Date: Mon, 9 Jan 2023 15:01:44 GMT From: Ed Maste <emaste@FreeBSD.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-branches@FreeBSD.org Subject: git: 04f8674aa748 - stable/13 - unionfs: implement VOP_SET_TEXT/VOP_UNSET_TEXT Message-ID: <202301091501.309F1iaQ033429@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch stable/13 has been updated by emaste: URL: https://cgit.FreeBSD.org/src/commit/?id=04f8674aa7486ec2d73d60c2247fe5e33b5399c5 commit 04f8674aa7486ec2d73d60c2247fe5e33b5399c5 Author: Jason A. Harmening <jah@FreeBSD.org> AuthorDate: 2021-12-21 23:51:51 +0000 Commit: Ed Maste <emaste@FreeBSD.org> CommitDate: 2023-01-09 02:31:15 +0000 unionfs: implement VOP_SET_TEXT/VOP_UNSET_TEXT The implementation simply passes the text ref to the appropriate underlying vnode. Without this, the default [un]set_text implementation will only manage the text ref on the unionfs vnode, causing it to be out of sync with the underlying filesystems and potentially allowing corruption of executable file contents. On INVARIANTS kernels, it also readily produces a panic on process termination because the VM object representing the executable mapping is backed by the underlying vnode, not the unionfs vnode. PR: 251342 Reviewed by: kib Differential Revision: https://reviews.freebsd.org/D33611 (cherry picked from commit 9e891d43f586e91541bd61fb12550de296d76fd9) --- sys/fs/unionfs/union_vnops.c | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) diff --git a/sys/fs/unionfs/union_vnops.c b/sys/fs/unionfs/union_vnops.c index 09fbccb3d7ff..331235ab93d5 100644 --- a/sys/fs/unionfs/union_vnops.c +++ b/sys/fs/unionfs/union_vnops.c @@ -2503,6 +2503,38 @@ unionfs_add_writecount(struct vop_add_writecount_args *ap) return (error); } +static int +unionfs_set_text(struct vop_set_text_args *ap) +{ + struct vnode *tvp; + struct unionfs_node *unp; + int error; + + /* + * We assume text refs are managed against lvp/uvp through the + * executable mapping backed by its VM object. We therefore don't + * need to track leased text refs in the case of a forcible unmount. + */ + unp = VTOUNIONFS(ap->a_vp); + ASSERT_VOP_LOCKED(ap->a_vp, __func__); + tvp = unp->un_uppervp != NULL ? unp->un_uppervp : unp->un_lowervp; + error = VOP_SET_TEXT(tvp); + return (error); +} + +static int +unionfs_unset_text(struct vop_unset_text_args *ap) +{ + struct vnode *tvp; + struct unionfs_node *unp; + + ASSERT_VOP_LOCKED(ap->a_vp, __func__); + unp = VTOUNIONFS(ap->a_vp); + tvp = unp->un_uppervp != NULL ? unp->un_uppervp : unp->un_lowervp; + VOP_UNSET_TEXT_CHECKED(tvp); + return (0); +} + struct vop_vector unionfs_vnodeops = { .vop_default = &default_vnodeops, @@ -2553,5 +2585,7 @@ struct vop_vector unionfs_vnodeops = { .vop_write = unionfs_write, .vop_vptofh = unionfs_vptofh, .vop_add_writecount = unionfs_add_writecount, + .vop_set_text = unionfs_set_text, + .vop_unset_text = unionfs_unset_text, }; VFS_VOP_VECTOR_REGISTER(unionfs_vnodeops);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202301091501.309F1iaQ033429>