Don't use mem cgroups for pid accounting.
Commit b82bab66 introduced the use of memory cgroups for keeping track
of forked PIDs; it basically creates a separate memory cgroup for every
process forked from zygote.
Each such memory cgroup which also have its own LRU with (in)active file
and anonymous pages. The current theory is this could potentially introduce
two problems:
1) kswapd runs longer because it has to iterate over the LRUs of all mem
cgroups, instead of over the LRUs of a single root mem cgroup;
2) the way kswapd reclaims things will be different also - I think it will
tend to bias reclaim to smaller mem cgroups, and process private pages
will end up on ZRAM swap much sooner.
Until we figure this out, fall back to the CPU accounting cgroup for keeping
track of forked PIDs. This leaves us with a single root mem cgroup again. We
can also keep userspace lmkd enabled because it only requires the root mem
cgroup.
Bug: 27381069
Change-Id: Ife397a6ac232761f2adfe6f5056582be0d1b4ff1
diff --git a/libprocessgroup/processgroup.cpp b/libprocessgroup/processgroup.cpp
index 5ab957d..1bc1659 100644
--- a/libprocessgroup/processgroup.cpp
+++ b/libprocessgroup/processgroup.cpp
@@ -37,6 +37,9 @@
#include <processgroup/processgroup.h>
+// Uncomment line below use memory cgroups for keeping track of (forked) PIDs
+// #define USE_MEMCG 1
+
#define MEM_CGROUP_PATH "/dev/memcg/apps"
#define MEM_CGROUP_TASKS "/dev/memcg/apps/tasks"
#define ACCT_CGROUP_PATH "/acct"
@@ -67,6 +70,7 @@
};
static const char* getCgroupRootPath() {
+#ifdef USE_MEMCG
static const char* cgroup_root_path = NULL;
std::call_once(init_path_flag, [&]() {
// Check if mem cgroup is mounted, only then check for write-access to avoid
@@ -75,6 +79,9 @@
ACCT_CGROUP_PATH : MEM_CGROUP_PATH;
});
return cgroup_root_path;
+#else
+ return ACCT_CGROUP_PATH;
+#endif
}
static int convertUidToPath(char *path, size_t size, uid_t uid)