From nobody Sun May 21 20:06:34 2023 X-Original-To: dev-commits-src-branches@mlmmj.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mlmmj.nyi.freebsd.org (Postfix) with ESMTP id 4QPWn65qTxz4CQQ1; Sun, 21 May 2023 20:06:34 +0000 (UTC) (envelope-from git@FreeBSD.org) Received: from mxrelay.nyi.freebsd.org (mxrelay.nyi.freebsd.org [IPv6:2610:1c1:1:606c::19:3]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256 client-signature RSA-PSS (4096 bits) client-digest SHA256) (Client CN "mxrelay.nyi.freebsd.org", Issuer "R3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 4QPWn656JTz44Tw; Sun, 21 May 2023 20:06:34 +0000 (UTC) (envelope-from git@FreeBSD.org) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=freebsd.org; s=dkim; t=1684699594; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding; bh=cCzPUAouCVxvr1DIGjr/MPMmJlauXSXncxnLJFzhqoQ=; b=iNgWF69txJsq9OMfOc32+kC8MTWVIwozAZ2Vz04Eaqx+PTHBfl/W4c2JAMKq8TMELqmNVy HpcB/MPN7B0FS5rfJszDoZS2r+jelT4a/O59pk6cH3vAPK88rpddWPCT/4MDvoGqwls2KG T51USpXX2JvLlLRD7QauilK1j5dYdkV0NnXwfWrCpHIJj1Et+0LrcSQpR3RCXzsHDVSAUi fcsCX6IBILLayVKxuHaKYwSw4GF8MNUoXCDH2XkWGfylq2b6uKTStSKsKybqlsK6gXjWtL Vw9IJ4ssAV6GfvwmxlOSMO0DSwPiyPNf/JAiyAfSTC3IG6sk0U7bFBdj54V+8Q== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=freebsd.org; s=dkim; t=1684699594; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding; bh=cCzPUAouCVxvr1DIGjr/MPMmJlauXSXncxnLJFzhqoQ=; b=jimK0i2mHsEknVx1aElSCQzu322aTbgiHU2B5oX2ALRelpDSMhqFGocn9gDEibiYY7sPi8 LWY0p+kUZYtoN9MxJRi620CipDTe5Xd3Winj3iSIYM+Ekx/YW4HTqQ7SVjaVZTkHCyYVBN 12aWn2lpCjFqLJXjKra3s56HMFmT79Urztahc8cnH3qIaXN4Z5F/KYEX8pI7W+XgVBAdG6 kxJhwUzH6C8RlA6JDGhga+JNGkyMSNpDf8XQrITGkEV7Q1KEusQKo/GBQsZpt1HqBhoCVX rV3LtbFXX4O7bKfA+Ft2rG6TGBahh4FfyCxwi/Rx6bXFrEbnwMmdE33XjXZVIw== ARC-Authentication-Results: i=1; mx1.freebsd.org; none ARC-Seal: i=1; s=dkim; d=freebsd.org; t=1684699594; a=rsa-sha256; cv=none; b=fn1y8i65ye1IPGBvvftK1LTaNdyoAS0nnGM/MQ/nG+3z98BhNs5d23ctKJLUsAWpJWt55F q/76aJrTpGo85FaTJNi5p5oUALWdl9SayTJ9Z6oIepdwYpByARZcYns/nGn7AL6pGCdoLM 1bqrRnq5JfHAUkdws4DFaTaxFKzKAjBtg1w8dFS7u7xjxdvCfELWXl/ncU/3Ej7N/2ArnX D+cNA7tKgkmeZshSeV3pwHNV0fT7tSn1mx+9pWLVKppJNEiSHHCRq7K6SOcU6n0hKCDNec hqtvs8O2MwmCk3o/rpi6G7+3d1RTaaLWqBtKBebC9PqRLJYrKmx/GyzL6wMiyg== Received: from gitrepo.freebsd.org (gitrepo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:5]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (Client did not present a certificate) by mxrelay.nyi.freebsd.org (Postfix) with ESMTPS id 4QPWn644h5z14t7; Sun, 21 May 2023 20:06:34 +0000 (UTC) (envelope-from git@FreeBSD.org) Received: from gitrepo.freebsd.org ([127.0.1.44]) by gitrepo.freebsd.org (8.16.1/8.16.1) with ESMTP id 34LK6Y5v029398; Sun, 21 May 2023 20:06:34 GMT (envelope-from git@gitrepo.freebsd.org) Received: (from git@localhost) by gitrepo.freebsd.org (8.16.1/8.16.1/Submit) id 34LK6YIn029397; Sun, 21 May 2023 20:06:34 GMT (envelope-from git) Date: Sun, 21 May 2023 20:06:34 GMT Message-Id: <202305212006.34LK6YIn029397@gitrepo.freebsd.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-branches@FreeBSD.org From: Rick Macklem Subject: git: dc52667a928a - stable/13 - vfs_mount.c: Allow mountd(8) to do exports in a vnet prison List-Id: Commits to the stable branches of the FreeBSD src repository List-Archive: https://lists.freebsd.org/archives/dev-commits-src-branches List-Help: List-Post: List-Subscribe: List-Unsubscribe: Sender: owner-dev-commits-src-branches@freebsd.org X-BeenThere: dev-commits-src-branches@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit X-Git-Committer: rmacklem X-Git-Repository: src X-Git-Refname: refs/heads/stable/13 X-Git-Reftype: branch X-Git-Commit: dc52667a928a40d85bf781e8fa9d9f2d0674ed52 Auto-Submitted: auto-generated X-ThisMailContainsUnwantedMimeParts: N The branch stable/13 has been updated by rmacklem: URL: https://cgit.FreeBSD.org/src/commit/?id=dc52667a928a40d85bf781e8fa9d9f2d0674ed52 commit dc52667a928a40d85bf781e8fa9d9f2d0674ed52 Author: Rick Macklem AuthorDate: 2023-03-02 21:09:01 +0000 Commit: Rick Macklem CommitDate: 2023-05-21 20:03:48 +0000 vfs_mount.c: Allow mountd(8) to do exports in a vnet prison To run mountd in a vnet prison, three checks in vfs_domount() and vfs_domount_update() related to doing exports needed to be changed, so that a file system visible within the prison but mounted outside the prison can be exported. I did all three in a minimal way, only changing the checks for the specific case of a process (typically mountd) doing exports within a vnet prison and not updating the mount point in other ways. The changes are: - Ignore the error return from vfs_suser(), since the file system being mounted outside the prison will cause it to fail. - Use the priv_check(PRIV_NFS_DAEMON) for this specific case within a prison. - Skip the call to VFS_MOUNT(), since it will return an error, due to the "from" argument not being set correctly. VFS_MOUNT() does not appear to do anything for the case of doing exports only. (cherry picked from commit 4bbbd5875d32f3cbe76235d90243f713eff9b9d0) --- sys/kern/vfs_mount.c | 107 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 93 insertions(+), 14 deletions(-) diff --git a/sys/kern/vfs_mount.c b/sys/kern/vfs_mount.c index c43385b9736b..0b4b3d31cb10 100644 --- a/sys/kern/vfs_mount.c +++ b/sys/kern/vfs_mount.c @@ -76,7 +76,8 @@ __FBSDID("$FreeBSD$"); #define VFS_MOUNTARG_SIZE_MAX (1024 * 64) static int vfs_domount(struct thread *td, const char *fstype, char *fspath, - uint64_t fsflags, struct vfsoptlist **optlist); + uint64_t fsflags, bool jail_export, + struct vfsoptlist **optlist); static void free_mntarg(struct mntarg *ma); static int usermount = 0; @@ -663,7 +664,7 @@ vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions) struct vfsopt *opt, *tmp_opt; char *fstype, *fspath, *errmsg; int error, fstypelen, fspathlen, errmsg_len, errmsg_pos; - bool autoro; + bool autoro, has_nonexport, jail_export; errmsg = fspath = NULL; errmsg_len = fspathlen = 0; @@ -699,15 +700,37 @@ vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions) goto bail; } + /* + * Check to see that "export" is only used with the "update", "fstype", + * "fspath", "from" and "errmsg" options when in a vnet jail. + * These are the ones used to set/update exports by mountd(8). + * If only the above options are set in a jail that can run mountd(8), + * then the jail_export argument of vfs_domount() will be true. + * When jail_export is true, the vfs_suser() check does not cause + * failure, but limits the update to exports only. + * This allows mountd(8) running within the vnet jail + * to export file systems visible within the jail, but + * mounted outside of the jail. + */ /* * We need to see if we have the "update" option * before we call vfs_domount(), since vfs_domount() has special * logic based on MNT_UPDATE. This is very important * when we want to update the root filesystem. */ + has_nonexport = false; + jail_export = false; TAILQ_FOREACH_SAFE(opt, optlist, link, tmp_opt) { int do_freeopt = 0; + if (jailed(td->td_ucred) && + strcmp(opt->name, "export") != 0 && + strcmp(opt->name, "update") != 0 && + strcmp(opt->name, "fstype") != 0 && + strcmp(opt->name, "fspath") != 0 && + strcmp(opt->name, "from") != 0 && + strcmp(opt->name, "errmsg") != 0) + has_nonexport = true; if (strcmp(opt->name, "update") == 0) { fsflags |= MNT_UPDATE; do_freeopt = 1; @@ -790,9 +813,10 @@ vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions) fsflags |= MNT_SYNCHRONOUS; else if (strcmp(opt->name, "union") == 0) fsflags |= MNT_UNION; - else if (strcmp(opt->name, "export") == 0) + else if (strcmp(opt->name, "export") == 0) { fsflags |= MNT_EXPORTED; - else if (strcmp(opt->name, "automounted") == 0) { + jail_export = true; + } else if (strcmp(opt->name, "automounted") == 0) { fsflags |= MNT_AUTOMOUNTED; do_freeopt = 1; } else if (strcmp(opt->name, "nocover") == 0) { @@ -822,7 +846,15 @@ vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions) goto bail; } - error = vfs_domount(td, fstype, fspath, fsflags, &optlist); + /* + * If has_nonexport is true or the caller is not running within a + * vnet prison that can run mountd(8), set jail_export false. + */ + if (has_nonexport || !jailed(td->td_ucred) || + !prison_check_nfsd(td->td_ucred)) + jail_export = false; + + error = vfs_domount(td, fstype, fspath, fsflags, jail_export, &optlist); if (error == ENOENT) { error = EINVAL; if (errmsg != NULL) @@ -840,7 +872,8 @@ vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions) printf("%s: R/W mount failed, possibly R/O media," " trying R/O mount\n", __func__); fsflags |= MNT_RDONLY; - error = vfs_domount(td, fstype, fspath, fsflags, &optlist); + error = vfs_domount(td, fstype, fspath, fsflags, jail_export, + &optlist); } bail: /* copyout the errmsg */ @@ -1114,6 +1147,7 @@ vfs_domount_update( struct thread *td, /* Calling thread. */ struct vnode *vp, /* Mount point vnode. */ uint64_t fsflags, /* Flags common to all filesystems. */ + bool jail_export, /* Got export option in vnet prison. */ struct vfsoptlist **optlist /* Options local to the filesystem. */ ) { @@ -1125,6 +1159,7 @@ vfs_domount_update( int error, export_error, i, len; uint64_t flag; gid_t *grps; + bool vfs_suser_failed; ASSERT_VOP_ELOCKED(vp, __func__); KASSERT((fsflags & MNT_UPDATE) != 0, ("MNT_UPDATE should be here")); @@ -1153,7 +1188,20 @@ vfs_domount_update( * Only privileged root, or (if MNT_USER is set) the user that * did the original mount is permitted to update it. */ + /* + * For the case