perf: prevent kill(0, SIGTERM);
in [Kernel]
|
Prev: linux-next: manual merge of the net tree with the net-current tree
Next: [PATCH 3/6] KVM: MMU: introduce gfn_to_page_atomic() and gfn_to_pfn_atomic()
From: Ian Munsie on 14 Jun 2010 22:40 Hi Ingo, Please consider merging this patch, it's a trivial fix to prevent perf from killing too many processes in certain unusual situations - such as killing the other intern's entire X session when called from eclipse with a typo in the program being profiled. Cheers, -Ian Excerpts from Ian Munsie's message of Wed Jun 09 18:38:00 +1000 2010: > From: Ian Munsie <imunsie(a)au1.ibm.com> > > At exit, perf record will kill the process it was profiling by sending a > SIGTERM to child_pid (if it had been initialised), but in certain > situations child_pid may be 0 and perf would mistakenly kill more > processes than intended. > > child_pid is set to the return of fork() to either 0 or the pid of the > child. Ordinarily this would not present an issue as the child calls > execvp to spawn the process to be profilled and would therefore never > run it's sig_atexit and never attempt to kill pid 0. > > However, if a nonexistant binary had been passed in to perf record the > call to execvp would fail and child_pid would be left set to 0. The > child would then exit and it's atexit handler, finding that child_pid > was initialised to 0, would call kill(0, SIGTERM), resulting in every > process within it's process group being killed. > > In the case that perf was being run directly from the shell this > typically would not be an issue as the shell isolates the process. > However, if perf was being called from another program it could kill > unexpected processes, which may even include X. > > This patch changes the logic of the test for whether child_pid was > initialised to only considder positive pids as valid, thereby never > attempting to kill pid 0. > > Signed-off-by: Ian Munsie <imunsie(a)au1.ibm.com> > --- > tools/perf/builtin-record.c | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c > index 5e5c640..300da82 100644 > --- a/tools/perf/builtin-record.c > +++ b/tools/perf/builtin-record.c > @@ -193,7 +193,7 @@ static void sig_handler(int sig) > > static void sig_atexit(void) > { > - if (child_pid != -1) > + if (child_pid > 0) > kill(child_pid, SIGTERM); > > if (signr == -1) -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Arnaldo Carvalho de Melo on 17 Jun 2010 16:30
Em Wed, Jun 09, 2010 at 06:38:00PM +1000, Ian Munsie escreveu: > At exit, perf record will kill the process it was profiling by sending a > SIGTERM to child_pid (if it had been initialised), but in certain > situations child_pid may be 0 and perf would mistakenly kill more > processes than intended. > > child_pid is set to the return of fork() to either 0 or the pid of the > child. Ordinarily this would not present an issue as the child calls > execvp to spawn the process to be profilled and would therefore never > run it's sig_atexit and never attempt to kill pid 0. before: [root(a)emilia mingo]# perf record bla bla: No such file or directory Terminated after: [root(a)emilia mingo]# perf record bla bla: No such file or directory [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.006 MB perf.data (~259 samples) ] [root(a)emilia mingo]# Got it, queuing it in perf/urgent, Thanks! - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |