From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/13259 Path: news.gmane.org!.POSTED!not-for-mail From: Rich Felker Newsgroups: gmane.linux.lib.musl.general Subject: Re: [PATCH] improve error handling of ttyname_r and isatty Date: Thu, 13 Sep 2018 14:25:15 -0400 Message-ID: <20180913182515.GK1878@brightrain.aerifal.cx> References: <20180913162755.18911-1-benjamin@python.org> Reply-To: musl@lists.openwall.com NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: blaine.gmane.org 1536863002 18860 195.159.176.226 (13 Sep 2018 18:23:22 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Thu, 13 Sep 2018 18:23:22 +0000 (UTC) User-Agent: Mutt/1.5.21 (2010-09-15) To: musl@lists.openwall.com Original-X-From: musl-return-13275-gllmg-musl=m.gmane.org@lists.openwall.com Thu Sep 13 20:23:18 2018 Return-path: Envelope-to: gllmg-musl@m.gmane.org Original-Received: from mother.openwall.net ([195.42.179.200]) by blaine.gmane.org with smtp (Exim 4.84_2) (envelope-from ) id 1g0WGs-0004of-LM for gllmg-musl@m.gmane.org; Thu, 13 Sep 2018 20:23:18 +0200 Original-Received: (qmail 28549 invoked by uid 550); 13 Sep 2018 18:25:27 -0000 Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-ID: Original-Received: (qmail 28531 invoked from network); 13 Sep 2018 18:25:27 -0000 Content-Disposition: inline In-Reply-To: <20180913162755.18911-1-benjamin@python.org> Original-Sender: Rich Felker Xref: news.gmane.org gmane.linux.lib.musl.general:13259 Archived-At: On Thu, Sep 13, 2018 at 09:27:55AM -0700, Benjamin Peterson wrote: > POSIX allows ttyname(_r) and isatty to return EBADF if passed file descriptor is > invalid. > --- > src/unistd/isatty.c | 6 +++++- > src/unistd/ttyname_r.c | 5 ++++- > 2 files changed, 9 insertions(+), 2 deletions(-) > > diff --git a/src/unistd/isatty.c b/src/unistd/isatty.c > index c8badaf5..3d5213c6 100644 > --- a/src/unistd/isatty.c > +++ b/src/unistd/isatty.c > @@ -1,9 +1,13 @@ > #include > +#include > #include > #include "syscall.h" > > int isatty(int fd) > { > struct winsize wsz; > - return !__syscall(SYS_ioctl, fd, TIOCGWINSZ, &wsz); > + unsigned long r = __syscall_ret(__syscall(SYS_ioctl, fd, TIOCGWINSZ, &wsz)); > + if (r == 0) return 1; > + if (errno != EBADF) errno = ENOTTY; > + return 0; > } __syscall_ret(__syscall(...)) is just an ugly way of writing syscall(). :-) But the whole reason __syscall_ret exists is for the sake of setting errno via a tailcall after the __syscall returns. When you have to patch up the value, rather than shuffling it in and out of errno and breaking the tailcall, a cleaner method is something like: int r = __syscall(...); if (r && r!=-EBADF) r=-ENOTTY; return __syscall_ret(r); > diff --git a/src/unistd/ttyname_r.c b/src/unistd/ttyname_r.c > index 33aa4ae1..3f836a6a 100644 > --- a/src/unistd/ttyname_r.c > +++ b/src/unistd/ttyname_r.c > @@ -10,7 +10,10 @@ int ttyname_r(int fd, char *name, size_t size) > char procname[sizeof "/proc/self/fd/" + 3*sizeof(int) + 2]; > ssize_t l; > > - if (!isatty(fd)) return ENOTTY; > + if (!isatty(fd)) { > + if (errno != EBADF) errno = ENOTTY; > + return errno; > + } > > __procfdname(procname, fd); > l = readlink(procname, name, size); > -- > 2.17.1 With isatty setting errno reasonably, there's no need for errno fixup here. Just:: if (!isatty(fd)) return errno; suffices. Rich