[PATCH 00/13] rasdaemon: cxl: Update CXL event logging and recording to CXL spec rev 3.1

[PATCH 00/13] rasdaemon: cxl: Update CXL event logging and recording to CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

1. Update CXL event logging and recording for CXL spec rev 3.1 and for the
following and corresponding kernel CXL trace events changes.
https://lore.kernel.org/lkml/20241120093745.1847-1-shiju.jose@huawei.com/

2. Add following fixes.
 - Fix logging of memory event type field of DRAM trace event.
 - Fix mismatch in 'region' field's name with that in kernel DRAM trace
   event.

Shiju Jose (13):
  rasdaemon: cxl: Fix logging of memory event type of DRAM trace event
  rasdaemon: cxl: Fix mismatch in region field's name with kernel DRAM
    trace event
  rasdaemon: cxl: Add automatic indexing for storing CXL fields in
    SQLite database
  rasdaemon: cxl: Update common event to CXL spec rev 3.1
  rasdaemon: cxl: Add Component Identifier formatting for CXL spec rev
    3.1
  rasdaemon: cxl: Update CXL general media event to CXL spec rev 3.1
  rasdaemon: cxl: Update CXL DRAM event to CXL spec rev 3.1
  rasdaemon: cxl: Update memory module event to CXL spec rev 3.1
  rasdaemon: ras-mc-ctl: Fix logging of memory event type in CXL DRAM
    error table
  rasdaemon: ras-mc-ctl: Update logging of common event data to align
    with CXL spec rev 3.1
  rasdaemon: ras-mc-ctl: Update logging of CXL general media event data
    to align with CXL spec rev 3.1
  rasdaemon: ras-mc-ctl: Update logging of CXL DRAM event data to align
    with CXL spec rev 3.1
  rasdaemon: ras-mc-ctl: Update logging of CXL memory module data to
    align with CXL spec rev 3.1

 ras-cxl-handler.c  | 262 ++++++++++++++++++++++++++++++++++++++++++---
 ras-record.c       | 181 ++++++++++++++++++++-----------
 ras-record.h       |  21 ++++
 ras-report.c       |  30 ++++--
 util/ras-mc-ctl.in | 194 ++++++++++++++++++++++++++++-----
 5 files changed, 583 insertions(+), 105 deletions(-)

-- 
2.43.0

[PATCH 01/13] rasdaemon: cxl: Fix logging of memory event type of DRAM trace event

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec rev 3.0 section 8.2.9.2.1.2 defines the DRAM Event Record.

Fix logging of memory event type field of DRAM trace event.
For e.g. if value is 0x1 it will be reported as an Invalid Address
(General Media Event Record - Memory Event Type) instead of Scrub Media
ECC Error (DRAM Event Record - Memory Event Type) and so on.

Fixes: 9a2f6186db26 ("rasdaemon: Add support for the CXL dram events")
Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index 67ce32e..8733b2b 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -898,6 +898,13 @@ int ras_cxl_general_media_event_handler(struct trace_seq *s,
 #define CXL_DER_VALID_COLUMN			BIT(6)
 #define CXL_DER_VALID_CORRECTION_MASK		BIT(7)
 
+static const char * const cxl_der_mem_event_type[] = {
+	"Media ECC Error",
+	"Scrub Media ECC Error",
+	"Invalid Address",
+	"Data Path Error",
+};
+
 int ras_cxl_dram_event_handler(struct trace_seq *s,
 			       struct tep_record *record,
 			       struct tep_event *event, void *context)
@@ -938,9 +945,9 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
 	if (tep_get_field_val(s,  event, "type", record, &val, 1) < 0)
 		return -1;
 	ev.type = val;
-	if (trace_seq_printf(s, "type:%s ",
-			     get_cxl_type_str(cxl_gmer_mem_event_type,
-					      ARRAY_SIZE(cxl_gmer_mem_event_type),
+	if (trace_seq_printf(s, "memory_event_type:%s ",
+			     get_cxl_type_str(cxl_der_mem_event_type,
+					      ARRAY_SIZE(cxl_der_mem_event_type),
 					      ev.type)) <= 0)
 		return -1;
 
-- 
2.43.0

Re: [PATCH 01/13] rasdaemon: cxl: Fix logging of memory event type of DRAM trace event

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:11 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec rev 3.0 section 8.2.9.2.1.2 defines the DRAM Event Record.
> 
> Fix logging of memory event type field of DRAM trace event.
> For e.g. if value is 0x1 it will be reported as an Invalid Address
> (General Media Event Record - Memory Event Type) instead of Scrub Media
> ECC Error (DRAM Event Record - Memory Event Type) and so on.
> 
> Fixes: 9a2f6186db26 ("rasdaemon: Add support for the CXL dram events")
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 02/13] rasdaemon: cxl: Fix mismatch in region field's name with kernel DRAM trace event

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

Fix mismatch in 'region' field's name with kernel DRAM trace event.

Fixes: ea224ad58b37 ("rasdaemon: CXL: Extract, log and record region info from cxl_general_media and cxl_dram events")

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index 8733b2b..7d4fc9f 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -966,7 +966,7 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
 	if (trace_seq_printf(s, "hpa:0x%llx ", (unsigned long long)ev.hpa) <= 0)
 		return -1;
 
-	ev.region = tep_get_field_raw(s, event, "region", record, &len, 1);
+	ev.region = tep_get_field_raw(s, event, "region_name", record, &len, 1);
 	if (!ev.region)
 		return -1;
 	if (trace_seq_printf(s, "region:%s ", ev.region) <= 0)
-- 
2.43.0

Re: [PATCH 02/13] rasdaemon: cxl: Fix mismatch in region field's name with kernel DRAM trace event

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:12 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> Fix mismatch in 'region' field's name with kernel DRAM trace event.
> 
> Fixes: ea224ad58b37 ("rasdaemon: CXL: Extract, log and record region info from cxl_general_media and cxl_dram events")
> 
No line break here (though this is rasdaemon so maybe kernel rules don't apply?)
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

> ---
>  ras-cxl-handler.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
> index 8733b2b..7d4fc9f 100644
> --- a/ras-cxl-handler.c
> +++ b/ras-cxl-handler.c
> @@ -966,7 +966,7 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
>  	if (trace_seq_printf(s, "hpa:0x%llx ", (unsigned long long)ev.hpa) <= 0)
>  		return -1;
>  
> -	ev.region = tep_get_field_raw(s, event, "region", record, &len, 1);
> +	ev.region = tep_get_field_raw(s, event, "region_name", record, &len, 1);
>  	if (!ev.region)
>  		return -1;
>  	if (trace_seq_printf(s, "region:%s ", ev.region) <= 0)

RE: [PATCH 02/13] rasdaemon: cxl: Fix mismatch in region field's name with kernel DRAM trace event

[Shiju Jose]

Hi Jonathan,

Thanks for reviewing. 
>-----Original Message-----
>From: Jonathan Cameron <jonathan.cameron@huawei.com>
>Sent: 21 November 2024 15:13
>To: Shiju Jose <shiju.jose@huawei.com>
>Cc: linux-edac@vger.kernel.org; linux-cxl@vger.kernel.org;
>mchehab@kernel.org; dave.jiang@intel.com; dan.j.williams@intel.com;
>alison.schofield@intel.com; nifan.cxl@gmail.com; vishal.l.verma@intel.com;
>ira.weiny@intel.com; dave@stgolabs.net; linux-kernel@vger.kernel.org;
>Linuxarm <linuxarm@huawei.com>; tanxiaofei <tanxiaofei@huawei.com>;
>Zengtao (B) <prime.zeng@hisilicon.com>
>Subject: Re: [PATCH 02/13] rasdaemon: cxl: Fix mismatch in region field's name
>with kernel DRAM trace event
>
>On Wed, 20 Nov 2024 09:59:12 +0000
><shiju.jose@huawei.com> wrote:
>
>> From: Shiju Jose <shiju.jose@huawei.com>
>>
>> Fix mismatch in 'region' field's name with kernel DRAM trace event.
>>
>> Fixes: ea224ad58b37 ("rasdaemon: CXL: Extract, log and record region
>> info from cxl_general_media and cxl_dram events")
>>
>No line break here (though this is rasdaemon so maybe kernel rules don't
>apply?)
Will fix.

>> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
>
>Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
>
>> ---
>>  ras-cxl-handler.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c index
>> 8733b2b..7d4fc9f 100644
>> --- a/ras-cxl-handler.c
>> +++ b/ras-cxl-handler.c
>> @@ -966,7 +966,7 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
>>  	if (trace_seq_printf(s, "hpa:0x%llx ", (unsigned long long)ev.hpa) <= 0)
>>  		return -1;
>>
>> -	ev.region = tep_get_field_raw(s, event, "region", record, &len, 1);
>> +	ev.region = tep_get_field_raw(s, event, "region_name", record, &len,
>> +1);
>>  	if (!ev.region)
>>  		return -1;
>>  	if (trace_seq_printf(s, "region:%s ", ev.region) <= 0)
Thanks,
Shiju

[PATCH 03/13] rasdaemon: cxl: Add automatic indexing for storing CXL fields in SQLite database

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

When the CXL specification adds new fields to the common header of
CXL event records, manual updates to the indexing are required to
store these CXL fields in the SQLite database. This update introduces
automatic indexing to facilitate the storage of CXL fields in the
SQLite database, eliminating the need for manual update to indexing.

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-record.c | 134 ++++++++++++++++++++++++++++-----------------------
 1 file changed, 74 insertions(+), 60 deletions(-)

diff --git a/ras-record.c b/ras-record.c
index b4a1abd..9e68158 100644
--- a/ras-record.c
+++ b/ras-record.c
@@ -780,23 +780,25 @@ int ras_store_cxl_overflow_event(struct ras_events *ras, struct ras_cxl_overflow
 
 static int ras_store_cxl_common_hdr(sqlite3_stmt *stmt, struct ras_cxl_event_common_hdr *hdr)
 {
+	int idx = 1;
+
 	if (!stmt || !hdr)
 		return 0;
 
-	sqlite3_bind_text(stmt, 1, hdr->timestamp, -1, NULL);
-	sqlite3_bind_text(stmt, 2, hdr->memdev, -1, NULL);
-	sqlite3_bind_text(stmt, 3, hdr->host, -1, NULL);
-	sqlite3_bind_int64(stmt, 4, hdr->serial);
-	sqlite3_bind_text(stmt, 5, hdr->log_type, -1, NULL);
-	sqlite3_bind_text(stmt, 6, hdr->hdr_uuid, -1, NULL);
-	sqlite3_bind_int(stmt, 7, hdr->hdr_flags);
-	sqlite3_bind_int(stmt, 8, hdr->hdr_handle);
-	sqlite3_bind_int(stmt, 9, hdr->hdr_related_handle);
-	sqlite3_bind_text(stmt, 10, hdr->hdr_timestamp, -1, NULL);
-	sqlite3_bind_int(stmt, 11, hdr->hdr_length);
-	sqlite3_bind_int(stmt, 12, hdr->hdr_maint_op_class);
-
-	return 0;
+	sqlite3_bind_text(stmt, idx++, hdr->timestamp, -1, NULL);
+	sqlite3_bind_text(stmt, idx++, hdr->memdev, -1, NULL);
+	sqlite3_bind_text(stmt, idx++, hdr->host, -1, NULL);
+	sqlite3_bind_int64(stmt, idx++, hdr->serial);
+	sqlite3_bind_text(stmt, idx++, hdr->log_type, -1, NULL);
+	sqlite3_bind_text(stmt, idx++, hdr->hdr_uuid, -1, NULL);
+	sqlite3_bind_int(stmt, idx++, hdr->hdr_flags);
+	sqlite3_bind_int(stmt, idx++, hdr->hdr_handle);
+	sqlite3_bind_int(stmt, idx++, hdr->hdr_related_handle);
+	sqlite3_bind_text(stmt, idx++, hdr->hdr_timestamp, -1, NULL);
+	sqlite3_bind_int(stmt, idx++, hdr->hdr_length);
+	sqlite3_bind_int(stmt, idx++, hdr->hdr_maint_op_class);
+
+	return idx;
 }
 
 /*
@@ -827,15 +829,18 @@ static const struct db_table_descriptor cxl_generic_event_tab = {
 
 int ras_store_cxl_generic_event(struct ras_events *ras, struct ras_cxl_generic_event *ev)
 {
-	int rc;
+	int rc, idx;
 	struct sqlite3_priv *priv = ras->db_priv;
 
 	if (!priv || !priv->stmt_cxl_generic_event)
 		return 0;
 	log(TERM, LOG_INFO, "cxl_generic_event store: %p\n", priv->stmt_cxl_generic_event);
 
-	ras_store_cxl_common_hdr(priv->stmt_cxl_generic_event, &ev->hdr);
-	sqlite3_bind_blob(priv->stmt_cxl_generic_event, 13, ev->data,
+	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_generic_event, &ev->hdr);
+	if (!idx)
+		return 0;
+
+	sqlite3_bind_blob(priv->stmt_cxl_generic_event, idx++, ev->data,
 			  CXL_EVENT_RECORD_DATA_LENGTH, NULL);
 
 	rc = sqlite3_step(priv->stmt_cxl_generic_event);
@@ -891,7 +896,7 @@ static const struct db_table_descriptor cxl_general_media_event_tab = {
 int ras_store_cxl_general_media_event(struct ras_events *ras,
 				      struct ras_cxl_general_media_event *ev)
 {
-	int rc;
+	int rc, idx;
 	struct sqlite3_priv *priv = ras->db_priv;
 
 	if (!priv || !priv->stmt_cxl_general_media_event)
@@ -899,20 +904,23 @@ int ras_store_cxl_general_media_event(struct ras_events *ras,
 	log(TERM, LOG_INFO, "cxl_general_media_event store: %p\n",
 	    priv->stmt_cxl_general_media_event);
 
-	ras_store_cxl_common_hdr(priv->stmt_cxl_general_media_event, &ev->hdr);
-	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, 13, ev->dpa);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 14, ev->dpa_flags);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 15, ev->descriptor);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 16, ev->type);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 17, ev->transaction_type);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 18, ev->channel);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 19, ev->rank);
-	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 20, ev->device);
-	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, 21, ev->comp_id,
+	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_general_media_event, &ev->hdr);
+	if (!idx)
+		return 0;
+
+	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev->dpa);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->dpa_flags);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->descriptor);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->type);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->transaction_type);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->channel);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->rank);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->device);
+	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, idx++, ev->comp_id,
 			  CXL_EVENT_GEN_MED_COMP_ID_SIZE, NULL);
-	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, 22, ev->hpa);
-	sqlite3_bind_text(priv->stmt_cxl_general_media_event, 23, ev->region, -1, NULL);
-	sqlite3_bind_text(priv->stmt_cxl_general_media_event, 24, ev->region_uuid, -1, NULL);
+	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev->hpa);
+	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev->region, -1, NULL);
+	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev->region_uuid, -1, NULL);
 
 	rc = sqlite3_step(priv->stmt_cxl_general_media_event);
 	if (rc != SQLITE_OK && rc != SQLITE_DONE)
@@ -970,7 +978,7 @@ static const struct db_table_descriptor cxl_dram_event_tab = {
 
 int ras_store_cxl_dram_event(struct ras_events *ras, struct ras_cxl_dram_event *ev)
 {
-	int rc;
+	int rc, idx;
 	struct sqlite3_priv *priv = ras->db_priv;
 
 	if (!priv || !priv->stmt_cxl_dram_event)
@@ -978,24 +986,27 @@ int ras_store_cxl_dram_event(struct ras_events *ras, struct ras_cxl_dram_event *
 	log(TERM, LOG_INFO, "cxl_dram_event store: %p\n",
 	    priv->stmt_cxl_dram_event);
 
-	ras_store_cxl_common_hdr(priv->stmt_cxl_dram_event, &ev->hdr);
-	sqlite3_bind_int64(priv->stmt_cxl_dram_event, 13, ev->dpa);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 14, ev->dpa_flags);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 15, ev->descriptor);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 16, ev->type);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 17, ev->transaction_type);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 18, ev->channel);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 19, ev->rank);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 20, ev->nibble_mask);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 21, ev->bank_group);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 22, ev->bank);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 23, ev->row);
-	sqlite3_bind_int(priv->stmt_cxl_dram_event, 24, ev->column);
-	sqlite3_bind_blob(priv->stmt_cxl_dram_event, 25, ev->cor_mask,
+	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_dram_event, &ev->hdr);
+	if (!idx)
+		return 0;
+
+	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->dpa);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->dpa_flags);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->descriptor);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->type);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->transaction_type);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->channel);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->rank);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->nibble_mask);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->bank_group);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->bank);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->row);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->column);
+	sqlite3_bind_blob(priv->stmt_cxl_dram_event, idx++, ev->cor_mask,
 			  CXL_EVENT_DER_CORRECTION_MASK_SIZE, NULL);
-	sqlite3_bind_int64(priv->stmt_cxl_dram_event, 26, ev->hpa);
-	sqlite3_bind_text(priv->stmt_cxl_dram_event, 27, ev->region, -1, NULL);
-	sqlite3_bind_text(priv->stmt_cxl_dram_event, 28, ev->region_uuid, -1, NULL);
+	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->hpa);
+	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region, -1, NULL);
+	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region_uuid, -1, NULL);
 
 	rc = sqlite3_step(priv->stmt_cxl_dram_event);
 	if (rc != SQLITE_OK && rc != SQLITE_DONE)
@@ -1047,7 +1058,7 @@ static const struct db_table_descriptor cxl_memory_module_event_tab = {
 int ras_store_cxl_memory_module_event(struct ras_events *ras,
 				      struct ras_cxl_memory_module_event *ev)
 {
-	int rc;
+	int rc, idx;
 	struct sqlite3_priv *priv = ras->db_priv;
 
 	if (!priv || !priv->stmt_cxl_memory_module_event)
@@ -1055,16 +1066,19 @@ int ras_store_cxl_memory_module_event(struct ras_events *ras,
 	log(TERM, LOG_INFO, "cxl_memory_module_event store: %p\n",
 	    priv->stmt_cxl_memory_module_event);
 
-	ras_store_cxl_common_hdr(priv->stmt_cxl_memory_module_event, &ev->hdr);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 13, ev->event_type);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 14, ev->health_status);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 15, ev->media_status);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 16, ev->life_used);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 17, ev->dirty_shutdown_cnt);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 18, ev->cor_vol_err_cnt);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 19, ev->cor_per_err_cnt);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 20, ev->device_temp);
-	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 21, ev->add_status);
+	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_memory_module_event, &ev->hdr);
+	if (!idx)
+		return 0;
+
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->event_type);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->health_status);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->media_status);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->life_used);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->dirty_shutdown_cnt);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->cor_vol_err_cnt);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->cor_per_err_cnt);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->device_temp);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->add_status);
 
 	rc = sqlite3_step(priv->stmt_cxl_memory_module_event);
 	if (rc != SQLITE_OK && rc != SQLITE_DONE)
-- 
2.43.0

Re: [PATCH 03/13] rasdaemon: cxl: Add automatic indexing for storing CXL fields in SQLite database

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:13 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> When the CXL specification adds new fields to the common header of
> CXL event records, manual updates to the indexing are required to
> store these CXL fields in the SQLite database. This update introduces
> automatic indexing to facilitate the storage of CXL fields in the
> SQLite database, eliminating the need for manual update to indexing.
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>

Using 0 as an error code seems odd, maybe a negative instead?
With that changed to say -1 then this looks good to me.

