity: Medium]
Will this emit an event with an unchanged or zero counter value for
correctable SOC errors?

In soc_hw_error_handler(), if the error is HARDWARE_ERROR_CORRECTABLE,
it clears the registers and jumps to unmask_gsysevtctl without
incrementing the error counter:

soc_hw_error_handler() {
    ...
    if (hw_err == HARDWARE_ERROR_CORRECTABLE) {
        ...
        goto unmask_gsysevtctl;
    }
    ...
}

However, here in hw_error_source_handler(), xe_drm_ras_event() is
called unconditionally after soc_hw_error_handler() returns.
Does this result in userspace receiving an error event where the
counter value hasn't actually changed?

>  	}
>  
>  clear_reg:

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260622101716.3313496-5-riana.tauro@intel.com?part=2

Re: [PATCH v3 3/3] drm/xe/xe_ras: Add error-event support for CRI

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] ras_send_error_event() passes internal hardware-specific XE enums to xe_ras_get_counter() instead of the expected DRM enums, resulting in queries for the wrong error counters.
--

commit dc2a656220167a5ce967cead702288ba6f6f06e4
Author: Riana Tauro <riana.tauro@intel.com>

drm/xe/xe_ras: Add error-event support for CRI

This commit adds error-event support for correctable errors in CRI,
reporting errors that cross a threshold to userspace via interrupts.

> diff --git a/drivers/gpu/drm/xe/xe_ras.c b/drivers/gpu/drm/xe/xe_ras.c
> --- a/drivers/gpu/drm/xe/xe_ras.c
> +++ b/drivers/gpu/drm/xe/xe_ras.c

[ ... ]

> @@ -131,6 +161,27 @@ static inline const char *comp_to_str(u8 component)
>  	return xe_ras_components[component];
>  }
>  
> +static void ras_send_error_event(struct xe_device *xe, u8 severity, u8 component)
> +{
> +	u8 drm_severity, drm_component;
> +	u32 value;
> +	int ret;
> +
> +	drm_severity = xe_to_drm_ras_severity(severity);
> +	if (drm_severity == DRM_XE_RAS_ERR_SEV_MAX)
> +		return;
> +
> +	drm_component = xe_to_drm_ras_component(component);
> +	if (drm_component == DRM_XE_RAS_ERR_COMP_MAX)
> +		return;
> +
> +	ret = xe_ras_get_counter(xe, severity, component, &value);

[Severity: Medium]
Does this pass the wrong enum type to xe_ras_get_counter?

In this function, severity and component are the raw XE hardware enums. We
convert them to DRM enums just above as drm_severity and drm_component.

Looking at xe_ras_get_counter(), it expects the DRM enums because it converts
them back to XE enums internally:

int xe_ras_get_counter(struct xe_device *xe, u8 severity, u8 component, u32 *value)
{
        struct xe_ras_error_class counter = {0};

        counter.common.severity = drm_to_xe_ras_severity(severity);
        counter.common.component = drm_to_xe_ras_component(component);
        ...
}

Should this be passing drm_severity and drm_component instead to ensure the
correct telemetry counters are read?

> +	if (ret)
> +		return;
> +
> +	xe_drm_ras_event(xe, drm_component, drm_severity, value, GFP_KERNEL);
> +}
> +

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260622101716.3313496-5-riana.tauro@intel.com?part=3

Re: [PATCH v3 3/3] drm/xe/xe_ras: Add error-event support for CRI

[Raag Jadav]

On Mon, Jun 22, 2026 at 03:47:20PM +0530, Riana Tauro wrote:
> Add error-event support for Correctable errors in CRI.
> error-event is reported to  userspace for all errors that crossed
> threshold on receiving an interrupt for correctable errors.

...

> +static void ras_send_error_event(struct xe_device *xe, u8 severity, u8 component)
> +{
> +	u8 drm_severity, drm_component;
> +	u32 value;
> +	int ret;
> +
> +	drm_severity = xe_to_drm_ras_severity(severity);
> +	if (drm_severity == DRM_XE_RAS_ERR_SEV_MAX)
> +		return;
> +
> +	drm_component = xe_to_drm_ras_component(component);
> +	if (drm_component == DRM_XE_RAS_ERR_COMP_MAX)
> +		return;
> +
> +	ret = xe_ras_get_counter(xe, severity, component, &value);
> +	if (ret)
> +		return;
> +
> +	xe_drm_ras_event(xe, drm_component, drm_severity, value, GFP_KERNEL);
> +}
> +
>  void xe_ras_counter_threshold_crossed(struct xe_device *xe,
>  				      struct xe_sysctrl_event_response *response)
>  {
> @@ -152,6 +203,8 @@ void xe_ras_counter_threshold_crossed(struct xe_device *xe,
>  		severity = errors[id].common.severity;
>  		component = errors[id].common.component;
>  
> +		ras_send_error_event(xe, severity, component);

We already have error-value as part of the event, so this looks like
a lot of redundant duplicate events if threshold is set high enough.
I'm wondering if we use a local variable as a bitmap for components
(which we set here) and only send event for_each_set_bit() after the
loop?

Raag

>  		xe_warn(xe, "[RAS]: %s %s detected\n",
>  			comp_to_str(component), sev_to_str(severity));
>  	}
> -- 
> 2.47.1
> 

Re: [PATCH v3 3/3] drm/xe/xe_ras: Add error-event support for CRI

[Tauro, Riana]

On 30-06-2026 10:50, Raag Jadav wrote:
> On Mon, Jun 22, 2026 at 03:47:20PM +0530, Riana Tauro wrote:
>> Add error-event support for Correctable errors in CRI.
>> error-event is reported to  userspace for all errors that crossed
>> threshold on receiving an interrupt for correctable errors.
> ...
>
>> +static void ras_send_error_event(struct xe_device *xe, u8 severity, u8 component)
>> +{
>> +	u8 drm_severity, drm_component;
>> +	u32 value;
>> +	int ret;
>> +
>> +	drm_severity = xe_to_drm_ras_severity(severity);
>> +	if (drm_severity == DRM_XE_RAS_ERR_SEV_MAX)
>> +		return;
>> +
>> +	drm_component = xe_to_drm_ras_component(component);
>> +	if (drm_component == DRM_XE_RAS_ERR_COMP_MAX)
>> +		return;
>> +
>> +	ret = xe_ras_get_counter(xe, severity, component, &value);
>> +	if (ret)
>> +		return;
>> +
>> +	xe_drm_ras_event(xe, drm_component, drm_severity, value, GFP_KERNEL);
>> +}
>> +
>>   void xe_ras_counter_threshold_crossed(struct xe_device *xe,
>>   				      struct xe_sysctrl_event_response *response)
>>   {
>> @@ -152,6 +203,8 @@ void xe_ras_counter_threshold_crossed(struct xe_device *xe,
>>   		severity = errors[id].common.severity;
>>   		component = errors[id].common.component;
>>   
>> +		ras_send_error_event(xe, severity, component);
> We already have error-value as part of the event, so this looks like
> a lot of redundant duplicate events if threshold is set high enough.

Yeah makes sense based on testing. Will add it once per error type.

Thanks
Riana

> I'm wondering if we use a local variable as a bitmap for components
> (which we set here) and only send event for_each_set_bit() after the
> loop?
>
> Raag
>
>>   		xe_warn(xe, "[RAS]: %s %s detected\n",
>>   			comp_to_str(component), sev_to_str(severity));
>>   	}
>> -- 
>> 2.47.1
>>