diff options
Diffstat (limited to 'sys/kern')
-rw-r--r-- | sys/kern/kern_descrip.c | 114 | ||||
-rw-r--r-- | sys/kern/subr_asan.c | 3 | ||||
-rw-r--r-- | sys/kern/subr_pctrie.c | 36 | ||||
-rw-r--r-- | sys/kern/subr_trap.c | 5 | ||||
-rw-r--r-- | sys/kern/sys_generic.c | 6 | ||||
-rw-r--r-- | sys/kern/vfs_cache.c | 13 | ||||
-rw-r--r-- | sys/kern/vfs_inotify.c | 7 | ||||
-rw-r--r-- | sys/kern/vfs_syscalls.c | 17 |
8 files changed, 146 insertions, 55 deletions
diff --git a/sys/kern/kern_descrip.c b/sys/kern/kern_descrip.c index 406236fc2723..a27ab33b34da 100644 --- a/sys/kern/kern_descrip.c +++ b/sys/kern/kern_descrip.c @@ -480,6 +480,92 @@ kern_fcntl_freebsd(struct thread *td, int fd, int cmd, intptr_t arg) return (error); } +struct flags_trans_elem { + u_int f; + u_int t; +}; + +static u_int +flags_trans(const struct flags_trans_elem *ftes, int nitems, u_int from_flags) +{ + u_int res; + int i; + + res = 0; + for (i = 0; i < nitems; i++) { + if ((from_flags & ftes[i].f) != 0) + res |= ftes[i].t; + } + return (res); +} + +static uint8_t +fd_to_fde_flags(int fd_flags) +{ + static const struct flags_trans_elem fd_to_fde_flags_s[] = { + { .f = FD_CLOEXEC, .t = UF_EXCLOSE }, + { .f = FD_CLOFORK, .t = UF_FOCLOSE }, + { .f = FD_RESOLVE_BENEATH, .t = UF_RESOLVE_BENEATH }, + }; + + return (flags_trans(fd_to_fde_flags_s, nitems(fd_to_fde_flags_s), + fd_flags)); +} + +static int +fde_to_fd_flags(uint8_t fde_flags) +{ + static const struct flags_trans_elem fde_to_fd_flags_s[] = { + { .f = UF_EXCLOSE, .t = FD_CLOEXEC }, + { .f = UF_FOCLOSE, .t = FD_CLOFORK }, + { .f = UF_RESOLVE_BENEATH, .t = FD_RESOLVE_BENEATH }, + }; + + return (flags_trans(fde_to_fd_flags_s, nitems(fde_to_fd_flags_s), + fde_flags)); +} + +static uint8_t +fddup_to_fde_flags(int fddup_flags) +{ + static const struct flags_trans_elem fddup_to_fde_flags_s[] = { + { .f = FDDUP_FLAG_CLOEXEC, .t = UF_EXCLOSE }, + { .f = FDDUP_FLAG_CLOFORK, .t = UF_FOCLOSE }, + }; + + return (flags_trans(fddup_to_fde_flags_s, nitems(fddup_to_fde_flags_s), + fddup_flags)); +} + +static uint8_t +close_range_to_fde_flags(int close_range_flags) +{ + static const struct flags_trans_elem close_range_to_fde_flags_s[] = { + { .f = CLOSE_RANGE_CLOEXEC, .t = UF_EXCLOSE }, + { .f = CLOSE_RANGE_CLOFORK, .t = UF_FOCLOSE }, + }; + + return (flags_trans(close_range_to_fde_flags_s, + nitems(close_range_to_fde_flags_s), close_range_flags)); +} + +static uint8_t +open_to_fde_flags(int open_flags, bool sticky_orb) +{ + static const struct flags_trans_elem open_to_fde_flags_s[] = { + { .f = O_CLOEXEC, .t = UF_EXCLOSE }, + { .f = O_CLOFORK, .t = UF_FOCLOSE }, + { .f = O_RESOLVE_BENEATH, .t = UF_RESOLVE_BENEATH }, + }; +#if defined(__clang__) && __clang_major__ >= 19 + _Static_assert(open_to_fde_flags_s[nitems(open_to_fde_flags_s) - 1].f == + O_RESOLVE_BENEATH, "O_RESOLVE_BENEATH must be last, for sticky_orb"); +#endif + + return (flags_trans(open_to_fde_flags_s, nitems(open_to_fde_flags_s) - + (sticky_orb ? 