Determine memory load based on cgroup usage. (#19518)

cgroup usage is used to trigger oom kills. It includes rss and file cache
of the cgroup.

The implementation was only using the process rss to determine memory load.
This is less than the cgroup usage and leads to oom kills due to GC not
being triggered soon enough.

2 files changed, 30 insertions, 2 deletions

diff --git a/src/pal/inc/pal.h b/src/pal/inc/pal.h
index 97fee605e0..9676b437b0 100644
--- a/src/pal/inc/pal.h
+++ b/src/pal/inc/pal.h
@@ -2396,7 +2396,7 @@ PAL_GetRestrictedPhysicalMemoryLimit(VOID);
 PALIMPORT
 BOOL
 PALAPI
-PAL_GetWorkingSetSize(size_t* val);
+PAL_GetPhysicalMemoryUsed(size_t* val);
 
 PALIMPORT
 BOOL
diff --git a/src/pal/src/misc/cgroup.cpp b/src/pal/src/misc/cgroup.cpp
index 7a3a9261a1..145586a0b9 100644
--- a/src/pal/src/misc/cgroup.cpp
+++ b/src/pal/src/misc/cgroup.cpp
@@ -23,6 +23,7 @@ SET_DEFAULT_DEBUG_CHANNEL(MISC);
 #define PROC_CGROUP_FILENAME "/proc/self/cgroup"
 #define PROC_STATM_FILENAME "/proc/self/statm"
 #define MEM_LIMIT_FILENAME "/memory.limit_in_bytes"
+#define MEM_USAGE_FILENAME "/memory.usage_in_bytes"
 #define CFS_QUOTA_FILENAME "/cpu.cfs_quota_us"
 #define CFS_PERIOD_FILENAME "/cpu.cfs_period_us"
 class CGroup
@@ -63,6 +64,27 @@ public:
         return result;
     }
 
+    bool GetPhysicalMemoryUsage(size_t *val)
+    {
+        char *mem_usage_filename = nullptr;
+        bool result = false;
+
+        if (m_memory_cgroup_path == nullptr)
+            return result;
+
+        size_t len = strlen(m_memory_cgroup_path);
+        len += strlen(MEM_USAGE_FILENAME);
+        mem_usage_filename = (char*)malloc(len+1);
+        if (mem_usage_filename == nullptr)
+            return result;
+
+        strcpy(mem_usage_filename, m_memory_cgroup_path);
+        strcat(mem_usage_filename, MEM_USAGE_FILENAME);
+        result = ReadMemoryValueFromFile(mem_usage_filename, val);
+        free(mem_usage_filename);
+        return result;
+    }
+
     bool GetCpuLimit(UINT *val)
     {
         long long quota;
@@ -384,15 +406,21 @@ PAL_GetRestrictedPhysicalMemoryLimit()
 
 BOOL
 PALAPI
-PAL_GetWorkingSetSize(size_t* val)
+PAL_GetPhysicalMemoryUsed(size_t* val)
 {
     BOOL result = false;
     size_t linelen;
     char* line = nullptr;
+    CGroup cgroup;
 
     if (val == nullptr)
         return FALSE;
 
+    // Linux uses cgroup usage to trigger oom kills.
+    if (cgroup.GetPhysicalMemoryUsage(val))
+        return TRUE;
+
+    // process resident set size.
     FILE* file = fopen(PROC_STATM_FILENAME, "r");
     if (file != nullptr && getline(&line, &linelen, file) != -1)
     {