From cf949731294dba0a85d3e7ec505fb6742f5bedeb Mon Sep 17 00:00:00 2001
From: Aldo Lacuku <aldo@lacuku.eu>
Date: Wed, 6 Nov 2024 14:59:40 +0100
Subject: [PATCH] fix(proc/scan): scan all existing threads

By scanning only /proc we missed the threads of a given process.
It scans all threads under a given process in /proc/pid/task.

Signed-off-by: Aldo Lacuku <aldo@lacuku.eu>
---
 plugins/k8smeta/src/plugin.cpp           | 75 ++++++++++++++++--------
 plugins/k8smeta/test/src/parsing_pod.cpp |  3 +-
 2 files changed, 51 insertions(+), 27 deletions(-)

diff --git a/plugins/k8smeta/src/plugin.cpp b/plugins/k8smeta/src/plugin.cpp
index eccc6e08..d7349117 100644
--- a/plugins/k8smeta/src/plugin.cpp
+++ b/plugins/k8smeta/src/plugin.cpp
@@ -302,42 +302,65 @@ void my_plugin::do_initial_proc_scan()
             continue;
         }
 
-        // Example of the path: `/proc/1/cgroup`
-        proc_path = std::string(proc_root)
-                            .append("/")
-                            .append(file_name.c_str())
-                            .append("/cgroup");
+        // Now scan /proc/pid/task for threads
+        std::string task_dir = proc_root + "/" + file_name.c_str() + "/task";
+        std::filesystem::directory_iterator task_iter(task_dir);
 
-        std::ifstream file(proc_path);
-
-        if(file.is_open())
+        for(const auto& task_entry : task_iter)
         {
-            // Read the first line from the file
-            if(std::getline(file, cgroup_line))
+            auto task_file_name = task_entry.path().filename();
+            tid = strtol(task_file_name.c_str(), NULL, 10);
+
+            if(tid == 0 || !task_entry.is_directory())
             {
-                std::string pod_uid =
-                        get_pod_uid_from_cgroup_string(cgroup_line);
-                if(!pod_uid.empty())
-                {
-                    m_thread_id_pod_uid_map[tid] = pod_uid;
-                    SPDLOG_TRACE("Found thread with tid '{}' and pod uid '{}'",
-                                 tid, pod_uid);
-                }
+                SPDLOG_WARN("Found task entry `{}` in process `{}` that is not "
+                            "a number",
+                            task_file_name.c_str(), file_name.c_str());
+                continue; // skip if not a thread id directory
             }
-            else
+            // Example of the path: `/proc/1/task/200/cgroup`
+            proc_path = std::string(proc_root)
+                                .append("/")
+                                .append(file_name.c_str())
+                                .append("/task/")
+                                .append(task_file_name.c_str())
+                                .append("/cgroup");
+
+            std::ifstream file(proc_path);
+
+            if(file.is_open())
             {
-                SPDLOG_WARN("cannot retrieve the cgroup first line for '{}'. "
+                // Read the first line from the file
+                if(std::getline(file, cgroup_line))
+                {
+                    std::string pod_uid =
+                            get_pod_uid_from_cgroup_string(cgroup_line);
+                    if(!pod_uid.empty())
+                    {
+                        m_thread_id_pod_uid_map[tid] = pod_uid;
+                        SPDLOG_TRACE(
+                                "Found thread with tid '{}' and pod uid '{}'",
+                                tid, pod_uid);
+                    }
+                }
+                else
+                {
+                    SPDLOG_WARN(
+                            "Cannot retrieve the cgroup first line for '{}'. "
                             "Error: {}. Skip it",
                             proc_path,
                             file.eof() ? "Empty file" : strerror(errno));
+                }
+                file.close();
+            }
+            else
+            {
+                SPDLOG_WARN("Cannot open '{}'. Error: {}. Skip it.", proc_path,
+                            strerror(errno));
             }
-            file.close();
-        }
-        else
-        {
-            SPDLOG_WARN("cannot open '{}'. Error: {}. Skip it.", proc_path,
-                        strerror(errno));
         }
+        SPDLOG_DEBUG("Thread scan correctly completed for process `{}`",
+                     file_name.c_str());
     }
     SPDLOG_INFO(
             "Process scan correctly completed. Found '{}' threads inside pods.",
diff --git a/plugins/k8smeta/test/src/parsing_pod.cpp b/plugins/k8smeta/test/src/parsing_pod.cpp
index ae878104..0fdb8b51 100644
--- a/plugins/k8smeta/test/src/parsing_pod.cpp
+++ b/plugins/k8smeta/test/src/parsing_pod.cpp
@@ -424,7 +424,8 @@ TEST_F(sinsp_with_test_input, plugin_k8s_proc_scan_in_the_plugin)
 
 #define PLUGIN_PROC_SCAN_UNDER_TMP_PROC                                        \
     int32_t mock_tid = (1 << 16) - 1;                                          \
-    std::string mock_proc_dir = "/tmp/proc/" + std::to_string(mock_tid);       \
+    std::string mock_proc_dir = "/tmp/proc/" + std::to_string(mock_tid) +      \
+                                "/task/" + std::to_string(mock_tid);           \
     std::string mock_proc_cgroup = mock_proc_dir + "/cgroup";                  \
     std::string expected_pod_uid = "1d34c7bb-7d94-4f00-bed9-fe4eca61d446";     \
                                                                                \