From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:43622) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1dV0Az-0000uE-J9 for qemu-devel@nongnu.org; Tue, 11 Jul 2017 14:46:26 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1dV0Av-0005ZS-M7 for qemu-devel@nongnu.org; Tue, 11 Jul 2017 14:46:25 -0400 Received: from mx1.redhat.com ([209.132.183.28]:33266) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1dV0Av-0005YY-DW for qemu-devel@nongnu.org; Tue, 11 Jul 2017 14:46:21 -0400 Date: Tue, 11 Jul 2017 19:46:16 +0100 From: "Dr. David Alan Gilbert" Message-ID: <20170711184615.GH2223@work-vm> References: <1496679576-14336-1-git-send-email-peter.maydell@linaro.org> <1496679576-14336-3-git-send-email-peter.maydell@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1496679576-14336-3-git-send-email-peter.maydell@linaro.org> Subject: Re: [Qemu-devel] [PATCH 2/2] slirp: Handle error returns from sosendoob() List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: qemu-devel@nongnu.org, patches@linaro.org, Samuel Thibault , Jan Kiszka * Peter Maydell (peter.maydell@linaro.org) wrote: > sosendoob() can return a failure code, but all its callers ignore it. > This is OK in sbappend(), as the comment there states -- we will try > again later in sowrite(). Add a (void) cast to tell Coverity so. > In sowrite() we do need to check the return value -- we should handle > a write failure in sosendoob() the same way we handle a write failure > for the normal data. > > Signed-off-by: Peter Maydell I think this is OK, I do have one worry, which is perhaps there were errors previously that would just loose OOB but get silently ignored that perhaps we survived OK. There's a comment there about seeing EAGAIN or EINTR in the normal data path and not erroring; hopefully we don't in the OOB case? However, it generally seems to be sane, so: Reviewed-by: Dr. David Alan Gilbert > --- > slirp/sbuf.c | 2 +- > slirp/socket.c | 22 ++++++++++++++++------ > 2 files changed, 17 insertions(+), 7 deletions(-) > > diff --git a/slirp/sbuf.c b/slirp/sbuf.c > index 10119d3..912f235 100644 > --- a/slirp/sbuf.c > +++ b/slirp/sbuf.c > @@ -91,7 +91,7 @@ sbappend(struct socket *so, struct mbuf *m) > if (so->so_urgc) { > sbappendsb(&so->so_rcv, m); > m_free(m); > - sosendoob(so); > + (void)sosendoob(so); > return; > } > > diff --git a/slirp/socket.c b/slirp/socket.c > index a17caa9..84cf13a 100644 > --- a/slirp/socket.c > +++ b/slirp/socket.c > @@ -404,7 +404,14 @@ sowrite(struct socket *so) > DEBUG_ARG("so = %p", so); > > if (so->so_urgc) { > - sosendoob(so); > + if (sosendoob(so) < so->so_urgc) { > + /* Treat a short write as a fatal error too, > + * rather than continuing on and sending the urgent > + * data as if it were non-urgent and leaving the > + * so_urgc count wrong. > + */ > + goto err_disconnected; > + } > if (sb->sb_cc == 0) > return 0; > } > @@ -448,11 +455,7 @@ sowrite(struct socket *so) > return 0; > > if (nn <= 0) { > - DEBUG_MISC((dfd, " --- sowrite disconnected, so->so_state = %x, errno = %d\n", > - so->so_state, errno)); > - sofcantsendmore(so); > - tcp_sockclosed(sototcpcb(so)); > - return -1; > + goto err_disconnected; > } > > #ifndef HAVE_READV > @@ -479,6 +482,13 @@ sowrite(struct socket *so) > sofcantsendmore(so); > > return nn; > + > +err_disconnected: > + DEBUG_MISC((dfd, " --- sowrite disconnected, so->so_state = %x, errno = %d\n", > + so->so_state, errno)); > + sofcantsendmore(so); > + tcp_sockclosed(sototcpcb(so)); > + return -1; > } > > /* > -- > 2.7.4 > > -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK