rpms/ppc64-diag
7
0
Fork 0
Code Issues Pull Requests Releases Activity

Resolves: RHEL-102474, Update to 2.7.11

Browse Source
This commit is contained in:
Than Ngo 2026-01-27 17:08:20 +01:00
parent 4f4f98a09e
commit 5f14218b3e
16 changed files with 7 additions and 948 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@ -16,3 +16,4 @@
/ppc64-diag-2.7.7.tar.gz
/ppc64-diag-2.7.8.tar.gz
/ppc64-diag-2.7.9.tar.gz
/ppc64-diag-2.7.11.tar.gz

71
ppc64-diag-2.7.9-call_home-fail-on-nvmf-device.patch
View File

@ -1,71 +0,0 @@
commit db0c6d7974d7f8909878384d77ec02457759d6df
Author: Nilay Shroff <nilay@linux.ibm.com>
Date: Tue Jan 16 13:55:03 2024 +0530
diags/diag_nvme: call_home command fails on nvmf drive
The diag_nvme command needs to retrieve the VPD log page from NVMe for
filling in the product data while generating the call-home event.
However, call-home feature is supported for directly attached NVMe
module. In the current diag_nvme implementation, if user doesn't
provide NVMe device name for diagnostics then it(diag_nvme) loops
through each NVMe moudle (directly connected to the system/LPAR as
well as discovered over fabrics) and attempt retrieving the SMART log
page as well as VPD page. Unfortunately, diag_nvme fails to retrieve
the VPD page for NVMe connected over fabrics and that causes the
diag_nvme to print "not-so-nice" failure messages on console.
Henec fixed the diag_nvme code so that for call-home event reporting,
it skips the NVMe which is connected over fabrics and prints a
"nice-message" informing the user that it's skipping diagnosting for
NVMe module connected over fabrics. In a nutshell, with this fix now
diag_nvme would only diagnose the NVMe module which is directtly
attached (over PCIe) to the system.
Signed-off-by: Nilay Shroff <nilay@linux.ibm.com>
diff --git a/diags/diag_nvme.c b/diags/diag_nvme.c
index c1c0a20..e86786c 100644
--- a/diags/diag_nvme.c
+++ b/diags/diag_nvme.c
@@ -375,9 +375,40 @@ static int diagnose_nvme(char *device_name, struct notify *notify, char *file_pa
char endurance_s[sizeof(vpd.endurance) + 1], capacity_s[sizeof(vpd.capacity)+1];
uint64_t event_id;
uint8_t severity;
+ FILE *fp;
+ char tr_file_path[PATH_MAX];
uint32_t raw_data_len = 0;
unsigned char *raw_data = NULL;
+ /*
+ * Skip diag test if NVMe is connected over fabric
+ */
+ snprintf(tr_file_path, sizeof(tr_file_path),
+ NVME_SYS_PATH"/%s/%s", device_name, "transport");
+ fp = fopen(tr_file_path, "r");
+ if (fp) {
+ char buf[12];
+ int n = fread(buf, 1, sizeof(buf), fp);
+
+ if (n) {
+ /*
+ * If NVMe transport is anything but pcie then skip the diag test
+ */
+ if (strncmp(buf, "pcie", 4) != 0) {
+ fprintf(stdout, "Skipping diagnostics for nvmf : %s\n",
+ device_name);
+ fclose(fp);
+ return 0;
+ }
+ }
+ fclose(fp);
+ } else {
+ fprintf(stderr, "Skipping diagnostics for %s:\n"
+ "Unable to find the nvme transport type\n",
+ device_name);
+ return -1;
+ }
+
tmp_rc = regex_controller(controller_name, device_name);
if (tmp_rc != 0)
return -1;

43
ppc64-diag-2.7.9-compiler-warnings.patch
View File

@ -1,43 +0,0 @@
diff --git a/lpd/usysident.c b/lpd/usysident.c
index 8fee736..11a512b 100644
--- a/lpd/usysident.c
+++ b/lpd/usysident.c
@@ -429,7 +429,7 @@ retry:
* next elements in the list (which may match the device name);
* for this we need to retry, re-starting on the next element.
*/
- if (dvalue && current->devname &&
+ if (dvalue && current->devname[0] != '\0' &&
strncmp(dvalue, current->devname, DEV_LENGTH)) {
list_start = current->next;
goto retry;
diff --git a/opal_errd/extract_opal_dump.c b/opal_errd/extract_opal_dump.c
index b7f365f..656899d 100644
--- a/opal_errd/extract_opal_dump.c
+++ b/opal_errd/extract_opal_dump.c
@@ -77,8 +77,10 @@ static void dump_get_file_name(char *buf, int bsize, char *dfile,
if (bsize >= DUMP_HDR_PREFIX_OFFSET + sizeof(uint16_t))
*prefix_size = be16toh(*(uint16_t *)(buf + DUMP_HDR_PREFIX_OFFSET));
- if (bsize >= DUMP_HDR_FNAME_OFFSET + DUMP_MAX_FNAME_LEN)
- strncpy(dfile, buf + DUMP_HDR_FNAME_OFFSET, dfile_size);
+ if (bsize >= DUMP_HDR_FNAME_OFFSET + DUMP_MAX_FNAME_LEN) {
+ strncpy(dfile, buf + DUMP_HDR_FNAME_OFFSET, dfile_size - 1);
+ dfile[dfile_size - 1] = '\0';
+ }
else
strncpy(dfile, "platform.dumpid.PARTIAL", dfile_size);
diff --git a/rtas_errd/extract_platdump.c b/rtas_errd/extract_platdump.c
index 831e57e..50bac99 100644
--- a/rtas_errd/extract_platdump.c
+++ b/rtas_errd/extract_platdump.c
@@ -220,7 +220,7 @@ extract_platform_dump(uint64_t dump_tag)
char filename[DUMP_MAX_FNAME_LEN + 1];
char pathname[PATH_MAX];
char dump_err[RTAS_ERROR_LOG_MAX];
- char dumpid[5];
+ char dumpid[5] = "";
int out=-1, rc, librtas_rc, dump_complete=0, ret=0;
msg("Dump tag: 0x%016LX", dump_tag);