of mountd(8) doing exports in a jail, the vfs_suser() + * call does not cause failure. vfs_domount() has already checked + * that "root" is doing this and vfs_suser() will fail when + * the file system has been mounted outside the jail. + * jail_export set true indicates that "export" is not mixed + * with other options that change mount behaviour. + */ + vfs_suser_failed = false; error = vfs_suser(mp, td); + if (jail_export && error != 0) { + error = 0; + vfs_suser_failed = true; + } if (error != 0) { vput(vp); return (error); @@ -1183,11 +1231,26 @@ vfs_domount_update( error = EBUSY; goto end; } - mp->mnt_flag &= ~MNT_UPDATEMASK; - mp->mnt_flag |= fsflags & (MNT_RELOAD | MNT_FORCE | MNT_UPDATE | - MNT_SNAPSHOT | MNT_ROOTFS | MNT_UPDATEMASK | MNT_RDONLY); - if ((mp->mnt_flag & MNT_ASYNC) == 0) - mp->mnt_kern_flag &= ~MNTK_ASYNC; + if (vfs_suser_failed) { + KASSERT((fsflags & (MNT_EXPORTED | MNT_UPDATE)) == + (MNT_EXPORTED | MNT_UPDATE), + ("%s: jailed export did not set expected fsflags", + __func__)); + /* + * For this case, only MNT_UPDATE and + * MNT_EXPORTED have been set in fsflags + * by the options. Only set MNT_UPDATE, + * since that is the one that would be set + * when set in fsflags, below. + */ + mp->mnt_flag |= MNT_UPDATE; + } else { + mp->mnt_flag &= ~MNT_UPDATEMASK; + mp->mnt_flag |= fsflags & (MNT_RELOAD | MNT_FORCE | MNT_UPDATE | + MNT_SNAPSHOT | MNT_ROOTFS | MNT_UPDATEMASK | MNT_RDONLY); + if ((mp->mnt_flag & MNT_ASYNC) == 0) + mp->mnt_kern_flag &= ~MNTK_ASYNC; + } rootvp = vfs_cache_root_clear(mp); MNT_IUNLOCK(mp); mp->mnt_optnew = *optlist; @@ -1198,7 +1261,17 @@ vfs_domount_update( * XXX The final recipients of VFS_MOUNT just overwrite the ndp they * get. No freeing of cn_pnbuf. */ - error = VFS_MOUNT(mp); + /* + * For the case of mountd(8) doing exports from within a vnet jail, + * "from" is typically not set correctly such that VFS_MOUNT() will + * return ENOENT. It is not obvious that VFS_MOUNT() ever needs to be + * called when mountd is doing exports, but this check only applies to + * the specific case where it is running inside a vnet jail, to + * avoid any POLA violation. + */ + error = 0; + if (!jail_export) + error = VFS_MOUNT(mp); export_error = 0; /* Process the export option. */ @@ -1336,6 +1409,7 @@ vfs_domount( const char *fstype, /* Filesystem type. */ char *fspath, /* Mount path. */ uint64_t fsflags, /* Flags common to all filesystems. */ + bool jail_export, /* Got export option in vnet prison. */ struct vfsoptlist **optlist /* Options local to the filesystem. */ ) { @@ -1353,7 +1427,11 @@ vfs_domount( if (strlen(fstype) >= MFSNAMELEN || strlen(fspath) >= MNAMELEN) return (ENAMETOOLONG); - if (jailed(td->td_ucred) || usermount == 0) { + if (jail_export) { + error = priv_check(td, PRIV_NFS_DAEMON); + if (error) + return (error); + } else if (jailed(td->td_ucred) || usermount == 0) { if ((error = priv_check(td, PRIV_VFS_MOUNT)) != 0) return (error); } @@ -1434,7 +1512,8 @@ vfs_domount( } free(pathbuf, M_TEMP); } else - error = vfs_domount_update(td, vp, fsflags, optlist); + error = vfs_domount_update(td, vp, fsflags, jail_export, + optlist); out: NDFREE(&nd, NDF_ONLY_PNBUF);