> From: Paolo Bonzini [mailto:paolo.bonz...@gmail.com] On Behalf Of Paolo > Bonzini > Sent: Friday, September 07, 2012 5:25 PM > To: Joachim Schmitz > Cc: bug-gnulib@gnu.org; rsbec...@nexbridge.com > Subject: Re: poll() emulation in git/gnulib > > Il 07/09/2012 17:01, Joachim Schmitz ha scritto: > >> From: Paolo Bonzini [mailto:paolo.bonz...@gmail.com] On Behalf Of Paolo > >> Bonzini > >> Sent: Friday, September 07, 2012 4:47 PM > >> To: Joachim Schmitz > >> Cc: bug-gnulib@gnu.org; rsbec...@nexbridge.com > >> Subject: Re: poll() emulation in git/gnulib > >> > >> Il 07/09/2012 12:58, Joachim Schmitz ha scritto: > >>>> From: Paolo Bonzini [mailto:paolo.bonz...@gmail.com] On Behalf Of Paolo > >>>> Bonzini > >>>> Sent: Friday, September 07, 2012 11:41 AM > >>>> To: Joachim Schmitz > >>>> Cc: g...@vger.kernel.org; 'Junio C Hamano'; 'Erik Faye-Lund'; > >>>> bug-gnulib@gnu.org; rsbec...@nexbridge.com > >>>> Subject: Re: poll() emulation in git > >>>> > >>>> Il 07/09/2012 09:39, Joachim Schmitz ha scritto: > >>>>>>> I suppose it works to always handle ENOTSOCK that way, even on > >>>>>>> non-__TANDEM systems. > >>>>> Will you be fixing this in gnulib? How? > >>>> > >>>> I don't have access to the system, so it's best if you post the patches > >>>> yourself to bug-gnulib and git mailing lists (separately, since the code > >>>> is cross-pollinated but forked). > >>> > >>> Here's the patch hat fixed the problems for me. > >>> Whether or not to keep the #ifdef __TANDEM and/or to enable the commented > >>> additional check, to make it usable for non HP > NonStop > > I > >>> don't know... > >> > >> Please remove it, > > > > Remove what, just the "#ifdef __TANDEM" and "#endif" or the comment for the > > "(r == -1) &&" too? > > > >> and change the comment to "some systems can't use > >> recv() on non-socket, including HP NonStop". > > > > OK. > > > >> Also please add a ChangeLog. I'll commit the patch then. > > > > An entry in the ChangeLog file, right? > > > > Assuming this and that you just want the #ifdef __TANDEM removed: > > > > diff --git a/ChangeLog b/ChangeLog > > index 282e060..8e51746 100644 > > --- a/ChangeLog > > +++ b/ChangeLog > > @@ -1,3 +1,8 @@ > > +2012-09-06 Joachim Schmitz <j...@schmitz-digital.de> > > + > > + poll: fix for systems that can't recv() on a non-socket > > + poll: don't exit early if NULL is the 1st arg to poll() > > + > > 2012-09-06 Eric Blake <ebl...@redhat.com> > > > > net_if: give more details about the bug being fixeddiff --git > > a/lib/poll.c b/lib/poll.c > > index 5ad9d86..62430bb 100644 > > --- a/lib/poll.c > > +++ b/lib/poll.c > > @@ -303,6 +303,10 @@ compute_revents (int fd, int sought, fd_set *rfds, > > fd_set * > > || socket_errno == ECONNABORTED || socket_errno == > > ENETRESET) > > happened |= POLLHUP; > > > > + /* some systems can't use recv() on non-socket, including HP NonStop > > */ > > + else if (/* (r == -1) && */ socket_errno == ENOTSOCK) > > + happened |= (POLLIN | POLLRDNORM) & sought; > > + > > else > > happened |= POLLERR; > > } > > @@ -350,7 +354,7 @@ poll (struct pollfd *pfd, nfds_t nfd, int timeout) > > > > /* EFAULT is not necessary to implement, but let's do it in the > > simplest case. */ > > - if (!pfd) > > + if (!pfd && nfd) > > { > > errno = EFAULT; > > return -1; > > > > > > Ok, thanks (will get to it on Monday).
Which Monday? ;-) Bye, Jojo
