Date: Mon, 15 Mar 2021 20:36:24 GMT From: Rick Macklem <rmacklem@FreeBSD.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-branches@FreeBSD.org Subject: git: 3481445f20e2 - stable/13 - nfsclient: add nfs node locking around uses of n_direofoffset Message-ID: <202103152036.12FKaOW9093776@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch stable/13 has been updated by rmacklem: URL: https://cgit.FreeBSD.org/src/commit/?id=3481445f20e2e5b5013fb9b61706032d0e9f3816 commit 3481445f20e2e5b5013fb9b61706032d0e9f3816 Author: Rick Macklem <rmacklem@FreeBSD.org> AuthorDate: 2021-02-28 22:53:54 +0000 Commit: Rick Macklem <rmacklem@FreeBSD.org> CommitDate: 2021-03-15 19:34:13 +0000 nfsclient: add nfs node locking around uses of n_direofoffset During code inspection I noticed that the n_direofoffset field of the NFS node was being manipulated without any lock being held to make it SMP safe. This patch adds locking of the NFS node's mutex around handling of n_direofoffset to make it SMP safe. I have not seen any failure that could be attributed to n_direofoffset being manipulated concurrently by multiple processors, but I think this is possible, since directories are read with shared vnode locking, plus locks only on individual buffer cache blocks. However, there have been as yet unexplained issues w.r.t reading large directories over NFS that could have conceivably been caused by concurrent manipulation of n_direofoffset. (cherry picked from commit 15bed8c46b32dec19e922cb89e12c8970867a303) --- sys/fs/nfsclient/nfs_clbio.c | 10 ++++++++++ sys/fs/nfsclient/nfs_clsubs.c | 3 ++- sys/fs/nfsclient/nfs_clvnops.c | 21 ++++++++++++++++----- 3 files changed, 28 insertions(+), 6 deletions(-) diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c index 09fedaa47eb8..67bc3b7ce4d5 100644 --- a/sys/fs/nfsclient/nfs_clbio.c +++ b/sys/fs/nfsclient/nfs_clbio.c @@ -584,11 +584,14 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred) break; case VDIR: NFSINCRGLOBAL(nfsstatsv1.biocache_readdirs); + NFSLOCKNODE(np); if (np->n_direofoffset && uio->uio_offset >= np->n_direofoffset) { + NFSUNLOCKNODE(np); error = 0; goto out; } + NFSUNLOCKNODE(np); lbn = (uoff_t)uio->uio_offset / NFS_DIRBLKSIZ; on = uio->uio_offset & (NFS_DIRBLKSIZ - 1); bp = nfs_getcacheblk(vp, lbn, NFS_DIRBLKSIZ, td); @@ -620,11 +623,14 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred) * NFSERR_BAD_COOKIE (double yuch!). */ for (i = 0; i <= lbn && !error; i++) { + NFSLOCKNODE(np); if (np->n_direofoffset && (i * NFS_DIRBLKSIZ) >= np->n_direofoffset) { + NFSUNLOCKNODE(np); error = 0; goto out; } + NFSUNLOCKNODE(np); bp = nfs_getcacheblk(vp, i, NFS_DIRBLKSIZ, td); if (!bp) { error = newnfs_sigintr(nmp, td); @@ -667,11 +673,13 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred) * (You need the current block first, so that you have the * directory offset cookie of the next block.) */ + NFSLOCKNODE(np); if (nmp->nm_readahead > 0 && (bp->b_flags & B_INVAL) == 0 && (np->n_direofoffset == 0 || (lbn + 1) * NFS_DIRBLKSIZ < np->n_direofoffset) && incore(&vp->v_bufobj, lbn + 1) == NULL) { + NFSUNLOCKNODE(np); rabp = nfs_getcacheblk(vp, lbn + 1, NFS_DIRBLKSIZ, td); if (rabp) { if ((rabp->b_flags & (B_CACHE|B_DELWRI)) == 0) { @@ -688,6 +696,7 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred) brelse(rabp); } } + NFSLOCKNODE(np); } /* * Unlike VREG files, whos buffer size ( bp->b_bcount ) is @@ -704,6 +713,7 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred) n = lmin(uio->uio_resid, NFS_DIRBLKSIZ - bp->b_resid - on); if (np->n_direofoffset && n > np->n_direofoffset - uio->uio_offset) n = np->n_direofoffset - uio->uio_offset; + NFSUNLOCKNODE(np); break; default: printf(" ncl_bioread: type %x unexpected\n", vp->v_type); diff --git a/sys/fs/nfsclient/nfs_clsubs.c b/sys/fs/nfsclient/nfs_clsubs.c index f26ad0452e07..d361c175aa8e 100644 --- a/sys/fs/nfsclient/nfs_clsubs.c +++ b/sys/fs/nfsclient/nfs_clsubs.c @@ -118,6 +118,7 @@ ncl_uninit(struct vfsconf *vfsp) #endif } +/* Returns with NFSLOCKNODE() held. */ void ncl_dircookie_lock(struct nfsnode *np) { @@ -125,7 +126,6 @@ ncl_dircookie_lock(struct nfsnode *np) while (np->n_flag & NDIRCOOKIELK) (void) msleep(&np->n_flag, &np->n_mtx, PZERO, "nfsdirlk", 0); np->n_flag |= NDIRCOOKIELK; - NFSUNLOCKNODE(np); } void @@ -330,6 +330,7 @@ ncl_invaldir(struct vnode *vp) KASSERT(vp->v_type == VDIR, ("nfs: invaldir not dir")); ncl_dircookie_lock(np); np->n_direofoffset = 0; + NFSUNLOCKNODE(np); np->n_cookieverf.nfsuquad[0] = 0; np->n_cookieverf.nfsuquad[1] = 0; if (LIST_FIRST(&np->n_cookies)) diff --git a/sys/fs/nfsclient/nfs_clvnops.c b/sys/fs/nfsclient/nfs_clvnops.c index fc5445ef1e76..217290b080b3 100644 --- a/sys/fs/nfsclient/nfs_clvnops.c +++ b/sys/fs/nfsclient/nfs_clvnops.c @@ -2369,8 +2369,10 @@ nfs_readdir(struct vop_readdir_args *ap) /* * First, check for hit on the EOF offset cache */ + NFSLOCKNODE(np); if (np->n_direofoffset > 0 && uio->uio_offset >= np->n_direofoffset && (np->n_flag & NMODIFIED) == 0) { + NFSUNLOCKNODE(np); if (VOP_GETATTR(vp, &vattr, ap->a_cred) == 0) { NFSLOCKNODE(np); if ((NFS_ISV4(vp) && np->n_change == vattr.va_filerev) || @@ -2383,7 +2385,8 @@ nfs_readdir(struct vop_readdir_args *ap) } else NFSUNLOCKNODE(np); } - } + } else + NFSUNLOCKNODE(np); /* * NFS always guarantees that directory entries don't straddle @@ -2436,6 +2439,7 @@ ncl_readdirrpc(struct vnode *vp, struct uio *uiop, struct ucred *cred, * If there is no cookie, assume directory was stale. */ ncl_dircookie_lock(dnp); + NFSUNLOCKNODE(dnp); cookiep = ncl_getcookie(dnp, uiop->uio_offset, 0); if (cookiep) { cookie = *cookiep; @@ -2458,12 +2462,15 @@ ncl_readdirrpc(struct vnode *vp, struct uio *uiop, struct ucred *cred, * We are now either at the end of the directory or have filled * the block. */ - if (eof) + if (eof) { + NFSLOCKNODE(dnp); dnp->n_direofoffset = uiop->uio_offset; - else { + NFSUNLOCKNODE(dnp); + } else { if (uiop->uio_resid > 0) printf("EEK! readdirrpc resid > 0\n"); ncl_dircookie_lock(dnp); + NFSUNLOCKNODE(dnp); cookiep = ncl_getcookie(dnp, uiop->uio_offset, 1); *cookiep = cookie; ncl_dircookie_unlock(dnp); @@ -2496,6 +2503,7 @@ ncl_readdirplusrpc(struct vnode *vp, struct uio *uiop, struct ucred *cred, * If there is no cookie, assume directory was stale. */ ncl_dircookie_lock(dnp); + NFSUNLOCKNODE(dnp); cookiep = ncl_getcookie(dnp, uiop->uio_offset, 0); if (cookiep) { cookie = *cookiep; @@ -2517,12 +2525,15 @@ ncl_readdirplusrpc(struct vnode *vp, struct uio *uiop, struct ucred *cred, * We are now either at end of the directory or have filled the * the block. */ - if (eof) + if (eof) { + NFSLOCKNODE(dnp); dnp->n_direofoffset = uiop->uio_offset; - else { + NFSUNLOCKNODE(dnp); + } else { if (uiop->uio_resid > 0) printf("EEK! readdirplusrpc resid > 0\n"); ncl_dircookie_lock(dnp); + NFSUNLOCKNODE(dnp); cookiep = ncl_getcookie(dnp, uiop->uio_offset, 1); *cookiep = cookie; ncl_dircookie_unlock(dnp);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202103152036.12FKaOW9093776>