On Tue, Apr 18, 2023 at 9:49 PM Adhemerval Zanella Netto <adhemerval.zane...@linaro.org> wrote: > > > > On 18/04/23 10:48, Cristian Rodríguez wrote: > > > > > > On Tue, Apr 18, 2023 at 8:02 AM Adhemerval Zanella Netto via Libc-alpha > > <libc-al...@sourceware.org <mailto:libc-al...@sourceware.org>> wrote: > > > > > > > > On 17/04/23 19:58, Sergey Bugaev via Libc-alpha wrote: > > > This is nicer, and is going to be required for the following changes > > > to reasonably stay within the 79 column limit. > > > > > > No functional change. > > > > > > Signed-off-by: Sergey Bugaev <buga...@gmail.com > > <mailto:buga...@gmail.com>> > > > > LGTM, some minor nits below. > > > > > --- > > > misc/daemon.c | 88 > > ++++++++++++++++++++++++++++----------------------- > > > 1 file changed, 49 insertions(+), 39 deletions(-) > > > > > > diff --git a/misc/daemon.c b/misc/daemon.c > > > index 3c73ac2a..61da49b7 100644 > > > --- a/misc/daemon.c > > > +++ b/misc/daemon.c > > > @@ -43,50 +43,60 @@ static char sccsid[] = "@(#)daemon.c 8.1 > > (Berkeley) 6/4/93"; > > > > > > I think there is no need to ignore return code. > > > > > > Also This code clearly comes from freeBSD.. which has since updated the > > code to ignore SIGHUP when the parent exits. > > > > https://web.mit.edu/freebsd/head/lib/libc/gen/daemon.c > > <https://web.mit.edu/freebsd/head/lib/libc/gen/daemon.c> > > I think such change should be in a different patch though.
I'm going to send out a v2 of this series with the proposed changes to daemon () included (in a separate patch from reformatting it). It is my understanding, though, that setting and resetting the signal handler like that is racy, since signal delivery is asynchronous. We may get a SIGHUP some time after daemon () completes, even. Sergey
