Fix problem with wait_for_gdb.
When someone enables wait_for_gdb, activity manager will kill the
stopped process before a developer can attach to the process. To
allow debugging in this case, change the code to only contact the
activity manager right before continuing the process that is
crashing.
Also, modify the conditions under which to do a gdb attach. The previous
code did a partial attach if perform_dump failed. The new version simply
allows an attach regardless of whether perform_dump passes or fails.
Bug: 28409358
(cherry picked from commit 9818bd2bbee00c387247210eeae2a0a0d3912c8b)
Change-Id: I42f464b69332748e16b07d9d00f44b3aa26ce8b7
diff --git a/debuggerd/utility.h b/debuggerd/utility.h
index ed08ddc..cd01188 100644
--- a/debuggerd/utility.h
+++ b/debuggerd/utility.h
@@ -21,6 +21,8 @@
#include <stdbool.h>
#include <sys/types.h>
+#include <string>
+
#include <backtrace/Backtrace.h>
// Figure out the abi based on defined macros.
@@ -42,10 +44,10 @@
struct log_t{
- /* tombstone file descriptor */
+ // Tombstone file descriptor.
int tfd;
- /* Activity Manager socket file descriptor */
- int amfd;
+ // Data to be sent to the Activity Manager.
+ std::string* amfd_data;
// The tid of the thread that crashed.
pid_t crashed_tid;
// The tid of the thread we are currently working with.
@@ -54,7 +56,8 @@
bool should_retrieve_logcat;
log_t()
- : tfd(-1), amfd(-1), crashed_tid(-1), current_tid(-1), should_retrieve_logcat(true) {}
+ : tfd(-1), amfd_data(nullptr), crashed_tid(-1), current_tid(-1),
+ should_retrieve_logcat(true) {}
};
// List of types of logs to simplify the logging decision in _LOG