On Thu, 24 Jun 2004, Jun Kuriyama wrote: > I don't know how to fix correctly. But it has disappeared with > following patch. Grrr. I think this is not quite correct; you will need to unlock the nfsd_mtx regardless of the state of RC_LOCKED, as that's a structure sleep lock and nfsd_mtx should be locked at that point regardless of RC_LOCKED. Assuming this code was safe on 4.x and performed a blocking socket addr copy here, it should also be safe on 5.x to drop nfsd_mtx here. Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert_at_fledge.watson.org Senior Research Scientist, McAfee Research > > Index: nfs_srvcache.c > =================================================================== > RCS file: /home/ncvs/src/sys/nfsserver/nfs_srvcache.c,v > retrieving revision 1.38 > diff -u -r1.38 nfs_srvcache.c > --- nfs_srvcache.c 24 May 2004 04:06:14 -0000 1.38 > +++ nfs_srvcache.c 22 Jun 2004 06:28:12 -0000 > _at__at_ -247,10 +247,19 _at__at_ > break; > /* case AF_INET6: */ > /* case AF_ISO: */ > - default: > + default: { > + struct sockaddr *sa2; > rp->rc_flag |= RC_NAM; > - rp->rc_nam = sodupsockaddr(nd->nd_nam, M_WAITOK); > + if (rp->rc_flag | RC_LOCKED) { > + NFSD_UNLOCK(); > + } > + sa2 = sodupsockaddr(nd->nd_nam, M_WAITOK); > + if (rp->rc_flag | RC_LOCKED) { > + NFSD_LOCK(); > + } > + rp->rc_nam = sa2; > break; > + } > }; > rp->rc_proc = nd->nd_procnum; > LIST_INSERT_HEAD(NFSRCHASH(nd->nd_retxid), rp, rc_hash); > > > -- > Jun Kuriyama <kuriyama_at_imgsrc.co.jp> // IMG SRC, Inc. > <kuriyama_at_FreeBSD.org> // FreeBSD Project > _______________________________________________ > freebsd-current_at_freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org" >Received on Thu Jun 24 2004 - 12:11:46 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:58 UTC