(trafficserver) branch master updated: Fix false crash logs with regression tests (#12783)

[bneradt]

This is an automated email from the ASF dual-hosted git repository.

bneradt pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/trafficserver.git


The following commit(s) were added to refs/heads/master by this push:
     new dd3455cffa Fix false crash logs with regression tests (#12783)
dd3455cffa is described below

commit dd3455cffa13c3c802ec4bf711f6e2f4ad5b4f10
Author: Brian Neradt <brian.ner...@gmail.com>
AuthorDate: Tue Jan 6 14:02:17 2026 -0600

    Fix false crash logs with regression tests (#12783)
    
    When traffic_server exits normally (e.g., after regression tests
    complete), traffic_crashlog was incorrectly logging a crash because it
    detected its parent process had terminated. This happened because
    traffic_crashlog uses PR_SET_PDEATHSIG to wake up when traffic_server
    exits, but it couldn't distinguish between a crash (where
    crash_logger_invoke sends signal info via the pipe) and a normal exit
    (where the pipe is simply closed). This fix adds a poll() check on stdin
    to verify that crash data was actually sent before logging a crash,
    preventing false positive crash logs.
---
 src/traffic_crashlog/traffic_crashlog.cc | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/src/traffic_crashlog/traffic_crashlog.cc 
b/src/traffic_crashlog/traffic_crashlog.cc
index 9354c5f838..68c94acb6f 100644
--- a/src/traffic_crashlog/traffic_crashlog.cc
+++ b/src/traffic_crashlog/traffic_crashlog.cc
@@ -32,6 +32,7 @@
 #include "tscore/BaseLogFile.h"
 #include "tscore/runroot.h"
 #include "iocore/eventsystem/RecProcess.h"
+#include <poll.h>
 #include <unistd.h>
 
 static int   syslog_mode  = false;
@@ -171,6 +172,25 @@ main(int /* argc ATS_UNUSED */, const char **argv)
     return 0;
   }
 
+  // In wait mode, we need to verify this is a real crash by checking if 
crash_logger_invoke
+  // sent us signal info via the pipe. If traffic_server just exited normally, 
the pipe will be
+  // closed with no data, and we should exit without logging a false "crash".
+  if (wait_mode) {
+    // Use poll to check if there's data available on stdin without blocking 
indefinitely.
+    struct pollfd pfd;
+    pfd.fd     = STDIN_FILENO;
+    pfd.events = POLLIN;
+
+    // Wait briefly for data. If crash_logger_invoke was called, data should 
already be there.
+    int poll_result = poll(&pfd, 1, 100); // 100ms timeout
+
+    // POLLHUP means the write end of the pipe was closed - normal exit, not 
crash.
+    // No data or error also means no crash occurred.
+    if (poll_result <= 0 || (pfd.revents & POLLHUP) || !(pfd.revents & 
POLLIN)) {
+      return 0;
+    }
+  }
+
   runroot_handler(argv);
   Layout::create();
   RecProcessInit(nullptr /* diags */);