72
ppc64-diag-2.7.9-diag_encl-man-and-help-page.patch
View File

@ -1,72 +0,0 @@
commit 565035cf98e97c7adda523221e109c676daf3430
Author: Barnali Guha Thakurata <barnali@linux.ibm.com>
Date: Mon Jan 8 20:16:19 2024 +0530
diag_encl help output update
This patch changes "scsi_enclosure" in the help output of diag_encl to
"sgn ..." so that it matches with the naming convention in the
manual page
Before patch:
diags/diag_encl --help
Usage: diags/diag_encl [-h] [-V] [...] [<scsi_enclosure>]
[...]
<scsi_enclosure>: the sg device on which to operate, such
as sg7; if not specified, all such
devices will be diagnosed
man diags/man/diag_encl.8
...
SYNOPSIS
diag_encl [-s [-c][-l]] [-v] [-f <path.pg2>] [sgn ...]
diag_encl -d
diag_encl --version
diag_encl --help
...
After patch:
diags/diag_encl --help
Usage: diags/diag_encl [-h] [-V] [...] [-f <path.pg2>] [<sgn ...>]
[...]
<sgn ...>: the scsi device on which to operate, such
as sg7; if not specified, all such
devices will be diagnosed
man diags/man/diag_encl.8
...
SYNOPSIS
diag_encl [-s [-c][-l]] [-v] [-f <path.pg2>] [sgn ...]
diag_encl -d
diag_encl --version
diag_encl --help
...
Signed-off-by: Barnali Guha Thakurata <barnali@linux.ibm.com>
diff --git a/diags/diag_encl.c b/diags/diag_encl.c
index 449b097..e0d86ce 100644
--- a/diags/diag_encl.c
+++ b/diags/diag_encl.c
@@ -75,8 +75,7 @@ static struct {
*/
static void
print_usage(const char *name) {
- printf("Usage: %s [-h] [-V] [-s [-c][-l]] [-v] [-f <path.pg2>]"
- " [<scsi_enclosure>]\n"
+ printf("Usage: %s [-h] [-V] [-s [-c][-l]] [-v] [-f <path.pg2>] [<sgn ...>]\n"
"\n\t-h: print this help message\n"
"\t-s: generate serviceable events for any failures and\n"
"\t write events to the servicelog\n"
@@ -87,7 +86,7 @@ print_usage(const char *name) {
"\t-V: print the version of the command and exit\n"
"\t-f: for testing, read SES data from path.pg2 and VPD\n"
"\t from path.vpd\n"
- "\t<scsi_enclosure>: the sg device on which to operate, such\n"
+ "\t<sgn ...>: the scsi device on which to operate, such\n"
"\t as sg7; if not specified, all such\n"
"\t devices will be diagnosed\n", name);
}

62
ppc64-diag-2.7.9-diag_nvme-man-and-help-page.patch
View File

@ -1,62 +0,0 @@
commit 420d5145230041e0256f0f950a36c4ecaed4d584
Author: Barnali Guha Thakurata <barnali@linux.ibm.com>
Date: Mon Jan 8 16:59:33 2024 +0530
diag_nvme help output update
This patch changes "nvme_devices" in the help output of diag_nvme to
"nvmen ..." so that it matches with the naming convention in the
manual page
Before patch:
diags/diag_nvme --help
Usage: diags/diag_nvme [-h] [<nvme_devices>]
-h or --help: print this help message
<nvme_devices>: the NVMe devices on which to operate, for
example nvme0; if not specified, all detected
nvme devices will be diagnosed
man diags/man/diag_nvme.8
...
SYNOPSIS
diag_nvme [<nvmen ...>]
diag_nvme <nvmen>
diag_nvme --help
...
After patch:
diags/diag_nvme --help
Usage: diags/diag_nvme [-h] [<nvmen ...>]
-h or --help: print this help message
<nvmen ...>: the NVMe devices on which to operate, for
example nvme0; if not specified, all detected
nvme devices will be diagnosed
man diags/man/diag_nvme.8
...
SYNOPSIS
diag_nvme [<nvmen ...>]
diag_nvme <nvmen>
diag_nvme --help
...
Signed-off-by: Barnali Guha Thakurata <barnali@linux.ibm.com>
diff --git a/diags/diag_nvme.c b/diags/diag_nvme.c
index c1c0a20..2a74007 100644
--- a/diags/diag_nvme.c
+++ b/diags/diag_nvme.c
@@ -989,9 +989,9 @@ extern int open_nvme(char *dev_path) {
}
static void print_usage(char *command) {
- printf("Usage: %s [-h] [<nvme_devices>]\n"
+ printf("Usage: %s [-h] [<nvmen ...>]\n"
"\t-h or --help: print this help message\n"
- "\t<nvme_devices>: the NVMe devices on which to operate, for\n"
+ "\t<nvmen ...>: the NVMe devices on which to operate, for\n"
"\t example nvme0; if not specified, all detected\n"
"\t nvme devices will be diagnosed\n", command);
}

55
ppc64-diag-2.7.9-diag_vnme-add-power11-support.patch
View File

@ -1,55 +0,0 @@
commit 6bfa2ec7a7f424b018fd2e21a70e9d2f04041832
Author: Barnali Guha Thakurata <barnali@linux.ibm.com>
Date: Tue Oct 22 12:05:39 2024 +0530
Added Power11 support for diag_nvme
Without Patch:
# diag_nvme nvme0
diag_nvme is only supported in PowerVM LPARs and at least Power10 processors
With Patch:
# diag_nvme nvme0
Running diagnostics for nvme0
NVMe diag command completed successfully
Signed-off-by: Barnali Guha Thakurata <barnali@linux.ibm.com>
Signed-off-by: Nilay Shroff <nilay@linux.ibm.com>
Reviewed-by: Sathvika Vasireddy <sv@linux.ibm.com>
Tested-by: Sathvika Vasireddy <sv@linux.ibm.com>
diff --git a/common/platform.c b/common/platform.c
index 1359adb..801b924 100644
--- a/common/platform.c
+++ b/common/platform.c
@@ -96,6 +96,9 @@ get_processor(void)
case PVR_POWER10:
rc = POWER10;
break;
+ case PVR_POWER11:
+ rc = POWER11;
+ break;
default:
rc = PROCESSOR_UNKNOWN;
break;
diff --git a/common/platform.h b/common/platform.h
index b265d1e..ae3ea21 100644
--- a/common/platform.h
+++ b/common/platform.h
@@ -36,6 +36,7 @@
#define PVR_POWER8 0x004D
#define PVR_POWER9 0x004E
#define PVR_POWER10 0x0080
+#define PVR_POWER11 0x0082
enum {
PLATFORM_UNKNOWN = 0,
@@ -56,6 +57,7 @@ enum {
POWER8,
POWER9,
POWER10,
+ POWER11,
};
extern const char *__platform_name[];

63
ppc64-diag-2.7.9-enable-light-path-diagnostics-for-RTAS-events.patch
View File

@ -1,63 +0,0 @@
commit 0fa486dbe800bea05c81fc33eee197873573fefb
Author: Sathvika Vasireddy <sv@linux.ibm.com>
Date: Fri Sep 8 12:35:14 2023 +0530
ppc64-diag/lp_diag: Enable light path diagnostics for RTAS events
Currently, Light Path Diagnostics support is enabled only for OS and
Enclosure type events. Enable light path diagnostics support for RTAS
type events by turning on only the high priority FRU callouts.
Signed-off-by: Sathvika Vasireddy <sv@linux.ibm.com>
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
diff --git a/lpd/lp_diag.c b/lpd/lp_diag.c
index e6f5d3c..e67db02 100644
--- a/lpd/lp_diag.c
+++ b/lpd/lp_diag.c
@@ -37,6 +37,8 @@
#include "lp_util.h"
#include "utils.h"
+static int rtas_event;
+
/* FRU callout priority as defined in PAPR+
*
* Note: Order of the priority is important!
@@ -173,8 +175,10 @@ service_event_supported(struct sl_event *event)
return 0;
}
break;
- case SL_TYPE_BMC:
case SL_TYPE_RTAS:
+ rtas_event = 1;
+ break;
+ case SL_TYPE_BMC:
case SL_TYPE_BASIC:
default:
return 0;
@@ -446,14 +450,20 @@ parse_service_event(int event_id)
attn_loc = &list[0];
if (operating_mode == LED_MODE_LIGHT_PATH) {
- if (event->callouts)
+ if (event->callouts) {
/* Run over FRU callout priority in order and
* enable fault indicator
*/
- for (i = 0; FRU_CALLOUT_PRIORITY[i]; i++)
+ if (!rtas_event) {
+ for (i = 0; FRU_CALLOUT_PRIORITY[i]; i++)
+ rc = event_fru_callout(event->callouts, list,
+ FRU_CALLOUT_PRIORITY[i],
+ &attn_on);
+ } else {
rc = event_fru_callout(event->callouts, list,
- FRU_CALLOUT_PRIORITY[i],
- &attn_on);
+ 'H', &attn_on);
+ }
+ }
else {
/* No callout list, enable check log indicator */
indicator_log_write("Empty callout list");

100
ppc64-diag-2.7.9-handle_multiple_platform_dumps.patch
View File

@ -1,100 +0,0 @@
commit d05654e5ec6f37cf6caa491fc7d95b336f9603e2
Author: Sathvika Vasireddy <sv@linux.ibm.com>
Date: Mon Jul 10 13:43:21 2023 +0530
rtas_errd: Handle multiple platform dumps
Currently, whenever a new dump arrives, old dump file of that specific dump
type is removed before writing the new dump out. Any dump file with the
same prefix (dump type) gets deleted. This means only one set of dump files
is saved, since only one dump file per dump type is saved.
Handle multiple dumps on Linux by allowing as many dumps to be offloaded
until disk space is available. To do this, remove the function that checks
for prefix size and removes old dump files. In the event of not enough
disk space available, log an error to the user along with the dump tag.
User will free up space and run extract_platdump tool using the dump tag
provided in the error message to offload the dump. Error log can be viewed
by the user by issuing 'journalctl -p err -t rtas_errd' command.
Signed-off-by: Sathvika Vasireddy <sv@linux.ibm.com>
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
diff --git a/rtas_errd/dump.c b/rtas_errd/dump.c
index cc50d91..494c322 100644
--- a/rtas_errd/dump.c
+++ b/rtas_errd/dump.c
@@ -30,8 +30,10 @@
#include <fcntl.h>
#include <librtas.h>
#include <librtasevent.h>
+#include <syslog.h>
#include <sys/stat.h>
#include <sys/wait.h>
+#include <sys/statvfs.h>
#include "utils.h"
#include "rtas_errd.h"
@@ -284,7 +286,9 @@ void
check_platform_dump(struct event *event)
{
struct rtas_dump_scn *dump_scn;
+ struct statvfs vfs;
uint64_t dump_tag;
+ uint64_t dump_size;
char filename[DUMP_MAX_FNAME_LEN + 20], *pos;
char *pathname = NULL;
FILE *f;
@@ -306,11 +310,34 @@ check_platform_dump(struct event *event)
return;
}
- /* Retrieve the dump */
+ /* Retrieve the dump tag */
dump_tag = dump_scn->id;
dump_tag |= ((uint64_t)dump_scn->v6hdr.subtype << 32);
dbg("Dump ID: 0x%016LX", dump_tag);
+ if (statvfs(d_cfg.platform_dump_path, &vfs) == -1) {
+ log_msg(event, "statvfs() failed on %s: %s",
+ d_cfg.platform_dump_path, strerror(errno));
+ return;
+ }
+
+ /* Retrieve the size of the platform dump */
+ dump_size = dump_scn->size_hi;
+ dump_size <<= 32;
+ dump_size |= dump_scn->size_lo;
+
+ /* Check if there is sufficient space in the file system to store the dump */
+ if (vfs.f_bavail * vfs.f_frsize < dump_size) {
+ syslog(LOG_ERR, "Insufficient space in %s to store platform dump for dump ID: "
+ "0x%016lX (required: %lu bytes, available: %lu bytes)",
+ d_cfg.platform_dump_path, dump_tag, dump_size,
+ (vfs.f_bavail * vfs.f_frsize));
+ syslog(LOG_ERR, "After clearing space, run 'extract_platdump "
+ "0x%016lX'.\n", dump_tag);
+ return;
+ }
+
+ /* Retrieve the dump */
snprintf(tmp_sys_arg, 60, "0x%016LX", (long long unsigned int)dump_tag);
system_args[0] = EXTRACT_PLATDUMP_CMD;
system_args[1] = tmp_sys_arg;
diff --git a/rtas_errd/extract_platdump.c b/rtas_errd/extract_platdump.c
index fbe65b2..831e57e 100644
--- a/rtas_errd/extract_platdump.c
+++ b/rtas_errd/extract_platdump.c
@@ -290,12 +290,6 @@ extract_platform_dump(uint64_t dump_tag)
}
}
- /*
- * Before writing the new dump out, we need to see if any older
- * dumps need to be removed first
- */
- remove_old_dumpfiles(filename, prefix_size);
-
/* Copy the dump off to the filesystem */
pathname[0] = '\0';
strcpy(pathname, d_cfg.platform_dump_path);

37
ppc64-diag-2.7.9-improvements_status_and_err_messages.patch
View File

@ -1,37 +0,0 @@
commit 3ad587d20868f154bb7ab972ee7812add1380d7a
Author: Greg Joyce <gjoyce@linux.vnet.ibm.com>
Date: Wed Oct 11 12:10:40 2023 -0500
diag_nvme: improvements to status and err messages
Signed-off-by: Greg Joyce <gjoyce@linux.vnet.ibm.com>
diff --git a/diags/diag_nvme.c b/diags/diag_nvme.c
index df191f2..00efec2 100644
--- a/diags/diag_nvme.c
+++ b/diags/diag_nvme.c
@@ -166,9 +166,9 @@ int main(int argc, char *argv[]) {
}
if (rc == 0)
- fprintf(stdout, "Command completed successfully\n");
+ fprintf(stdout, "NVMe diag command completed successfully\n");
else
- fprintf(stderr, "Command failed, exiting with rc %d\n", rc);
+ fprintf(stderr, "NVMe diag command failed with rc %d\n", rc);
return rc;
}
@@ -724,8 +724,11 @@ extern int get_smart_file(char *file_path, struct nvme_smart_log_page *log) {
int num_elements = 0;
struct dictionary dict[MAX_DICT_ELEMENTS];
- if ((num_elements = read_file_dict(file_path, dict, MAX_DICT_ELEMENTS)) < 0)
+ if ((num_elements = read_file_dict(file_path, dict, MAX_DICT_ELEMENTS)) < 0) {
+ fprintf(stderr, "read_file_dict failed: %s, rc % d\n",
+ file_path, num_elements);
return num_elements;
+ }
return set_smart_log_field(log, dict, num_elements);
}

