POLLHUP polling
This page is a work in progress
Recently I've been trying to listen on multiple sockets at once for events like reading or writing. This has worked fine, but as I've tried to listen for a socket close event I've found the documentation poor and often contradictory. This page is my attempt to figure this topic out.
POSIX
POSIX specifies two functions for waiting on multiple file descriptors:
- select which originates from the BSD sockets API
- poll which originates from the System V STREAMS API
select doesn't specify anything to do with sockets being closed, so that's not too useful.
poll initially looks like it's better but it leaves a lot of questions after reading:
- Does POLLHUP apply to sockets?
- Can the events field be 0 and still get POLLHUP events?
- Under what conditions do we even get return events?
I highly recommend trying to read the poll standard and sussing out these details.
As far as I can tell nobody has brought up either of these questions during the standards process. At least in these sources:
- Aardvark defect reports for The Austin Group Specification
- Austin Group Defect Tracker
- Austin Group Document Register
- austin-group-l mailing list
The only talk of this issue I've found online is the 2001 email "poll() and events==0" which has no clear answers.
General POLLHUP lore
Richard Kettlewell's poll() and EOF page has some interesting
- events = POLLIN
- doesn't write to the socket first
- may indicate with POLLIN
- may indicate with POLLHUP
- may indicate with BOTH
- On socket disconnection, POLLIN set in poll()->revents and recv() returning 0 is the only portable and reliable method.
https://www.illumos.org/issues/4627 not generating POLLHUP dangers
https://github.com/illumos/illumos-gate/commit/68846fd00135fb0b10944e7806025cbefcfd6546
- stuff to read still in socket
- may indicate with POLLOUT if the socket has failed to connect in the first place
https://lkml.iu.edu/hypermail/linux/kernel/0605.1/1420.html
https://jira.mongodb.org/browse/CDRIVER-2996 POLLOUThttps://lists.freebsd.org/pipermail/freebsd-net/2011-September/029712.html
freebsd different ideas
POLLIN but no data, only POLLHUP https://sourceware.org/bugzilla/show_bug.cgi?id=13660
https://developer.illumos.narkive.com/dUe1v0Ya/poll-not-returning-pollhup-for-tcp-sockets-closed-by-the-other-end "ie you get POLLIN with a zero length read(). SO_KEEPALIVE won't be involved. Now if you down the network interface you should eventually get the POLLHUP."
https://bugs.dragonflybsd.org/issues/3268 pipes don't do POLLHUP BUG
Test code
With that in mind, I wrote this quick program that polls for POLLHUP and nothing else:
/* compile and run with gcc test.c -oTEST && ./TEST */ #include <stdio.h> #include <poll.h> #include <sys/socket.h> #include <unistd.h> int main(void) { /* create a socket pair, one end for parent, one end for child */ int sockets[2]; int rc; rc = socketpair(AF_UNIX, SOCK_STREAM, 0, sockets); /* fork in to to a parent in child */ int proc = fork(); if(proc == -1) { printf("fork err\n"); return 1; } if(proc != 0) { /* parent: poll for POLLHUP */ rc = close(sockets[1]); /* close child end to avoid holding open */ if(rc == -1) { printf("close err 1 \n"); return 1; } struct pollfd fd; fd.fd = sockets[0]; /* watch parent end */ fd.events = 0; fd.revents = 0; rc = poll(&fd, 1, 10000); if(rc < 1) { printf("poll err\n"); return 1; } else if(!(fd.revents & POLLHUP)) { printf("no POLLHUP?\n"); return 1; } printf("got POLLHUP, all good\n"); return 0; } else { /* child: close child socket end and quit */ rc = close(sockets[1]); /* close child end */ if(rc == -1) { printf("close err 3\n"); return 1; } return 0; } }
It's a bit nasty but it gets the job done.
TODO: test for pre-connect
TODO: don't fork
TODO: write to buffer before closing
Linux
https://man7.org/linux/man-pages/man2/poll.2.html
"The field fd contains a file descriptor for an open file."
"The field events is an input parameter, a bit mask specifying the events the application is interested in for the file descriptor fd. This field may be specified as zero, in which case the only events that can be returned in revents are POLLHUP, POLLERR, and POLLNVAL (see below)."
"POLLHUP - Hang up (only returned in revents; ignored in events). Note that when reading from a channel such as a pipe or a stream socket, this event merely indicates that the peer closed its end of the channel."
"If none of the events requested (and no error) has occurred for any of the file descriptors, then poll() blocks until one of the events occurs."
POLLRDHUP can be ignored
Linux's poll implementation
when did poll get added? how does it work?
Unix
- sysv
- sunos
BSDs
- netbsd
- freebsd
- openbsd
- illumos
- macos
Windows
wsapoll
Emulation
cygwin
minix
my dosbox code :(