Date: Mon, 28 Jun 2010 22:16:19 +0300 From: Kostik Belousov <kostikbel@gmail.com> To: John Baldwin <jhb@freebsd.org> Cc: threads@freebsd.org Subject: Re: SIGPIPE and threads Message-ID: <20100628191619.GF13238@deviant.kiev.zoral.com.ua> In-Reply-To: <201006281444.50021.jhb@freebsd.org> References: <201006280833.54224.jhb@freebsd.org> <20100628140534.GZ13238@deviant.kiev.zoral.com.ua> <201006281444.50021.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--7Y0jlJGlNwEFz8Iq Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, Jun 28, 2010 at 02:44:49PM -0400, John Baldwin wrote: > On Monday 28 June 2010 10:05:34 am Kostik Belousov wrote: > > On Mon, Jun 28, 2010 at 08:33:54AM -0400, John Baldwin wrote: > > > Currently when a thread performs a write(2) on a disconnected socket = or a FIFO=20 > > > with no readers the SIGPIPE signal is posted to the entire process vi= a=20 > > > psignal(). This means that the signal can be delivered to any thread= in the=20 > > > process. However, it seems more intuitive to me that SIGPIPE should = be sent=20 > > > to the "offending" thread similar to signals sent in response to trap= s via=20 > > > trapsignal(). POSIX seems to require this in that the description of= the=20 > > > EPIPE error return value for write(2) and fflush(3) in the Open Group= 's online=20 > > > manpages both say that SIGPIPE should be sent to the current thread i= n=20 > > > addition to returning EPIPE: > > >=20 > > > http://www.opengroup.org/onlinepubs/000095399/functions/write.html > > >=20 > > > http://www.opengroup.org/onlinepubs/000095399/functions/fflush.html > > >=20 > > > I have an untested (only compiled) patch below: > > > > I think the patch is right, but, as you note, having a dedicated > > function that wraps automatic ksi initialization and tdsignal() > > call would be even better. >=20 > Ok, what I've done is to rename tdsignal() to tdsendsignals() and make it > private to kern_sig.c. I then added 'tdsignal()' and 'tdksignal()' to the > public KPI to mirror the existing psignal() and pksignal() routines. >=20 > This patch can be found at http://www.freebsd.org/~jhb/patches/tdsignal.p= atch It seems that tdsendsignal() call in trapsignal() can be replaced by tdksignal(), unless I am mistaken. The same for psignal_event(). There is also a reference to tdsignal() in subr_sleepqueue.c comment, that is probably better to replace with tdsendsignal(). >=20 > I then reworked the sigpipe patch to just convert > calls to psignal() to tdsignal() instead. It is at > http://www.freebsd.org/~jhb/patches/sigpipe.patch Looks good. --7Y0jlJGlNwEFz8Iq Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (FreeBSD) iEYEARECAAYFAkwo9QMACgkQC3+MBN1Mb4g2uQCgoR6eJFZG7p/nAnavGvFR7zwm iqoAoLSKIG5ezIlVejXdB3/BINFaFJZl =FuA1 -----END PGP SIGNATURE----- --7Y0jlJGlNwEFz8Iq--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100628191619.GF13238>