On Fri, Oct 18, 2019 at 05:55:10PM +0800, Wang Yechao wrote:

Libvirtd has set SIGPIPE to ignored, and virFork resets all signal handlers to the defaults. But child process may write logs to stderr/stdout, that may generate SIGPIPE if journald has stopped.

So set SIGPIPE to a dummy no-op handler before unmask signals in virFork. And there is no need to set SIGPIPE ignored before running hooks in virExec. At last, set SIGPIPE to defaults before execve.

Signed-off-by: Wang Yechao <wang.yechao255@zte.com.cn> --- v3 patch: https://www.redhat.com/archives/libvir-list/2019-October/msg00934.html

Changes in v4: - don't block SIGPIPE, ignore it when invoke VIR_FORCE_CLOSE and virCommandMassClose

Changes in v5: - chang from SIG_IGN to a no-op handler in child process --- src/util/vircommand.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-)

diff --git a/src/util/vircommand.c b/src/util/vircommand.c index 93b3dd2..c13739c 100644 --- a/src/util/vircommand.c +++ b/src/util/vircommand.c @@ -217,6 +217,8 @@ virCommandFDSet(virCommandPtr cmd,

#ifndef WIN32

+static void virDummyHandler(int sig G_GNUC_UNUSED) {} + /** * virFork: * @@ -312,6 +314,9 @@ virFork(void) ignore_value(sigaction(i, &sig_action, NULL)); }

Put a comment here /* Code that runs between fork & execve might trigger * SIG_PIPE, so we must explicitly set that to a no-op * handler. This handler will get reset to SIG_DFL when * execve() runs */

+ sig_action.sa_handler = virDummyHandler; + ignore_value(sigaction(SIGPIPE, &sig_action, NULL)); + /* Unmask all signals in child, since we've no idea what the * caller's done with their signal mask and don't want to * propagate that to children */ @@ -550,7 +555,7 @@ virExec(virCommandPtr cmd) g_autofree char *binarystr = NULL; const char *binary = NULL; int ret; - struct sigaction waxon, waxoff; + struct sigaction sig_action; g_autofree gid_t *groups = NULL; int ngroups;

@@ -718,21 +723,6 @@ virExec(virCommandPtr cmd) } }

- /* virFork reset all signal handlers to the defaults. - * This is good for the child process, but our hook - * risks running something that generates SIGPIPE, - * so we need to temporarily block that again - */ - memset(&waxoff, 0, sizeof(waxoff)); - waxoff.sa_handler = SIG_IGN; - sigemptyset(&waxoff.sa_mask); - memset(&waxon, 0, sizeof(waxon)); - if (sigaction(SIGPIPE, &waxoff, &waxon) < 0) { - virReportSystemError(errno, "%s", - _("Could not disable SIGPIPE")); - goto fork_error; - } - if (virProcessSetMaxMemLock(0, cmd->maxMemLock) < 0) goto fork_error; if (virProcessSetMaxProcesses(0, cmd->maxProcesses) < 0)

@@ -783,7 +773,10 @@ virExec(virCommandPtr cmd) if (virCommandHandshakeChild(cmd) < 0) goto fork_error;

- if (sigaction(SIGPIPE, &waxon, NULL) < 0) { + memset(&sig_action, 0, sizeof(sig_action)); + sig_action.sa_handler = SIG_DFL; + sigemptyset(&sig_action.sa_mask); + if (sigaction(SIGPIPE, &sig_action, NULL) < 0) { virReportSystemError(errno, "%s", _("Could not re-enable SIGPIPE")); goto fork_error;

We can just delete this sigaction() call entirely. The execve() call will purge the no-op handler function we registered in virFork(), resetting it back to SIG_DFL, so no need todo it manually here. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|