!240 fix period task some bugs and fix env config

From: @tong_1001 Reviewed-by: @hubin95 Signed-off-by: @hubin95
fix period task some bugs and fix env config
2025-03-29 07:43:30 +00:00 · 2025-03-29 14:49:26 +08:00 · 2025-03-14 08:42:52 +00:00 · 2025-03-14 16:16:25 +08:00 · 2025-03-14 02:18:58 +00:00 · 2025-03-14 10:05:39 +08:00
7 changed files with 1014 additions and 1 deletions
--- a/add-log-utils-for-c.patch
+++ b/add-log-utils-for-c.patch
@ -0,0 +1,203 @@
 From 0ee8307d556c200733270fdffd8db2d48869724a Mon Sep 17 00:00:00 2001
 From: shixuantong <shixuantong1@huawei.com>
 Date: Fri, 14 Mar 2025 14:55:56 +0800
 Subject: [PATCH] add log utils for c
 ---
 Makefile                           | 14 ++++++-
 src/libsentry/c/log/CMakeLists.txt |  7 ++++
 src/libsentry/c/log/log_utils.c    | 45 +++++++++++++++++++++
 src/libsentry/c/log/log_utils.h    | 64 ++++++++++++++++++++++++++++++
 4 files changed, 129 insertions(+), 1 deletion(-)
 create mode 100644 src/libsentry/c/log/CMakeLists.txt
 create mode 100644 src/libsentry/c/log/log_utils.c
 create mode 100644 src/libsentry/c/log/log_utils.h
 diff --git a/Makefile b/Makefile
 index 29c4a53..73ada63 100644
 --- a/Makefile
 +++ b/Makefile
@@ -29,12 +29,16 @@ PKGVEREGG := syssentry-$(VERSION)-py$(PYTHON_VERSION).egg-info
 all: lib ebpf hbm_online_repair
 -lib:libxalarm
 +lib:libxalarm log
 libxalarm:
 	cd $(CURLIBDIR) && cmake . -DXD_INSTALL_BINDIR=$(LIBINSTALLDIR) -B build
 	cd $(CURLIBDIR)/build && make
 +log:
 +	cd $(CURSRCDIR)/libsentry/c/log && cmake . -B build
 +	cd $(CURSRCDIR)/libsentry/c/log/build && make
 +
 ebpf:
 	@if [ -d "$(CURSRCDIR)/services/sentryCollector/ebpf_collector/" ]; then \
 		cd $(CURSRCDIR)/services/sentryCollector/ebpf_collector/ && make;	\
@@ -134,6 +138,11 @@ isentry:
 	# pyxalarm
 	install -m 550 src/libs/pyxalarm/register_xalarm.py $(PYDIR)/xalarm
 +	
 +	# log utils
 +	install -d -m 700 $(INCLUDEDIR)/libsentry
 +	install -m 644 $(CURSRCDIR)/libsentry/c/log/log_utils.h $(INCLUDEDIR)/libsentry/
 +	install -m 550 $(CURSRCDIR)/libsentry/c/log/build/libsentry_log.so $(LIBINSTALLDIR)
 ebpf_clean:
 	cd $(CURSRCDIR)/services/sentryCollector/ebpf_collector && make clean
@@ -144,6 +153,7 @@ hbm_clean:
 clean: ebpf_clean hbm_clean
 	rm -rf $(CURLIBDIR)/build
 	rm -rf $(CURSRCDIR)/build
 +	rm -rf $(CURSRCDIR)/libsentry/c/log/build
 	rm -rf $(CURSRCDIR)/syssentry.egg-info
 	rm -rf $(CURSRCDIR)/SENTRY_FILES
@@ -156,6 +166,8 @@ uninstall:
 	rm -rf $(BINDIR)/ebpf_collector
 	rm -rf $(LIBINSTALLDIR)/libxalarm.so
 	rm -rf $(INCLUDEDIR)/xalarm
 +	rm -rf $(LIBINSTALLDIR)/libsentry_log.so
 +	rm -rf $(INCLUDEDIR)/libsentry
 	rm -rf $(ETCDIR)/sysSentry
 	rm -rf $(ETCDIR)/hbm_online_repair.env
 	rm -rf $(LOGSAVEDIR)/sysSentry
 diff --git a/src/libsentry/c/log/CMakeLists.txt b/src/libsentry/c/log/CMakeLists.txt
 new file mode 100644
 index 0000000..6488195
 --- /dev/null
 +++ b/src/libsentry/c/log/CMakeLists.txt
@@ -0,0 +1,7 @@
 +# Copyright (c) Huawei Technologies Co., Ltd. 2023-2025. All rights reserved.
 +# Description: cmake file for log_utils
 +project(sentry_log)
 +cmake_minimum_required(VERSION 3.22)
 +add_library(sentry_log SHARED log_utils.c)
 +set_target_properties(sentry_log PROPERTIES LINK_FLAGS "-Wl,-z,relro -Wl,-z,now -Wl,-z,noexecstack -Wtrampolines")
 +set_target_properties(sentry_log PROPERTIES CMAKE_C_FLAGS "-shared -fPIC -fstack-protector-strong -D_FORTIFY_SOURCE=2 -O2 -Wall -Werror -g")
 diff --git a/src/libsentry/c/log/log_utils.c b/src/libsentry/c/log/log_utils.c
 new file mode 100644
 index 0000000..935e6d6
 --- /dev/null
 +++ b/src/libsentry/c/log/log_utils.c
@@ -0,0 +1,45 @@
 +/*
 + * Copyright (c) Huawei Technologies Co., Ltd. 2023-2025. All rights reserved.
 + * Description: log utils for sysSentry
 + * Author: sxt1001
 + * Create: 2025-2-16
 + */
 +
 +#include "log_utils.h"
 +
 +static LogLevel currentLogLevel = LOG_INFO;
 +
 +void logMessage(LogLevel level, char* file, int line, const char *format, ...)