136
ppc64-diag-2.7.9-moving-trim_trail_space-function.patch
View File

@ -1,136 +0,0 @@
commit c507319d1b5f0286d67e08a3598949ca4144f475
Author: Sathvika Vasireddy <sv@linux.ibm.com>
Date: Fri Sep 8 12:35:12 2023 +0530
ppc64-diag: Move trim_trail_space() function to common/utils.c
Currently, trim_trail_space() function is used in diags/diag_nvme.c file
to be able to trim trailing white spaces from a given location code. Allow
code reusability by moving the trim_trail_space() function from
diags/diag_nvme.c to common/utils.c.
Signed-off-by: Sathvika Vasireddy <sv@linux.ibm.com>
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
diff --git a/common/utils.c b/common/utils.c
index 0312943..2349878 100644
--- a/common/utils.c
+++ b/common/utils.c
@@ -24,9 +24,34 @@
#include <fcntl.h>
#include <string.h>
#include <assert.h>
+#include <ctype.h>
#include "utils.h"
+/* trim_trail_space - Trim trailing white spaces from string
+ * @string - Null terminated string to remove white spaces from
+ *
+ * This function will alter the passed string by removing any trailing white spaces and null
+ * terminating it at that point.
+ */
+void trim_trail_space(char *string)
+{
+ char *end;
+ size_t length;
+
+ if (string == NULL)
+ return;
+
+ length = strlen(string);
+ if (length == 0)
+ return;
+
+ end = string + length - 1;
+ while (end >= string && isspace(*end))
+ end--;
+ *(end + 1) = '\0';
+}
+
static int process_child(char *argv[], int pipefd[])
{
int nullfd;
diff --git a/common/utils.h b/common/utils.h
index ec2072d..2459b5b 100644
--- a/common/utils.h
+++ b/common/utils.h
@@ -18,6 +18,7 @@
#ifndef UTILS_H
#define UTILS_H
+void trim_trail_space(char *string);
FILE *spopen(char **, pid_t *);
int spclose(FILE *, pid_t);
diff --git a/diags/Makefile.am b/diags/Makefile.am
index 4ac81b8..dea0a79 100644
--- a/diags/Makefile.am
+++ b/diags/Makefile.am
@@ -13,7 +13,8 @@ encl_led_h_files = diags/encl_led.h \
$(diag_common_h_files)
diag_nvme_h_files = diags/diag_nvme.h \
- common/platform.h
+ common/platform.h \
+ common/utils.h
sbin_PROGRAMS += diags/diag_encl diags/encl_led diags/diag_nvme
@@ -41,6 +42,7 @@ diags_encl_led_SOURCES = diags/encl_led.c \
diags_diag_nvme_SOURCES = diags/diag_nvme.c \
common/platform.c \
+ common/utils.c \
$(diag_nvme_h_files)
diags_diag_nvme_LDADD = -lservicelog -lm
diags_diag_nvme_CFLAGS = $(AM_CFLAGS) -Wno-stringop-truncation
diff --git a/diags/diag_nvme.c b/diags/diag_nvme.c
index 2a78034..2606f2c 100644
--- a/diags/diag_nvme.c
+++ b/diags/diag_nvme.c
@@ -27,6 +27,7 @@
#include <sys/utsname.h>
#include "diag_nvme.h"
#include "platform.h"
+#include "utils.h"
#define ITEM_DATA_LENGTH 255
#define MIN_HOURS_ON 720
@@ -71,7 +72,6 @@ static int raw_data_smart(unsigned char **raw_data, uint32_t *raw_data_len, stru
static int raw_data_vpd(unsigned char **raw_data, uint32_t *raw_data_len, struct nvme_ibm_vpd *vpd);
static int regex_controller(char *controller_name, char *device_name);
static void set_notify(struct notify *notify, struct dictionary *dict, int num_elements);
-static void trim_trail_space(char *string);
static long double uint128_to_long_double(uint8_t *data);
int main(int argc, char *argv[]) {
@@ -1426,28 +1426,6 @@ extern void set_vpd_pcie_field(const char *keyword, const char *vpd_data, struct
strncpy(vpd->firmware_level, vpd_data, sizeof(vpd->firmware_level));
}
-/* trim_trail_space - Trim trailing white spaces from string
- * @string - Null terminated string to remove white spaces from
- *
- * This function will alter the passed string by removing any trailing white spaces and null
- * terminating it at that point.
- */
-static void trim_trail_space(char *string) {
- char *end;
- size_t length;
-
- if (string == NULL)
- return;
-
- if ((length = strlen(string)) == 0)
- return;
-
- end = string + length - 1;
- while (end >= string && isspace(*end))
- end--;
- *(end + 1) = '\0';
-}
-
static long double uint128_to_long_double(uint8_t *data) {
int i;
long double value = 0;

