Date: Wed, 1 Jan 2014 22:43:17 +0000 (UTC) From: Rick Macklem <rmacklem@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-9@freebsd.org Subject: svn commit: r260173 - stable/9/sys/fs/nfsserver Message-ID: <201401012243.s01MhHcS070874@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: rmacklem Date: Wed Jan 1 22:43:16 2014 New Revision: 260173 URL: http://svnweb.freebsd.org/changeset/base/260173 Log: MFC: r259845 An intermittent problem with NFSv4 exporting of ZFS snapshots was reported to the freebsd-fs mailing list. I believe the problem was caused by the Readdir operation using VFS_VGET() for a snapshot file entry instead of VOP_LOOKUP(). This would not occur for NFSv3, since it will do a VFS_VGET() of "." which fails with ENOTSUPP at the beginning of the directory, whereas NFSv4 does not check "." or "..". This patch adds a call to VFS_VGET() for the directory being read to check for ENOTSUPP. I also observed that the mount_on_fileid and fsid attributes were not correct at the snapshot's auto mountpoints when looking at packet traces for the Readdir. This patch fixes the attributes by doing a check for different v_mount structure, even if the vnode v_mountedhere is not set. Modified: stable/9/sys/fs/nfsserver/nfs_nfsdport.c Directory Properties: stable/9/sys/ (props changed) stable/9/sys/fs/ (props changed) Modified: stable/9/sys/fs/nfsserver/nfs_nfsdport.c ============================================================================== --- stable/9/sys/fs/nfsserver/nfs_nfsdport.c Wed Jan 1 22:35:11 2014 (r260172) +++ stable/9/sys/fs/nfsserver/nfs_nfsdport.c Wed Jan 1 22:43:16 2014 (r260173) @@ -1985,6 +1985,27 @@ again: } /* + * Check to see if entries in this directory can be safely acquired + * via VFS_VGET() or if a switch to VOP_LOOKUP() is required. + * ZFS snapshot directories need VOP_LOOKUP(), so that any + * automount of the snapshot directory that is required will + * be done. + * This needs to be done here for NFSv4, since NFSv4 never does + * a VFS_VGET() for "." or "..". + */ + if (not_zfs == 0) { + r = VFS_VGET(mp, at.na_fileid, LK_SHARED, &nvp); + if (r == EOPNOTSUPP) { + usevget = 0; + cn.cn_nameiop = LOOKUP; + cn.cn_lkflags = LK_SHARED | LK_RETRY; + cn.cn_cred = nd->nd_cred; + cn.cn_thread = p; + } else if (r == 0) + vput(nvp); + } + + /* * Save this position, in case there is an error before one entry * is created. */ @@ -2122,6 +2143,22 @@ again: if (!r) r = nfsvno_getattr(nvp, nvap, nd->nd_cred, p, 1); + if (r == 0 && not_zfs == 0 && + nfsrv_enable_crossmntpt != 0 && + (nd->nd_flag & ND_NFSV4) != 0 && + nvp->v_type == VDIR && + vp->v_mount != nvp->v_mount) { + /* + * For a ZFS snapshot, there is a + * pseudo mount that does not set + * v_mountedhere, so it needs to + * be detected via a different + * mount structure. + */ + at_root = 1; + if (new_mp == mp) + new_mp = nvp->v_mount; + } } } else { nvp = NULL;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201401012243.s01MhHcS070874>