 +{
 +    if (level >= currentLogLevel) {
 +        PRINT_LOG_PREFIX(level, file, line);
 +        va_list args;
 +        va_start(args, format);
 +        vfprintf(LOG_FD(level), format, args);
 +        va_end(args);
 +        fflush(LOG_FD(level));
 +    }
 +}
 +
 +void setLogLevel()
 +{
 +    currentLogLevel = LOG_INFO;
 +    char* levelStr = getenv(LOG_LEVEL_ENV);
 +    if (levelStr == NULL) {
 +        logMessage(LOG_WARN, __FILE__, __LINE__, "getenv('%s') is NULL, use default log level : %s\n", LOG_LEVEL_ENV, LOG_LEVEL_STRING(LOG_INFO));
 +    } else if (strcmp(levelStr, "info") == 0) {
 +        currentLogLevel = LOG_INFO;
 +        logMessage(LOG_INFO, __FILE__, __LINE__, "Set log level : %s\n", LOG_LEVEL_STRING(LOG_INFO));
 +    } else if (strcmp(levelStr, "warning") == 0) { 
 +        currentLogLevel = LOG_WARN;
 +        logMessage(LOG_INFO,__FILE__, __LINE__,"Set log level : %s\n", LOG_LEVEL_STRING(LOG_WARN));
 +    } else if (strcmp(levelStr, "error") == 0) {
 +        currentLogLevel = LOG_ERROR;
 +        logMessage(LOG_INFO,__FILE__, __LINE__,"Set log level : %s\n", LOG_LEVEL_STRING(LOG_ERROR));
 +    } else if (strcmp(levelStr, "debug") == 0) { 
 +        currentLogLevel = LOG_DEBUG;
 +        logMessage(LOG_INFO,__FILE__, __LINE__,"Set log level : %s\n", LOG_LEVEL_STRING(LOG_DEBUG));
 +    } else {
 +        logMessage(LOG_WARN, __FILE__, __LINE__, "unknown log level : %s,  use default log level : %s\n", levelStr, LOG_LEVEL_STRING(LOG_INFO));
 +    }
 +}
 diff --git a/src/libsentry/c/log/log_utils.h b/src/libsentry/c/log/log_utils.h
 new file mode 100644
 index 0000000..8a56520
 --- /dev/null
 +++ b/src/libsentry/c/log/log_utils.h
@@ -0,0 +1,64 @@
 +/*
 + * Copyright (c) Huawei Technologies Co., Ltd. 2023-2025. All rights reserved.
 + * Description: log utils for sysSentry
 + * Author: sxt1001
 + * Create: 2025-2-16
 + */
 +
 +#ifndef _SYSSENTRY_LOG_H
 +#define _SYSSENTRY_LOG_H
 +
 +#include "stdio.h"
 +#include <stdlib.h>
 +#include "string.h"
 +#include "stdarg.h"
 +#include "time.h"
 +#include "libgen.h"
 +
 +typedef enum {
 +    LOG_DEBUG = 0,
 +    LOG_INFO,
 +    LOG_WARN,
 +    LOG_ERROR,
 +} LogLevel;
 +
 +#define LOG_FD(level) (level == LOG_ERROR ? stderr : stdout)
 +
 +#define LOG_LEVEL_STRING(level) \
 +    (level == LOG_DEBUG ? "DEBUG":   \
 +     level == LOG_INFO ? "INFO" :    \
 +     level == LOG_WARN ? "WARNING" : \
 +     level == LOG_ERROR ? "ERROR" :  \
 +     "UNKNOWN_LEVEL")
 +
 +#define PRINT_LOG_PREFIX(level, file, line) do {                    \
 +    time_t t = time(NULL);                                          \
 +    struct tm *local_time = localtime(&t);                          \
 +    fprintf(LOG_FD(level), "%d-%02d-%02d %02d:%02d:%02d,000 - %s - [%s:%d] - ", \
 +            local_time->tm_year + 1900,                             \
 +            local_time->tm_mon + 1,                                 \
 +            local_time->tm_mday,                                    \
 +            local_time->tm_hour,                                    \
 +            local_time->tm_min,                                     \
 +            local_time->tm_sec,                                     \
 +            LOG_LEVEL_STRING(level),                                \
 +            basename(file),                                         \
 +            line);                                                  \
 +} while (0)
 +
 +// configure Env for log 
 +#define LOG_LEVEL_ENV "LOG_LEVEL"
 +
 +// print msg
 +void logMessage(LogLevel level, char* file, int line, const char *format, ...);
 +
 +// set log level
 +void setLogLevel();
 +
 +// log function
 +#define logging_debug(...)       logMessage(LOG_DEBUG, __FILE__, __LINE__, __VA_ARGS__)
 +#define logging_info(...)        logMessage(LOG_INFO, __FILE__, __LINE__, __VA_ARGS__)
 +#define logging_warn(...)        logMessage(LOG_WARN, __FILE__, __LINE__, __VA_ARGS__)
 +#define logging_error(...)       logMessage(LOG_ERROR, __FILE__, __LINE__, __VA_ARGS__)
 +
 +#endif
 -- 
 2.27.0
--- a/add-new-func-for-ebpf-in-the-rq_driver-stage.patch
+++ b/add-new-func-for-ebpf-in-the-rq_driver-stage.patch
@ -0,0 +1,586 @@
 From 966e539d3b0c0eaaa94fdd1fb21dd29e97e48bee Mon Sep 17 00:00:00 2001
 From: zhuofeng <zhuofeng2@huawei.com>
 Date: Sat, 22 Feb 2025 18:42:08 +0800
 Subject: [PATCH] add new func for ebpf in the rq_driver stage
 ---
 src/sentryPlugins/ai_block_io/detector.py     |   3 +-
 .../ebpf_collector/ebpf_collector.bpf.c       | 459 ++++--------------
 .../ebpf_collector/ebpf_collector.c           |   2 +-
 .../ebpf_collector/ebpf_collector.h           |  16 +-
 4 files changed, 116 insertions(+), 364 deletions(-)
 diff --git a/src/sentryPlugins/ai_block_io/detector.py b/src/sentryPlugins/ai_block_io/detector.py
 index 27fb7f7..2688cb1 100644
 --- a/src/sentryPlugins/ai_block_io/detector.py
 +++ b/src/sentryPlugins/ai_block_io/detector.py
@@ -55,11 +55,12 @@ class Detector:
         detection_result = self._slidingWindow.is_slow_io_event(metric_value)
         # 检测到慢周期，由Detector负责打印info级别日志
         if detection_result[0][1]:
 +            ai_threshold = "None" if detection_result[2] is None else round(detection_result[2], 3)