24
ppc64-diag-2.7.9-remove-d-and-f-from-usage.patch
View File

@ -1,24 +0,0 @@
commit 316d2baf5dba0d00447a3ac49f2f95339dbdf5cd
Author: Greg Joyce <gjoyce@linux.vnet.ibm.com>
Date: Wed Nov 15 11:28:05 2023 -0600
nvme_call_home: remove -d and -f from usage message
The -d and -f weren't removed from the first line of the
usage message.
Signed-off-by: Greg Joyce <gjoyce@linux.vnet.ibm.com>
diff --git a/diags/diag_nvme.c b/diags/diag_nvme.c
index 00efec2..164afe1 100644
--- a/diags/diag_nvme.c
+++ b/diags/diag_nvme.c
@@ -989,7 +989,7 @@ extern int open_nvme(char *dev_path) {
}
static void print_usage(char *command) {
- printf("Usage: %s [-h] [-d <file>] [-f <file>] [<nvme_devices>]\n"
+ printf("Usage: %s [-h] [<nvme_devices>]\n"
"\t-h or --help: print this help message\n"
"\t<nvme_devices>: the NVMe devices on which to operate, for\n"
"\t example nvme0; if not specified, all detected\n"

143
ppc64-diag-2.7.9-remove_d_and_f_option.patch
View File

@ -1,143 +0,0 @@
commit 11cb2a44a59b63bdc23c94e386c4e2f43ea7eb61
Author: Greg Joyce <gjoyce@linux.vnet.ibm.com>
Date: Fri Sep 22 15:19:34 2023 -0500
nvme_call_home: remove -d and -f as visible options
The dump(-d) and file(-f) options have been removed as option in usage
and also in the man page. The options are for internal test only and
should not have been exposed. Also added some helpful status messages.
Signed-off-by: Greg Joyce <gjoyce@linux.vnet.ibm.com>
diff --git a/diags/diag_nvme.c b/diags/diag_nvme.c
index 2a78034..df191f2 100644
--- a/diags/diag_nvme.c
+++ b/diags/diag_nvme.c
@@ -165,6 +165,11 @@ int main(int argc, char *argv[]) {
optind++;
}
+ if (rc == 0)
+ fprintf(stdout, "Command completed successfully\n");
+ else
+ fprintf(stderr, "Command failed, exiting with rc %d\n", rc);
+
return rc;
}
@@ -310,6 +315,7 @@ extern int dump_smart_data(char *device_name, char *dump_path) {
int fd, rc;
FILE *fp;
struct nvme_smart_log_page smart_log = { 0 };
+ char ans;
/* Read SMART data from device */
snprintf(dev_path,sizeof(dev_path), "/dev/%s", device_name);
@@ -331,8 +337,16 @@ extern int dump_smart_data(char *device_name, char *dump_path) {
}
fp = fopen(dump_path, "wx");
if (fp == NULL) {
- fprintf(stderr, "%s open failed: %s\n", dump_path, strerror(errno));
- return -1;
+ if (errno == EEXIST) {
+ fprintf(stdout, "File %s exists. Overwrite (y/n)? ", dump_path);
+ rc = scanf("%c", &ans);
+ if (ans == 'y' || ans == 'Y')
+ fp = fopen(dump_path, "w");
+ }
+ if (fp == NULL) {
+ fprintf(stderr, "%s open failed: %s\n", dump_path, strerror(errno));
+ return -1;
+ }
}
write_smart_file(fp, &smart_log);
fclose(fp);
@@ -974,11 +988,6 @@ extern int open_nvme(char *dev_path) {
static void print_usage(char *command) {
printf("Usage: %s [-h] [-d <file>] [-f <file>] [<nvme_devices>]\n"
"\t-h or --help: print this help message\n"
- "\t-d or --dump: dump SMART data to the specified path and file name <file>\n"
- "\t one <nvme_device> is expected with this option\n"
- "\t-f or --file: use SMART data from the specified path and file name <file>\n"
- "\t instead of device, one <nvme_device> is expected with\n"
- "\t this option\n"
"\t<nvme_devices>: the NVMe devices on which to operate, for\n"
"\t example nvme0; if not specified, all detected\n"
"\t nvme devices will be diagnosed\n", command);
diff --git a/diags/man/diag_nvme.8 b/diags/man/diag_nvme.8
index 45567ac..c023ed6 100644
--- a/diags/man/diag_nvme.8
+++ b/diags/man/diag_nvme.8
@@ -1,7 +1,7 @@
.\"
-.\" Copyright (C) 2022 IBM Corporation
+.\" Copyright (C) 2022, 2023 IBM Corporation
.\"
-.TH "DIAG_NVME" "8" "June 2022" "Linux" "PowerLinux Diagnostic Tools"
+.TH "DIAG_NVME" "8" "September 2023" "Linux" "PowerLinux Diagnostic Tools"
.hy
.SH NAME
.PP
@@ -12,7 +12,7 @@ diag_nvme \- diagnose NVMe devices
.PD 0
.P
.PD
-\f[B]diag_nvme\f[] [\-d <\f[I]file\f[]>] [\-f <\f[I]file\f[]>]
+\f[B]diag_nvme\f[]
<\f[B]nvme\f[]\f[I]n\f[]>
.PD 0
.P
@@ -29,52 +29,10 @@ go through the diagnostics procedure.
The user can control which events will be reported through the
configuration file \f[I]/etc/ppc64\-diag/diag_nvme.config\f[]
.SH OPTIONS
-.TP
-.B \f[B]\-d\f[], \f[B]--dump\f[] \f[I]file\f[]
-Dump SMART data to the specified path and file name \f[I]file\f[].
-The SMART data is extracted from an NVMe device, so specifying one is
-mandatory if this option is selected.
-File created is in a simple key=value format.
-.RS
-.RE
-.TP
-.B \f[B]\-f\f[], \f[B]--file\f[] \f[I]file\f[]
-This option usage is for testing only.
-Use SMART data from the specified path and file name \f[I]file\f[]
-instead of device, one NVMe is mandatory if this option is selected.
-The expected format of the file is a simple key=value that is the same
-one provided with the \-d / --dump option.
-If \f[I]file\f[] is missing from the filesystem it will be treated as a
-failure to retrieve SMART data and an event will be reported.
-.RS
-.RE
-.TP
-.B \f[B]\-h\f[], \f[B]--help\f[]
-Print a help message and exit
-.RS
-.RE
-.SH EXAMPLES
-.TP
-.B \f[B]diag_nvme\f[]
-Run diagnostics in all NVMe devices detected in the system.
-.RS
-.RE
-.TP
.B \f[B]diag_nvme nvme0 nvme1\f[]
Run diagnostics only in nvme0 and nvme1 devices.
.RS
.RE
-.TP
-.B \f[B]diag_nvme \-d smart.txt nvme0\f[]
-Dump SMART data from nvme0 into file smart.txt.
-.RS
-.RE
-.TP
-.B \f[B]diag_nvme \-f smart.txt nvme0\f[]
-Read SMART data from file smart.txt and use it as health information for
-diagnostics of device nvme0.
-.RS
-.RE
.SH REPORTING BUGS
.PP
Patches and issues may be submitted at

89
ppc64-diag-2.7.9-rtas_errd_crash.patch
View File

@ -1,89 +0,0 @@
commit 357923dafda766429acbe68470c3bd6028d5ebf5
Author: Sathvika Vasireddy <sv@linux.ibm.com>
Date: Sun May 5 02:48:25 2024 -0400
rtas_errd: Fix core dump with large filename passed to -f option
When a filename passed to rtas_errd's -f option is excessively long,
it causes a stack buffer overflow in the _log_msg() function, leading
to stack smashing.
The _log_msg() function, utilized by log_msg(), operates with a buffer
limited to 4096 characters. Additional space is consumed by _dbg() and
re-formatting in _log_msg(), increasing the risk of buffer overflow.
To address this, in case of an error, if the filename exceeds the
length of 2048 characters, truncate it and log a message indicating the
truncation. Use 2048 characters as limit for filename length to allow
space for further modifications.
Additionally, add a check when processing the -f option to ensure that
the length of the specified filename does not exceed 4096 characters.
Reported-by: R Nageswara Sastry <rnsastry@linux.ibm.com>
Closes: https://github.com/power-ras/ppc64-diag/issues/31
Signed-off-by: Sathvika Vasireddy <sv@linux.ibm.com>
Tested-by: R Nageswara Sastry<rnsastry@linux.ibm.com>
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
diff --git a/rtas_errd/files.c b/rtas_errd/files.c
index 4be75c6..3750bdc 100644
--- a/rtas_errd/files.c
+++ b/rtas_errd/files.c
@@ -284,10 +284,27 @@ init_files(void)
proc_error_log_fd = open(proc_error_log2, O_RDONLY);
if (proc_error_log_fd < 0) {
- log_msg(NULL, "Could not open error log file at either %s or "
- "%s, %s\nThe rtas_errd daemon cannot continue and will "
- "exit", proc_error_log1, proc_error_log2,
- strerror(errno));
+
+ if (strlen(proc_error_log1) <= 2048) {
+ log_msg(NULL, "Could not open error log file at either %s or "
+ "%s, %s\nThe rtas_errd daemon cannot continue and will "
+ "exit", proc_error_log1, proc_error_log2,
+ strerror(errno));
+ } else {
+ /* If the filename is too long, truncate it */
+ char truncated_filename[2048];
+
+ strncpy(truncated_filename, proc_error_log1,
+ sizeof(truncated_filename) - 1);
+ truncated_filename[sizeof(truncated_filename) - 1] = '\0';
+
+ log_msg(NULL, "Truncating the filename since it is too long "
+ "(%d characters)", strlen(proc_error_log1));
+ log_msg(NULL, "Could not open error log file at either %s or "
+ "%s, %s\nThe rtas_errd daemon cannot continue and will "
+ "exit", truncated_filename, proc_error_log2,
+ strerror(errno));
+ }
return -1;
}
diff --git a/rtas_errd/rtas_errd.c b/rtas_errd/rtas_errd.c
index 10c422c..f6c6a4a 100644
--- a/rtas_errd/rtas_errd.c
+++ b/rtas_errd/rtas_errd.c
@@ -30,6 +30,7 @@
#include <sys/stat.h>
#include <sys/wait.h>
#include <librtas.h>
+#include <limits.h>
#include "rtas_errd.h"
#include "platform.h"
@@ -474,6 +475,12 @@ main(int argc, char *argv[])
f_flag++;
proc_error_log1 = optarg;
+ if (strlen(proc_error_log1) > PATH_MAX) {
+ fprintf(stderr, "The specified file path %s exceeds"
+ " the maximum supported file length of 4096 characters.\n",
+ proc_error_log1);
+ goto error_out;
+ }
proc_error_log2 = NULL;
break;