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>


> ---
>  ras-record.c | 134 ++++++++++++++++++++++++++++-----------------------
>  1 file changed, 74 insertions(+), 60 deletions(-)
> 
> diff --git a/ras-record.c b/ras-record.c
> index b4a1abd..9e68158 100644
> --- a/ras-record.c
> +++ b/ras-record.c
> @@ -780,23 +780,25 @@ int ras_store_cxl_overflow_event(struct ras_events *ras, struct ras_cxl_overflow
>  
>  static int ras_store_cxl_common_hdr(sqlite3_stmt *stmt, struct ras_cxl_event_common_hdr *hdr)
>  {
> +	int idx = 1;
> +
>  	if (!stmt || !hdr)
>  		return 0;
>  
> -	sqlite3_bind_text(stmt, 1, hdr->timestamp, -1, NULL);
> -	sqlite3_bind_text(stmt, 2, hdr->memdev, -1, NULL);
> -	sqlite3_bind_text(stmt, 3, hdr->host, -1, NULL);
> -	sqlite3_bind_int64(stmt, 4, hdr->serial);
> -	sqlite3_bind_text(stmt, 5, hdr->log_type, -1, NULL);
> -	sqlite3_bind_text(stmt, 6, hdr->hdr_uuid, -1, NULL);
> -	sqlite3_bind_int(stmt, 7, hdr->hdr_flags);
> -	sqlite3_bind_int(stmt, 8, hdr->hdr_handle);
> -	sqlite3_bind_int(stmt, 9, hdr->hdr_related_handle);
> -	sqlite3_bind_text(stmt, 10, hdr->hdr_timestamp, -1, NULL);
> -	sqlite3_bind_int(stmt, 11, hdr->hdr_length);
> -	sqlite3_bind_int(stmt, 12, hdr->hdr_maint_op_class);
> -
> -	return 0;
> +	sqlite3_bind_text(stmt, idx++, hdr->timestamp, -1, NULL);
> +	sqlite3_bind_text(stmt, idx++, hdr->memdev, -1, NULL);
> +	sqlite3_bind_text(stmt, idx++, hdr->host, -1, NULL);
> +	sqlite3_bind_int64(stmt, idx++, hdr->serial);
> +	sqlite3_bind_text(stmt, idx++, hdr->log_type, -1, NULL);
> +	sqlite3_bind_text(stmt, idx++, hdr->hdr_uuid, -1, NULL);
> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_flags);
> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_handle);
> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_related_handle);
> +	sqlite3_bind_text(stmt, idx++, hdr->hdr_timestamp, -1, NULL);
> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_length);
> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_maint_op_class);
> +
> +	return idx;
>  }
>  
>  /*
> @@ -827,15 +829,18 @@ static const struct db_table_descriptor cxl_generic_event_tab = {
>  
>  int ras_store_cxl_generic_event(struct ras_events *ras, struct ras_cxl_generic_event *ev)
>  {
> -	int rc;
> +	int rc, idx;
>  	struct sqlite3_priv *priv = ras->db_priv;
>  
>  	if (!priv || !priv->stmt_cxl_generic_event)
>  		return 0;
>  	log(TERM, LOG_INFO, "cxl_generic_event store: %p\n", priv->stmt_cxl_generic_event);
>  
> -	ras_store_cxl_common_hdr(priv->stmt_cxl_generic_event, &ev->hdr);
> -	sqlite3_bind_blob(priv->stmt_cxl_generic_event, 13, ev->data,
> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_generic_event, &ev->hdr);
> +	if (!idx)
> +		return 0;
Seems like an odd form of error code as without looking at implementation it
might seem reasonable for that call to return 0 because it didn't add anythin
to idx?

> +
> +	sqlite3_bind_blob(priv->stmt_cxl_generic_event, idx++, ev->data,
>  			  CXL_EVENT_RECORD_DATA_LENGTH, NULL);
>  
>  	rc = sqlite3_step(priv->stmt_cxl_generic_event);
> @@ -891,7 +896,7 @@ static const struct db_table_descriptor cxl_general_media_event_tab = {
>  int ras_store_cxl_general_media_event(struct ras_events *ras,
>  				      struct ras_cxl_general_media_event *ev)
>  {
> -	int rc;
> +	int rc, idx;
>  	struct sqlite3_priv *priv = ras->db_priv;
>  
>  	if (!priv || !priv->stmt_cxl_general_media_event)
> @@ -899,20 +904,23 @@ int ras_store_cxl_general_media_event(struct ras_events *ras,
>  	log(TERM, LOG_INFO, "cxl_general_media_event store: %p\n",
>  	    priv->stmt_cxl_general_media_event);
>  
> -	ras_store_cxl_common_hdr(priv->stmt_cxl_general_media_event, &ev->hdr);
> -	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, 13, ev->dpa);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 14, ev->dpa_flags);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 15, ev->descriptor);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 16, ev->type);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 17, ev->transaction_type);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 18, ev->channel);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 19, ev->rank);
> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 20, ev->device);
> -	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, 21, ev->comp_id,
> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_general_media_event, &ev->hdr);
> +	if (!idx)
As above,
> +		return 0;
> +
> +	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev->dpa);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->dpa_flags);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->descriptor);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->type);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->transaction_type);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->channel);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->rank);
> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->device);
> +	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, idx++, ev->comp_id,
>  			  CXL_EVENT_GEN_MED_COMP_ID_SIZE, NULL);
> -	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, 22, ev->hpa);
> -	sqlite3_bind_text(priv->stmt_cxl_general_media_event, 23, ev->region, -1, NULL);
> -	sqlite3_bind_text(priv->stmt_cxl_general_media_event, 24, ev->region_uuid, -1, NULL);
> +	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev->hpa);
> +	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev->region, -1, NULL);
> +	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev->region_uuid, -1, NULL);
>  
>  	rc = sqlite3_step(priv->stmt_cxl_general_media_event);
>  	if (rc != SQLITE_OK && rc != SQLITE_DONE)
> @@ -970,7 +978,7 @@ static const struct db_table_descriptor cxl_dram_event_tab = {
>  
>  int ras_store_cxl_dram_event(struct ras_events *ras, struct ras_cxl_dram_event *ev)
>  {
> -	int rc;
> +	int rc, idx;
>  	struct sqlite3_priv *priv = ras->db_priv;
>  
>  	if (!priv || !priv->stmt_cxl_dram_event)
> @@ -978,24 +986,27 @@ int ras_store_cxl_dram_event(struct ras_events *ras, struct ras_cxl_dram_event *
>  	log(TERM, LOG_INFO, "cxl_dram_event store: %p\n",
>  	    priv->stmt_cxl_dram_event);
>  
> -	ras_store_cxl_common_hdr(priv->stmt_cxl_dram_event, &ev->hdr);
> -	sqlite3_bind_int64(priv->stmt_cxl_dram_event, 13, ev->dpa);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 14, ev->dpa_flags);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 15, ev->descriptor);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 16, ev->type);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 17, ev->transaction_type);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 18, ev->channel);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 19, ev->rank);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 20, ev->nibble_mask);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 21, ev->bank_group);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 22, ev->bank);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 23, ev->row);
> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 24, ev->column);
> -	sqlite3_bind_blob(priv->stmt_cxl_dram_event, 25, ev->cor_mask,
> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_dram_event, &ev->hdr);
> +	if (!idx)
As above.

> +		return 0;
> +
> +	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->dpa);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->dpa_flags);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->descriptor);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->type);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->transaction_type);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->channel);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->rank);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->nibble_mask);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->bank_group);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->bank);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->row);
> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->column);
> +	sqlite3_bind_blob(priv->stmt_cxl_dram_event, idx++, ev->cor_mask,
>  			  CXL_EVENT_DER_CORRECTION_MASK_SIZE, NULL);
> -	sqlite3_bind_int64(priv->stmt_cxl_dram_event, 26, ev->hpa);
> -	sqlite3_bind_text(priv->stmt_cxl_dram_event, 27, ev->region, -1, NULL);
> -	sqlite3_bind_text(priv->stmt_cxl_dram_event, 28, ev->region_uuid, -1, NULL);
> +	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->hpa);
> +	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region, -1, NULL);
> +	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region_uuid, -1, NULL);
>  
>  	rc = sqlite3_step(priv->stmt_cxl_dram_event);
>  	if (rc != SQLITE_OK && rc != SQLITE_DONE)
> @@ -1047,7 +1058,7 @@ static const struct db_table_descriptor cxl_memory_module_event_tab = {
>  int ras_store_cxl_memory_module_event(struct ras_events *ras,
>  				      struct ras_cxl_memory_module_event *ev)
>  {
> -	int rc;
> +	int rc, idx;
>  	struct sqlite3_priv *priv = ras->db_priv;
>  
>  	if (!priv || !priv->stmt_cxl_memory_module_event)
> @@ -1055,16 +1066,19 @@ int ras_store_cxl_memory_module_event(struct ras_events *ras,
>  	log(TERM, LOG_INFO, "cxl_memory_module_event store: %p\n",
>  	    priv->stmt_cxl_memory_module_event);
>  
> -	ras_store_cxl_common_hdr(priv->stmt_cxl_memory_module_event, &ev->hdr);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 13, ev->event_type);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 14, ev->health_status);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 15, ev->media_status);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 16, ev->life_used);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 17, ev->dirty_shutdown_cnt);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 18, ev->cor_vol_err_cnt);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 19, ev->cor_per_err_cnt);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 20, ev->device_temp);
> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 21, ev->add_status);
> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_memory_module_event, &ev->hdr);
> +	if (!idx)
as above
> +		return 0;
> +
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->event_type);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->health_status);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->media_status);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->life_used);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->dirty_shutdown_cnt);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->cor_vol_err_cnt);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->cor_per_err_cnt);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->device_temp);
> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->add_status);
>  
>  	rc = sqlite3_step(priv->stmt_cxl_memory_module_event);
>  	if (rc != SQLITE_OK && rc != SQLITE_DONE)

RE: [PATCH 03/13] rasdaemon: cxl: Add automatic indexing for storing CXL fields in SQLite database

[Shiju Jose]

>-----Original Message-----
>From: Jonathan Cameron <jonathan.cameron@huawei.com>
>Sent: 21 November 2024 15:17
>To: Shiju Jose <shiju.jose@huawei.com>
>Cc: linux-edac@vger.kernel.org; linux-cxl@vger.kernel.org;
>mchehab@kernel.org; dave.jiang@intel.com; dan.j.williams@intel.com;
>alison.schofield@intel.com; nifan.cxl@gmail.com; vishal.l.verma@intel.com;
>ira.weiny@intel.com; dave@stgolabs.net; linux-kernel@vger.kernel.org;
>Linuxarm <linuxarm@huawei.com>; tanxiaofei <tanxiaofei@huawei.com>;
>Zengtao (B) <prime.zeng@hisilicon.com>
>Subject: Re: [PATCH 03/13] rasdaemon: cxl: Add automatic indexing for storing
>CXL fields in SQLite database
>
>On Wed, 20 Nov 2024 09:59:13 +0000
><shiju.jose@huawei.com> wrote:
>
>> From: Shiju Jose <shiju.jose@huawei.com>
>>
>> When the CXL specification adds new fields to the common header of CXL
>> event records, manual updates to the indexing are required to store
>> these CXL fields in the SQLite database. This update introduces
>> automatic indexing to facilitate the storage of CXL fields in the
>> SQLite database, eliminating the need for manual update to indexing.
>>
>> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
>
>Using 0 as an error code seems odd, maybe a negative instead?
>With that changed to say -1 then this looks good to me.

