commit 4dd5d00eccd8a70862f9129601ebe1e0e1bebcc3
author Daniel Colascione <dancol@google.com> Wed Jan 03 12:01:02 2018 -0800
committer Daniel Colascione <dancol@google.com> Wed Jan 03 12:32:14 2018 -0800
tree 100c03a29f612370eac6f0f90c1ca160e7d1ed2d
parent 93d344d98cd02d66c3aac8067718be828ea283f9

Pin lmkd for real

We pin lmkd in memory so that we don't take page faults (and thus
requisition memory) while we're in the process of responding to a
low-memory condition. mlockall(2) is the right primitive for this
pinning. Previously, we used the MCL_FUTURE flag to mlockall: used
this way, mlockall doesn't actually pin all pages in memory, since
MCL_FUTURE affects only the default flags for future mappings and
doesn't affect mapping already in existence at the time of the
mlockall call --- like the lmkd executable itself.

This patch adds the MCL_CURRENT flag, which also pins all pages
already mapped.

Test: code inspection
Change-Id: I4563959367a2f0a9cadc3ea41731b7f311326685

lmkd/lmkd.c

diff --git a/lmkd/lmkd.c b/lmkd/lmkd.c
index 5cfa2c8..fd83ecc 100644
--- a/lmkd/lmkd.c
+++ b/lmkd/lmkd.c
@@ -900,7 +900,9 @@
     downgrade_pressure = (int64_t)property_get_int32("ro.lmk.downgrade_pressure", 60);
     is_go_device = property_get_bool("ro.config.low_ram", false);
 
-    mlockall(MCL_FUTURE);
+    if (mlockall(MCL_CURRENT | MCL_FUTURE))
+        ALOGW("mlockall failed: errno=%d", errno);
+
     sched_setscheduler(0, SCHED_FIFO, &param);
     if (!init())
         mainloop();