             logging.info(f'[abnormal_period]: disk: {self._metric_name.disk_name}, '
                          f'stage: {self._metric_name.stage_name}, '
                          f'iotype: {self._metric_name.io_access_type_name}, '
                          f'type: {self._metric_name.metric_name}, '
 -                         f'ai_threshold: {round(detection_result[2], 3)}, '
 +                         f'ai_threshold: {ai_threshold}, '
                          f'curr_val: {metric_value}')
         else:
             logging.debug(f'Detection result: {str(detection_result)}')
 diff --git a/src/services/sentryCollector/ebpf_collector/ebpf_collector.bpf.c b/src/services/sentryCollector/ebpf_collector/ebpf_collector.bpf.c
 index 978b114..ece8c93 100644
 --- a/src/services/sentryCollector/ebpf_collector/ebpf_collector.bpf.c
 +++ b/src/services/sentryCollector/ebpf_collector/ebpf_collector.bpf.c
@@ -115,7 +115,7 @@ struct {
 } ringbuf SEC(".maps");
 -static void log_event(u32 stage, u32 period, u32 err) {
 +static void log_event(int stage, int period, int err) {
     struct event *e;
     void *data = bpf_ringbuf_reserve(&ringbuf, sizeof(struct event), 0);
     if (!data)
@@ -338,6 +338,105 @@ int kprobe_blk_mq_start_request(struct pt_regs *regs) {
     return 0;
 }
 +SEC("kprobe/blk_mq_end_request_batch")
 +int kprobe_blk_mq_end_request_batch(struct pt_regs *regs) {
 +    struct io_comp_batch *iob = (struct io_comp_batch *)PT_REGS_PARM1(regs);
 +    struct request *rq;
 +    struct request_queue *q;
 +    struct gendisk *curr_rq_disk;
 +    int major, first_minor;
 +    unsigned int cmd_flags;
 +    struct io_counter *counterp;
 +    struct stage_data *curr_data;
 +    rq = BPF_CORE_READ(iob, req_list);
 +
 +    for (int i = 0; i <= BATCH_COUT; i++) {
 +        bpf_core_read(&q, sizeof(q), &rq->q);
 +        bpf_core_read(&curr_rq_disk, sizeof(curr_rq_disk), &q->disk);
 +        bpf_core_read(&major, sizeof(major), &curr_rq_disk->major);
 +        bpf_core_read(&first_minor, sizeof(first_minor), &curr_rq_disk->first_minor);
 +        bpf_core_read(&cmd_flags, sizeof(cmd_flags), &rq->cmd_flags);
 +
 +        if (major == 0) {
 +            log_event(STAGE_RQ_DRIVER, PERIOD_END, ERROR_MAJOR_ZERO);
 +            continue;
 +        }
 +
 +        u32 key = find_matching_key_rq_driver(major, first_minor);
 +        if (key >= MAP_SIZE) {
 +            continue;
 +        }
 +
 +        counterp = bpf_map_lookup_elem(&blk_map, &rq);
 +        if (!counterp) {
 +            continue;
 +        }
 +
 +        u64 duration = bpf_ktime_get_ns() - counterp->start_time;
 +        u64 curr_start_range = counterp->start_time / THRESHOLD;
 +
 +        struct update_params params = {
 +            .major = major,
 +            .first_minor = first_minor,
 +            .cmd_flags = cmd_flags,
 +            .curr_start_range = curr_start_range,
 +        };
 +
 +        curr_data = bpf_map_lookup_elem(&blk_res, &key);
 +        if (curr_data == NULL && duration > DURATION_THRESHOLD) { 
 +            struct stage_data new_data = {
 +                .start_count = 1,
 +                .finish_count = 1,
 +                .finish_over_time = 1,
 +                .duration = 0,
 +                .major = major,
 +                .first_minor = first_minor,
 +                .io_type = "",
 +            };
 +            blk_fill_rwbs(new_data.io_type, cmd_flags);
 +            bpf_map_update_elem(&blk_res, &key, &new_data, 0);
 +        } else if (curr_data == NULL) { 
 +            struct stage_data new_data = {
 +                .start_count = 1,
 +                .finish_count = 1,
 +                .finish_over_time = 0,
 +                .duration = 0,
 +                .major = major,
 +                .first_minor = first_minor,
 +                .io_type = "",
 +            };
 +            blk_fill_rwbs(new_data.io_type, cmd_flags);
 +            bpf_map_update_elem(&blk_res, &key, &new_data, 0); 
 +        } else {
 +            curr_data->duration += duration; 
 +            update_curr_data_in_finish(curr_data, &params, duration);
 +        }
 +
 +        struct time_range_io_count *curr_data_time_range;
 +        curr_data_time_range = bpf_map_lookup_elem(&blk_res_2, &curr_start_range);
 +        if (curr_data_time_range == NULL) { 
 +            struct time_range_io_count new_data = {.count = 0};
 +            bpf_map_update_elem(&blk_res_2, &curr_start_range, &new_data, 0);
 +        } else {
 +            if (key < MAP_SIZE && key >= 0) {
 +                __sync_fetch_and_add(&curr_data_time_range->count[key], -1);
 +            }
 +        }
 +        struct request *rq_next = BPF_CORE_READ(rq, rq_next);
 +        bpf_map_delete_elem(&blk_map, &rq); 
 +        rq = rq_next;
 +        
 +        if (!rq) {
 +            break;
 +        }
 +
 +        if (i >= BATCH_COUT) {
 +            log_event(STAGE_RQ_DRIVER, PERIOD_END, i);
 +        }
 +    }
 +    return 0;
 +}
 +
 // finish rq_driver
 SEC("kprobe/blk_mq_free_request")
 int kprobe_blk_mq_free_request(struct pt_regs *regs)
@@ -418,7 +517,7 @@ int kprobe_blk_mq_free_request(struct pt_regs *regs)
         struct time_range_io_count new_data = { .count = {0} };
         bpf_map_update_elem(&blk_res_2, &curr_start_range, &new_data, 0);
     } else {
 -        if (key < MAP_SIZE && curr_data_time_range->count[key] > 0) {
 +        if (key < MAP_SIZE && key >= 0) {
             __sync_fetch_and_add(&curr_data_time_range->count[key], -1);
         }
     }
@@ -463,7 +562,6 @@ int kprobe_blk_mq_submit_bio(struct pt_regs *regs)
     long err = bpf_map_update_elem(&bio_map, &bio, &zero, BPF_NOEXIST); 
     if (err) {
 -        log_event(STAGE_BIO, PERIOD_START, ERROR_UPDATE_FAIL);
         return 0;
     }
@@ -597,359 +695,4 @@ int kprobe_bio_endio(struct pt_regs *regs)
     return 0; 
 }
 -// start get_tag 
 -SEC("kprobe/blk_mq_get_tag") 
 -int kprobe_blk_mq_get_tag(struct pt_regs *regs) 
 -{
 -    u64 tagkey = bpf_get_current_task();
 -    u64 value = (u64)PT_REGS_PARM1(regs);
 -    (void)bpf_map_update_elem(&tag_args, &tagkey, &value, BPF_ANY);
 -
 -    struct blk_mq_alloc_data *bd;
 -    struct request_queue *q;
 -    struct gendisk *curr_rq_disk;
 -    int major, first_minor;
 -    unsigned int cmd_flags = 0;
 -
 -    bd = (struct blk_mq_alloc_data *)value;
 -    bpf_core_read(&q, sizeof(q), &bd->q);
 -    bpf_core_read(&curr_rq_disk, sizeof(curr_rq_disk), &q->disk);
 -    bpf_core_read(&major, sizeof(major), &curr_rq_disk->major);
 -    bpf_core_read(&first_minor, sizeof(first_minor), &curr_rq_disk->first_minor);
 -
 -    if (major == 0) {
 -        log_event(STAGE_GET_TAG, PERIOD_START, ERROR_MAJOR_ZERO);
 -        return 0;
 -    }
 -
 -    u32 key = find_matching_key_get_tag(major, first_minor);
 -    if (key >= MAP_SIZE) {
 -        return 0;
 -    }
 -
 -    struct io_counter *counterp, zero = {}; 
 -    init_io_counter(&zero, major, first_minor);
 -    counterp = bpf_map_lookup_elem(&tag_map, &tagkey);
 -    if (counterp) {
 -        return 0;
 -    }
 -    long err = bpf_map_update_elem(&tag_map, &tagkey, &zero, BPF_NOEXIST); 
 -    if (err) {
 -        log_event(STAGE_GET_TAG, PERIOD_START, ERROR_UPDATE_FAIL);
 -        return 0;
 -    }
 -
 -    u64 curr_start_range = zero.start_time / THRESHOLD;
 -
 -    struct update_params params = {
 -        .major = major,
 -        .first_minor = first_minor,
 -        .cmd_flags = cmd_flags,
 -        .curr_start_range = curr_start_range,
 -    };
 -
 -    struct stage_data *curr_data; 
 -    curr_data = bpf_map_lookup_elem(&tag_res, &key);
 -    if (!curr_data) { 
 -        struct stage_data new_data = {
 -            .start_count = 1,
 -            .finish_count = 0,
 -            .finish_over_time = 0,
 -            .duration = 0,
 -            .major = major,
 -            .first_minor = first_minor,
 -            .io_type = "",
 -        };
 -        blk_fill_rwbs(new_data.io_type, cmd_flags);
 -        bpf_map_update_elem(&tag_res, &key, &new_data, 0);
 -    } else { 
 -        update_curr_data_in_start(curr_data, &params);
 -    }
 -
 -    struct time_range_io_count *curr_data_time_range;
 -    curr_data_time_range = bpf_map_lookup_elem(&tag_res_2, &curr_start_range);
 -    if (curr_data_time_range == NULL) {
 -        struct time_range_io_count new_data = { .count = {0} };
 -        bpf_map_update_elem(&tag_res_2, &curr_start_range, &new_data, 0); 
 -    } else {
 -        if (key < MAP_SIZE && key >= 0) {
 -            __sync_fetch_and_add(&curr_data_time_range->count[key], 1);
 -        }
 -    }
 -    return 0; 
 -}
 -
 -// finish get_tag 
 -SEC("kretprobe/blk_mq_get_tag") 
 -int kretprobe_blk_mq_get_tag(struct pt_regs *regs) 
 -{
 -    u64 tagkey = bpf_get_current_task();
 -    u64 *tagargs = (u64 *)bpf_map_lookup_elem(&tag_args, &tagkey);
 -    if (tagargs == NULL) {
 -        bpf_map_delete_elem(&tag_args, &tagkey);
 -        return 0;
 -    }
 -
 -    struct blk_mq_alloc_data *bd;
 -    struct request_queue *q;
 -    struct gendisk *curr_rq_disk;
 -    int major, first_minor;
 -    unsigned int cmd_flags = 0;
 -
 -    bd = (struct blk_mq_alloc_data *)*tagargs;
 -    bpf_core_read(&q, sizeof(q), &bd->q);
 -    bpf_core_read(&curr_rq_disk, sizeof(curr_rq_disk), &q->disk);
 -    bpf_core_read(&major, sizeof(major), &curr_rq_disk->major);
 -    bpf_core_read(&first_minor, sizeof(first_minor), &curr_rq_disk->first_minor);
 -
 -    if (major == 0) {
 -        log_event(STAGE_GET_TAG, PERIOD_END, ERROR_MAJOR_ZERO);
 -        return 0;
 -    }
 -
 -    u32 key = find_matching_key_get_tag(major, first_minor);
 -    if (key >= MAP_SIZE) {
 -        return 0;
 -    }
 -
 -    struct io_counter *counterp = bpf_map_lookup_elem(&tag_map, &tagkey);
 -    if (!counterp) {
 -        return 0;
 -    }
 -
 -    u64 duration = bpf_ktime_get_ns() - counterp->start_time; 
 -    u64 curr_start_range = counterp->start_time / THRESHOLD;
 -
 -    struct update_params params = {
 -        .major = major,
 -        .first_minor = first_minor,
 -        .cmd_flags = cmd_flags,
 -        .curr_start_range = curr_start_range,
 -    };
 -
 -    struct stage_data *curr_data; 
 -    curr_data = bpf_map_lookup_elem(&tag_res, &key); 
 -    if (curr_data == NULL && duration > DURATION_THRESHOLD) { 
 -        struct stage_data new_data = {
 -            .start_count = 1,
 -            .finish_count = 1,
 -            .finish_over_time = 1,
 -            .duration = 0,
 -            .major = major,
 -            .first_minor = first_minor,
 -            .io_type = "",
 -        };
 -        blk_fill_rwbs(new_data.io_type, cmd_flags);
 -        bpf_map_update_elem(&tag_res, &key, &new_data, 0); 
 -    } else if (curr_data == NULL) { 
 -        struct stage_data new_data = {
 -            .start_count = 1,
 -            .finish_count = 1,
 -            .finish_over_time = 0,
 -            .duration = 0,
 -            .major = major,
 -            .first_minor = first_minor,
 -            .io_type = "",
 -        };
 -        blk_fill_rwbs(new_data.io_type, cmd_flags);
 -        bpf_map_update_elem(&tag_res, &key, &new_data, 0); 
 -    } else {
 -        curr_data->duration += duration; 
 -        update_curr_data_in_finish(curr_data, &params, duration);
 -    }
 -
 -    struct time_range_io_count *curr_data_time_range;
 -    curr_data_time_range = bpf_map_lookup_elem(&tag_res_2, &curr_start_range);
 -    if (curr_data_time_range == NULL) { 
 -        struct time_range_io_count new_data = { .count = {0} };
 -        bpf_map_update_elem(&tag_res_2, &curr_start_range, &new_data, 0);
 -    } else {
 -        if (key < MAP_SIZE && curr_data_time_range->count[key] > 0) {
 -            __sync_fetch_and_add(&curr_data_time_range->count[key], -1);
 -        }
 -    }
 -
 -    bpf_map_delete_elem(&tag_map, &tagkey); 
 -    bpf_map_delete_elem(&tag_args, &tagkey);
 -    return 0; 
 -}
 -
 -// start wbt
 -SEC("kprobe/wbt_wait") 
 -int kprobe_wbt_wait(struct pt_regs *regs) 
 -{
 -    u64 wbtkey = bpf_get_current_task();
 -    u64 value = (u64)PT_REGS_PARM2(regs);
 -    (void)bpf_map_update_elem(&wbt_args, &wbtkey, &value, BPF_ANY);
 -    
 -    struct bio *bio;
 -    struct block_device *bd;
 -    struct gendisk *curr_rq_disk;
 -    int major, first_minor;
 -    unsigned int cmd_flags;
 -
 -    bio = (struct bio *)value;
 -    bpf_core_read(&bd, sizeof(bd), &bio->bi_bdev);
 -    bpf_core_read(&curr_rq_disk, sizeof(curr_rq_disk), &bd->bd_disk);
 -    bpf_core_read(&major, sizeof(major), &curr_rq_disk->major);
 -    bpf_core_read(&first_minor, sizeof(first_minor), &curr_rq_disk->first_minor);
 -    bpf_core_read(&cmd_flags, sizeof(cmd_flags), &bio->bi_opf);
 -
 -    if (major == 0) {
 -        log_event(STAGE_WBT, PERIOD_START, ERROR_MAJOR_ZERO);
 -        return 0;
 -    }
 -
 -    u32 key = find_matching_key_wbt(major, first_minor);
 -    if (key >= MAP_SIZE) {
 -        return 0;
 -    }
 -
 -    struct io_counter *counterp, zero = {}; 
 -    init_io_counter(&zero, major, first_minor);
 -    counterp = bpf_map_lookup_elem(&wbt_map, &wbtkey);
 -    if (counterp) {
 -        return 0;
 -    }
 -    long err = bpf_map_update_elem(&wbt_map, &wbtkey, &zero, BPF_NOEXIST); 
 -    if (err) {
 -        log_event(STAGE_WBT, PERIOD_START, ERROR_UPDATE_FAIL);
 -        return 0;
 -    }
 -
 -    u64 curr_start_range = zero.start_time / THRESHOLD;
 -
 -    struct update_params params = {
 -        .major = major,
 -        .first_minor = first_minor,
 -        .cmd_flags = cmd_flags,
 -        .curr_start_range = curr_start_range,
 -    }; 
 -
 -    struct stage_data *curr_data; 
 -    curr_data = bpf_map_lookup_elem(&wbt_res, &key);
 -    if (!curr_data) { 
 -        struct stage_data new_data = {
 -            .start_count = 1,
 -            .finish_count = 0,
 -            .finish_over_time = 0,
 -            .duration = 0,
 -            .major = major,
 -            .first_minor = first_minor,
 -            .io_type = "",
 -        };
 -        blk_fill_rwbs(new_data.io_type, cmd_flags);
 -        bpf_map_update_elem(&wbt_res, &key, &new_data, 0);
 -    } else {
 -        update_curr_data_in_start(curr_data, &params);
 -    }
 -
 -    struct time_range_io_count *curr_data_time_range;
 -    curr_data_time_range = bpf_map_lookup_elem(&wbt_res_2, &curr_start_range);
 -    if (curr_data_time_range == NULL) {
 -        struct time_range_io_count new_data = { .count = {0} };
 -        bpf_map_update_elem(&wbt_res_2, &curr_start_range, &new_data, 0); 
 -    } else {
 -        if (key < MAP_SIZE && key >= 0) {
 -            __sync_fetch_and_add(&curr_data_time_range->count[key], 1);
 -        }
 -    }
 -    return 0; 
 -}
 -
 -// finish wbt
 -SEC("kretprobe/wbt_wait") 
 -int kretprobe_wbt_wait(struct pt_regs *regs) 
 -{
 -    u64 wbtkey = bpf_get_current_task();
 -    u64 *wbtargs = (u64 *)bpf_map_lookup_elem(&wbt_args, &wbtkey);
 -    if (wbtargs == NULL) {
 -        bpf_map_delete_elem(&wbt_args, &wbtkey);
 -        return 0;
 -    }
 -    
 -    struct bio *bio;
 -    struct block_device *bd;
 -    struct gendisk *curr_rq_disk;
 -    int major, first_minor;
 -    unsigned int cmd_flags;
 -
 -    bio = (struct bio *)(*wbtargs);
 -    bpf_core_read(&bd, sizeof(bd), &bio->bi_bdev);
 -    bpf_core_read(&curr_rq_disk, sizeof(curr_rq_disk), &bd->bd_disk);
 -    bpf_core_read(&major, sizeof(major), &curr_rq_disk->major);
 -    bpf_core_read(&first_minor, sizeof(first_minor), &curr_rq_disk->first_minor);
 -    bpf_core_read(&cmd_flags, sizeof(cmd_flags), &bio->bi_opf);
 -
 -    if (major == 0) {
 -        log_event(STAGE_WBT, PERIOD_END, ERROR_MAJOR_ZERO);
 -        return 0;
 -    }
 -
 -    u32 key = find_matching_key_wbt(major, first_minor);
 -    if (key >= MAP_SIZE) {
 -        return 0;
 -    }
 -