Will change error code to -1

Thanks,
Shiju
>
>Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
>
>
>> ---
>>  ras-record.c | 134
>> ++++++++++++++++++++++++++++-----------------------
>>  1 file changed, 74 insertions(+), 60 deletions(-)
>>
>> diff --git a/ras-record.c b/ras-record.c index b4a1abd..9e68158 100644
>> --- a/ras-record.c
>> +++ b/ras-record.c
>> @@ -780,23 +780,25 @@ int ras_store_cxl_overflow_event(struct
>> ras_events *ras, struct ras_cxl_overflow
>>
>>  static int ras_store_cxl_common_hdr(sqlite3_stmt *stmt, struct
>> ras_cxl_event_common_hdr *hdr)  {
>> +	int idx = 1;
>> +
>>  	if (!stmt || !hdr)
>>  		return 0;
>>
>> -	sqlite3_bind_text(stmt, 1, hdr->timestamp, -1, NULL);
>> -	sqlite3_bind_text(stmt, 2, hdr->memdev, -1, NULL);
>> -	sqlite3_bind_text(stmt, 3, hdr->host, -1, NULL);
>> -	sqlite3_bind_int64(stmt, 4, hdr->serial);
>> -	sqlite3_bind_text(stmt, 5, hdr->log_type, -1, NULL);
>> -	sqlite3_bind_text(stmt, 6, hdr->hdr_uuid, -1, NULL);
>> -	sqlite3_bind_int(stmt, 7, hdr->hdr_flags);
>> -	sqlite3_bind_int(stmt, 8, hdr->hdr_handle);
>> -	sqlite3_bind_int(stmt, 9, hdr->hdr_related_handle);
>> -	sqlite3_bind_text(stmt, 10, hdr->hdr_timestamp, -1, NULL);
>> -	sqlite3_bind_int(stmt, 11, hdr->hdr_length);
>> -	sqlite3_bind_int(stmt, 12, hdr->hdr_maint_op_class);
>> -
>> -	return 0;
>> +	sqlite3_bind_text(stmt, idx++, hdr->timestamp, -1, NULL);
>> +	sqlite3_bind_text(stmt, idx++, hdr->memdev, -1, NULL);
>> +	sqlite3_bind_text(stmt, idx++, hdr->host, -1, NULL);
>> +	sqlite3_bind_int64(stmt, idx++, hdr->serial);
>> +	sqlite3_bind_text(stmt, idx++, hdr->log_type, -1, NULL);
>> +	sqlite3_bind_text(stmt, idx++, hdr->hdr_uuid, -1, NULL);
>> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_flags);
>> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_handle);
>> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_related_handle);
>> +	sqlite3_bind_text(stmt, idx++, hdr->hdr_timestamp, -1, NULL);
>> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_length);
>> +	sqlite3_bind_int(stmt, idx++, hdr->hdr_maint_op_class);
>> +
>> +	return idx;
>>  }
>>
>>  /*
>> @@ -827,15 +829,18 @@ static const struct db_table_descriptor
>> cxl_generic_event_tab = {
>>
>>  int ras_store_cxl_generic_event(struct ras_events *ras, struct
>> ras_cxl_generic_event *ev)  {
>> -	int rc;
>> +	int rc, idx;
>>  	struct sqlite3_priv *priv = ras->db_priv;
>>
>>  	if (!priv || !priv->stmt_cxl_generic_event)
>>  		return 0;
>>  	log(TERM, LOG_INFO, "cxl_generic_event store: %p\n",
>> priv->stmt_cxl_generic_event);
>>
>> -	ras_store_cxl_common_hdr(priv->stmt_cxl_generic_event, &ev->hdr);
>> -	sqlite3_bind_blob(priv->stmt_cxl_generic_event, 13, ev->data,
>> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_generic_event, &ev-
>>hdr);
>> +	if (!idx)
>> +		return 0;
>Seems like an odd form of error code as without looking at implementation it
>might seem reasonable for that call to return 0 because it didn't add anythin to
>idx?
>
>> +
>> +	sqlite3_bind_blob(priv->stmt_cxl_generic_event, idx++, ev->data,
>>  			  CXL_EVENT_RECORD_DATA_LENGTH, NULL);
>>
>>  	rc = sqlite3_step(priv->stmt_cxl_generic_event);
>> @@ -891,7 +896,7 @@ static const struct db_table_descriptor
>> cxl_general_media_event_tab = {  int
>ras_store_cxl_general_media_event(struct ras_events *ras,
>>  				      struct ras_cxl_general_media_event *ev)  {
>> -	int rc;
>> +	int rc, idx;
>>  	struct sqlite3_priv *priv = ras->db_priv;
>>
>>  	if (!priv || !priv->stmt_cxl_general_media_event)
>> @@ -899,20 +904,23 @@ int ras_store_cxl_general_media_event(struct
>ras_events *ras,
>>  	log(TERM, LOG_INFO, "cxl_general_media_event store: %p\n",
>>  	    priv->stmt_cxl_general_media_event);
>>
>> -	ras_store_cxl_common_hdr(priv->stmt_cxl_general_media_event, &ev-
>>hdr);
>> -	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, 13, ev->dpa);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 14, ev-
>>dpa_flags);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 15, ev-
>>descriptor);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 16, ev->type);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 17, ev-
>>transaction_type);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 18, ev->channel);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 19, ev->rank);
>> -	sqlite3_bind_int(priv->stmt_cxl_general_media_event, 20, ev->device);
>> -	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, 21, ev-
>>comp_id,
>> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_general_media_event,
>&ev->hdr);
>> +	if (!idx)
>As above,
>> +		return 0;
>> +
>> +	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev-
>>dpa);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev-
>>dpa_flags);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev-
>>descriptor);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->type);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev-
>>transaction_type);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev-
>>channel);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->rank);
>> +	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev-
>>device);
>> +	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, idx++,
>> +ev->comp_id,
>>  			  CXL_EVENT_GEN_MED_COMP_ID_SIZE, NULL);
>> -	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, 22, ev->hpa);
>> -	sqlite3_bind_text(priv->stmt_cxl_general_media_event, 23, ev->region,
>-1, NULL);
>> -	sqlite3_bind_text(priv->stmt_cxl_general_media_event, 24, ev-
>>region_uuid, -1, NULL);
>> +	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev-
>>hpa);
>> +	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev-
>>region, -1, NULL);
>> +	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++,
>> +ev->region_uuid, -1, NULL);
>>
>>  	rc = sqlite3_step(priv->stmt_cxl_general_media_event);
>>  	if (rc != SQLITE_OK && rc != SQLITE_DONE) @@ -970,7 +978,7 @@
>static
>> const struct db_table_descriptor cxl_dram_event_tab = {
>>
>>  int ras_store_cxl_dram_event(struct ras_events *ras, struct
>> ras_cxl_dram_event *ev)  {
>> -	int rc;
>> +	int rc, idx;
>>  	struct sqlite3_priv *priv = ras->db_priv;
>>
>>  	if (!priv || !priv->stmt_cxl_dram_event) @@ -978,24 +986,27 @@ int
>> ras_store_cxl_dram_event(struct ras_events *ras, struct ras_cxl_dram_event
>*
>>  	log(TERM, LOG_INFO, "cxl_dram_event store: %p\n",
>>  	    priv->stmt_cxl_dram_event);
>>
>> -	ras_store_cxl_common_hdr(priv->stmt_cxl_dram_event, &ev->hdr);
>> -	sqlite3_bind_int64(priv->stmt_cxl_dram_event, 13, ev->dpa);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 14, ev->dpa_flags);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 15, ev->descriptor);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 16, ev->type);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 17, ev->transaction_type);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 18, ev->channel);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 19, ev->rank);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 20, ev->nibble_mask);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 21, ev->bank_group);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 22, ev->bank);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 23, ev->row);
>> -	sqlite3_bind_int(priv->stmt_cxl_dram_event, 24, ev->column);
>> -	sqlite3_bind_blob(priv->stmt_cxl_dram_event, 25, ev->cor_mask,
>> +	idx = ras_store_cxl_common_hdr(priv->stmt_cxl_dram_event, &ev-
>>hdr);
>> +	if (!idx)
>As above.
>
>> +		return 0;
>> +
>> +	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->dpa);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->dpa_flags);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->descriptor);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->type);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev-
>>transaction_type);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->channel);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->rank);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->nibble_mask);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->bank_group);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->bank);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->row);
>> +	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->column);
>> +	sqlite3_bind_blob(priv->stmt_cxl_dram_event, idx++, ev->cor_mask,
>>  			  CXL_EVENT_DER_CORRECTION_MASK_SIZE, NULL);
>> -	sqlite3_bind_int64(priv->stmt_cxl_dram_event, 26, ev->hpa);
>> -	sqlite3_bind_text(priv->stmt_cxl_dram_event, 27, ev->region, -1, NULL);
>> -	sqlite3_bind_text(priv->stmt_cxl_dram_event, 28, ev->region_uuid, -1,
>NULL);
>> +	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->hpa);
>> +	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region, -1,
>NULL);
>> +	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region_uuid,
>> +-1, NULL);
>>
>>  	rc = sqlite3_step(priv->stmt_cxl_dram_event);
>>  	if (rc != SQLITE_OK && rc != SQLITE_DONE) @@ -1047,7 +1058,7 @@
>> static const struct db_table_descriptor cxl_memory_module_event_tab =
>> {  int ras_store_cxl_memory_module_event(struct ras_events *ras,
>>  				      struct ras_cxl_memory_module_event *ev)
>{
>> -	int rc;
>> +	int rc, idx;
>>  	struct sqlite3_priv *priv = ras->db_priv;
>>
>>  	if (!priv || !priv->stmt_cxl_memory_module_event)
>> @@ -1055,16 +1066,19 @@ int ras_store_cxl_memory_module_event(struct
>ras_events *ras,
>>  	log(TERM, LOG_INFO, "cxl_memory_module_event store: %p\n",
>>  	    priv->stmt_cxl_memory_module_event);
>>
>> -	ras_store_cxl_common_hdr(priv->stmt_cxl_memory_module_event,
>&ev->hdr);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 13, ev-
>>event_type);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 14, ev-
>>health_status);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 15, ev-
>>media_status);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 16, ev-
>>life_used);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 17, ev-
>>dirty_shutdown_cnt);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 18, ev-
>>cor_vol_err_cnt);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 19, ev-
>>cor_per_err_cnt);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 20, ev-
>>device_temp);
>> -	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, 21, ev-
>>add_status);
>> +	idx = ras_store_cxl_common_hdr(priv-
>>stmt_cxl_memory_module_event, &ev->hdr);
>> +	if (!idx)
>as above
>> +		return 0;
>> +
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>event_type);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>health_status);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>media_status);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>life_used);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>dirty_shutdown_cnt);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>cor_vol_err_cnt);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>cor_per_err_cnt);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev-
>>device_temp);
>> +	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++,
>> +ev->add_status);
>>
>>  	rc = sqlite3_step(priv->stmt_cxl_memory_module_event);
>>  	if (rc != SQLITE_OK && rc != SQLITE_DONE)

[PATCH 04/13] rasdaemon: cxl: Update common event to CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec 3.1 section 8.2.9.2.1 Table 8-42, Common Event Record format has
updated with Maintenance Operation Subclass information.

Add updates in rasdaemon CXL event handler for the above spec change
and for the corresponding changes in kernel CXL common trace event
implementation.

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 10 +++++++++-
 ras-record.c      |  5 +++++
 ras-record.h      |  1 +
 ras-report.c      |  6 ++++--
 4 files changed, 19 insertions(+), 3 deletions(-)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index 7d4fc9f..d16eaef 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -575,18 +575,20 @@ int ras_cxl_overflow_event_handler(struct trace_seq *s,
 
 /*
  * Common Event Record Format
- * CXL 3.0 section 8.2.9.2.1; Table 8-42
+ * CXL 3.1 section 8.2.9.2.1; Table 8-43
  */
 #define CXL_EVENT_RECORD_FLAG_PERMANENT		BIT(2)
 #define CXL_EVENT_RECORD_FLAG_MAINT_NEEDED	BIT(3)
 #define CXL_EVENT_RECORD_FLAG_PERF_DEGRADED	BIT(4)
 #define CXL_EVENT_RECORD_FLAG_HW_REPLACE	BIT(5)
+#define CXL_EVENT_RECORD_FLAG_MAINT_OP_SUB_CLASS_VALID	BIT(6)
 
 static const struct  cxl_event_flags cxl_hdr_flags[] = {
 	{ .bit = CXL_EVENT_RECORD_FLAG_PERMANENT, .flag = "PERMANENT_CONDITION" },
 	{ .bit = CXL_EVENT_RECORD_FLAG_MAINT_NEEDED, .flag = "MAINTENANCE_NEEDED" },
 	{ .bit = CXL_EVENT_RECORD_FLAG_PERF_DEGRADED, .flag = "PERFORMANCE_DEGRADED" },
 	{ .bit = CXL_EVENT_RECORD_FLAG_HW_REPLACE, .flag = "HARDWARE_REPLACEMENT_NEEDED" },
+	{ .bit = CXL_EVENT_RECORD_FLAG_MAINT_OP_SUB_CLASS_VALID, .flag = "MAINT_OP_SUB_CLASS_VALID" },
 };
 
 static int handle_ras_cxl_common_hdr(struct trace_seq *s,
@@ -670,6 +672,12 @@ static int handle_ras_cxl_common_hdr(struct trace_seq *s,
 	if (trace_seq_printf(s, "hdr_maint_op_class:%u ", hdr->hdr_maint_op_class) <= 0)
 		return -1;
 
+	if (tep_get_field_val(s,  event, "hdr_maint_op_sub_class", record, &val, 1) < 0)
+		return -1;
+	hdr->hdr_maint_op_sub_class = val;
+	if (trace_seq_printf(s, "hdr_maint_op_sub_class:%u ", hdr->hdr_maint_op_sub_class) <= 0)
+		return -1;
+
 	return 0;
 }
 
diff --git a/ras-record.c b/ras-record.c
index 9e68158..b52973f 100644
--- a/ras-record.c
+++ b/ras-record.c
@@ -797,6 +797,7 @@ static int ras_store_cxl_common_hdr(sqlite3_stmt *stmt, struct ras_cxl_event_com
 	sqlite3_bind_text(stmt, idx++, hdr->hdr_timestamp, -1, NULL);
 	sqlite3_bind_int(stmt, idx++, hdr->hdr_length);
 	sqlite3_bind_int(stmt, idx++, hdr->hdr_maint_op_class);
+	sqlite3_bind_int(stmt, idx++, hdr->hdr_maint_op_sub_class);
 
 	return idx;
 }
@@ -818,6 +819,7 @@ static const struct db_fields cxl_generic_event_fields[] = {
 	{ .name = "hdr_ts",		.type = "TEXT" },
 	{ .name = "hdr_length",		.type = "INTEGER" },
 	{ .name = "hdr_maint_op_class",	.type = "INTEGER" },
+	{ .name = "hdr_maint_op_sub_class",	.type = "INTEGER" },
 	{ .name = "data",		.type = "BLOB" },
 };
 
@@ -873,6 +875,7 @@ static const struct db_fields cxl_general_media_event_fields[] = {
 	{ .name = "hdr_ts",		.type = "TEXT" },
 	{ .name = "hdr_length",		.type = "INTEGER" },
 	{ .name = "hdr_maint_op_class",	.type = "INTEGER" },
+	{ .name = "hdr_maint_op_sub_class",	.type = "INTEGER" },
 	{ .name = "dpa",		.type = "INTEGER" },
 	{ .name = "dpa_flags",		.type = "INTEGER" },
 	{ .name = "descriptor",		.type = "INTEGER" },
@@ -952,6 +955,7 @@ static const struct db_fields cxl_dram_event_fields[] = {
 	{ .name = "hdr_ts",		.type = "TEXT" },
 	{ .name = "hdr_length",		.type = "INTEGER" },
 	{ .name = "hdr_maint_op_class",	.type = "INTEGER" },
+	{ .name = "hdr_maint_op_sub_class",	.type = "INTEGER" },
 	{ .name = "dpa",		.type = "INTEGER" },
 	{ .name = "dpa_flags",		.type = "INTEGER" },
 	{ .name = "descriptor",		.type = "INTEGER" },
@@ -1038,6 +1042,7 @@ static const struct db_fields cxl_memory_module_event_fields[] = {
 	{ .name = "hdr_ts",		.type = "TEXT" },
 	{ .name = "hdr_length",		.type = "INTEGER" },
 	{ .name = "hdr_maint_op_class",	.type = "INTEGER" },
+	{ .name = "hdr_maint_op_sub_class",	.type = "INTEGER" },
 	{ .name = "event_type",		.type = "INTEGER" },
 	{ .name = "health_status",	.type = "INTEGER" },
 	{ .name = "media_status",	.type = "INTEGER" },
diff --git a/ras-record.h b/ras-record.h
index cd549a1..2a0124a 100644
--- a/ras-record.h
+++ b/ras-record.h
@@ -179,6 +179,7 @@ struct ras_cxl_event_common_hdr {
 	char hdr_timestamp[64];
 	uint8_t hdr_length;
 	uint8_t hdr_maint_op_class;
+	uint8_t hdr_maint_op_sub_class;
 };
 
 struct ras_cxl_generic_event {
diff --git a/ras-report.c b/ras-report.c
index bc77d04..7e974b9 100644
--- a/ras-report.c
+++ b/ras-report.c
@@ -507,7 +507,8 @@ static int set_cxl_generic_event_backtrace(char *buf, struct ras_cxl_generic_eve
 		"hdr_related_handle=0x%x\n"
 		"hdr_timestamp=%s\n"
 		"hdr_length=%u\n"
-		"hdr_maint_op_class=%u\n",
+		"hdr_maint_op_class=%u\n"
+		"hdr_maint_op_sub_class=%u\n",
 		ev->hdr.timestamp,
 		ev->hdr.memdev,
 		ev->hdr.host,
@@ -519,7 +520,8 @@ static int set_cxl_generic_event_backtrace(char *buf, struct ras_cxl_generic_eve
 		ev->hdr.hdr_related_handle,
 		ev->hdr.hdr_timestamp,
 		ev->hdr.hdr_length,
-		ev->hdr.hdr_maint_op_class);
+		ev->hdr.hdr_maint_op_class,
+		ev->hdr.hdr_maint_op_sub_class);
 
 	return 0;
 }
-- 
2.43.0

Re: [PATCH 04/13] rasdaemon: cxl: Update common event to CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:14 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec 3.1 section 8.2.9.2.1 Table 8-42, Common Event Record format has
> updated with Maintenance Operation Subclass information.
> 
> Add updates in rasdaemon CXL event handler for the above spec change
> and for the corresponding changes in kernel CXL common trace event
> implementation.
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
LGTM
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 05/13] rasdaemon: cxl: Add Component Identifier formatting for CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

Add Component Identifier formatting for CXL spec rev 3.1, Section
8.2.9.2.1, Table 8-44.

Add helper function to print component ID, parse and log PLDM entity ID
and resource ID.

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 41 +++++++++++++++++++++++++++++++++++++++++
 ras-record.h      |  3 +++
 2 files changed, 44 insertions(+)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index d16eaef..80afa9f 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -573,6 +573,47 @@ int ras_cxl_overflow_event_handler(struct trace_seq *s,
 	return 0;
 }
 
+/*
+ * Component ID Format
+ * CXL 3.1 section 8.2.9.2.1; Table 8-44
+ */
+#define CXL_PLDM_COMPONENT_ID_ENTITY_VALID	BIT(0)
+#define CXL_PLDM_COMPONENT_ID_RES_VALID		BIT(1)
+static const struct  cxl_event_flags cxl_pldm_comp_id_flags[] = {
+	{ .bit = CXL_PLDM_COMPONENT_ID_ENTITY_VALID, .flag = "PLDM Entity ID" },
+	{ .bit = CXL_PLDM_COMPONENT_ID_RES_VALID, .flag = "Resource ID" },
+};
+
+static int ras_cxl_print_component_id(struct trace_seq *s, uint8_t *comp_id,
+				      uint8_t *entity_id, uint8_t *res_id)
+{
+	int i;
+
+	if (comp_id[0] & CXL_PLDM_COMPONENT_ID_ENTITY_VALID) {
+		if (trace_seq_printf(s, "PLDM Entity ID:") <= 0)
+			return -1;
+		for (i = 1; i < 7; i++) {
+			if (trace_seq_printf(s, "%02x ", comp_id[i]) <= 0)
+				return -1;
+		}
+		if (entity_id)
+			memcpy(entity_id, &comp_id[1], CXL_PLDM_ENTITY_ID_LEN);
+	}
+
+	if (comp_id[0] & CXL_PLDM_COMPONENT_ID_RES_VALID) {
+		if (trace_seq_printf(s, "Resource ID:") <= 0)
+			return -1;
+		for (i = 7; i < 11; i++) {
+			if (trace_seq_printf(s, "%02x ", comp_id[i]) <= 0)
+				return -1;
+		}
+		if (res_id)
+			memcpy(res_id, &comp_id[7], CXL_PLDM_RES_ID_LEN);
+	}
+
+	return 0;
+}
+
 /*
  * Common Event Record Format
  * CXL 3.1 section 8.2.9.2.1; Table 8-43
diff --git a/ras-record.h b/ras-record.h
index 2a0124a..a3a88eb 100644
--- a/ras-record.h
+++ b/ras-record.h
@@ -137,6 +137,9 @@ struct ras_cxl_poison_event {
 #define CXL_EVENT_GEN_MED_COMP_ID_SIZE	0x10
 #define CXL_EVENT_DER_CORRECTION_MASK_SIZE	0x20
 
+#define CXL_PLDM_ENTITY_ID_LEN	6
+#define CXL_PLDM_RES_ID_LEN	4
+
 struct ras_cxl_aer_ue_event {
 	char timestamp[64];
 	const char *memdev;
-- 
2.43.0

Re: [PATCH 05/13] rasdaemon: cxl: Add Component Identifier formatting for CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:15 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> Add Component Identifier formatting for CXL spec rev 3.1, Section
> 8.2.9.2.1, Table 8-44.
> 
> Add helper function to print component ID, parse and log PLDM entity ID
> and resource ID.
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 06/13] rasdaemon: cxl: Update CXL general media event to CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec rev 3.1 section 8.2.9.2.1.1 Table 8-45, General Media Event
Record has updated with following new fields and new types for Memory
Event Type and Transaction Type fields.
1. Advanced Programmable Corrected Memory Error Threshold Event Flags
2. Corrected Memory Error Count at Event
3. Memory Event Sub-Type

The format of component identifier has changed (CXL spec 3.1 section
8.2.9.2.1 Table 8-44).

Update the parsing, logging and recording of general media event for
the above spec changes.

Example rasdaemon log for CXL general media event,

cxl_general_media 2024-11-20 00:00:49 +0000 memdev:mem1 host:0000:0f:00.0 \
serial:0x3 log type:Fatal hdr_uuid:fbcd0a77-c260-417f-85a9-088b1621eba6 \
hdr_handle:0x1 hdr_related_handle:0x0 hdr_timestamp:1970-01-01 00:00:49 +0000 \
hdr_length:128 hdr_maint_op_class:2 hdr_maint_op_sub_class:4 dpa:0x30d40 \
dpa_flags:descriptor:'UNCORRECTABLE EVENT' 'THRESHOLD EVENT' 'POISON LIST OVERFLOW' \
memory_event_type:TE State Violation memory_event_sub_type:Media Link Command \
Training Error transaction_type:Host Inject Poison hpa:0xffffffffffffffff \
region: region_uuid:00000000-0000-0000-0000-000000000000 channel:3 rank:33 \
device:5 comp_id:03 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
comp_id_pldm_valid_flags:'PLDM Entity ID' 'Resource ID' \
PLDM Entity ID:74 c5 08 9a 1a 0b Resource ID:fc d2 7e 2f \
Advanced Programmable CME threshold Event Flags:'Corrected Memory Errors in \
Multiple Media Components' 'Exceeded Programmable Threshold' \
Corrected Memory Error Count:0x78

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 73 +++++++++++++++++++++++++++++++++++++++++++++--
 ras-record.c      | 13 +++++++++
 ras-record.h      |  5 ++++
 ras-report.c      | 10 +++++--
 4 files changed, 96 insertions(+), 5 deletions(-)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index 80afa9f..c64c540 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -772,9 +772,32 @@ static const struct cxl_event_flags cxl_dpa_flags[] = {
 	{ .bit = CXL_DPA_NOT_REPAIRABLE, .flag = "NOT_REPAIRABLE" },
 };
 
+/* CXL rev 3.1 Section 8.2.9.2.1.1; Table 8-45 */
+static const char * const cxl_mem_event_sub_type[] = {
+	"Not Reported",
+	"Internal Datapath Error",
+	"Media Link Command Training Error",
+	"Media Link Control Training Error",
+	"Media Link Data Training Error",
+	"Media Link CRC Error",
+};
+
+#define CXL_CME_EV_FLAG_CME_MULTIPLE_MEDIA	BIT(0)
+#define CXL_CME_EV_FLAG_THRESHOLD_EXCEEDED	BIT(1)
+static const struct cxl_event_flags cxl_cme_threshold_ev_flags[] = {
+	{
+		.bit = CXL_CME_EV_FLAG_CME_MULTIPLE_MEDIA,
+		.flag = "Corrected Memory Errors in Multiple Media Components"
+	},
+	{
+		.bit = CXL_CME_EV_FLAG_THRESHOLD_EXCEEDED,
+		.flag = "Exceeded Programmable Threshold"
+	},
+};
+
 /*
  * General Media Event Record - GMER
- * CXL rev 3.0 Section 8.2.9.2.1.1; Table 8-43
+ * CXL rev 3.1 Section 8.2.9.2.1.1; Table 8-45
  */
 #define CXL_GMER_EVT_DESC_UNCORRECTABLE_EVENT		BIT(0)
 #define CXL_GMER_EVT_DESC_THRESHOLD_EVENT		BIT(1)
@@ -790,11 +813,16 @@ static const struct cxl_event_flags cxl_gmer_event_desc_flags[] = {
 #define CXL_GMER_VALID_RANK			BIT(1)
 #define CXL_GMER_VALID_DEVICE			BIT(2)
 #define CXL_GMER_VALID_COMPONENT		BIT(3)
+#define CXL_GMER_VALID_COMPONENT_ID_FORMAT	BIT(4)
 
 static const char * const cxl_gmer_mem_event_type[] = {
 	"ECC Error",
 	"Invalid Address",
 	"Data Path Error",
+	"TE State Violation",
+	"Scrub Media ECC Error",
+	"Advanced Programmable CME Counter Expiration",
+	"CKID Violation",
 };
 
 static const char * const cxl_gmer_trans_type[] = {
@@ -805,13 +833,15 @@ static const char * const cxl_gmer_trans_type[] = {
 	"Host Inject Poison",
 	"Internal Media Scrub",
 	"Internal Media Management",
+	"Internal Media Error Check Scrub",
+	"Media Initialization",
 };
 
 int ras_cxl_general_media_event_handler(struct trace_seq *s,
 					struct tep_record *record,
 					struct tep_event *event, void *context)
 {
-	int len, i;
+	int len, i, rc;
 	unsigned long long val;
 	struct ras_events *ras = context;
 	struct ras_cxl_general_media_event ev;
@@ -846,11 +876,20 @@ int ras_cxl_general_media_event_handler(struct trace_seq *s,
 	if (tep_get_field_val(s,  event, "type", record, &val, 1) < 0)
 		return -1;
 	ev.type = val;
-	if (trace_seq_printf(s, "type:%s ",
+	if (trace_seq_printf(s, "memory_event_type:%s ",
 			     get_cxl_type_str(cxl_gmer_mem_event_type,
 					      ARRAY_SIZE(cxl_gmer_mem_event_type), ev.type)) <= 0)
 		return -1;
 
+	if (tep_get_field_val(s,  event, "sub_type", record, &val, 1) < 0)
+		return -1;
+	ev.sub_type = val;
+	if (trace_seq_printf(s, "memory_event_sub_type:%s ",
+			     get_cxl_type_str(cxl_mem_event_sub_type,
+					      ARRAY_SIZE(cxl_mem_event_sub_type),
+					      ev.sub_type)) <= 0)
+		return -1;
+
 	if (tep_get_field_val(s,  event, "transaction_type", record, &val, 1) < 0)
 		return -1;
 	ev.transaction_type = val;
@@ -918,8 +957,36 @@ int ras_cxl_general_media_event_handler(struct trace_seq *s,
 			if (trace_seq_printf(s, "%02x ", ev.comp_id[i]) <= 0)
 				break;
 		}
+
+		if (ev.validity_flags & CXL_GMER_VALID_COMPONENT_ID_FORMAT) {
+			if (trace_seq_printf(s, "comp_id_pldm_valid_flags:") <= 0)
+				return -1;
+			if (decode_cxl_event_flags(s, ev.comp_id[0], cxl_pldm_comp_id_flags,
+						   ARRAY_SIZE(cxl_pldm_comp_id_flags)) < 0)
+				return -1;
+
+			rc = ras_cxl_print_component_id(s, ev.comp_id, ev.entity_id, ev.res_id);
+			if (rc)
+				return rc;
+		}
 	}
 
+	if (tep_get_field_val(s,  event, "cme_threshold_ev_flags", record, &val, 1) < 0)
+		return -1;
+	ev.cme_threshold_ev_flags = val;
+	if (trace_seq_printf(s, "Advanced Programmable CME threshold Event Flags:") <= 0)
+		return -1;
+	if (decode_cxl_event_flags(s, ev.cme_threshold_ev_flags,
+				   cxl_cme_threshold_ev_flags,
+				   ARRAY_SIZE(cxl_cme_threshold_ev_flags)) < 0)
+		return -1;
+
+	if (tep_get_field_val(s,  event, "cme_count", record, &val, 1) < 0)
+		return -1;
+	ev.cme_count = val;
+	if (trace_seq_printf(s, "Corrected Memory Error Count:0x%x ", ev.cme_count) <= 0)
+		return -1;
+
 	/* Insert data into the SGBD */
 #ifdef HAVE_SQLITE3
 	ras_store_cxl_general_media_event(ras, &ev);
diff --git a/ras-record.c b/ras-record.c
index b52973f..e24562e 100644
--- a/ras-record.c
+++ b/ras-record.c
@@ -888,6 +888,11 @@ static const struct db_fields cxl_general_media_event_fields[] = {
 	{ .name = "hpa",		.type = "INTEGER" },
 	{ .name = "region",		.type = "TEXT" },
 	{ .name = "region_uuid",	.type = "TEXT" },
+	{ .name = "pldm_entity_id",	.type = "BLOB" },
+	{ .name = "pldm_resource_id",	.type = "BLOB" },
+	{ .name = "sub_type",		.type = "INTEGER" },
+	{ .name = "cme_threshold_ev_flags",	.type = "INTEGER" },
+	{ .name = "cme_count",		.type = "INTEGER" },
 };
 
 static const struct db_table_descriptor cxl_general_media_event_tab = {
@@ -924,6 +929,14 @@ int ras_store_cxl_general_media_event(struct ras_events *ras,
 	sqlite3_bind_int64(priv->stmt_cxl_general_media_event, idx++, ev->hpa);
 	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev->region, -1, NULL);
 	sqlite3_bind_text(priv->stmt_cxl_general_media_event, idx++, ev->region_uuid, -1, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, idx++, ev->entity_id,
+			  CXL_PLDM_ENTITY_ID_LEN, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_general_media_event, idx++, ev->res_id,
+			  CXL_PLDM_RES_ID_LEN, NULL);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->sub_type);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++,
+			 ev->cme_threshold_ev_flags);
+	sqlite3_bind_int(priv->stmt_cxl_general_media_event, idx++, ev->cme_count);
 
 	rc = sqlite3_step(priv->stmt_cxl_general_media_event);
 	if (rc != SQLITE_OK && rc != SQLITE_DONE)
diff --git a/ras-record.h b/ras-record.h
index a3a88eb..12e693b 100644
--- a/ras-record.h
+++ b/ras-record.h
@@ -196,15 +196,20 @@ struct ras_cxl_general_media_event {
 	uint8_t dpa_flags;
 	uint8_t descriptor;
 	uint8_t type;
+	uint8_t sub_type;
 	uint8_t transaction_type;
 	uint8_t channel;
 	uint8_t rank;
 	uint32_t device;
 	uint8_t *comp_id;
+	uint8_t entity_id[CXL_PLDM_ENTITY_ID_LEN];
+	uint8_t res_id[CXL_PLDM_RES_ID_LEN];
 	uint16_t validity_flags;
 	uint64_t hpa;
 	const char *region;
 	const char *region_uuid;
+	uint8_t cme_threshold_ev_flags;
+	uint32_t cme_count;
 };
 
 struct ras_cxl_dram_event {
diff --git a/ras-report.c b/ras-report.c
index 7e974b9..ed1f4b8 100644
--- a/ras-report.c
+++ b/ras-report.c
@@ -555,13 +555,16 @@ static int set_cxl_general_media_event_backtrace(char *buf, struct ras_cxl_gener
 		"dpa_flags=%u\n"
 		"descriptor=%u\n"
 		"type=%u\n"
+		"sub_type=0x%x\n"
 		"transaction_type=%u\n"
 		"hpa=0x%lx\n"
 		"region=%s\n"
 		"region_uuid=%s\n"
 		"channel=%u\n"
 		"rank=%u\n"
-		"device=0x%x\n",
+		"device=0x%x\n"
+		"cme_threshold_ev_flags=0x%x\n"
+		"cme_count=0x%x\n",
 		ev->hdr.timestamp,
 		ev->hdr.memdev,
 		ev->hdr.host,
@@ -578,13 +581,16 @@ static int set_cxl_general_media_event_backtrace(char *buf, struct ras_cxl_gener
 		ev->dpa_flags,
 		ev->descriptor,
 		ev->type,
+		ev->sub_type,
 		ev->transaction_type,
 		ev->hpa,
 		ev->region,
 		ev->region_uuid,
 		ev->channel,
 		ev->rank,
-		ev->device);
+		ev->device,
+		ev->cme_threshold_ev_flags,
+		ev->cme_count);
 
 	return 0;
 }
-- 
2.43.0

Re: [PATCH 06/13] rasdaemon: cxl: Update CXL general media event to CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:16 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec rev 3.1 section 8.2.9.2.1.1 Table 8-45, General Media Event
> Record has updated with following new fields and new types for Memory
> Event Type and Transaction Type fields.
> 1. Advanced Programmable Corrected Memory Error Threshold Event Flags
> 2. Corrected Memory Error Count at Event
> 3. Memory Event Sub-Type
> 
> The format of component identifier has changed (CXL spec 3.1 section
> 8.2.9.2.1 Table 8-44).
> 
> Update the parsing, logging and recording of general media event for
> the above spec changes.
> 
> Example rasdaemon log for CXL general media event,
> 
> cxl_general_media 2024-11-20 00:00:49 +0000 memdev:mem1 host:0000:0f:00.0 \
> serial:0x3 log type:Fatal hdr_uuid:fbcd0a77-c260-417f-85a9-088b1621eba6 \
> hdr_handle:0x1 hdr_related_handle:0x0 hdr_timestamp:1970-01-01 00:00:49 +0000 \
> hdr_length:128 hdr_maint_op_class:2 hdr_maint_op_sub_class:4 dpa:0x30d40 \
> dpa_flags:descriptor:'UNCORRECTABLE EVENT' 'THRESHOLD EVENT' 'POISON LIST OVERFLOW' \
> memory_event_type:TE State Violation memory_event_sub_type:Media Link Command \
> Training Error transaction_type:Host Inject Poison hpa:0xffffffffffffffff \
> region: region_uuid:00000000-0000-0000-0000-000000000000 channel:3 rank:33 \
> device:5 comp_id:03 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
> comp_id_pldm_valid_flags:'PLDM Entity ID' 'Resource ID' \
> PLDM Entity ID:74 c5 08 9a 1a 0b Resource ID:fc d2 7e 2f \
> Advanced Programmable CME threshold Event Flags:'Corrected Memory Errors in \
> Multiple Media Components' 'Exceeded Programmable Threshold' \
> Corrected Memory Error Count:0x78
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
LGTM
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 07/13] rasdaemon: cxl: Update CXL DRAM event to CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec 3.1 section 8.2.9.2.1.2 Table 8-46, DRAM Event Record has updated
with following new fields and new types for Memory Event Type, Transaction
Type and Validity Flags fields.
1. Component Identifier
2. Sub-channel
3. Advanced Programmable Corrected Memory Error Threshold Event Flags
4. Corrected Memory Error Count at Event
5. Memory Event Sub-Type

Update the parsing, logging and recording of DRAM event for the above
spec rev 3.1 changes.

Example rasdaemon log for CXL DRAM event,

cxl_dram 2024-11-20 00:18:53 +0000 memdev:mem0 host:0000:0f:00.0 serial:0x3 \
log type:Informational hdr_uuid:601dcbb3-9c06-4eab-b8af-4e9bfb5c9624 \
hdr_handle:0x1 hdr_related_handle:0x0 hdr_timestamp:1970-01-01 00:00:58 +0000 \
hdr_length:128 hdr_maint_op_class:1 hdr_maint_op_sub_class:3 dpa:0x18680 \
dpa_flags:descriptor:'UNCORRECTABLE EVENT' 'THRESHOLD EVENT' \
memory_event_type:Data Path Error memory_event_sub_type:Media Link CRC Error \
transaction_type:Internal Media Scrub hpa:0xffffffffffffffff region: \
region_uuid:00000000-0000-0000-0000-000000000000 channel:3 rank:17 \
nibble_mask:3866802 bank_group:7 bank:11 row:2 column:77
correction_mask:21 00 00 00 00 00 00 00 2c 00 00 00 00 00 00 00 37 00 00 \
00 00 00 00 00 42 00 00 00 00 00 00 00 comp_id:01 74 c5 08 9a 1a 0b fc d2 \
7e 2f 31 9b 3c 81 4d comp_id_pldm_valid_flags:'PLDM Entity ID' \
PLDM Entity ID:74 c5 08 9a 1a 0b \
Advanced Programmable CME threshold Event Flags:'Corrected Memory Errors \
in Multiple Media Components' 'Exceeded Programmable Threshold' \
CVME Count:0x94

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 67 +++++++++++++++++++++++++++++++++++++++++++++--
 ras-record.c      | 18 +++++++++++++
 ras-record.h      |  7 +++++
 ras-report.c      | 12 +++++++--
 4 files changed, 100 insertions(+), 4 deletions(-)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index c64c540..c028203 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -1003,7 +1003,7 @@ int ras_cxl_general_media_event_handler(struct trace_seq *s,
 /*
  * DRAM Event Record - DER
  *
- * CXL rev 3.0 section 8.2.9.2.1.2; Table 8-44
+ * CXL rev 3.1 section 8.2.9.2.1.2; Table 8-46
  */
 #define CXL_DER_VALID_CHANNEL			BIT(0)
 #define CXL_DER_VALID_RANK			BIT(1)
@@ -1013,19 +1013,25 @@ int ras_cxl_general_media_event_handler(struct trace_seq *s,
 #define CXL_DER_VALID_ROW			BIT(5)
 #define CXL_DER_VALID_COLUMN			BIT(6)
 #define CXL_DER_VALID_CORRECTION_MASK		BIT(7)
+#define CXL_DER_VALID_COMPONENT_ID		BIT(8)
+#define CXL_DER_VALID_COMPONENT_ID_FORMAT	BIT(9)
+#define CXL_DER_VALID_SUB_CHANNEL		BIT(10)
 
 static const char * const cxl_der_mem_event_type[] = {
 	"Media ECC Error",
 	"Scrub Media ECC Error",
 	"Invalid Address",
 	"Data Path Error",
+	"TE State Violation",
+	"Advanced Programmable CME Counter Expiration",
+	"CKID Violation",
 };
 
 int ras_cxl_dram_event_handler(struct trace_seq *s,
 			       struct tep_record *record,
 			       struct tep_event *event, void *context)
 {
-	int len, i;
+	int len, i, rc;
 	unsigned long long val;
 	struct ras_events *ras = context;
 	struct ras_cxl_dram_event ev;
@@ -1067,6 +1073,15 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
 					      ev.type)) <= 0)
 		return -1;
 
+	if (tep_get_field_val(s,  event, "sub_type", record, &val, 1) < 0)
+		return -1;
+	ev.sub_type = val;
+	if (trace_seq_printf(s, "memory_event_sub_type:%s ",
+			     get_cxl_type_str(cxl_mem_event_sub_type,
+					      ARRAY_SIZE(cxl_mem_event_sub_type),
+					      ev.sub_type)) <= 0)
+		return -1;
+
 	if (tep_get_field_val(s,  event, "transaction_type", record, &val, 1) < 0)
 		return -1;
 	ev.transaction_type = val;
@@ -1108,6 +1123,14 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
 			return -1;
 	}
 
