debuggerd: call setsid in our children.
There appears to be a kernel bug that causes SIGHUP and SIGCONT to be
sent to the parent process group we spawn from if the process group
contains stopped jobs (e.g. the parent itself, because of wait_for_gdb).
Call setsid in all of our children to prevent this from happening.
Bug: http://b/31124563
Test: adb shell 'setprop debug.debuggerd.wait_for_gdb 1; killall -ABRT surfaceflinger'
Change-Id: I1a48d70886880a5bfbe2deb80d48deece55faf09
diff --git a/debuggerd/crash_dump.cpp b/debuggerd/crash_dump.cpp
index 82ba0a1..c608a8c 100644
--- a/debuggerd/crash_dump.cpp
+++ b/debuggerd/crash_dump.cpp
@@ -363,6 +363,12 @@
DefuseSignalHandlers();
InstallSigPipeHandler();
+ // There appears to be a bug in the kernel where our death causes SIGHUP to
+ // be sent to our process group if we exit while it has stopped jobs (e.g.
+ // because of wait_for_gdb). Use setsid to create a new process group to
+ // avoid hitting this.
+ setsid();
+
atrace_begin(ATRACE_TAG, "before reparent");
pid_t target_process = getppid();
diff --git a/debuggerd/handler/debuggerd_handler.cpp b/debuggerd/handler/debuggerd_handler.cpp
index bca5e36..598ea85 100644
--- a/debuggerd/handler/debuggerd_handler.cpp
+++ b/debuggerd/handler/debuggerd_handler.cpp
@@ -268,8 +268,15 @@
_exit(errno);
}
- // Exit immediately on both sides of the fork.
- // crash_dump is ptracing us, so it'll get to do whatever it wants in between.
+ // crash_dump is ptracing both sides of the fork; it'll let the parent exit,
+ // but keep the orphan stopped to peek at its memory.
+
+ // There appears to be a bug in the kernel where our death causes SIGHUP to
+ // be sent to our process group if we exit while it has stopped jobs (e.g.
+ // because of wait_for_gdb). Use setsid to create a new process group to
+ // avoid hitting this.
+ setsid();
+
_exit(0);
}
