Date: Thu, 10 May 2012 20:35:50 +0000 (UTC) From: Daichi GOTO <daichi@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: r235243 - stable/9/sys/fs/unionfs Message-ID: <201205102035.q4AKZog1050756@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: daichi Date: Thu May 10 20:35:50 2012 New Revision: 235243 URL: http://svn.freebsd.org/changeset/base/235243 Log: MFC: 234867 and 234944 - fixed a vnode lock hang-up issue. - fixed an incorrect lock status issue. - fixed an incorrect lock issue of unionfs root vnode removed. (pointed out by keith) - fixed an infinity loop issue. (pointed out by dumbbell) - changed to do LK_RELEASE expressly when unlocked. - fixed a unionfs_readdir math issue Submitted by: ozawa@ongs.co.jp, Matthew Fleming <mfleming@isilon.com> Modified: stable/9/sys/fs/unionfs/union_subr.c stable/9/sys/fs/unionfs/union_vfsops.c stable/9/sys/fs/unionfs/union_vnops.c Modified: stable/9/sys/fs/unionfs/union_subr.c ============================================================================== --- stable/9/sys/fs/unionfs/union_subr.c Thu May 10 20:31:08 2012 (r235242) +++ stable/9/sys/fs/unionfs/union_subr.c Thu May 10 20:35:50 2012 (r235243) @@ -2,8 +2,8 @@ * Copyright (c) 1994 Jan-Simon Pendry * Copyright (c) 1994 * The Regents of the University of California. All rights reserved. - * Copyright (c) 2005, 2006 Masanori Ozawa <ozawa@ongs.co.jp>, ONGS Inc. - * Copyright (c) 2006 Daichi Goto <daichi@freebsd.org> + * Copyright (c) 2005, 2006, 2012 Masanori Ozawa <ozawa@ongs.co.jp>, ONGS Inc. + * Copyright (c) 2006, 2012 Daichi Goto <daichi@freebsd.org> * * This code is derived from software contributed to Berkeley by * Jan-Simon Pendry. @@ -350,19 +350,22 @@ unionfs_noderem(struct vnode *vp, struct uvp = unp->un_uppervp; dvp = unp->un_dvp; unp->un_lowervp = unp->un_uppervp = NULLVP; - vp->v_vnlock = &(vp->v_lock); vp->v_data = NULL; - lockmgr(vp->v_vnlock, LK_EXCLUSIVE | LK_INTERLOCK, VI_MTX(vp)); + vp->v_object = NULL; + VI_UNLOCK(vp); + if (lvp != NULLVP) - VOP_UNLOCK(lvp, 0); + VOP_UNLOCK(lvp, LK_RELEASE); if (uvp != NULLVP) - VOP_UNLOCK(uvp, 0); - vp->v_object = NULL; + VOP_UNLOCK(uvp, LK_RELEASE); if (dvp != NULLVP && unp->un_hash.le_prev != NULL) unionfs_rem_cached_vnode(unp, dvp); + if (lockmgr(vp->v_vnlock, LK_EXCLUSIVE, VI_MTX(vp)) != 0) + panic("the lock for deletion is unacquirable."); + if (lvp != NULLVP) { vfslocked = VFS_LOCK_GIANT(lvp->v_mount); vrele(lvp); @@ -550,7 +553,7 @@ unionfs_relookup(struct vnode *dvp, stru cn->cn_flags |= (cnp->cn_flags & SAVESTART); vref(dvp); - VOP_UNLOCK(dvp, 0); + VOP_UNLOCK(dvp, LK_RELEASE); if ((error = relookup(dvp, vpp, cn))) { uma_zfree(namei_zone, cn->cn_pnbuf); @@ -955,7 +958,7 @@ unionfs_vn_create_on_upper(struct vnode *vpp = vp; unionfs_vn_create_on_upper_free_out1: - VOP_UNLOCK(udvp, 0); + VOP_UNLOCK(udvp, LK_RELEASE); unionfs_vn_create_on_upper_free_out2: if (cn.cn_flags & HASBUF) { Modified: stable/9/sys/fs/unionfs/union_vfsops.c ============================================================================== --- stable/9/sys/fs/unionfs/union_vfsops.c Thu May 10 20:31:08 2012 (r235242) +++ stable/9/sys/fs/unionfs/union_vfsops.c Thu May 10 20:35:50 2012 (r235243) @@ -1,8 +1,8 @@ /*- * Copyright (c) 1994, 1995 The Regents of the University of California. * Copyright (c) 1994, 1995 Jan-Simon Pendry. - * Copyright (c) 2005, 2006 Masanori Ozawa <ozawa@ongs.co.jp>, ONGS Inc. - * Copyright (c) 2006 Daichi Goto <daichi@freebsd.org> + * Copyright (c) 2005, 2006, 2012 Masanori Ozawa <ozawa@ongs.co.jp>, ONGS Inc. + * Copyright (c) 2006, 2012 Daichi Goto <daichi@freebsd.org> * All rights reserved. * * This code is derived from software donated to Berkeley by @@ -165,7 +165,7 @@ unionfs_domount(struct mount *mp) uid = va.va_uid; gid = va.va_gid; } - VOP_UNLOCK(mp->mnt_vnodecovered, 0); + VOP_UNLOCK(mp->mnt_vnodecovered, LK_RELEASE); if (error) return (error); @@ -250,7 +250,7 @@ unionfs_domount(struct mount *mp) * Save reference */ if (below) { - VOP_UNLOCK(upperrootvp, 0); + VOP_UNLOCK(upperrootvp, LK_RELEASE); vn_lock(lowerrootvp, LK_EXCLUSIVE | LK_RETRY); ump->um_lowervp = upperrootvp; ump->um_uppervp = lowerrootvp; @@ -281,7 +281,7 @@ unionfs_domount(struct mount *mp) /* * Unlock the node */ - VOP_UNLOCK(ump->um_uppervp, 0); + VOP_UNLOCK(ump->um_uppervp, LK_RELEASE); /* * Get the unionfs root vnode. Modified: stable/9/sys/fs/unionfs/union_vnops.c ============================================================================== --- stable/9/sys/fs/unionfs/union_vnops.c Thu May 10 20:31:08 2012 (r235242) +++ stable/9/sys/fs/unionfs/union_vnops.c Thu May 10 20:35:50 2012 (r235243) @@ -2,8 +2,8 @@ * Copyright (c) 1992, 1993, 1994, 1995 Jan-Simon Pendry. * Copyright (c) 1992, 1993, 1994, 1995 * The Regents of the University of California. - * Copyright (c) 2005, 2006 Masanori Ozawa <ozawa@ongs.co.jp>, ONGS Inc. - * Copyright (c) 2006 Daichi Goto <daichi@freebsd.org> + * Copyright (c) 2005, 2006, 2012 Masanori Ozawa <ozawa@ongs.co.jp>, ONGS Inc. + * Copyright (c) 2006, 2012 Daichi Goto <daichi@freebsd.org> * All rights reserved. * * This code is derived from software contributed to Berkeley by @@ -75,21 +75,6 @@ KASSERT(((vp)->v_op == &unionfs_vnodeops), \ ("unionfs: it is not unionfs-vnode")) -/* lockmgr lock <-> reverse table */ -struct lk_lr_table { - int lock; - int revlock; -}; - -static struct lk_lr_table un_llt[] = { - {LK_SHARED, LK_RELEASE}, - {LK_EXCLUSIVE, LK_RELEASE}, - {LK_UPGRADE, LK_DOWNGRADE}, - {LK_DOWNGRADE, LK_UPGRADE}, - {0, 0} -}; - - static int unionfs_lookup(struct vop_cachedlookup_args *ap) { @@ -141,7 +126,7 @@ unionfs_lookup(struct vop_cachedlookup_a if (udvp != NULLVP) { dtmpvp = udvp; if (ldvp != NULLVP) - VOP_UNLOCK(ldvp, 