 -    struct io_counter *counterp = bpf_map_lookup_elem(&wbt_map, &wbtkey); 
 -    if (!counterp) {
 -        return 0;
 -    }
 -
 -    u64 duration = bpf_ktime_get_ns() - counterp->start_time; 
 -    u64 curr_start_range = counterp->start_time / THRESHOLD;
 -
 -    struct update_params params = {
 -        .major = major,
 -        .first_minor = first_minor,
 -        .cmd_flags = cmd_flags,
 -        .curr_start_range = curr_start_range,
 -    }; 
 -
 -    struct stage_data *curr_data; 
 -    curr_data = bpf_map_lookup_elem(&wbt_res, &key);
 -    if (curr_data == NULL && duration > DURATION_THRESHOLD) { 
 -       struct stage_data new_data = {
 -            .start_count = 1,
 -            .finish_count = 1,
 -            .finish_over_time = 1,
 -            .duration = 0,
 -            .major = major,
 -            .first_minor = first_minor,
 -            .io_type = "",
 -        };
 -        blk_fill_rwbs(new_data.io_type, cmd_flags);
 -        bpf_map_update_elem(&wbt_res, &key, &new_data, 0); 
 -    } else if (curr_data == NULL) { 
 -        struct stage_data new_data = {
 -            .start_count = 1,
 -            .finish_count = 1,
 -            .finish_over_time = 0,
 -            .duration = 0,
 -            .io_type = "",
 -            .major = major,
 -            .first_minor = first_minor,
 -        };
 -        blk_fill_rwbs(new_data.io_type, cmd_flags);
 -        bpf_map_update_elem(&wbt_res, &key, &new_data, 0); 
 -    } else {
 -        curr_data->duration += duration; 
 -        update_curr_data_in_finish(curr_data, &params, duration);
 -    }
 -
 -    struct time_range_io_count *curr_data_time_range;
 -    curr_data_time_range = bpf_map_lookup_elem(&wbt_res_2, &curr_start_range);
 -    if (curr_data_time_range == NULL) { 
 -        struct time_range_io_count new_data = { .count = {0} };
 -        bpf_map_update_elem(&wbt_res_2, &curr_start_range, &new_data, 0);
 -    } else {
 -        if (key < MAP_SIZE && curr_data_time_range->count[key] > 0) {
 -            __sync_fetch_and_add(&curr_data_time_range->count[key], -1);
 -        }
 -    }
 -    bpf_map_delete_elem(&wbt_map, &wbtkey);
 -    bpf_map_delete_elem(&wbt_args, &wbtkey);
 -    return 0; 
 -}
 -
 char _license[] SEC("license") = "GPL";
 diff --git a/src/services/sentryCollector/ebpf_collector/ebpf_collector.c b/src/services/sentryCollector/ebpf_collector/ebpf_collector.c
 index 5a2528b..445cce7 100644
 --- a/src/services/sentryCollector/ebpf_collector/ebpf_collector.c
 +++ b/src/services/sentryCollector/ebpf_collector/ebpf_collector.c
@@ -155,7 +155,7 @@ static void update_io_dump(int fd, int *io_dump, int map_size, char *stage) {
         if ((curr_time - io_dump_key) >= 2) {
             int isempty = 1;
             for (int key = 0; key < map_size; key++) {
 -                if (time_count.count[key] > 0) {
 +                if (time_count.count[key] > IO_DUMP_DIFF) {
                     io_dump[key] += time_count.count[key];
                     isempty = 0;
                 }
 diff --git a/src/services/sentryCollector/ebpf_collector/ebpf_collector.h b/src/services/sentryCollector/ebpf_collector/ebpf_collector.h
 index adf926b..e9de49e 100644
 --- a/src/services/sentryCollector/ebpf_collector/ebpf_collector.h
 +++ b/src/services/sentryCollector/ebpf_collector/ebpf_collector.h
@@ -39,6 +39,10 @@ typedef unsigned int u32;
 #define MAJOR(dev)    ((unsigned int) ((dev) >> MINORBITS))
 #define MINOR(dev)    ((unsigned int) ((dev) & MINORMASK))
 +#ifndef NULL
 +#define NULL ((void *)0)
 +#endif
 +
 // 阶段
 #define STAGE_RQ_DRIVER 1
 #define STAGE_BIO       2
@@ -55,6 +59,10 @@ typedef unsigned int u32;
 #define ERROR_KEY_EXIST 3
 #define ERROR_UPDATE_FAIL 4
 #define ERROR_KEY_NOEXIST 5
 +#define ERROR_DELETE_FAIL 6
 +
 +#define BATCH_COUT 100
 +#define IO_DUMP_DIFF 3
 enum stage_type {
     BIO=0,
@@ -94,13 +102,13 @@ struct update_params {
 struct time_range_io_count
 {
 -    u32 count[MAP_SIZE];
 +    int count[MAP_SIZE];
 };
 struct event {
 -    u32 stage;
 -    u64 period;
 -    u32 err; 
 +    int stage;
 +    int period;
 +    int err; 
 };
 #endif /* __EBPFCOLLECTOR_H */
 -- 
 2.43.0
--- a/ai-block-io-exit-when-stage-is-not-supported.patch
+++ b/ai-block-io-exit-when-stage-is-not-supported.patch
@ -0,0 +1,69 @@
 From b462c193e8b6bb7b8f252b9ef8931d91831e1321 Mon Sep 17 00:00:00 2001
 From: luckky <guodashun1@huawei.com>
 Date: Thu, 13 Mar 2025 11:55:15 +0800
 Subject: [PATCH] ai block io: exit when stage is not supported
 ---
 .../ai_block_io/config_parser.py              | 32 +++++++++++++++++--
 1 file changed, 30 insertions(+), 2 deletions(-)
 diff --git a/src/sentryPlugins/ai_block_io/config_parser.py b/src/sentryPlugins/ai_block_io/config_parser.py
 index 1bbb609..612fe9f 100644
 --- a/src/sentryPlugins/ai_block_io/config_parser.py
 +++ b/src/sentryPlugins/ai_block_io/config_parser.py
@@ -32,6 +32,12 @@ ALL_STAGE_LIST = [
     "rq_driver",
     "bio",
 ]
 +EBPF_STAGE_LIST = [
 +    "wbt",
 +    "rq_driver",
 +    "bio",
 +    "gettag"
 +]
 ALL_IOTPYE_LIST = ["read", "write"]
 DISK_TYPE_MAP = {
     0: "nvme_ssd",
@@ -312,15 +318,37 @@ class ConfigParser:
         if len(stage_list) == 1 and stage_list[0] == "":
             logging.critical("stage value not allow is empty, exiting...")
