Re: [Libguestfs] [PATCH libnbd] states: connect_command: Don't set O_NONBLOCK on socket passed to child.

On 5/25/19 1:33 PM, Richard W.M. Jones wrote: > I also made the code a bit more robust about closing the socket along > error paths. > --- > generator/states-connect.c | 21 ++++++++++++++++++--- > 1 file changed, 18 insertions(+), 3 deletions(-) > > assert (!h->sock); > assert (h->argv); > assert (h->argv[0]); > - if (socketpair (AF_UNIX, SOCK_STREAM|SOCK_NONBLOCK|SOCK_CLOEXEC, 0, > - sv) == -1) { > + if (socketpair (AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC, 0, sv) == -1) { Is it any easier to keep SOCK_NONBLOCK here and then clear O_NONBLOCK in the child process? It may matter if we try to port to a system that lacks SOCK_CLOEXEC (SOCK_NONBLOCK and SOCK_CLOEXEC were added at the same time). But we can deal with portability when someone reports a problem. For now the patch looks fine.