Kill logcat with SEGV so we get a tombstone on timeout.

Looking at the history of this code, we've had hard to reproduce issues
here since at least 2016. The most recent bug doesn't even have any
information to let us work out whether the problem is with logcat, with
timeout, or with the Java code calling them. Let's use SIGSEGV so we'll
at least have a tombstone if timeout has to kill logcat.

Bug: http://b/187005648
Test: treehugger
Change-Id: I224555c631d784647dafc30641c2c8a447884971
This commit is contained in:
Elliott Hughes
2021-05-04 16:21:58 -07:00
parent c4520f1398
commit 2046d54ed6

View File

@@ -10272,11 +10272,13 @@ public class ActivityManagerService extends IActivityManager.Stub
if (lines > 0) {
sb.append("\n");
// Merge several logcat streams, and take the last N lines
InputStreamReader input = null;
try {
java.lang.Process logcat = new ProcessBuilder(
"/system/bin/timeout", "-k", "15s", "10s",
// Time out after 10s, but kill logcat with SEGV
// so we can investigate why it didn't finish.
"/system/bin/timeout", "-s", "SEGV", "10s",
// Merge several logcat streams, and take the last N lines.
"/system/bin/logcat", "-v", "threadtime", "-b", "events", "-b", "system",
"-b", "main", "-b", "crash", "-t", String.valueOf(lines))
.redirectErrorStream(true).start();