             exit(1)
 +
 +        # check if kernel or ebpf is supported (code is from collector)
 +        valid_stage_list = ALL_STAGE_LIST
 +        base_path = '/sys/kernel/debug/block'
 +        all_disk = []
 +        for disk_name in os.listdir(base_path):
 +            disk_path = os.path.join(base_path, disk_name)
 +            blk_io_hierarchy_path = os.path.join(disk_path, 'blk_io_hierarchy')
 +
 +            if not os.path.exists(blk_io_hierarchy_path):
 +                logging.warning("no blk_io_hierarchy directory found in %s, skipping.", disk_name)
 +                continue
 +
 +            for file_name in os.listdir(blk_io_hierarchy_path):
 +                if file_name == 'stats':
 +                    all_disk.append(disk_name)
 +        
 +        if len(all_disk) == 0:
 +            logging.debug("no blk_io_hierarchy disk, it is not lock-free collection")
 +            valid_stage_list = EBPF_STAGE_LIST
 +
         if len(stage_list) == 1 and stage_list[0] == "default":
             logging.warning(
                 "stage will enable default value: %s",
                 self.DEFAULT_CONF["common"]["stage"],
             )
 -            self._conf["common"]["stage"] = ALL_STAGE_LIST
 +            self._conf["common"]["stage"] = valid_stage_list
             return
 +
         for stage in stage_list:
 -            if stage not in ALL_STAGE_LIST:
 +            if stage not in valid_stage_list:
                 logging.critical(
                     "stage: %s is not valid stage, ai_block_io will exit...", stage
                 )
 -- 
 2.43.0
--- a/fix-env-for-subprocess.Popen.patch
+++ b/fix-env-for-subprocess.Popen.patch
@ -0,0 +1,25 @@
 From 8d6ca181b85ee32837e8891a1003d24826902f08 Mon Sep 17 00:00:00 2001
 From: shixuantong <shixuantong1@huawei.com>
 Date: Sat, 29 Mar 2025 10:58:17 +0800
 Subject: [PATCH] fix env for subprocess.Popen
 ---
 src/services/syssentry/global_values.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
 diff --git a/src/services/syssentry/global_values.py b/src/services/syssentry/global_values.py
 index 931b8ab..7cb99e0 100644
 --- a/src/services/syssentry/global_values.py
 +++ b/src/services/syssentry/global_values.py
@@ -76,7 +76,7 @@ class InspectTask:
         # ccnfig env_file
         self.env_file = ""
         # env conf to popen arg
 -        self.environ_conf = {}
 +        self.environ_conf = None
         # start mode
         self.conflict = "up"
         # alarm id
 -- 
 2.27.0
--- a/fix-period-task-some-bugs.patch
+++ b/fix-period-task-some-bugs.patch
@ -0,0 +1,54 @@
 From b0af4890cd131f33ab85708c75742f0a9680705c Mon Sep 17 00:00:00 2001
 From: shixuantong <shixuantong1@huawei.com>
 Date: Sat, 22 Mar 2025 10:38:30 +0800
 Subject: [PATCH] fix period task some bugs
 ---
 src/services/syssentry/cron_process.py | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
 diff --git a/src/services/syssentry/cron_process.py b/src/services/syssentry/cron_process.py
 index 204d4f3..fab350e 100644
 --- a/src/services/syssentry/cron_process.py
 +++ b/src/services/syssentry/cron_process.py
@@ -59,7 +59,6 @@ class PeriodTask(InspectTask):
         self.result_info["details"] = {}
         if not self.period_enabled:
             self.period_enabled = True
 -            self.upgrade_period_timestamp()
         if self.conflict != 'up':
             ret = self.check_conflict()
@@ -87,6 +86,7 @@ class PeriodTask(InspectTask):
             self.runtime_status = FAILED_STATUS
             return False, "period task start popen failed, invalid command"
         finally:
 +            self.upgrade_period_timestamp()
             if isinstance(logfile, io.TextIOWrapper) and not logfile.closed:
                 logfile.close()
@@ -127,7 +127,6 @@ class PeriodTask(InspectTask):
         res, _ = self.start()
         if res:
             set_runtime_status(self.name, RUNNING_STATUS)
 -        self.upgrade_period_timestamp()
 def period_tasks_handle():
@@ -142,7 +141,7 @@ def period_tasks_handle():
             logging.debug("period not enabled")
             continue
 -        if not task.onstart:
 +        if not task.onstart and task.last_exec_timestamp == 0:
             logging.debug("period onstart not enabled, task: %s", task.name)
             task.runtime_status = EXITED_STATUS
             continue
@@ -153,4 +152,3 @@ def period_tasks_handle():
             res, _ = task.start()
             if res:
                 set_runtime_status(task.name, RUNNING_STATUS)
 -            task.upgrade_period_timestamp()
 -- 
 2.27.0
--- a/fix-the-sentryCollector-service-can-t-be-stopped-for.patch
+++ b/fix-the-sentryCollector-service-can-t-be-stopped-for.patch
@ -0,0 +1,37 @@
 From 93a58d1e0faee37cbc27981c85f41c50e81bdc66 Mon Sep 17 00:00:00 2001
 From: zhuofeng <1107893276@qq.com>
 Date: Mon, 24 Feb 2025 03:00:57 +0000
 Subject: [PATCH] fix the sentryCollector service can't be stopped for a long
 time.