+	if (ev.validity_flags & CXL_DER_VALID_SUB_CHANNEL) {
+		if (tep_get_field_val(s,  event, "sub_channel", record, &val, 1) < 0)
+			return -1;
+		ev.sub_channel = val;
+		if (trace_seq_printf(s, "sub_channel:%u ", ev.sub_channel) <= 0)
+			return -1;
+	}
+
 	if (ev.validity_flags & CXL_DER_VALID_RANK) {
 		if (tep_get_field_val(s,  event, "rank", record, &val, 1) < 0)
 			return -1;
@@ -1168,6 +1191,46 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
 		}
 	}
 
+	if (ev.validity_flags & CXL_DER_VALID_COMPONENT_ID) {
+		ev.comp_id = tep_get_field_raw(s, event, "comp_id", record, &len, 1);
+		if (!ev.comp_id)
+			return -1;
+		if (trace_seq_printf(s, "comp_id:") <= 0)
+			return -1;
+		for (i = 0; i < CXL_EVENT_GEN_MED_COMP_ID_SIZE; i++) {
+			if (trace_seq_printf(s, "%02x ", ev.comp_id[i]) <= 0)
+				break;
+		}
+
+		if (ev.validity_flags & CXL_DER_VALID_COMPONENT_ID_FORMAT) {
+			if (trace_seq_printf(s, "comp_id_pldm_valid_flags:") <= 0)
+				return -1;
+			if (decode_cxl_event_flags(s, ev.comp_id[0], cxl_pldm_comp_id_flags,
+						   ARRAY_SIZE(cxl_pldm_comp_id_flags)) < 0)
+				return -1;
+
+			rc = ras_cxl_print_component_id(s, ev.comp_id, ev.entity_id, ev.res_id);
+			if (rc)
+				return rc;
+		}
+	}
+
+	if (tep_get_field_val(s,  event, "cme_threshold_ev_flags", record, &val, 1) < 0)
+		return -1;
+	ev.cme_threshold_ev_flags = val;
+	if (trace_seq_printf(s, "Advanced Programmable CME threshold Event Flags:") <= 0)
+		return -1;
+	if (decode_cxl_event_flags(s, ev.cme_threshold_ev_flags,
+				   cxl_cme_threshold_ev_flags,
+				   ARRAY_SIZE(cxl_cme_threshold_ev_flags)) < 0)
+		return -1;
+
+	if (tep_get_field_val(s,  event, "cvme_count", record, &val, 1) < 0)
+		return -1;
+	ev.cvme_count = val;
+	if (trace_seq_printf(s, "CVME Count:0x%x ", ev.cvme_count) <= 0)
+		return -1;
+
 	/* Insert data into the SGBD */
 #ifdef HAVE_SQLITE3
 	ras_store_cxl_dram_event(ras, &ev);
diff --git a/ras-record.c b/ras-record.c
index e24562e..82939e7 100644
--- a/ras-record.c
+++ b/ras-record.c
@@ -985,6 +985,13 @@ static const struct db_fields cxl_dram_event_fields[] = {
 	{ .name = "hpa",		.type = "INTEGER" },
 	{ .name = "region",		.type = "TEXT" },
 	{ .name = "region_uuid",	.type = "TEXT" },
+	{ .name = "comp_id",		.type = "BLOB" },
+	{ .name = "pldm_entity_id",	.type = "BLOB" },
+	{ .name = "pldm_resource_id",	.type = "BLOB" },
+	{ .name = "sub_type",		.type = "INTEGER" },
+	{ .name = "sub_channel",	.type = "INTEGER" },
+	{ .name = "cme_threshold_ev_flags",	.type = "INTEGER" },
+	{ .name = "cvme_count",		.type = "INTEGER" },
 };
 
 static const struct db_table_descriptor cxl_dram_event_tab = {
@@ -1024,6 +1031,17 @@ int ras_store_cxl_dram_event(struct ras_events *ras, struct ras_cxl_dram_event *
 	sqlite3_bind_int64(priv->stmt_cxl_dram_event, idx++, ev->hpa);
 	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region, -1, NULL);
 	sqlite3_bind_text(priv->stmt_cxl_dram_event, idx++, ev->region_uuid, -1, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_dram_event, idx++, ev->comp_id,
+			  CXL_EVENT_GEN_MED_COMP_ID_SIZE, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_dram_event, idx++, ev->entity_id,
+			  CXL_PLDM_ENTITY_ID_LEN, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_dram_event, idx++, ev->res_id,
+			  CXL_PLDM_RES_ID_LEN, NULL);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->sub_type);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->sub_channel);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++,
+			 ev->cme_threshold_ev_flags);
+	sqlite3_bind_int(priv->stmt_cxl_dram_event, idx++, ev->cvme_count);
 
 	rc = sqlite3_step(priv->stmt_cxl_dram_event);
 	if (rc != SQLITE_OK && rc != SQLITE_DONE)
diff --git a/ras-record.h b/ras-record.h
index 12e693b..3aec063 100644
--- a/ras-record.h
+++ b/ras-record.h
@@ -218,8 +218,10 @@ struct ras_cxl_dram_event {
 	uint8_t dpa_flags;
 	uint8_t descriptor;
 	uint8_t type;
+	uint8_t sub_type;
 	uint8_t transaction_type;
 	uint8_t channel;
+	uint8_t sub_channel;
 	uint8_t rank;
 	uint32_t nibble_mask;
 	uint8_t bank_group;
@@ -231,6 +233,11 @@ struct ras_cxl_dram_event {
 	uint64_t hpa;
 	const char *region;
 	const char *region_uuid;
+	uint8_t *comp_id;
+	uint8_t entity_id[CXL_PLDM_ENTITY_ID_LEN];
+	uint8_t res_id[CXL_PLDM_RES_ID_LEN];
+	uint8_t cme_threshold_ev_flags;
+	uint32_t cvme_count;
 };
 
 struct ras_cxl_memory_module_event {
diff --git a/ras-report.c b/ras-report.c
index ed1f4b8..8e343fc 100644
--- a/ras-report.c
+++ b/ras-report.c
@@ -624,17 +624,21 @@ static int set_cxl_dram_event_backtrace(char *buf, struct ras_cxl_dram_event *ev
 		"dpa_flags=%u\n"
 		"descriptor=%u\n"
 		"type=%u\n"
+		"sub_type=0x%x\n"
 		"transaction_type=%u\n"
 		"hpa=0x%lx\n"
 		"region=%s\n"
 		"region_uuid=%s\n"
 		"channel=%u\n"
+		"sub_channel=%u\n"
 		"rank=%u\n"
 		"nibble_mask=%u\n"
 		"bank_group=%u\n"
 		"bank=%u\n"
 		"row=%u\n"
-		"column=%u\n",
+		"column=%u\n"
+		"cme_threshold_ev_flags=0x%x\n"
+		"cvme_count=0x%x\n",
 		ev->hdr.timestamp,
 		ev->hdr.memdev,
 		ev->hdr.host,
@@ -651,17 +655,21 @@ static int set_cxl_dram_event_backtrace(char *buf, struct ras_cxl_dram_event *ev
 		ev->dpa_flags,
 		ev->descriptor,
 		ev->type,
+		ev->sub_type,
 		ev->transaction_type,
 		ev->hpa,
 		ev->region,
 		ev->region_uuid,
 		ev->channel,
+		ev->sub_channel,
 		ev->rank,
 		ev->nibble_mask,
 		ev->bank_group,
 		ev->bank,
 		ev->row,
-		ev->column);
+		ev->column,
+		ev->cme_threshold_ev_flags,
+		ev->cvme_count);
 
 	return 0;
 }
-- 
2.43.0

Re: [PATCH 07/13] rasdaemon: cxl: Update CXL DRAM event to CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:17 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec 3.1 section 8.2.9.2.1.2 Table 8-46, DRAM Event Record has updated
> with following new fields and new types for Memory Event Type, Transaction
> Type and Validity Flags fields.
> 1. Component Identifier
> 2. Sub-channel
> 3. Advanced Programmable Corrected Memory Error Threshold Event Flags
> 4. Corrected Memory Error Count at Event
> 5. Memory Event Sub-Type
> 
> Update the parsing, logging and recording of DRAM event for the above
> spec rev 3.1 changes.
> 
> Example rasdaemon log for CXL DRAM event,
> 
> cxl_dram 2024-11-20 00:18:53 +0000 memdev:mem0 host:0000:0f:00.0 serial:0x3 \
> log type:Informational hdr_uuid:601dcbb3-9c06-4eab-b8af-4e9bfb5c9624 \
> hdr_handle:0x1 hdr_related_handle:0x0 hdr_timestamp:1970-01-01 00:00:58 +0000 \
> hdr_length:128 hdr_maint_op_class:1 hdr_maint_op_sub_class:3 dpa:0x18680 \
> dpa_flags:descriptor:'UNCORRECTABLE EVENT' 'THRESHOLD EVENT' \
> memory_event_type:Data Path Error memory_event_sub_type:Media Link CRC Error \
> transaction_type:Internal Media Scrub hpa:0xffffffffffffffff region: \
> region_uuid:00000000-0000-0000-0000-000000000000 channel:3 rank:17 \
> nibble_mask:3866802 bank_group:7 bank:11 row:2 column:77
> correction_mask:21 00 00 00 00 00 00 00 2c 00 00 00 00 00 00 00 37 00 00 \
> 00 00 00 00 00 42 00 00 00 00 00 00 00 comp_id:01 74 c5 08 9a 1a 0b fc d2 \
> 7e 2f 31 9b 3c 81 4d comp_id_pldm_valid_flags:'PLDM Entity ID' \
> PLDM Entity ID:74 c5 08 9a 1a 0b \
> Advanced Programmable CME threshold Event Flags:'Corrected Memory Errors \
> in Multiple Media Components' 'Exceeded Programmable Threshold' \
> CVME Count:0x94
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Also LGTM
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 08/13] rasdaemon: cxl: Update memory module event to CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec 3.1 section 8.2.9.2.1.3 Table 8-47, Memory Module Event Record
has updated with following new fields and new info for Device Event Type
and Device Health Information fields.
1. Validity Flags
2. Component Identifier
3. Device Event Sub-Type

Update the parsing, logging and recording of memory module event for the
above spec rev 3.1 changes.

Example rasdaemon log for CXL memory module event,

cxl_memory_module 2024-11-20 00:22:33 +0000 memdev:mem0 host:0000:0f:00.0 \
serial:0x3 log type:Fatal hdr_uuid:fe927475-dd59-4339-a586-79bab113b774 \
hdr_handle:0x1 hdr_related_handle:0x0 hdr_timestamp:1970-01-01 00:04:38 \
+0000 hdr_length:128 hdr_maint_op_class:0 hdr_maint_op_sub_class:1 \
event_type:Temperature Change event_sub_type:Unsupported Config Data \
health_status:'MAINTENANCE_NEEDED' 'REPLACEMENT_NEEDED' media_status:All \
Data Loss in Event of Power Loss as_life_used:Unknown as_dev_temp:Normal \
as_cor_vol_err_cnt:Normal as_cor_per_err_cnt:Normal life_used:8 \
device_temp:3 dirty_shutdown_cnt:33 cor_vol_err_cnt:25 cor_per_err_cnt:45 \
comp_id:02 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
comp_id_pldm_valid_flags:'Resource ID' Resource ID:fc d2 7e 2f

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 ras-cxl-handler.c | 56 +++++++++++++++++++++++++++++++++++++++++++++--
 ras-record.c      | 11 ++++++++++
 ras-record.h      |  5 +++++
 ras-report.c      |  2 ++
 4 files changed, 72 insertions(+), 2 deletions(-)

diff --git a/ras-cxl-handler.c b/ras-cxl-handler.c
index c028203..a16a095 100644
--- a/ras-cxl-handler.c
+++ b/ras-cxl-handler.c
@@ -1247,7 +1247,7 @@ int ras_cxl_dram_event_handler(struct trace_seq *s,
 /*
  * Memory Module Event Record - MMER
  *
- * CXL res 3.0 section 8.2.9.2.1.3; Table 8-45
+ * CXL res 3.1 section 8.2.9.2.1.3; Table 8-47
  */
 static const char * const cxl_dev_evt_type[] = {
 	"Health Status Change",
@@ -1256,21 +1256,33 @@ static const char * const cxl_dev_evt_type[] = {
 	"Temperature Change",
 	"Data Path Error",
 	"LSA Error",
+	"Unrecoverable Internal Sideband Bus Error",
+	"Memory Media FRU Error",
+	"Power Management Fault",
+};
+
+static const char * const cxl_dev_evt_sub_type[] = {
+	"Not Reported",
+	"Invalid Config Data",
+	"Unsupported Config Data",
+	"Unsupported Memory Media FRU",
 };
 
 /*
  * Device Health Information - DHI
  *
- * CXL res 3.0 section 8.2.9.8.3.1; Table 8-100
+ * CXL res 3.1 section 8.2.9.9.3.1; Table 8-133
  */
 #define CXL_DHI_HS_MAINTENANCE_NEEDED				BIT(0)
 #define CXL_DHI_HS_PERFORMANCE_DEGRADED				BIT(1)
 #define CXL_DHI_HS_HW_REPLACEMENT_NEEDED			BIT(2)
+#define CXL_DHI_HS_MEM_CAPACITY_DEGRADED			BIT(3)
 
 static const struct cxl_event_flags cxl_health_status[] = {
 	{ .bit = CXL_DHI_HS_MAINTENANCE_NEEDED, .flag = "MAINTENANCE_NEEDED" },
 	{ .bit = CXL_DHI_HS_PERFORMANCE_DEGRADED, .flag = "PERFORMANCE_DEGRADED" },
 	{ .bit = CXL_DHI_HS_HW_REPLACEMENT_NEEDED, .flag = "REPLACEMENT_NEEDED" },
+	{ .bit = CXL_DHI_HS_MEM_CAPACITY_DEGRADED, .flag = "MEM_CAPACITY_DEGRADED" },
 };
 
 static const char * const cxl_media_status[] = {
@@ -1302,10 +1314,14 @@ static const char * const cxl_one_bit_status[] = {
 #define CXL_DHI_AS_COR_VOL_ERR_CNT(as)	(((as) & 0x10) >> 4)
 #define CXL_DHI_AS_COR_PER_ERR_CNT(as)	(((as) & 0x20) >> 5)
 
+#define CXL_MMER_VALID_COMPONENT_ID		BIT(0)
+#define CXL_MMER_VALID_COMPONENT_ID_FORMAT	BIT(1)
+
 int ras_cxl_memory_module_event_handler(struct trace_seq *s,
 					struct tep_record *record,
 					struct tep_event *event, void *context)
 {
+	int len, i, rc;
 	unsigned long long val;
 	struct ras_events *ras = context;
 	struct ras_cxl_memory_module_event ev;
@@ -1323,6 +1339,15 @@ int ras_cxl_memory_module_event_handler(struct trace_seq *s,
 					      ev.event_type)) <= 0)
 		return -1;
 
+	if (tep_get_field_val(s, event, "event_sub_type", record, &val, 1) < 0)
+		return -1;
+	ev.event_sub_type = val;
+	if (trace_seq_printf(s, "event_sub_type:%s ",
+			     get_cxl_type_str(cxl_dev_evt_sub_type,
+					      ARRAY_SIZE(cxl_dev_evt_sub_type),
+					      ev.event_sub_type)) <= 0)
+		return -1;
+
 	if (tep_get_field_val(s, event, "health_status", record, &val, 1) < 0)
 		return -1;
 	ev.health_status = val;
@@ -1395,6 +1420,33 @@ int ras_cxl_memory_module_event_handler(struct trace_seq *s,
 	if (trace_seq_printf(s, "cor_per_err_cnt:%u ", ev.cor_per_err_cnt) <= 0)
 		return -1;
 
