Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com Date: Sun, 2 Feb 2003 23:15:08 -0800 Subject: Re: 1.3.19-1:poll bug. Patch included. Content-Type: text/plain; charset=US-ASCII; format=flowed Mime-Version: 1.0 (Apple Message framework v551) From: Peter Rehley To: cygwin AT cygwin DOT com Content-Transfer-Encoding: 7bit In-Reply-To: <20030203030307.GA16149@redhat.com> Message-Id: <39DA1DA2-3747-11D7-87B3-00306540AD5E@rehley.net> Hello, Just to make sure that we are both clear. Change recvfrom so that if the buffer is too small it returns the number of bytes that can be written into the buffer instead of setting errno. Also clear errno since there is no error. Now in poll since we will now be returning the number of bytes instead of -1, do we will need to check for WSAEMSGSIZE? I don't think so because that will now be handled in recvfrom. Is this correct? Peter On Sunday, February 2, 2003, at 07:03 PM, Christopher Faylor wrote: > On Sun, Feb 02, 2003 at 06:37:34PM -0800, Peter Rehley wrote: >> Hello, >> >> While working on another program (syslogd in inetutils) I found that >> the poll function was returning POLLERR when something arrived on a >> UDP >> socket; it should be return POLLIN. Inside the poll function, >> recvfrom >> was returning -1 with errno = EMSGSIZE, but the only error value being >> check for in WSAENOTCONN. However EMSGSIZE is a reasonable return >> value since the buffer to recvfrom was only 1 byte while the actual >> message size was larger than this. Basically saying your buffer is >> too >> small for the message, but since we are only peeking, this shouldn't >> be >> returned as a POLLERR. >> >> The fix should be to add EMSGSIZE as part of the condition so this >> error returns POLLIN. I'm including the patch below and a changelog >> following that. Please review these changes and make sure that they >> seem reasonable. I've done tests on my system (windows XP) and have >> verified that the change works. > > Thanks for the patch. Would you mind reworking it so that it doesn't > set > errno at all in the case that you describe and then check against the > correct > WSAGetLastError value? > > cgf > >> Patch >> ----------------------------------- >> Index: winsup/cygwin/poll.cc >> =================================================================== >> RCS file: /cvs/src/src/winsup/cygwin/poll.cc,v >> retrieving revision 1.34 >> diff -u -p -r1.34 poll.cc >> --- winsup/cygwin/poll.cc 20 Nov 2002 11:00:15 -0000 1.34 >> +++ winsup/cygwin/poll.cc 3 Feb 2003 01:52:10 -0000 >> @@ -112,7 +112,8 @@ poll (struct pollfd *fds, unsigned int n >> sense then. It returns WSAENOTCONN in that >> case. Since that's not actually an error, >> we must not set POLLERR but POLLIN. */ >> - if (WSAGetLastError () != WSAENOTCONN) >> + if (WSAGetLastError () != WSAENOTCONN && >> + errno!=EMSGSIZE) >> fds[i].revents |= POLLERR; >> else >> fds[i].revents |= POLLIN; >> ---------------------- >> Changelog >> 2003-02-02 Peter Rehley >> >> * poll.cc: EMSGSIZE being ignore returning incorrect POLLERR on UDP >> socket receive >> if recvfrom returns error, return POLLIN if errno = EMSGSIZE >> ----------------------- >> >> Peter >> ===== >> Infinity Softworks. Making scientific, financial and realtor >> calculators for Palm Pilots and other PDAs since 1997 >> >> Visit them today at http://www.infinitysw.com >> >> >> -- >> Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple >> Bug reporting: http://cygwin.com/bugs.html >> Documentation: http://cygwin.com/docs.html >> FAQ: http://cygwin.com/faq/ > > -- > Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple > Bug reporting: http://cygwin.com/bugs.html > Documentation: http://cygwin.com/docs.html > FAQ: http://cygwin.com/faq/ > > > ===== Infinity Softworks. Making scientific, financial and realtor calculators for Palm Pilots and other PDAs since 1997 Visit them today at http://www.infinitysw.com -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/