 Signed-off-by: zhuofeng <1107893276@qq.com>
 ---
 src/services/sentryCollector/collect_io.py | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
 diff --git a/src/services/sentryCollector/collect_io.py b/src/services/sentryCollector/collect_io.py
 index 10446d9..6db28ec 100644
 --- a/src/services/sentryCollector/collect_io.py
 +++ b/src/services/sentryCollector/collect_io.py
@@ -405,10 +405,17 @@ class CollectIo():
         self
     ) -> None:
         global EBPF_PROCESS
 -        if EBPF_PROCESS:
 +        if not EBPF_PROCESS:
 +            logging.debug("No eBPF process to stop")
 +            return
 +        try:
             EBPF_PROCESS.terminate()
 +            EBPF_PROCESS.wait(timeout=3)
 +        except subprocess.TimeoutExpired:
 +            logging.debug("eBPF process did not exit within timeout. Forcing kill.")
 +            EBPF_PROCESS.kill()
             EBPF_PROCESS.wait()
 -            logging.info("ebpf collector thread exit")
 +        logging.info("ebpf collector thread exit")
     def main_loop(self):
         global IO_GLOBAL_DATA
 -- 
 2.43.0
--- a/sysSentry.spec
+++ b/sysSentry.spec
@ -4,7 +4,7 @@
 Summary: System Inspection Framework
 Name: sysSentry
 Version: 1.0.3
-Release: 5
+Release: 10
 License: Mulan PSL v2
 Group: System Environment/Daemons
 Source0: https://gitee.com/openeuler/sysSentry/releases/download/v%{version}/%{name}-%{version}.tar.gz
@ -13,6 +13,12 @@ Patch1:    add-bidirectional-communication-for-xalarm.patch
 Patch2:    fix-some-test-cases.patch
 Patch3:    add-log-for-xalarmd-and-fix-delete-on-iter-problem.patch
 Patch4:    fix-xalarm-log-not-print-and-add-on-iter-problem.patch
 Patch5:    add-new-func-for-ebpf-in-the-rq_driver-stage.patch
 Patch6:    fix-the-sentryCollector-service-can-t-be-stopped-for.patch
 Patch7:    ai-block-io-exit-when-stage-is-not-supported.patch
 Patch8:    add-log-utils-for-c.patch
 Patch9:    fix-env-for-subprocess.Popen.patch
 Patch10:   fix-period-task-some-bugs.patch
 BuildRequires: cmake gcc-c++
 BuildRequires: python3 python3-setuptools
@ -168,7 +174,9 @@ rm -rf /var/run/sysSentry | :
 %attr(0550,root,root) %{_bindir}/sentryCollector
 %attr(0600,root,root) %{_sysconfdir}/sysSentry/collector.conf
 %attr(0600,root,root) %{_unitdir}/sentryCollector.service
 %attr(0550,root,root) %{_libdir}/libsentry_log.so
 %exclude %{_includedir}/libsentry/log_utils.h
 %exclude %{_sysconfdir}/sysSentry/tasks/hbm_online_repair.mod
 %exclude %{python3_sitelib}/syssentry/bmc_*
 %exclude %{python3_sitelib}/syssentry/*/bmc_*
@ -209,6 +217,37 @@ rm -rf /var/run/sysSentry | :
 %attr(0550,root,root) %{python3_sitelib}/syssentry/bmc_alarm.py
 %changelog
 * Sat Mar 29 2025 shixuantong <shixuantong1@huawei.com> - 1.0.3-10
 - Type:bugfix
 - CVE:NA
 - SUG:NA
 - DESC:fix period task some bugs
       fix env_file and environ_conf
 * Fri Mar 14 2025 shixuantong <shixuantong1@huawei.com> - 1.0.3-9
 - Type:bugfix
 - CVE:NA
 - SUG:NA
 - DESC:add log utils for c
 * Thu Mar 13 2025 luckky <guodashun1@huawei.com> - 1.0.3-8
 - Type:bugfix
 - CVE:NA
 - SUG:NA
 - DESC: fix an issue with printing error
 * Mon Feb 24 2025 zhuofeng <zhuofeng2@huawei.com> - 1.0.3-7
 - Type:bugfix
 - CVE:NA
 - SUG:NA
 - DESC: fix the sentryCollector service can't be stopped for a long
 * Sat Feb 22 2025 zhuofeng <zhuofeng2@huawei.com> - 1.0.3-6
 - Type:bugfix
 - CVE:NA
 - SUG:NA
 - DESC: add new func for ebpf in the rq_driver stage
 * Tue Feb 18 2025 caixiaomeng <caixiaomeng2@huawei.com> - 1.0.3-5
 - Type:bugfix
 - CVE:NA
Author	SHA1	Message	Date
openeuler-ci-bot	91b26141b0	!240 fix period task some bugs and fix env config From: @tong_1001 Reviewed-by: @hubin95 Signed-off-by: @hubin95	2025-03-29 07:43:30 +00:00
shixuantong	b1b02ba203	fix period task some bugs and fix env config	2025-03-29 14:49:26 +08:00
openeuler-ci-bot	87260fb362	!236 add log utils for c From: @tong_1001 Reviewed-by: @znzjugod Signed-off-by: @znzjugod	2025-03-14 08:42:52 +00:00
shixuantong	d82730b2e6	add log utils for c	2025-03-14 16:16:25 +08:00
openeuler-ci-bot	23cd062a57	!221 ai block io: exit when stage is not supported From: @luckky7 Reviewed-by: @znzjugod Signed-off-by: @znzjugod	2025-03-14 02:18:58 +00:00
luckky	c6b8404ef4	ai block io: exit when stage is not supported	2025-03-14 10:05:39 +08:00
openeuler-ci-bot	26ec8137f8	!230 [sync] PR-213: fix the sentryCollector service can't be stopped for a long From: @openeuler-sync-bot Reviewed-by: @znzjugod Signed-off-by: @znzjugod	2025-03-14 02:00:13 +00:00
zhuofeng	c962ee5f32	fix the sentryCollector service can't be stopped for a long Signed-off-by: zhuofeng <1107893276@qq.com> (cherry picked from commit 885bbbd319adfe367baf6e4874b796657fba230e)	2025-03-13 20:39:42 +08:00
openeuler-ci-bot	8474d69ed0	!228 [sync] PR-212: add new func for ebpf in the rq_driver stage From: @openeuler-sync-bot Reviewed-by: @znzjugod Signed-off-by: @znzjugod	2025-03-13 12:38:57 +00:00
zhuofeng	52606113b3	add new func for ebpf in the rq_driver stage Signed-off-by: zhuofeng <1107893276@qq.com> (cherry picked from commit ec94b36256cf018893e543b4b57e95e6c78e3a4a)	2025-03-13 20:18:05 +08:00