+	if (tep_get_field_val(s,  event, "validity_flags", record, &val, 1) < 0)
+		return -1;
+	ev.validity_flags = val;
+
+	if (ev.validity_flags & CXL_MMER_VALID_COMPONENT_ID) {
+		ev.comp_id = tep_get_field_raw(s, event, "comp_id", record, &len, 1);
+		if (!ev.comp_id)
+			return -1;
+		if (trace_seq_printf(s, "comp_id:") <= 0)
+			return -1;
+		for (i = 0; i < CXL_EVENT_GEN_MED_COMP_ID_SIZE; i++) {
+			if (trace_seq_printf(s, "%02x ", ev.comp_id[i]) <= 0)
+				break;
+		}
+
+		if (ev.validity_flags & CXL_MMER_VALID_COMPONENT_ID_FORMAT) {
+			if (trace_seq_printf(s, "comp_id_pldm_valid_flags:") <= 0)
+				return -1;
+			if (decode_cxl_event_flags(s, ev.comp_id[0], cxl_pldm_comp_id_flags,
+						   ARRAY_SIZE(cxl_pldm_comp_id_flags)) < 0)
+				return -1;
+
+			rc = ras_cxl_print_component_id(s, ev.comp_id, ev.entity_id, ev.res_id);
+			if (rc)
+				return rc;
+		}
+	}
 	/* Insert data into the SGBD */
 #ifdef HAVE_SQLITE3
 	ras_store_cxl_memory_module_event(ras, &ev);
diff --git a/ras-record.c b/ras-record.c
index 82939e7..d87a70f 100644
--- a/ras-record.c
+++ b/ras-record.c
@@ -1083,6 +1083,10 @@ static const struct db_fields cxl_memory_module_event_fields[] = {
 	{ .name = "cor_per_err_cnt",	.type = "INTEGER" },
 	{ .name = "device_temp",	.type = "INTEGER" },
 	{ .name = "add_status",		.type = "INTEGER" },
+	{ .name = "event_sub_type",	.type = "INTEGER" },
+	{ .name = "comp_id",		.type = "BLOB" },
+	{ .name = "pldm_entity_id",	.type = "BLOB" },
+	{ .name = "pldm_resource_id",	.type = "BLOB" },
 };
 
 static const struct db_table_descriptor cxl_memory_module_event_tab = {
@@ -1115,6 +1119,13 @@ int ras_store_cxl_memory_module_event(struct ras_events *ras,
 	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->cor_per_err_cnt);
 	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->device_temp);
 	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->add_status);
+	sqlite3_bind_int(priv->stmt_cxl_memory_module_event, idx++, ev->event_sub_type);
+	sqlite3_bind_blob(priv->stmt_cxl_memory_module_event, idx++, ev->comp_id,
+			  CXL_EVENT_GEN_MED_COMP_ID_SIZE, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_memory_module_event, idx++, ev->entity_id,
+			  CXL_PLDM_ENTITY_ID_LEN, NULL);
+	sqlite3_bind_blob(priv->stmt_cxl_memory_module_event, idx++, ev->res_id,
+			  CXL_PLDM_RES_ID_LEN, NULL);
 
 	rc = sqlite3_step(priv->stmt_cxl_memory_module_event);
 	if (rc != SQLITE_OK && rc != SQLITE_DONE)
diff --git a/ras-record.h b/ras-record.h
index 3aec063..5eab62c 100644
--- a/ras-record.h
+++ b/ras-record.h
@@ -243,6 +243,7 @@ struct ras_cxl_dram_event {
 struct ras_cxl_memory_module_event {
 	struct ras_cxl_event_common_hdr hdr;
 	uint8_t event_type;
+	uint8_t event_sub_type;
 	uint8_t health_status;
 	uint8_t media_status;
 	uint8_t life_used;
@@ -251,6 +252,10 @@ struct ras_cxl_memory_module_event {
 	uint32_t cor_per_err_cnt;
 	int16_t device_temp;
 	uint8_t add_status;
+	uint16_t validity_flags;
+	uint8_t *comp_id;
+	uint8_t entity_id[CXL_PLDM_ENTITY_ID_LEN];
+	uint8_t res_id[CXL_PLDM_RES_ID_LEN];
 };
 
 struct ras_mc_event;
diff --git a/ras-report.c b/ras-report.c
index 8e343fc..4535421 100644
--- a/ras-report.c
+++ b/ras-report.c
@@ -700,6 +700,7 @@ static int set_cxl_memory_module_event_backtrace(char *buf, struct ras_cxl_memor
 		"hdr_length=%u\n"
 		"hdr_maint_op_class=%u\n"
 		"event_type=%u\n"
+		"event_sub_type=0x%x\n"
 		"health_status=%u\n"
 		"media_status=%u\n"
 		"life_used=%u\n"
@@ -721,6 +722,7 @@ static int set_cxl_memory_module_event_backtrace(char *buf, struct ras_cxl_memor
 		ev->hdr.hdr_length,
 		ev->hdr.hdr_maint_op_class,
 		ev->event_type,
+		ev->event_sub_type,
 		ev->health_status,
 		ev->media_status,
 		ev->life_used,
-- 
2.43.0

Re: [PATCH 08/13] rasdaemon: cxl: Update memory module event to CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:18 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec 3.1 section 8.2.9.2.1.3 Table 8-47, Memory Module Event Record
> has updated with following new fields and new info for Device Event Type
> and Device Health Information fields.
> 1. Validity Flags
> 2. Component Identifier
> 3. Device Event Sub-Type
> 
> Update the parsing, logging and recording of memory module event for the
> above spec rev 3.1 changes.
> 
> Example rasdaemon log for CXL memory module event,
> 
> cxl_memory_module 2024-11-20 00:22:33 +0000 memdev:mem0 host:0000:0f:00.0 \
> serial:0x3 log type:Fatal hdr_uuid:fe927475-dd59-4339-a586-79bab113b774 \
> hdr_handle:0x1 hdr_related_handle:0x0 hdr_timestamp:1970-01-01 00:04:38 \
> +0000 hdr_length:128 hdr_maint_op_class:0 hdr_maint_op_sub_class:1 \
> event_type:Temperature Change event_sub_type:Unsupported Config Data \
> health_status:'MAINTENANCE_NEEDED' 'REPLACEMENT_NEEDED' media_status:All \
> Data Loss in Event of Power Loss as_life_used:Unknown as_dev_temp:Normal \
> as_cor_vol_err_cnt:Normal as_cor_per_err_cnt:Normal life_used:8 \
> device_temp:3 dirty_shutdown_cnt:33 cor_vol_err_cnt:25 cor_per_err_cnt:45 \
> comp_id:02 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
> comp_id_pldm_valid_flags:'Resource ID' Resource ID:fc d2 7e 2f
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 09/13] rasdaemon: ras-mc-ctl: Fix logging of memory event type in CXL DRAM error table

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec rev 3.0 section 8.2.9.2.1.2 defines the DRAM Event Record.

Fix decoding of memory event type in the CXL DRAM error table in RAS
SQLite database.
For e.g. if value is 0x1 it will be logged as an Invalid Address
(General Media Event Record - Memory Event Type) instead of Scrub Media
ECC Error (DRAM Event Record - Memory Event Type) and so on.

Fixes: c38c14afc5d7 ("rasdaemon: ras-mc-ctl: Add support for CXL DRAM trace events")
Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 util/ras-mc-ctl.in | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)

diff --git a/util/ras-mc-ctl.in b/util/ras-mc-ctl.in
index c24941f..3f9bad0 100755
--- a/util/ras-mc-ctl.in
+++ b/util/ras-mc-ctl.in
@@ -1339,7 +1339,7 @@ sub get_cxl_descriptor_flags_text
     return join (", ", @out);
 }
 
-sub get_cxl_mem_event_type
+sub get_cxl_gmer_mem_event_type
 {
     my @types;
 
@@ -1354,6 +1354,22 @@ sub get_cxl_mem_event_type
     return $types[$_[0]];
 }
 
+sub get_cxl_der_mem_event_type
+{
+    my @types;
+
+    if ($_[0] < 0 || $_[0] > 3) {
+	return "unknown-type";
+    }
+
+    @types = ("Media ECC Error",
+	      "Scrub Media ECC Error",
+	      "Invalid Address",
+	      "Data Path Error");
+
+    return $types[$_[0]];
+}
+
 sub get_cxl_transaction_type
 {
     my @types;
@@ -1978,7 +1994,7 @@ sub errors
 	    $out .= sprintf "dpa=0x%llx, ", $dpa if (defined $dpa && length $dpa);
 	    $out .= sprintf "dpa_flags: %s, ", get_cxl_dpa_flags_text($dpa_flags) if (defined $dpa_flags && length $dpa_flags);
 	    $out .= sprintf "descriptor_flags: %s, ", get_cxl_descriptor_flags_text($descriptor) if (defined $descriptor && length $descriptor);
-	    $out .= sprintf "memory event type: %s, ", get_cxl_mem_event_type($mem_event_type) if (defined $mem_event_type && length $mem_event_type);
+	    $out .= sprintf "memory event type: %s, ", get_cxl_gmer_mem_event_type($mem_event_type) if (defined $mem_event_type && length $mem_event_type);
 	    $out .= sprintf "transaction_type: %s, ", get_cxl_transaction_type($transaction_type) if (defined $transaction_type && length $transaction_type);
 	    $out .= sprintf "channel=%u, ", $channel if (defined $channel && length $channel);
 	    $out .= sprintf "rank=%u, ", $rank if (defined $rank && length $rank);
@@ -2024,7 +2040,7 @@ sub errors
 	    $out .= sprintf "dpa=0x%llx, ", $dpa if (defined $dpa && length $dpa);
 	    $out .= sprintf "dpa_flags: %s, ", get_cxl_dpa_flags_text($dpa_flags) if (defined $dpa_flags && length $dpa_flags);
 	    $out .= sprintf "descriptor_flags: %s, ", get_cxl_descriptor_flags_text($descriptor) if (defined $descriptor && length $descriptor);
-	    $out .= sprintf "memory event type: %s, ", get_cxl_mem_event_type($type) if (defined $type && length $type);
+	    $out .= sprintf "memory event type: %s, ", get_cxl_der_mem_event_type($type) if (defined $type && length $type);
 	    $out .= sprintf "transaction_type: %s, ", get_cxl_transaction_type($transaction_type) if (defined $transaction_type && length $transaction_type);
 	    $out .= sprintf "channel=%u, ", $channel if (defined $channel && length $channel);
 	    $out .= sprintf "rank=%u, ", $rank if (defined $rank && length $rank);
-- 
2.43.0

Re: [PATCH 09/13] rasdaemon: ras-mc-ctl: Fix logging of memory event type in CXL DRAM error table

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:19 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec rev 3.0 section 8.2.9.2.1.2 defines the DRAM Event Record.
> 
> Fix decoding of memory event type in the CXL DRAM error table in RAS
> SQLite database.
> For e.g. if value is 0x1 it will be logged as an Invalid Address
> (General Media Event Record - Memory Event Type) instead of Scrub Media
> ECC Error (DRAM Event Record - Memory Event Type) and so on.
> 
> Fixes: c38c14afc5d7 ("rasdaemon: ras-mc-ctl: Add support for CXL DRAM trace events")
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Though note I don't really understand this code so only
reviewing based on changes looking correct given what was there
before.

J

[PATCH 10/13] rasdaemon: ras-mc-ctl: Update logging of common event data to align with CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

The Common Event Record format in the CXL spec 3.1, section 8.2.9.2.1,
Table 8-42, has been updated to include Maintenance Operation Subclass
information.

This update modifies ras-mc-ctl to log CXL common event data in the RAS
SQLite database tables, reflecting the specification changes introduced
in revision 3.1.

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 util/ras-mc-ctl.in | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/util/ras-mc-ctl.in b/util/ras-mc-ctl.in
index 3f9bad0..e075ad8 100755
--- a/util/ras-mc-ctl.in
+++ b/util/ras-mc-ctl.in
@@ -1755,7 +1755,7 @@ sub errors
     my ($memdev, $host, $serial, $error_status, $first_error, $header_log);
     my ($log_type, $first_ts, $last_ts);
     my ($trace_type, $region, $region_uuid, $hpa, $dpa, $dpa_length, $source, $flags, $overflow_ts);
-    my ($hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $data);
+    my ($hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $hdr_maint_op_sub_class, $data);
     my ($dpa_flags, $descriptor, $mem_event_type, $transaction_type, $channel, $rank, $device, $comp_id);
     my ($nibble_mask, $bank_group, $row, $column, $cor_mask);
     my ($event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status);
@@ -1935,10 +1935,10 @@ sub errors
 
 	# CXL generic errors
 	use constant CXL_EVENT_RECORD_DATA_LENGTH => 0x50;
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, data from cxl_generic_event$conf{opt}{since} order by id";
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, data, hdr_maint_op_sub_class from cxl_generic_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $data));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $data, $hdr_maint_op_sub_class));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -1953,6 +1953,7 @@ sub errors
 	    $out .= "hdr_timestamp=$hdr_ts, " if (defined $hdr_ts && length $hdr_ts);
 	    $out .= sprintf "hdr_length=%u, ", $hdr_length if (defined $hdr_length && length $hdr_length);
 	    $out .= sprintf "hdr_maint_op_class=%u, ", $hdr_maint_op_class if (defined $hdr_maint_op_class && length $hdr_maint_op_class);
+	    $out .= sprintf "hdr_maint_op_sub_class=%u, ", $hdr_maint_op_sub_class if (defined $hdr_maint_op_sub_class && length $hdr_maint_op_sub_class);
 	    if (defined $data && length $data) {
 		$out .= sprintf "data:\n";
 		my @bytes = unpack "C*", $data;
@@ -1973,10 +1974,10 @@ sub errors
 
 	# CXL general media errors
 	use constant CXL_EVENT_GEN_MED_COMP_ID_SIZE => 0x10;
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, device, comp_id, hpa, region, region_uuid from cxl_general_media_event$conf{opt}{since} order by id";
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, device, comp_id, hpa, region, region_uuid, hdr_maint_op_sub_class from cxl_general_media_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $mem_event_type, $transaction_type, $channel, $rank, $device, $comp_id, $hpa, $region, $region_uuid));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $mem_event_type, $transaction_type, $channel, $rank, $device, $comp_id, $hpa, $region, $region_uuid, $hdr_maint_op_sub_class));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -1991,6 +1992,7 @@ sub errors
 	    $out .= "hdr_timestamp=$hdr_ts, " if (defined $hdr_ts && length $hdr_ts);
 	    $out .= sprintf "hdr_length=%u, ", $hdr_length if (defined $hdr_length && length $hdr_length);
 	    $out .= sprintf "hdr_maint_op_class=%u, ", $hdr_maint_op_class if (defined $hdr_maint_op_class && length $hdr_maint_op_class);
+	    $out .= sprintf "hdr_maint_op_sub_class=%u, ", $hdr_maint_op_sub_class if (defined $hdr_maint_op_sub_class && length $hdr_maint_op_sub_class);
 	    $out .= sprintf "dpa=0x%llx, ", $dpa if (defined $dpa && length $dpa);
 	    $out .= sprintf "dpa_flags: %s, ", get_cxl_dpa_flags_text($dpa_flags) if (defined $dpa_flags && length $dpa_flags);
 	    $out .= sprintf "descriptor_flags: %s, ", get_cxl_descriptor_flags_text($descriptor) if (defined $descriptor && length $descriptor);
@@ -2019,10 +2021,10 @@ sub errors
 
 	# CXL DRAM errors
 	use constant CXL_EVENT_DER_CORRECTION_MASK_SIZE => 0x20;
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, nibble_mask, bank_group, bank, row, column, cor_mask, hpa, region, region_uuid from cxl_dram_event$conf{opt}{since} order by id";
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, nibble_mask, bank_group, bank, row, column, cor_mask, hpa, region, region_uuid, hdr_maint_op_sub_class from cxl_dram_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $type, $transaction_type, $channel, $rank, $nibble_mask, $bank_group, $bank, $row, $column, $cor_mask, $hpa, $region, $region_uuid));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $type, $transaction_type, $channel, $rank, $nibble_mask, $bank_group, $bank, $row, $column, $cor_mask, $hpa, $region, $region_uuid, $hdr_maint_op_sub_class));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -2037,6 +2039,7 @@ sub errors
 	    $out .= "hdr_timestamp=$hdr_ts, " if (defined $hdr_ts && length $hdr_ts);
 	    $out .= sprintf "hdr_length=%u, ", $hdr_length if (defined $hdr_length && length $hdr_length);
 	    $out .= sprintf "hdr_maint_op_class=%u, ", $hdr_maint_op_class if (defined $hdr_maint_op_class && length $hdr_maint_op_class);
+	    $out .= sprintf "hdr_maint_op_sub_class=%u, ", $hdr_maint_op_sub_class if (defined $hdr_maint_op_sub_class && length $hdr_maint_op_sub_class);
 	    $out .= sprintf "dpa=0x%llx, ", $dpa if (defined $dpa && length $dpa);
 	    $out .= sprintf "dpa_flags: %s, ", get_cxl_dpa_flags_text($dpa_flags) if (defined $dpa_flags && length $dpa_flags);
 	    $out .= sprintf "descriptor_flags: %s, ", get_cxl_descriptor_flags_text($descriptor) if (defined $descriptor && length $descriptor);
@@ -2068,10 +2071,10 @@ sub errors
 	}
 
 	# CXL memory module errors
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, event_type, health_status, media_status, life_used, dirty_shutdown_cnt, cor_vol_err_cnt, cor_per_err_cnt, device_temp, add_status from cxl_memory_module_event$conf{opt}{since} order by id";
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, event_type, health_status, media_status, life_used, dirty_shutdown_cnt, cor_vol_err_cnt, cor_per_err_cnt, device_temp, add_status, hdr_maint_op_sub_class from cxl_memory_module_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status, $hdr_maint_op_sub_class));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -2086,6 +2089,7 @@ sub errors
 	    $out .= "hdr_timestamp=$hdr_ts, " if (defined $hdr_ts && length $hdr_ts);
 	    $out .= sprintf "hdr_length=%u, ", $hdr_length if (defined $hdr_length && length $hdr_length);
 	    $out .= sprintf "hdr_maint_op_class=%u, ", $hdr_maint_op_class if (defined $hdr_maint_op_class && length $hdr_maint_op_class);
+	    $out .= sprintf "hdr_maint_op_sub_class=%u, ", $hdr_maint_op_sub_class if (defined $hdr_maint_op_sub_class && length $hdr_maint_op_sub_class);
 	    $out .= sprintf "event_type: %s, ", get_cxl_dev_event_type($event_type)  if (defined $event_type && length $event_type);
 	    $out .= sprintf "health_status: %s, ", get_cxl_health_status_text($health_status)  if (defined $health_status && length $health_status);
 	    $out .= sprintf "media_status: %s, ", get_cxl_media_status($media_status)  if (defined $media_status && length $media_status);
-- 
2.43.0

Re: [PATCH 10/13] rasdaemon: ras-mc-ctl: Update logging of common event data to align with CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:20 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> The Common Event Record format in the CXL spec 3.1, section 8.2.9.2.1,
> Table 8-42, has been updated to include Maintenance Operation Subclass
> information.
> 
> This update modifies ras-mc-ctl to log CXL common event data in the RAS
> SQLite database tables, reflecting the specification changes introduced
> in revision 3.1.
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Matches what I'd expect to see assuming original code was fine!

Thanks,

Jonathan

[PATCH 11/13] rasdaemon: ras-mc-ctl: Update logging of CXL general media event data to align with CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec rev 3.1 section 8.2.9.2.1.1 Table 8-45, General Media Event
Record has updated with following new fields and new types for Memory
Event Type and Transaction Type fields.
1. Advanced Programmable Corrected Memory Error Threshold Event Flags
2. Corrected Memory Error Count at Event
3. Memory Event Sub-Type

The format of component identifier has changed (CXL spec 3.1 section
8.2.9.2.1 Table 8-44).

This update modifies ras-mc-ctl to parse and log CXL general media event
data stored in the RAS SQLite database table, reflecting the specification
changes introduced in revision 3.1.

Example output,

./util/ras-mc-ctl --errors
...
CXL general media events:
1 2024-11-20 00:00:49 +0000 error: memdev=mem1, host=0000:0f:00.0, \
serial=0x3, log=Fatal, hdr_uuid=00000000-0000-0000-0000-000000000000, \
hdr_flags=0x1 , hdr_handle=0x1, hdr_related_handle=0x0, \
hdr_timestamp=1970-01-01 00:00:49 +0000, hdr_length=128, \
hdr_maint_op_class=2, hdr_maint_op_sub_class=4, dpa=0x30d40, dpa_flags: , \
descriptor_flags: 'UNCORRECTABLE EVENT' , 'THRESHOLD EVENT' , \
'POISON LIST OVERFLOW' , memory event type: TE State Violation, \
memory event sub type: Media Link Command Training Error, transaction_type: \
Host Inject Poison, channel=3, rank=33, device=0x5, \
component_id:03 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
pldm_entity_id:74 c5 08 9a 1a 0b pldm_resource_id:fc d2 7e 2f \
hpa=0xffffffffffffffff, region_uuid=00000000-0000-0000-0000-000000000000, \
cme_threshold_ev_flags: 'Corrected Memory Errors in Multiple Media Components' , \
'Exceeded Programmable Threshold' , cme_count=0x78, 
...

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 util/ras-mc-ctl.in | 96 +++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 82 insertions(+), 14 deletions(-)

diff --git a/util/ras-mc-ctl.in b/util/ras-mc-ctl.in
index e075ad8..f8efa91 100755
--- a/util/ras-mc-ctl.in
+++ b/util/ras-mc-ctl.in
@@ -1339,21 +1339,65 @@ sub get_cxl_descriptor_flags_text
     return join (", ", @out);
 }
 
+#CXL rev 3.1 Section 8.2.9.2.1.1; Table 8-45
+#General Media Event Record - GMER
 sub get_cxl_gmer_mem_event_type
 {
     my @types;
 
-    if ($_[0] < 0 || $_[0] > 2) {
+    if ($_[0] < 0 || $_[0] > 6) {
 	return "unknown-type";
     }
 
     @types = ("ECC Error",
 	      "Invalid Address",
-	      "Data Path Error");
+	      "Data Path Error",
+	      "TE State Violation",
+	      "Scrub Media ECC Error",
+	      "Advanced Programmable CME Counter Expiration",
+	      "CKID Violation");
+
+    return $types[$_[0]];
+}
+
+# CXL rev 3.1 Section 8.2.9.2.1.1; Table 8-45
+sub get_cxl_mem_event_sub_type
+{
+    my @types;
+
+    if ($_[0] < 0 || $_[0] > 5) {
+	return "unknown-type";
+    }
+
+    @types = ("Not Reported",
+	      "Internal Datapath Error",
+	      "Media Link Command Training Error",
+	      "Media Link Control Training Error",
+	      "Media Link Data Training Error",
+	      "Media Link CRC Error");
 
     return $types[$_[0]];
 }
 
+use constant {
+    CXL_EVT_FLAG_CME_MULTIPLE_MEDIA => 0x0001,
+    CXL_EVT_FLAG_THRESHOLD_EXCEEDED => 0x0002,
+};
+sub get_cxl_cme_threshold_ev_flags_text
+{
+    my $flags = $_[0];
+    my @out;
+
+    if ($flags & CXL_EVT_FLAG_CME_MULTIPLE_MEDIA) {
+	push @out, (sprintf "\'Corrected Memory Errors in Multiple Media Components\' ");
+    }
+    if ($flags & CXL_EVT_FLAG_THRESHOLD_EXCEEDED) {
+	push @out, (sprintf "\'Exceeded Programmable Threshold\' ");
+    }
+
+    return join (", ", @out);
+}
+
 sub get_cxl_der_mem_event_type
 {
     my @types;
@@ -1370,11 +1414,12 @@ sub get_cxl_der_mem_event_type
     return $types[$_[0]];
 }
 
+# CXL rev 3.1 Section 8.2.9.2.1.1; Table 8-45
 sub get_cxl_transaction_type
 {
     my @types;
 
-    if ($_[0] < 0 || $_[0] > 6) {
+    if ($_[0] < 0 || $_[0] > 8) {
 	return "unknown-type";
     }
 
@@ -1384,7 +1429,9 @@ sub get_cxl_transaction_type
 	      "Host Scan Media",
 	      "Host Inject Poison",
 	      "Internal Media Scrub",
-	      "Internal Media Management");
+	      "Internal Media Management",
+	      "Internal Media Error Check Scrub",
+	      "Media Initialization");
 
     return $types[$_[0]];
 }
@@ -1453,6 +1500,19 @@ sub get_cxl_media_status
     return $types[$_[0]];
 }
 
+# arg0 - name of the id
+# arg1 - id
+# arg2 - size(bytes) of id
+# arg3 - out log buffer
+sub print_cxl_dev_id
+{
+    $_[3] .= sprintf "%s:", $_[0];
+    my @bytes = unpack "C*", $_[1];
+    for (my $i = 0; $i < $_[2]; $i++) {
+        $_[3] .= sprintf "%02x ", $bytes[$i];
+    }
+}
+
 sub summary
 {
     require DBI;
@@ -1756,9 +1816,10 @@ sub errors
     my ($log_type, $first_ts, $last_ts);
     my ($trace_type, $region, $region_uuid, $hpa, $dpa, $dpa_length, $source, $flags, $overflow_ts);
     my ($hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $hdr_maint_op_sub_class, $data);
-    my ($dpa_flags, $descriptor, $mem_event_type, $transaction_type, $channel, $rank, $device, $comp_id);
+    my ($dpa_flags, $descriptor, $mem_event_type, $mem_event_sub_type, $transaction_type, $channel, $rank, $device, $comp_id, $pldm_entity_id, $pldm_res_id);
     my ($nibble_mask, $bank_group, $row, $column, $cor_mask);
     my ($event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status);
+    my ($cme_threshold_ev_flags, $cme_count);
 
     my $dbh = DBI->connect("dbi:SQLite:dbname=$dbname", "", "", {});
 
@@ -1974,10 +2035,12 @@ sub errors
 
 	# CXL general media errors
 	use constant CXL_EVENT_GEN_MED_COMP_ID_SIZE => 0x10;
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, device, comp_id, hpa, region, region_uuid, hdr_maint_op_sub_class from cxl_general_media_event$conf{opt}{since} order by id";
+	use constant CXL_EVENT_GEN_PLDM_ENTITY_ID_SIZE => 0x6;
+	use constant CXL_EVENT_GEN_PLDM_RES_ID_SIZE => 0x4;
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, device, comp_id, hpa, region, region_uuid, hdr_maint_op_sub_class, pldm_entity_id, pldm_resource_id, sub_type, cme_threshold_ev_flags, cme_count from cxl_general_media_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $mem_event_type, $transaction_type, $channel, $rank, $device, $comp_id, $hpa, $region, $region_uuid, $hdr_maint_op_sub_class));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $mem_event_type, $transaction_type, $channel, $rank, $device, $comp_id, $hpa, $region, $region_uuid, $hdr_maint_op_sub_class, $pldm_entity_id, $pldm_res_id, $mem_event_sub_type, $cme_threshold_ev_flags, $cme_count));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -1997,20 +2060,25 @@ sub errors
 	    $out .= sprintf "dpa_flags: %s, ", get_cxl_dpa_flags_text($dpa_flags) if (defined $dpa_flags && length $dpa_flags);
 	    $out .= sprintf "descriptor_flags: %s, ", get_cxl_descriptor_flags_text($descriptor) if (defined $descriptor && length $descriptor);
 	    $out .= sprintf "memory event type: %s, ", get_cxl_gmer_mem_event_type($mem_event_type) if (defined $mem_event_type && length $mem_event_type);
+            $out .= sprintf "memory event sub type: %s, ", get_cxl_mem_event_sub_type($mem_event_sub_type) if (defined $mem_event_sub_type && length $mem_event_sub_type);
 	    $out .= sprintf "transaction_type: %s, ", get_cxl_transaction_type($transaction_type) if (defined $transaction_type && length $transaction_type);
 	    $out .= sprintf "channel=%u, ", $channel if (defined $channel && length $channel);
 	    $out .= sprintf "rank=%u, ", $rank if (defined $rank && length $rank);
 	    $out .= sprintf "device=0x%x, ", $device if (defined $device && length $device);
-	    if (defined $comp_id && length $comp_id) {
-		$out .= sprintf "component_id:";
-		my @bytes = unpack "C*", $comp_id;
-		for (my $i = 0; $i < CXL_EVENT_GEN_MED_COMP_ID_SIZE; $i++) {
-		    $out .= sprintf "%02x ", $bytes[$i];
-		}
-	    }
+            if (defined $comp_id && length $comp_id) {
+                print_cxl_dev_id("component_id", $comp_id, CXL_EVENT_GEN_MED_COMP_ID_SIZE, $out);
+            }
+            if (defined $pldm_entity_id && length $pldm_entity_id) {
+                print_cxl_dev_id("pldm_entity_id", $pldm_entity_id, CXL_EVENT_GEN_PLDM_ENTITY_ID_SIZE, $out);
+            }
+            if (defined $pldm_res_id && length $pldm_res_id) {
+                print_cxl_dev_id("pldm_resource_id", $pldm_res_id, CXL_EVENT_GEN_PLDM_RES_ID_SIZE, $out);
+            }
             $out .= sprintf "hpa=0x%llx, ", $hpa if (defined $hpa && length $hpa);
             $out .= "region=$region, " if (defined $region && length $region);
             $out .= "region_uuid=$region_uuid, " if (defined $region_uuid && length $region_uuid);
+            $out .= sprintf "cme_threshold_ev_flags: %s, ", get_cxl_cme_threshold_ev_flags_text($cme_threshold_ev_flags) if (defined $cme_threshold_ev_flags && length $cme_threshold_ev_flags);
+            $out .= sprintf "cme_count=0x%x, ", $cme_count if (defined $cme_count && length $cme_count);
 	    $out .= "\n";
 	}
 	if ($out ne "") {
-- 
2.43.0

Re: [PATCH 11/13] rasdaemon: ras-mc-ctl: Update logging of CXL general media event data to align with CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:21 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec rev 3.1 section 8.2.9.2.1.1 Table 8-45, General Media Event
> Record has updated with following new fields and new types for Memory
> Event Type and Transaction Type fields.
> 1. Advanced Programmable Corrected Memory Error Threshold Event Flags
> 2. Corrected Memory Error Count at Event
> 3. Memory Event Sub-Type
> 
> The format of component identifier has changed (CXL spec 3.1 section
> 8.2.9.2.1 Table 8-44).
> 
> This update modifies ras-mc-ctl to parse and log CXL general media event
> data stored in the RAS SQLite database table, reflecting the specification
> changes introduced in revision 3.1.
> 
> Example output,
> 
> ./util/ras-mc-ctl --errors
> ...
> CXL general media events:
> 1 2024-11-20 00:00:49 +0000 error: memdev=mem1, host=0000:0f:00.0, \
> serial=0x3, log=Fatal, hdr_uuid=00000000-0000-0000-0000-000000000000, \
> hdr_flags=0x1 , hdr_handle=0x1, hdr_related_handle=0x0, \
> hdr_timestamp=1970-01-01 00:00:49 +0000, hdr_length=128, \
> hdr_maint_op_class=2, hdr_maint_op_sub_class=4, dpa=0x30d40, dpa_flags: , \
> descriptor_flags: 'UNCORRECTABLE EVENT' , 'THRESHOLD EVENT' , \
> 'POISON LIST OVERFLOW' , memory event type: TE State Violation, \
> memory event sub type: Media Link Command Training Error, transaction_type: \
> Host Inject Poison, channel=3, rank=33, device=0x5, \
> component_id:03 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
> pldm_entity_id:74 c5 08 9a 1a 0b pldm_resource_id:fc d2 7e 2f \
> hpa=0xffffffffffffffff, region_uuid=00000000-0000-0000-0000-000000000000, \
> cme_threshold_ev_flags: 'Corrected Memory Errors in Multiple Media Components' , \
> 'Exceeded Programmable Threshold' , cme_count=0x78, 
> ...
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

[PATCH 12/13] rasdaemon: ras-mc-ctl: Update logging of CXL DRAM event data to align with CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec 3.1 section 8.2.9.2.1.2 Table 8-46, DRAM Event Record has updated
with following new fields and new types for Memory Event Type, Transaction
Type and Validity Flags fields.
1. Component Identifier
2. Sub-channel
3. Advanced Programmable Corrected Memory Error Threshold Event Flags
4. Corrected Volatile Memory Error Count at Event
5. Memory Event Sub-Type

This update modifies ras-mc-ctl to parse and log CXL DRAM event data
stored in the RAS SQLite database table, reflecting the specification
changes introduced in revision 3.1.

Example output,

./util/ras-mc-ctl --errors
...
CXL DRAM events:
1 2024-11-20 00:18:53 +0000 error: memdev=mem0, host=0000:0f:00.0, serial=0x3, \
log=Informational, hdr_uuid=00000000-0000-0000-0000-000000000000, \
hdr_flags=0x1, , hdr_handle=0x1, hdr_related_handle=0x0, \
hdr_timestamp=1970-01-01 00:00:58 +0000, hdr_length=128, hdr_maint_op_class=1, \
hdr_maint_op_sub_class=3, dpa=0x18680, dpa_flags: , \
descriptor_flags: 'UNCORRECTABLE EVENT' , 'THRESHOLD EVENT' , \
memory event type: Data Path Error, memory event sub type: Media Link CRC Error, \
transaction_type: Internal Media Scrub, channel=3, sub_channel=0, rank=17, \
nibble_mask=3866802, bank_group=7, bank=11, row=2, column=77, \
correction_mask:21 00 00 00 00 00 00 00 2c 00 00 00 00 00 00 00 37 00 00 \
00 00 00 00 00 42 00 00 00 00 00 00 00 hpa=0xffffffffffffffff, \
region_uuid=00000000-0000-0000-0000-000000000000, \
component_id:01 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
pldm_entity_id:74 c5 08 9a 1a 0b pldm_resource_id:00 00 00 00 \
cme_threshold_ev_flags: 'Corrected Memory Errors in Multiple Media Components' , \
'Exceeded Programmable Threshold' , cvme_count=0x94, 
...

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 util/ras-mc-ctl.in | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/util/ras-mc-ctl.in b/util/ras-mc-ctl.in
index f8efa91..dbb1607 100755
--- a/util/ras-mc-ctl.in
+++ b/util/ras-mc-ctl.in
@@ -1402,14 +1402,17 @@ sub get_cxl_der_mem_event_type
 {
     my @types;
 
-    if ($_[0] < 0 || $_[0] > 3) {
+    if ($_[0] < 0 || $_[0] > 6) {
 	return "unknown-type";
     }
 
     @types = ("Media ECC Error",
 	      "Scrub Media ECC Error",
 	      "Invalid Address",
-	      "Data Path Error");
+	      "Data Path Error",
+	      "TE State Violation",
+	      "Advanced Programmable CME Counter Expiration",
+	      "CKID Violation");
 
     return $types[$_[0]];
 }
@@ -1819,7 +1822,7 @@ sub errors
     my ($dpa_flags, $descriptor, $mem_event_type, $mem_event_sub_type, $transaction_type, $channel, $rank, $device, $comp_id, $pldm_entity_id, $pldm_res_id);
     my ($nibble_mask, $bank_group, $row, $column, $cor_mask);
     my ($event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status);
-    my ($cme_threshold_ev_flags, $cme_count);
+    my ($sub_channel, $cme_threshold_ev_flags, $cme_count, $cvme_count);
 
     my $dbh = DBI->connect("dbi:SQLite:dbname=$dbname", "", "", {});
 
@@ -2089,10 +2092,10 @@ sub errors
 
 	# CXL DRAM errors
 	use constant CXL_EVENT_DER_CORRECTION_MASK_SIZE => 0x20;
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, nibble_mask, bank_group, bank, row, column, cor_mask, hpa, region, region_uuid, hdr_maint_op_sub_class from cxl_dram_event$conf{opt}{since} order by id";
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, dpa, dpa_flags, descriptor, type, transaction_type, channel, rank, nibble_mask, bank_group, bank, row, column, cor_mask, hpa, region, region_uuid, hdr_maint_op_sub_class, comp_id, pldm_entity_id, pldm_resource_id, sub_type, sub_channel, cme_threshold_ev_flags, cvme_count from cxl_dram_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $type, $transaction_type, $channel, $rank, $nibble_mask, $bank_group, $bank, $row, $column, $cor_mask, $hpa, $region, $region_uuid, $hdr_maint_op_sub_class));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $dpa, $dpa_flags, $descriptor, $type, $transaction_type, $channel, $rank, $nibble_mask, $bank_group, $bank, $row, $column, $cor_mask, $hpa, $region, $region_uuid, $hdr_maint_op_sub_class, $comp_id, $pldm_entity_id, $pldm_res_id, $mem_event_sub_type, $sub_channel, $cme_threshold_ev_flags, $cvme_count));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -2112,8 +2115,10 @@ sub errors
 	    $out .= sprintf "dpa_flags: %s, ", get_cxl_dpa_flags_text($dpa_flags) if (defined $dpa_flags && length $dpa_flags);
 	    $out .= sprintf "descriptor_flags: %s, ", get_cxl_descriptor_flags_text($descriptor) if (defined $descriptor && length $descriptor);
 	    $out .= sprintf "memory event type: %s, ", get_cxl_der_mem_event_type($type) if (defined $type && length $type);
+	    $out .= sprintf "memory event sub type: %s, ", get_cxl_mem_event_sub_type($mem_event_sub_type) if (defined $mem_event_sub_type && length $mem_event_sub_type);
 	    $out .= sprintf "transaction_type: %s, ", get_cxl_transaction_type($transaction_type) if (defined $transaction_type && length $transaction_type);
 	    $out .= sprintf "channel=%u, ", $channel if (defined $channel && length $channel);
+	    $out .= sprintf "sub_channel=%u, ", $sub_channel if (defined $sub_channel && length $sub_channel);
 	    $out .= sprintf "rank=%u, ", $rank if (defined $rank && length $rank);
 	    $out .= sprintf "nibble_mask=%u, ", $nibble_mask if (defined $nibble_mask && length $nibble_mask);
 	    $out .= sprintf "bank_group=%u, ", $bank_group if (defined $bank_group && length $bank_group);
@@ -2130,6 +2135,17 @@ sub errors
             $out .= sprintf "hpa=0x%llx, ", $hpa if (defined $hpa && length $hpa);
             $out .= "region=$region, " if (defined $region && length $region);
             $out .= "region_uuid=$region_uuid, " if (defined $region_uuid && length $region_uuid);
+            if (defined $comp_id && length $comp_id) {
+                print_cxl_dev_id("component_id", $comp_id, CXL_EVENT_GEN_MED_COMP_ID_SIZE, $out);
+            }
+            if (defined $pldm_entity_id && length $pldm_entity_id) {
+                print_cxl_dev_id("pldm_entity_id", $pldm_entity_id, CXL_EVENT_GEN_PLDM_ENTITY_ID_SIZE, $out);
+            }
+            if (defined $pldm_res_id && length $pldm_res_id) {
+                print_cxl_dev_id("pldm_resource_id", $pldm_res_id, CXL_EVENT_GEN_PLDM_RES_ID_SIZE, $out);
+            }
+            $out .= sprintf "cme_threshold_ev_flags: %s, ", get_cxl_cme_threshold_ev_flags_text($cme_threshold_ev_flags) if (defined $cme_threshold_ev_flags && length $cme_threshold_ev_flags);
+            $out .= sprintf "cvme_count=0x%x, ", $cvme_count if (defined $cvme_count && length $cvme_count);
 	    $out .= "\n";
 	}
 	if ($out ne "") {
-- 
2.43.0