37
ppc64-diag-2.7.9-utilize-trim_trail_space-in-event_fru_callout.patch
View File

@ -1,37 +0,0 @@
commit 476b0af7516b86c4d98cfa229fb0c6b856eea31d
Author: Sathvika Vasireddy <sv@linux.ibm.com>
Date: Fri Sep 8 12:35:13 2023 +0530
ppc64-diag/lp_diag: Utilize trim_trail_space() function in event_fru_callout()
Update the event_fru_callout() function to use the trim_trail_space()
function to be able to remove any trailing spaces from the location code.
This change aims to address an issue where the presence of trailing spaces
in the location code results in failure to find an indicator for the given
location code. Use trim_trail_space() on the location to ensure that the
device location code is properly compared with the indicator list.
Signed-off-by: Sathvika Vasireddy <sv@linux.ibm.com>
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
diff --git a/lpd/lp_diag.c b/lpd/lp_diag.c
index 988a021..e6f5d3c 100644
--- a/lpd/lp_diag.c
+++ b/lpd/lp_diag.c
@@ -35,6 +35,7 @@
#include "servicelog.h"
#include "indicator.h"
#include "lp_util.h"
+#include "utils.h"
/* FRU callout priority as defined in PAPR+
*
@@ -344,6 +345,8 @@ event_fru_callout(struct sl_callout *callouts, struct loc_code *list,
/* get FRUs nearest fault indicator */
strncpy(location, callout->location, LOCATION_LENGTH);
location[LOCATION_LENGTH - 1] = '\0';
+ trim_trail_space(location);
+
loc_led = get_fru_indicator(list, location, &truncated);
if (!loc_led) { /* No indicator found for the given loc code */
*attn_on = 1;

20
ppc64-diag.spec
View File

@ -1,6 +1,6 @@
Name: ppc64-diag
Version: 2.7.9
Release: 6%{?dist}
Version: 2.7.11
Release: 1%{?dist}
Summary: PowerLinux Platform Diagnostics
URL: https://github.com/power-ras/%{name}
License: GPLv2
@ -35,19 +35,6 @@ Source5: rtas_errd.8
Patch0: ppc64-diag-2.7.9-fedora.patch
# upstream fixes
Patch10: ppc64-diag-2.7.9-handle_multiple_platform_dumps.patch
Patch11: ppc64-diag-2.7.9-remove_d_and_f_option.patch
Patch12: ppc64-diag-2.7.9-moving-trim_trail_space-function.patch
Patch13: ppc64-diag-2.7.9-utilize-trim_trail_space-in-event_fru_callout.patch
Patch14: ppc64-diag-2.7.9-enable-light-path-diagnostics-for-RTAS-events.patch
Patch15: ppc64-diag-2.7.9-improvements_status_and_err_messages.patch
Patch16: ppc64-diag-2.7.9-remove-d-and-f-from-usage.patch
Patch17: ppc64-diag-2.7.9-diag_encl-man-and-help-page.patch
Patch18: ppc64-diag-2.7.9-diag_nvme-man-and-help-page.patch
Patch19: ppc64-diag-2.7.9-call_home-fail-on-nvmf-device.patch
Patch20: ppc64-diag-2.7.9-compiler-warnings.patch
Patch21: ppc64-diag-2.7.9-rtas_errd_crash.patch
Patch22: ppc64-diag-2.7.9-diag_vnme-add-power11-support.patch
%description
This package contains various diagnostic tools for PowerLinux.
@ -184,6 +171,9 @@ if [ "$1" = "0" ]; then # last uninstall
fi
%changelog
* Tue Jan 27 2026 Than Ngo <than@redhat.com> - 2.7.11-1
- Resolves: RHEL-102474, Update to 2.7.11
* Wed Jan 08 2025 Than Ngo <than@redhat.com> - 2.7.9-6
- Resolves: RHEL-72515, Added Power11 support for diag_nvme

2
sources
View File

@ -1 +1 @@
SHA512 (ppc64-diag-2.7.9.tar.gz) = b6756c17e67e876ad3a945d05dbd2c2523cbf121c4aa658a4edaf33775c8165ff26a252bdf812600e2d47ae69cd68f6e22c0de6f2d3994856adcd3d949322673
SHA512 (ppc64-diag-2.7.11.tar.gz) = 651e8faca5bdf092625424239afc5d6f0062ad9d001f70a4759fe9179b021494516e129606d0ae3eedfd2a2d45ad21983fb7fa8836d5310afe9722d3f7b7cd15