OSDN Git Service

perf record: Filter out POLLHUP'ed file descriptors
authorArnaldo Carvalho de Melo <acme@redhat.com>
Wed, 13 Aug 2014 14:33:59 +0000 (11:33 -0300)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Thu, 25 Sep 2014 19:46:56 +0000 (16:46 -0300)
So that we don't continue polling on vanished file descriptors, i.e.
file descriptors for events monitoring threads that exited.

I.e. the following 'perf record' command now exits as expected, instead
of staying in an eternal loop:

  $ sleep 5s &
  $ perf record -p `pidof sleep`

Reported-by: Jiri Olsa <jolsa@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-8dg8o21t2ntzly2bfh53p3sg@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/builtin-record.c

index a8c2e9d..320b198 100644 (file)
@@ -308,7 +308,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
        struct record_opts *opts = &rec->opts;
        struct perf_data_file *file = &rec->file;
        struct perf_session *session;
-       bool disabled = false;
+       bool disabled = false, draining = false;
 
        rec->progname = argv[0];
 
@@ -457,7 +457,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
                }
 
                if (hits == rec->samples) {
-                       if (done)
+                       if (done || draining)
                                break;
                        err = perf_evlist__poll(rec->evlist, -1);
                        /*
@@ -467,6 +467,9 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
                        if (err > 0 || (err < 0 && errno == EINTR))
                                err = 0;
                        waking++;
+
+                       if (perf_evlist__filter_pollfd(rec->evlist, POLLERR | POLLHUP) == 0)
+                               draining = true;
                }
 
                /*