Re: [PATCH 12/13] rasdaemon: ras-mc-ctl: Update logging of CXL DRAM event data to align with CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:22 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec 3.1 section 8.2.9.2.1.2 Table 8-46, DRAM Event Record has updated
> with following new fields and new types for Memory Event Type, Transaction
> Type and Validity Flags fields.
> 1. Component Identifier
> 2. Sub-channel
> 3. Advanced Programmable Corrected Memory Error Threshold Event Flags
> 4. Corrected Volatile Memory Error Count at Event
> 5. Memory Event Sub-Type
> 
> This update modifies ras-mc-ctl to parse and log CXL DRAM event data
> stored in the RAS SQLite database table, reflecting the specification
> changes introduced in revision 3.1.
> 
> Example output,
> 
> ./util/ras-mc-ctl --errors
> ...
> CXL DRAM events:
> 1 2024-11-20 00:18:53 +0000 error: memdev=mem0, host=0000:0f:00.0, serial=0x3, \
> log=Informational, hdr_uuid=00000000-0000-0000-0000-000000000000, \
> hdr_flags=0x1, , hdr_handle=0x1, hdr_related_handle=0x0, \
> hdr_timestamp=1970-01-01 00:00:58 +0000, hdr_length=128, hdr_maint_op_class=1, \
> hdr_maint_op_sub_class=3, dpa=0x18680, dpa_flags: , \
> descriptor_flags: 'UNCORRECTABLE EVENT' , 'THRESHOLD EVENT' , \
> memory event type: Data Path Error, memory event sub type: Media Link CRC Error, \
> transaction_type: Internal Media Scrub, channel=3, sub_channel=0, rank=17, \
> nibble_mask=3866802, bank_group=7, bank=11, row=2, column=77, \
> correction_mask:21 00 00 00 00 00 00 00 2c 00 00 00 00 00 00 00 37 00 00 \
> 00 00 00 00 00 42 00 00 00 00 00 00 00 hpa=0xffffffffffffffff, \
> region_uuid=00000000-0000-0000-0000-000000000000, \
> component_id:01 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
> pldm_entity_id:74 c5 08 9a 1a 0b pldm_resource_id:00 00 00 00 \
> cme_threshold_ev_flags: 'Corrected Memory Errors in Multiple Media Components' , \
> 'Exceeded Programmable Threshold' , cvme_count=0x94, 
> ...
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Changes as expected...

[PATCH 13/13] rasdaemon: ras-mc-ctl: Update logging of CXL memory module data to align with CXL spec rev 3.1

[shiju.jose]

From: Shiju Jose <shiju.jose@huawei.com>

CXL spec 3.1 section 8.2.9.2.1.3 Table 8-47, Memory Module Event Record
has updated with following new fields and new info for Device Event Type
and Device Health Information fields.
1. Validity Flags
2. Component Identifier
3. Device Event Sub-Type

This update modifies ras-mc-ctl to parse and log CXL memory module event
data stored in the RAS SQLite database table, reflecting the
specification changes introduced in revision 3.1.

Example output,

./util/ras-mc-ctl --errors
...
CXL memory module events:
1 2024-11-20 00:22:33 +0000 error: memdev=mem0, host=0000:0f:00.0, serial=0x3, \
log=Fatal, hdr_uuid=fe927475-dd59-4339-a586-79bab113b774, hdr_flags=0x1, , \
hdr_handle=0x1, hdr_related_handle=0x0, hdr_timestamp=1970-01-01 00:04:38 +0000, \
hdr_length=128, hdr_maint_op_class=0, hdr_maint_op_sub_class=1, \
event_type: Temperature Change, event_sub_type: Unsupported Config Data, \
health_status: 'MAINTENANCE_NEEDED' , 'REPLACEMENT_NEEDED' , \
media_status: All Data Loss in Event of Power Loss, life_used=8, \
dirty_shutdown_cnt=33, cor_vol_err_cnt=25, cor_per_err_cnt=45, \
device_temp=3, add_status=3 \
component_id:02 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
pldm_entity_id:00 00 00 00 00 00 pldm_resource_id:fc d2 7e 2f 
...

Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
---
 util/ras-mc-ctl.in | 46 +++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 41 insertions(+), 5 deletions(-)

diff --git a/util/ras-mc-ctl.in b/util/ras-mc-ctl.in
index dbb1607..0990da9 100755
--- a/util/ras-mc-ctl.in
+++ b/util/ras-mc-ctl.in
@@ -1439,11 +1439,12 @@ sub get_cxl_transaction_type
     return $types[$_[0]];
 }
 
+# CXL rev 3.1 section 8.2.9.2.1.3; Table 8-47
 sub get_cxl_dev_event_type
 {
     my @types;
 
-    if ($_[0] < 0 || $_[0] > 5) {
+    if ($_[0] < 0 || $_[0] > 8) {
 	return "unknown-type";
     }
 
@@ -1452,15 +1453,37 @@ sub get_cxl_dev_event_type
 	      "Life Used Change",
 	      "Temperature Change",
 	      "Data Path Error",
-	      "LSA Error");
+	      "LSA Error",
+	      "Unrecoverable Internal Sideband Bus Error",
+	      "Memory Media FRU Error",
+	      "Power Management Fault");
 
     return $types[$_[0]];
 }
 
+sub get_cxl_dev_event_sub_type
+{
+    my @types;
+
+    if ($_[0] < 0 || $_[0] > 3) {
+	return "unknown-type";
+    }
+
+    @types = ("Not Reported",
+	      "Invalid Config Data",
+	      "Unsupported Config Data",
+	      "Unsupported Memory Media FRU");
+
+    return $types[$_[0]];
+}
+
+#CXL rev 3.1 section 8.2.9.9.3.1; Table 8-133
 use constant {
     CXL_DHI_HS_MAINTENANCE_NEEDED => 0x0001,
     CXL_DHI_HS_PERFORMANCE_DEGRADED => 0x0002,
     CXL_DHI_HS_HW_REPLACEMENT_NEEDED => 0x0004,
+    CXL_DHI_HS_HW_REPLACEMENT_NEEDED => 0x0004,
+    CXL_DHI_HS_MEM_CAPACITY_DEGRADED => 0x0008,
 };
 
 sub get_cxl_health_status_text
@@ -1477,6 +1500,9 @@ sub get_cxl_health_status_text
     if ($flags & CXL_DHI_HS_HW_REPLACEMENT_NEEDED) {
 	push @out, (sprintf "\'REPLACEMENT_NEEDED\' ");
     }
+    if ($flags & CXL_DHI_HS_MEM_CAPACITY_DEGRADED) {
+	push @out, (sprintf "\'MEM_CAPACITY_DEGRADED\' ");
+    }
 
     return join (", ", @out);
 }
@@ -1821,7 +1847,7 @@ sub errors
     my ($hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $hdr_maint_op_sub_class, $data);
     my ($dpa_flags, $descriptor, $mem_event_type, $mem_event_sub_type, $transaction_type, $channel, $rank, $device, $comp_id, $pldm_entity_id, $pldm_res_id);
     my ($nibble_mask, $bank_group, $row, $column, $cor_mask);
-    my ($event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status);
+    my ($event_type, $event_sub_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status);
     my ($sub_channel, $cme_threshold_ev_flags, $cme_count, $cvme_count);
 
     my $dbh = DBI->connect("dbi:SQLite:dbname=$dbname", "", "", {});
@@ -2155,10 +2181,10 @@ sub errors
 	}
 
 	# CXL memory module errors
-	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, event_type, health_status, media_status, life_used, dirty_shutdown_cnt, cor_vol_err_cnt, cor_per_err_cnt, device_temp, add_status, hdr_maint_op_sub_class from cxl_memory_module_event$conf{opt}{since} order by id";
+	$query = "select id, timestamp, memdev, host, serial, log_type, hdr_uuid, hdr_flags, hdr_handle, hdr_related_handle, hdr_ts, hdr_length, hdr_maint_op_class, event_type, health_status, media_status, life_used, dirty_shutdown_cnt, cor_vol_err_cnt, cor_per_err_cnt, device_temp, add_status, hdr_maint_op_sub_class, event_sub_type, comp_id, pldm_entity_id, pldm_resource_id from cxl_memory_module_event$conf{opt}{since} order by id";
 	$query_handle = $dbh->prepare($query);
 	$query_handle->execute();
-	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status, $hdr_maint_op_sub_class));
+	$query_handle->bind_columns(\($id, $timestamp, $memdev, $host, $serial, $log_type, $hdr_uuid, $hdr_flags, $hdr_handle, $hdr_related_handle, $hdr_ts, $hdr_length, $hdr_maint_op_class, $event_type, $health_status, $media_status, $life_used, $dirty_shutdown_cnt, $cor_vol_err_cnt, $cor_per_err_cnt, $device_temp, $add_status, $hdr_maint_op_sub_class, $event_sub_type, $comp_id, $pldm_entity_id, $pldm_res_id));
 	$out = "";
 	while($query_handle->fetch()) {
 	    $out .= "$id $timestamp error: ";
@@ -2175,6 +2201,7 @@ sub errors
 	    $out .= sprintf "hdr_maint_op_class=%u, ", $hdr_maint_op_class if (defined $hdr_maint_op_class && length $hdr_maint_op_class);
 	    $out .= sprintf "hdr_maint_op_sub_class=%u, ", $hdr_maint_op_sub_class if (defined $hdr_maint_op_sub_class && length $hdr_maint_op_sub_class);
 	    $out .= sprintf "event_type: %s, ", get_cxl_dev_event_type($event_type)  if (defined $event_type && length $event_type);
+            $out .= sprintf "event_sub_type: %s, ", get_cxl_dev_event_sub_type($event_sub_type)  if (defined $event_sub_type && length $event_sub_type);
 	    $out .= sprintf "health_status: %s, ", get_cxl_health_status_text($health_status)  if (defined $health_status && length $health_status);
 	    $out .= sprintf "media_status: %s, ", get_cxl_media_status($media_status)  if (defined $media_status && length $media_status);
 	    $out .= sprintf "life_used=%u, ", $life_used  if (defined $life_used && length $life_used);
@@ -2183,6 +2210,15 @@ sub errors
 	    $out .= sprintf "cor_per_err_cnt=%u, ", $cor_per_err_cnt  if (defined $cor_per_err_cnt && length $cor_per_err_cnt);
 	    $out .= sprintf "device_temp=%u, ", $device_temp  if (defined $device_temp && length $device_temp);
 	    $out .= sprintf "add_status=%u ", $add_status  if (defined $add_status && length $add_status);
+            if (defined $comp_id && length $comp_id) {
+                print_cxl_dev_id("component_id", $comp_id, CXL_EVENT_GEN_MED_COMP_ID_SIZE, $out);
+            }
+            if (defined $pldm_entity_id && length $pldm_entity_id) {
+                print_cxl_dev_id("pldm_entity_id", $pldm_entity_id, CXL_EVENT_GEN_PLDM_ENTITY_ID_SIZE, $out);
+            }
+            if (defined $pldm_res_id && length $pldm_res_id) {
+                print_cxl_dev_id("pldm_resource_id", $pldm_res_id, CXL_EVENT_GEN_PLDM_RES_ID_SIZE, $out);
+            }
 	    $out .= "\n";
 	}
 	if ($out ne "") {
-- 
2.43.0

Re: [PATCH 13/13] rasdaemon: ras-mc-ctl: Update logging of CXL memory module data to align with CXL spec rev 3.1

[Jonathan Cameron]

On Wed, 20 Nov 2024 09:59:23 +0000
<shiju.jose@huawei.com> wrote:

> From: Shiju Jose <shiju.jose@huawei.com>
> 
> CXL spec 3.1 section 8.2.9.2.1.3 Table 8-47, Memory Module Event Record
> has updated with following new fields and new info for Device Event Type
> and Device Health Information fields.
> 1. Validity Flags
> 2. Component Identifier
> 3. Device Event Sub-Type
> 
> This update modifies ras-mc-ctl to parse and log CXL memory module event
> data stored in the RAS SQLite database table, reflecting the
> specification changes introduced in revision 3.1.
> 
> Example output,
> 
> ./util/ras-mc-ctl --errors
> ...
> CXL memory module events:
> 1 2024-11-20 00:22:33 +0000 error: memdev=mem0, host=0000:0f:00.0, serial=0x3, \
> log=Fatal, hdr_uuid=fe927475-dd59-4339-a586-79bab113b774, hdr_flags=0x1, , \
> hdr_handle=0x1, hdr_related_handle=0x0, hdr_timestamp=1970-01-01 00:04:38 +0000, \
> hdr_length=128, hdr_maint_op_class=0, hdr_maint_op_sub_class=1, \
> event_type: Temperature Change, event_sub_type: Unsupported Config Data, \
> health_status: 'MAINTENANCE_NEEDED' , 'REPLACEMENT_NEEDED' , \
> media_status: All Data Loss in Event of Power Loss, life_used=8, \
> dirty_shutdown_cnt=33, cor_vol_err_cnt=25, cor_per_err_cnt=45, \
> device_temp=3, add_status=3 \
> component_id:02 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
> pldm_entity_id:00 00 00 00 00 00 pldm_resource_id:fc d2 7e 2f 
> ...
> 
> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
Feels like there is a lot of duplication in here, but you aren't
really making it any worse and maybe it is hard to reduce it.

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

RE: [PATCH 13/13] rasdaemon: ras-mc-ctl: Update logging of CXL memory module data to align with CXL spec rev 3.1

[Shiju Jose]

Hi Jonathan,

>-----Original Message-----
>From: Jonathan Cameron <jonathan.cameron@huawei.com>
>Sent: 21 November 2024 15:39
>To: Shiju Jose <shiju.jose@huawei.com>
>Cc: linux-edac@vger.kernel.org; linux-cxl@vger.kernel.org;
>mchehab@kernel.org; dave.jiang@intel.com; dan.j.williams@intel.com;
>alison.schofield@intel.com; nifan.cxl@gmail.com; vishal.l.verma@intel.com;
>ira.weiny@intel.com; dave@stgolabs.net; linux-kernel@vger.kernel.org;
>Linuxarm <linuxarm@huawei.com>; tanxiaofei <tanxiaofei@huawei.com>;
>Zengtao (B) <prime.zeng@hisilicon.com>
>Subject: Re: [PATCH 13/13] rasdaemon: ras-mc-ctl: Update logging of CXL
>memory module data to align with CXL spec rev 3.1
>
>On Wed, 20 Nov 2024 09:59:23 +0000
><shiju.jose@huawei.com> wrote:
>
>> From: Shiju Jose <shiju.jose@huawei.com>
>>
>> CXL spec 3.1 section 8.2.9.2.1.3 Table 8-47, Memory Module Event
>> Record has updated with following new fields and new info for Device
>> Event Type and Device Health Information fields.
>> 1. Validity Flags
>> 2. Component Identifier
>> 3. Device Event Sub-Type
>>
>> This update modifies ras-mc-ctl to parse and log CXL memory module
>> event data stored in the RAS SQLite database table, reflecting the
>> specification changes introduced in revision 3.1.
>>
>> Example output,
>>
>> ./util/ras-mc-ctl --errors
>> ...
>> CXL memory module events:
>> 1 2024-11-20 00:22:33 +0000 error: memdev=mem0, host=0000:0f:00.0,
>> serial=0x3, \ log=Fatal,
>> hdr_uuid=fe927475-dd59-4339-a586-79bab113b774, hdr_flags=0x1, , \
>> hdr_handle=0x1, hdr_related_handle=0x0, hdr_timestamp=1970-01-01
>> 00:04:38 +0000, \ hdr_length=128, hdr_maint_op_class=0,
>> hdr_maint_op_sub_class=1, \
>> event_type: Temperature Change, event_sub_type: Unsupported Config
>> Data, \
>> health_status: 'MAINTENANCE_NEEDED' , 'REPLACEMENT_NEEDED' , \
>> media_status: All Data Loss in Event of Power Loss, life_used=8, \
>> dirty_shutdown_cnt=33, cor_vol_err_cnt=25, cor_per_err_cnt=45, \
>> device_temp=3, add_status=3 \
>> component_id:02 74 c5 08 9a 1a 0b fc d2 7e 2f 31 9b 3c 81 4d \
>> pldm_entity_id:00 00 00 00 00 00 pldm_resource_id:fc d2 7e 2f ...
>>
>> Signed-off-by: Shiju Jose <shiju.jose@huawei.com>
>Feels like there is a lot of duplication in here, but you aren't really making it any
>worse and maybe it is hard to reduce it.
>
ras-mc-ctl is a tool(script), used offline, to read, decode and print  the error event's data stored
by rasdaemon into the SQLite data base. Thus logging here is similar to those done in the rasdaemon.

>Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Thanks,
Shiju