0); + VOP_UNLOCK(ldvp, LK_RELEASE); } else dtmpvp = ldvp; @@ -149,7 +134,7 @@ unionfs_lookup(struct vop_cachedlookup_a error = VOP_LOOKUP(dtmpvp, &vp, cnp); if (dtmpvp == udvp && ldvp != NULLVP) { - VOP_UNLOCK(udvp, 0); + VOP_UNLOCK(udvp, LK_RELEASE); vn_lock(dvp, LK_EXCLUSIVE | LK_RETRY); } @@ -161,10 +146,10 @@ unionfs_lookup(struct vop_cachedlookup_a */ if (nameiop == DELETE || nameiop == RENAME || (cnp->cn_lkflags & LK_TYPE_MASK)) - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); vrele(vp); - VOP_UNLOCK(dvp, 0); + VOP_UNLOCK(dvp, LK_RELEASE); *(ap->a_vpp) = dunp->un_dvp; vref(dunp->un_dvp); @@ -202,7 +187,7 @@ unionfs_lookup(struct vop_cachedlookup_a } if (nameiop == DELETE || nameiop == RENAME || (cnp->cn_lkflags & LK_TYPE_MASK)) - VOP_UNLOCK(uvp, 0); + VOP_UNLOCK(uvp, LK_RELEASE); } /* check whiteout */ @@ -246,7 +231,7 @@ unionfs_lookup(struct vop_cachedlookup_a return (lerror); } if (cnp->cn_lkflags & LK_TYPE_MASK) - VOP_UNLOCK(lvp, 0); + VOP_UNLOCK(lvp, LK_RELEASE); } } @@ -281,7 +266,7 @@ unionfs_lookup(struct vop_cachedlookup_a goto unionfs_lookup_out; if (LK_SHARED == (cnp->cn_lkflags & LK_TYPE_MASK)) - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); if (LK_EXCLUSIVE != VOP_ISLOCKED(vp)) { vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); lockflag = 1; @@ -289,7 +274,7 @@ unionfs_lookup(struct vop_cachedlookup_a error = unionfs_mkshadowdir(MOUNTTOUNIONFSMOUNT(dvp->v_mount), udvp, VTOUNIONFS(vp), cnp, td); if (lockflag != 0) - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); if (error != 0) { UNIONFSDEBUG("unionfs_lookup: Unable to create shadow dir."); if ((cnp->cn_lkflags & LK_TYPE_MASK) == LK_EXCLUSIVE) @@ -386,7 +371,7 @@ unionfs_create(struct vop_create_args *a if (vp->v_type == VSOCK) *(ap->a_vpp) = vp; else { - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); error = unionfs_nodeget(ap->a_dvp->v_mount, vp, NULLVP, ap->a_dvp, ap->a_vpp, cnp, curthread); vrele(vp); @@ -460,7 +445,7 @@ unionfs_mknod(struct vop_mknod_args *ap) if (vp->v_type == VSOCK) *(ap->a_vpp) = vp; else { - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); error = unionfs_nodeget(ap->a_dvp->v_mount, vp, NULLVP, ap->a_dvp, ap->a_vpp, cnp, curthread); vrele(vp); @@ -564,6 +549,7 @@ unionfs_close(struct vop_close_args *ap) struct unionfs_node_status *unsp; struct ucred *cred; struct thread *td; + struct vnode *vp; struct vnode *ovp; UNIONFS_INTERNAL_DEBUG("unionfs_close: enter\n"); @@ -571,12 +557,14 @@ unionfs_close(struct vop_close_args *ap) KASSERT_UNIONFS_VNODE(ap->a_vp); locked = 0; - unp = VTOUNIONFS(ap->a_vp); + vp = ap->a_vp; + unp = VTOUNIONFS(vp); cred = ap->a_cred; td = ap->a_td; - if (VOP_ISLOCKED(ap->a_vp) != LK_EXCLUSIVE) { - vn_lock(ap->a_vp, LK_EXCLUSIVE | LK_RETRY); + if (VOP_ISLOCKED(vp) != LK_EXCLUSIVE) { + if (vn_lock(vp, LK_UPGRADE) != 0) + vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); locked = 1; } unionfs_get_node_status(unp, td, &unsp); @@ -599,7 +587,7 @@ unionfs_close(struct vop_close_args *ap) if (error != 0) goto unionfs_close_abort; - ap->a_vp->v_object = ovp->v_object; + vp->v_object = ovp->v_object; if (ovp == unp->un_uppervp) { unsp->uns_upper_opencnt--; @@ -610,7 +598,7 @@ unionfs_close(struct vop_close_args *ap) unsp->uns_lower_opencnt--; } if (unsp->uns_lower_opencnt > 0) - ap->a_vp->v_object = unp->un_lowervp->v_object; + vp->v_object = unp->un_lowervp->v_object; } } else unsp->uns_lower_opencnt--; @@ -619,7 +607,7 @@ unionfs_close_abort: unionfs_tryrem_node_status(unp, unsp); if (locked != 0) - VOP_UNLOCK(ap->a_vp, 0); + vn_lock(vp, LK_DOWNGRADE | LK_RETRY); UNIONFS_INTERNAL_DEBUG("unionfs_close: leave (%d)\n", error); @@ -914,7 +902,7 @@ unionfs_ioctl(struct vop_ioctl_args *ap) unionfs_get_node_status(unp, ap->a_td, &unsp); ovp = (unsp->uns_upper_opencnt ? unp->un_uppervp : unp->un_lowervp); unionfs_tryrem_node_status(unp, unsp); - VOP_UNLOCK(ap->a_vp, 0); + VOP_UNLOCK(ap->a_vp, LK_RELEASE); if (ovp == NULLVP) return (EBADF); @@ -941,7 +929,7 @@ unionfs_poll(struct vop_poll_args *ap) unionfs_get_node_status(unp, ap->a_td, &unsp); ovp = (unsp->uns_upper_opencnt ? unp->un_uppervp : unp->un_lowervp); unionfs_tryrem_node_status(unp, unsp); - VOP_UNLOCK(ap->a_vp, 0); + VOP_UNLOCK(ap->a_vp, LK_RELEASE); if (ovp == NULLVP) return (EBADF); @@ -1001,7 +989,7 @@ unionfs_remove(struct vop_remove_args *a ump = NULL; vp = uvp = lvp = NULLVP; /* search vnode */ - VOP_UNLOCK(ap->a_vp, 0); + VOP_UNLOCK(ap->a_vp, LK_RELEASE); error = unionfs_relookup(udvp, &vp, cnp, &cn, td, cnp->cn_nameptr, strlen(cnp->cn_nameptr), DELETE); if (error != 0 && error != ENOENT) { @@ -1204,7 +1192,7 @@ unionfs_rename(struct vop_rename_args *a if ((error = vn_lock(fvp, LK_EXCLUSIVE)) != 0) goto unionfs_rename_abort; error = unionfs_copyfile(unp, 1, fcnp->cn_cred, td); - VOP_UNLOCK(fvp, 0); + VOP_UNLOCK(fvp, LK_RELEASE); if (error != 0) goto unionfs_rename_abort; break; @@ -1212,7 +1200,7 @@ unionfs_rename(struct vop_rename_args *a if ((error = vn_lock(fvp, LK_EXCLUSIVE)) != 0) goto unionfs_rename_abort; error = unionfs_mkshadowdir(ump, rfdvp, unp, fcnp, td); - VOP_UNLOCK(fvp, 0); + VOP_UNLOCK(fvp, LK_RELEASE); if (error != 0) goto unionfs_rename_abort; break; @@ -1269,13 +1257,13 @@ unionfs_rename(struct vop_rename_args *a if ((error = vn_lock(fdvp, LK_EXCLUSIVE)) != 0) goto unionfs_rename_abort; error = unionfs_relookup_for_delete(fdvp, fcnp, td); - VOP_UNLOCK(fdvp, 0); + VOP_UNLOCK(fdvp, LK_RELEASE); if (error != 0) goto unionfs_rename_abort; /* Locke of tvp is canceled in order to avoid recursive lock. */ if (tvp != NULLVP && tvp != tdvp) - VOP_UNLOCK(tvp, 0); + VOP_UNLOCK(tvp, LK_RELEASE); error = unionfs_relookup_for_rename(tdvp, tcnp, td); if (tvp != NULLVP && tvp != tdvp) vn_lock(tvp, LK_EXCLUSIVE | LK_RETRY); @@ -1293,11 +1281,11 @@ unionfs_rename(struct vop_rename_args *a } if (ltdvp != NULLVP) - VOP_UNLOCK(ltdvp, 0); + VOP_UNLOCK(ltdvp, LK_RELEASE); if (tdvp != rtdvp) vrele(tdvp); if (ltvp != NULLVP) - VOP_UNLOCK(ltvp, 0); + VOP_UNLOCK(ltvp, LK_RELEASE); if (tvp != rtvp && tvp != NULLVP) { if (rtvp == NULLVP) vput(tvp); @@ -1371,7 +1359,7 @@ unionfs_mkdir(struct vop_mkdir_args *ap) } if ((error = VOP_MKDIR(udvp, &uvp, cnp, ap->a_vap)) == 0) { - VOP_UNLOCK(uvp, 0); + VOP_UNLOCK(uvp, LK_RELEASE); cnp->cn_lkflags = LK_EXCLUSIVE; error = unionfs_nodeget(ap->a_dvp->v_mount, uvp, NULLVP, ap->a_dvp, ap->a_vpp, cnp, td); @@ -1427,7 +1415,9 @@ unionfs_rmdir(struct vop_rmdir_args *ap) ump = MOUNTTOUNIONFSMOUNT(ap->a_vp->v_mount); if (ump->um_whitemode == UNIONFS_WHITE_ALWAYS || lvp != NULLVP) cnp->cn_flags |= DOWHITEOUT; - error = VOP_RMDIR(udvp, uvp, cnp); + error = unionfs_relookup_for_delete(ap->a_dvp, cnp, td); + if (!error) + error = VOP_RMDIR(udvp, uvp, cnp); } else if (lvp != NULLVP) error = unionfs_mkwhiteout(udvp, cnp, td, unp->un_path); @@ -1467,7 +1457,7 @@ unionfs_symlink(struct vop_symlink_args if (udvp != NULLVP) { error = VOP_SYMLINK(udvp, &uvp, cnp, ap->a_vap, ap->a_target); if (error == 0) { - VOP_UNLOCK(uvp, 0); + VOP_UNLOCK(uvp, LK_RELEASE); cnp->cn_lkflags = LK_EXCLUSIVE; error = unionfs_nodeget(ap->a_dvp->v_mount, uvp, NULLVP, ap->a_dvp, ap->a_vpp, cnp, td); @@ -1487,9 +1477,11 @@ unionfs_readdir(struct vop_readdir_args int error; int eofflag; int locked; + int uio_offset_bk; struct unionfs_node *unp; struct unionfs_node_status *unsp; struct uio *uio; + struct vnode *vp; struct vnode *uvp; struct vnode *lvp; struct thread *td; @@ -1505,41 +1497,50 @@ unionfs_readdir(struct vop_readdir_args error = 0; eofflag = 0; locked = 0; - unp = VTOUNIONFS(ap->a_vp); + uio_offset_bk = 0; uio = ap->a_uio; - uvp = unp->un_uppervp; - lvp = unp->un_lowervp; + uvp = NULLVP; + lvp = NULLVP; td = uio->uio_td; ncookies_bk = 0; cookies_bk = NULL; - if (ap->a_vp->v_type != VDIR) + vp = ap->a_vp; + if (vp->v_type != VDIR) return (ENOTDIR); - /* check opaque */ - if (uvp != NULLVP && lvp != NULLVP) { - if ((error = VOP_GETATTR(uvp, &va, ap->a_cred)) != 0) - goto unionfs_readdir_exit; - if (va.va_flags & OPAQUE) - lvp = NULLVP; - } - /* check the open count. unionfs needs to open before readdir. */ - if (VOP_ISLOCKED(ap->a_vp) != LK_EXCLUSIVE) { - vn_lock(ap->a_vp, LK_UPGRADE | LK_RETRY); + if (VOP_ISLOCKED(vp) != LK_EXCLUSIVE) { + if (vn_lock(vp, LK_UPGRADE) != 0) + vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); locked = 1; } - unionfs_get_node_status(unp, td, &unsp); - if ((uvp != NULLVP && unsp->uns_upper_opencnt <= 0) || - (lvp != NULLVP && unsp->uns_lower_opencnt <= 0)) { - unionfs_tryrem_node_status(unp, unsp); + unp = VTOUNIONFS(vp); + if (unp == NULL) error = EBADF; + else { + uvp = unp->un_uppervp; + lvp = unp->un_lowervp; + unionfs_get_node_status(unp, td, &unsp); + if ((uvp != NULLVP && unsp->uns_upper_opencnt <= 0) || + (lvp != NULLVP && unsp->uns_lower_opencnt <= 0)) { + unionfs_tryrem_node_status(unp, unsp); + error = EBADF; + } } - if (locked == 1) - vn_lock(ap->a_vp, LK_DOWNGRADE | LK_RETRY); + if (locked) + vn_lock(vp, LK_DOWNGRADE | LK_RETRY); if (error != 0) goto unionfs_readdir_exit; + /* check opaque */ + if (uvp != NULLVP && lvp != NULLVP) { + if ((error = VOP_GETATTR(uvp, &va, ap->a_cred)) != 0) + goto unionfs_readdir_exit; + if (va.va_flags & OPAQUE) + lvp = NULLVP; + } + /* upper only */ if (uvp != NULLVP && lvp == NULLVP) { error = VOP_READDIR(uvp, uio, ap->a_cred, ap->a_eofflag, @@ -1576,7 +1577,7 @@ unionfs_readdir(struct vop_readdir_args unsp->uns_readdir_status = 1; /* - * ufs(and other fs) needs size of uio_resid larger than + * UFS(and other FS) needs size of uio_resid larger than * DIRBLKSIZ. * size of DIRBLKSIZ equals DEV_BSIZE. * (see: ufs/ufs/ufs_vnops.c ufs_readdir func , ufs/ufs/dir.h) @@ -1585,7 +1586,7 @@ unionfs_readdir(struct vop_readdir_args goto unionfs_readdir_exit; /* - * backup cookies + * Backup cookies. * It prepares to readdir in lower. */ if (ap->a_ncookies != NULL) { @@ -1601,6 +1602,11 @@ unionfs_readdir(struct vop_readdir_args /* initialize for readdir in lower */ if (unsp->uns_readdir_status == 1) { unsp->uns_readdir_status = 2; + /* + * Backup uio_offset. See the comment after the + * VOP_READDIR call on the lower layer. + */ + uio_offset_bk = uio->uio_offset; uio->uio_offset = 0; } @@ -1612,6 +1618,19 @@ unionfs_readdir(struct vop_readdir_args error = VOP_READDIR(lvp, uio, ap->a_cred, ap->a_eofflag, ap->a_ncookies, ap->a_cookies); + /* + * We can't return an uio_offset of 0: this would trigger an + * infinite loop, because the next call to unionfs_readdir would + * always restart with the upper layer (uio_offset == 0) and + * always return some data. + * + * This happens when the lower layer root directory is removed. + * (A root directory deleting of unionfs should not be permitted. + * But current VFS can not do it.) + */ + if (uio->uio_offset == 0) + uio->uio_offset = uio_offset_bk; + if (cookies_bk != NULL) { /* merge cookies */ int size; @@ -1623,7 +1642,7 @@ unionfs_readdir(struct vop_readdir_args pos = newcookies; memcpy(pos, cookies_bk, ncookies_bk * sizeof(u_long)); - pos += ncookies_bk * sizeof(u_long); + pos += ncookies_bk; memcpy(pos, *(ap->a_cookies), *(ap->a_ncookies) * sizeof(u_long)); free(cookies_bk, M_TEMP); free(*(ap->a_cookies), M_TEMP); @@ -1743,18 +1762,66 @@ unionfs_print(struct vop_print_args *ap) } static int -unionfs_get_llt_revlock(int flags) +unionfs_islocked(struct vop_islocked_args *ap) { - int count; + struct unionfs_node *unp; - flags &= LK_TYPE_MASK; - for (count = 0; un_llt[count].lock != 0; count++) { - if (flags == un_llt[count].lock) { - return un_llt[count].revlock; - } + KASSERT_UNIONFS_VNODE(ap->a_vp); + + unp = VTOUNIONFS(ap->a_vp); + if (unp == NULL) + return (vop_stdislocked(ap)); + + if (unp->un_uppervp != NULLVP) + return (VOP_ISLOCKED(unp->un_uppervp)); + if (unp->un_lowervp != NULLVP) + return (VOP_ISLOCKED(unp->un_lowervp)); + return (vop_stdislocked(ap)); +} + +static int +unionfs_get_llt_revlock(struct vnode *vp, int flags) +{ + int revlock; + + revlock = 0; + + switch (flags & LK_TYPE_MASK) { + case LK_SHARED: + if (VOP_ISLOCKED(vp) == LK_EXCLUSIVE) + revlock = LK_UPGRADE; + else + revlock = LK_RELEASE; + break; + case LK_EXCLUSIVE: + case LK_UPGRADE: + revlock = LK_RELEASE; + break; + case LK_DOWNGRADE: + revlock = LK_UPGRADE; + break; + default: + break; } - return 0; + return (revlock); +} + +/* + * The state of an acquired lock is adjusted similarly to + * the time of error generating. + * flags: LK_RELEASE or LK_UPGRADE + */ +static void +unionfs_revlock(struct vnode *vp, int flags) +{ + if (flags & LK_RELEASE) + VOP_UNLOCK(vp, flags); + else { + /* UPGRADE */ + if (vn_lock(vp, flags) != 0) + vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); + } } static int @@ -1763,6 +1830,7 @@ unionfs_lock(struct vop_lock1_args *ap) int error; int flags; int revlock; + int interlock; int uhold; struct mount *mp; struct unionfs_mount *ump; @@ -1774,15 +1842,13 @@ unionfs_lock(struct vop_lock1_args *ap) KASSERT_UNIONFS_VNODE(ap->a_vp); error = 0; + interlock = 1; uhold = 0; flags = ap->a_flags; vp = ap->a_vp; if (LK_RELEASE == (flags & LK_TYPE_MASK) || !(flags & LK_TYPE_MASK)) - return (VOP_UNLOCK(vp, flags)); - - if ((revlock = unionfs_get_llt_revlock(flags)) == 0) - panic("unknown lock type: 0x%x", flags & LK_TYPE_MASK); + return (VOP_UNLOCK(vp, flags | LK_RELEASE)); if ((flags & LK_INTERLOCK) == 0) VI_LOCK(vp); @@ -1798,6 +1864,9 @@ unionfs_lock(struct vop_lock1_args *ap) lvp = unp->un_lowervp; uvp = unp->un_uppervp; + if ((revlock = unionfs_get_llt_revlock(vp, flags)) == 0) + panic("unknown lock type: 0x%x", flags & LK_TYPE_MASK); + if ((mp->mnt_kern_flag & MNTK_MPSAFE) != 0 && (vp->v_iflag & VI_OWEINACT) != 0) flags |= LK_NOWAIT; @@ -1811,6 +1880,23 @@ unionfs_lock(struct vop_lock1_args *ap) flags |= LK_CANRECURSE; if (lvp != NULLVP) { + if (uvp != NULLVP && flags & LK_UPGRADE) { + /* Share Lock is once released and a deadlock is avoided. */ + VI_LOCK_FLAGS(uvp, MTX_DUPOK); + vholdl(uvp); + uhold = 1; + VI_UNLOCK(vp); + VOP_UNLOCK(uvp, LK_RELEASE | LK_INTERLOCK); + VI_LOCK(vp); + unp = VTOUNIONFS(vp); + if (unp == NULL) { + /* vnode is released. */ + VI_UNLOCK(vp); + VOP_UNLOCK(lvp, LK_RELEASE); + vdrop(uvp); + return (EBUSY); + } + } VI_LOCK_FLAGS(lvp, MTX_DUPOK); flags |= LK_INTERLOCK; vholdl(lvp); @@ -1823,19 +1909,28 @@ unionfs_lock(struct vop_lock1_args *ap) VI_LOCK(vp); unp = VTOUNIONFS(vp); if (unp == NULL) { + /* vnode is released. */ VI_UNLOCK(vp); if (error == 0) - VOP_UNLOCK(lvp, 0); + VOP_UNLOCK(lvp, LK_RELEASE); vdrop(lvp); + if (uhold != 0) + vdrop(uvp); return (vop_stdlock(ap)); } } if (error == 0 && uvp != NULLVP) { + if (uhold && flags & LK_UPGRADE) { + flags &= ~LK_TYPE_MASK; + flags |= LK_EXCLUSIVE; + } VI_LOCK_FLAGS(uvp, MTX_DUPOK); flags |= LK_INTERLOCK; - vholdl(uvp); - uhold = 1; + if (uhold == 0) { + vholdl(uvp); + uhold = 1; + } VI_UNLOCK(vp); ap->a_flags &= ~LK_INTERLOCK; @@ -1845,30 +1940,27 @@ unionfs_lock(struct vop_lock1_args *ap) VI_LOCK(vp); unp = VTOUNIONFS(vp); if (unp == NULL) { + /* vnode is released. */ VI_UNLOCK(vp); - if (error == 0) { - VOP_UNLOCK(uvp, 0); - if (lvp != NULLVP) - VOP_UNLOCK(lvp, 0); - } - if (lvp != NULLVP) - vdrop(lvp); + if (error == 0) + VOP_UNLOCK(uvp, LK_RELEASE); vdrop(uvp); + if (lvp != NULLVP) { + VOP_UNLOCK(lvp, LK_RELEASE); + vdrop(lvp); + } return (vop_stdlock(ap)); } - if (error != 0 && lvp != NULLVP) { + /* rollback */ VI_UNLOCK(vp); - if ((revlock & LK_TYPE_MASK) == LK_RELEASE) - VOP_UNLOCK(lvp, revlock); - else - vn_lock(lvp, revlock | LK_RETRY); - goto unionfs_lock_abort; + unionfs_revlock(lvp, revlock); + interlock = 0; } } - VI_UNLOCK(vp); -unionfs_lock_abort: + if (interlock) + VI_UNLOCK(vp); if (lvp != NULLVP) vdrop(lvp); if (uhold != 0) @@ -2013,7 +2105,7 @@ unionfs_advlock(struct vop_advlock_args unionfs_tryrem_node_status(unp, unsp); } - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); error = VOP_ADVLOCK(uvp, ap->a_id, ap->a_op, ap->a_fl, ap->a_flags); @@ -2022,7 +2114,7 @@ unionfs_advlock(struct vop_advlock_args return error; unionfs_advlock_abort: - VOP_UNLOCK(vp, 0); + VOP_UNLOCK(vp, LK_RELEASE); UNIONFS_INTERNAL_DEBUG("unionfs_advlock: leave (%d)\n", error); @@ -2150,7 +2242,8 @@ unionfs_openextattr(struct vop_openextat error = VOP_OPENEXTATTR(tvp, ap->a_cred, ap->a_td); if (error == 0) { - vn_lock(vp, LK_UPGRADE | LK_RETRY); + if (vn_lock(vp, LK_UPGRADE) != 0) + vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); if (tvp == unp->un_uppervp) unp->un_flag |= UNIONFS_OPENEXTU; else @@ -2186,7 +2279,8 @@ unionfs_closeextattr(struct vop_closeext error = VOP_CLOSEEXTATTR(tvp, ap->a_commit, ap->a_cred, ap->a_td); if (error == 0) { - vn_lock(vp, LK_UPGRADE | LK_RETRY); + if (vn_lock(vp, LK_UPGRADE) != 0) + vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); if (tvp == unp->un_uppervp) unp->un_flag &= ~UNIONFS_OPENEXTU; else @@ -2435,6 +2529,7 @@ struct vop_vector unionfs_vnodeops = { .vop_getextattr = unionfs_getextattr, .vop_getwritemount = unionfs_getwritemount, .vop_inactive = unionfs_inactive, + .vop_islocked = unionfs_islocked, .vop_ioctl = unionfs_ioctl, .vop_link = unionfs_link, .vop_listextattr = unionfs_listextattr,
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201205102035.q4AKZog1050756>