0 : 1), open_flags)); +} + int kern_fcntl(struct thread *td, int fd, int cmd, intptr_t arg) { @@ -534,11 +620,7 @@ kern_fcntl(struct thread *td, int fd, int cmd, intptr_t arg) FILEDESC_SLOCK(fdp); fde = fdeget_noref(fdp, fd); if (fde != NULL) { - td->td_retval[0] = - ((fde->fde_flags & UF_EXCLOSE) ? FD_CLOEXEC : 0) | - ((fde->fde_flags & UF_FOCLOSE) ? FD_CLOFORK : 0) | - ((fde->fde_flags & UF_RESOLVE_BENEATH) ? - FD_RESOLVE_BENEATH : 0); + td->td_retval[0] = fde_to_fd_flags(fde->fde_flags); error = 0; } FILEDESC_SUNLOCK(fdp); @@ -552,11 +634,8 @@ kern_fcntl(struct thread *td, int fd, int cmd, intptr_t arg) /* * UF_RESOLVE_BENEATH is sticky and cannot be cleared. */ - fde->fde_flags = (fde->fde_flags & ~UF_EXCLOSE) | - ((arg & FD_CLOEXEC) != 0 ? UF_EXCLOSE : 0) | - ((arg & FD_CLOFORK) != 0 ? UF_FOCLOSE : 0) | - ((arg & FD_RESOLVE_BENEATH) != 0 ? - UF_RESOLVE_BENEATH : 0); + fde->fde_flags = (fde->fde_flags & + ~(UF_EXCLOSE | UF_FOCLOSE)) | fd_to_fde_flags(arg); error = 0; } FILEDESC_XUNLOCK(fdp); @@ -991,10 +1070,7 @@ kern_dup(struct thread *td, u_int mode, int flags, int old, int new) goto unlock; if (mode == FDDUP_FIXED && old == new) { td->td_retval[0] = new; - if ((flags & FDDUP_FLAG_CLOEXEC) != 0) - fdp->fd_ofiles[new].fde_flags |= UF_EXCLOSE; - if ((flags & FDDUP_FLAG_CLOFORK) != 0) - fdp->fd_ofiles[new].fde_flags |= UF_FOCLOSE; + fdp->fd_ofiles[new].fde_flags |= fddup_to_fde_flags(flags); error = 0; goto unlock; } @@ -1070,8 +1146,7 @@ kern_dup(struct thread *td, u_int mode, int flags, int old, int new) filecaps_copy_finish(&oldfde->fde_caps, &newfde->fde_caps, nioctls); newfde->fde_flags = (oldfde->fde_flags & ~(UF_EXCLOSE | UF_FOCLOSE)) | - ((flags & FDDUP_FLAG_CLOEXEC) != 0 ? UF_EXCLOSE : 0) | - ((flags & FDDUP_FLAG_CLOFORK) != 0 ? UF_FOCLOSE : 0); + fddup_to_fde_flags(flags); #ifdef CAPABILITIES seqc_write_end(&newfde->fde_seqc); #endif @@ -1444,8 +1519,7 @@ close_range_flags(struct thread *td, u_int lowfd, u_int highfd, int flags) struct filedescent *fde; int fd, fde_flags; - fde_flags = ((flags & CLOSE_RANGE_CLOEXEC) != 0 ? UF_EXCLOSE : 0) | - ((flags & CLOSE_RANGE_CLOFORK) != 0 ? UF_FOCLOSE : 0); + fde_flags = close_range_to_fde_flags(flags); fdp = td->td_proc->p_fd; FILEDESC_XLOCK(fdp); fdt = atomic_load_ptr(&fdp->fd_files); @@ -2194,9 +2268,7 @@ _finstall(struct filedesc *fdp, struct file *fp, int fd, int flags, seqc_write_begin(&fde->fde_seqc); #endif fde->fde_file = fp; - fde->fde_flags = ((flags & O_CLOEXEC) != 0 ? UF_EXCLOSE : 0) | - ((flags & O_CLOFORK) != 0 ? UF_FOCLOSE : 0) | - ((flags & O_RESOLVE_BENEATH) != 0 ? UF_RESOLVE_BENEATH : 0); + fde->fde_flags = open_to_fde_flags(flags, true); if (fcaps != NULL) filecaps_move(fcaps, &fde->fde_caps); else diff --git a/sys/kern/subr_asan.c b/sys/kern/subr_asan.c index 0edb631d1475..464efda1e91a 100644 --- a/sys/kern/subr_asan.c +++ b/sys/kern/subr_asan.c @@ -263,8 +263,7 @@ kasan_mark(const void *addr, size_t size, size_t redzsize, uint8_t code) if (__predict_false(!kasan_enabled)) return; - if ((vm_offset_t)addr >= DMAP_MIN_ADDRESS && - (vm_offset_t)addr < DMAP_MAX_ADDRESS) + if (kasan_md_unsupported((vm_offset_t)addr)) return; KASSERT((vm_offset_t)addr >= VM_MIN_KERNEL_ADDRESS && diff --git a/sys/kern/subr_pctrie.c b/sys/kern/subr_pctrie.c index 3a3548bad52b..bb86c779b936 100644 --- a/sys/kern/subr_pctrie.c +++ b/sys/kern/subr_pctrie.c @@ -691,21 +691,23 @@ _pctrie_lookup_ge(struct pctrie *ptree, struct pctrie_node *node, */ if (node == PCTRIE_NULL || *pctrie_toval(node) < index) { /* Climb the path to find a node with a descendant > index. */ - for (node = parent; node != NULL; node = pctrie_parent(node)) { - slot = pctrie_slot(node, index) + 1; - if ((node->pn_popmap >> slot) != 0) + node = NULL; + while (parent != NULL) { + slot = pctrie_slot(parent, index) + 1; + if ((parent->pn_popmap >> slot) != 0) break; + node = parent; + parent = pctrie_parent(node); } - if (node == NULL) { + if (parent == NULL) { if (parent_out != NULL) - *parent_out = NULL; + *parent_out = node; return (NULL); } /* Step to the least child with a descendant > index. */ - slot += ffs(node->pn_popmap >> slot) - 1; - parent = node; - node = pctrie_node_load(&node->pn_child[slot], NULL, + slot += ffs(parent->pn_popmap >> slot) - 1; + node = pctrie_node_load(&parent->pn_child[slot], NULL, PCTRIE_LOCKED); } /* Descend to the least leaf of the subtrie. */ @@ -785,21 +787,23 @@ _pctrie_lookup_le(struct pctrie *ptree, struct pctrie_node *node, */ if (node == PCTRIE_NULL || *pctrie_toval(node) > index) { /* Climb the path to find a node with a descendant < index. */ - for (node = parent; node != NULL; node = pctrie_parent(node)) { - slot = pctrie_slot(node, index); - if ((node->pn_popmap & ((1 << slot) - 1)) != 0) + node = NULL; + while (parent != NULL) { + slot = pctrie_slot(parent, index); + if ((parent->pn_popmap & ((1 << slot) - 1)) != 0) break; + node = parent; + parent = pctrie_parent(node); } - if (node == NULL) { + if (parent == NULL) { if (parent_out != NULL) - *parent_out = NULL; + *parent_out = node; return (NULL); } /* Step to the greatest child with a descendant < index. */ - slot = ilog2(node->pn_popmap & ((1 << slot) - 1)); - parent = node; - node = pctrie_node_load(&node->pn_child[slot], NULL, + slot = ilog2(parent->pn_popmap & ((1 << slot) - 1)); + node = pctrie_node_load(&parent->pn_child[slot], NULL, PCTRIE_LOCKED); } /* Descend to the greatest leaf of the subtrie. */ diff --git a/sys/kern/subr_trap.c b/sys/kern/subr_trap.c index 18388ae5f232..bac7d0080c71 100644 --- a/sys/kern/subr_trap.c +++ b/sys/kern/subr_trap.c @@ -338,8 +338,9 @@ ast_handler(struct thread *td, struct trapframe *framep, bool dtor) td->td_ast = 0; } - CTR3(KTR_SYSC, "ast: thread %p (pid %d, %s)", td, td->td_proc->p_pid, - td->td_proc->p_comm); + CTR3(KTR_SYSC, "ast: thread %p (pid %d, %s)", td, + td->td_proc == NULL ? -1 : td->td_proc->p_pid, + td->td_proc == NULL ? "" : td->td_proc->p_comm); KASSERT(framep == NULL || TRAPF_USERMODE(framep), ("ast in kernel mode")); diff --git a/sys/kern/sys_generic.c b/sys/kern/sys_generic.c index 94e44d888181..b472aaea89e6 100644 --- a/sys/kern/sys_generic.c +++ b/sys/kern/sys_generic.c @@ -2309,6 +2309,12 @@ sys_exterrctl(struct thread *td, struct exterrctl_args *uap) return (EINVAL); td->td_pflags2 &= ~TDP2_UEXTERR; return (0); + case EXTERRCTL_UD: + /* + * Important: this code must always return EINVAL and never any + * extended error, for testing purposes. + */ + /* FALLTHROUGH */ default: return (EINVAL); } diff --git a/sys/kern/vfs_cache.c b/sys/kern/vfs_cache.c index 3d455b3874cc..89c1d779f04c 100644 --- a/sys/kern/vfs_cache.c +++ b/sys/kern/vfs_cache.c @@ -332,7 +332,8 @@ SDT_PROBE_DEFINE2(vfs, namecache, evict_negative, done, "struct vnode *", "char *"); SDT_PROBE_DEFINE1(vfs, namecache, symlink, alloc__fail, "size_t"); -SDT_PROBE_DEFINE3(vfs, fplookup, lookup, done, "struct nameidata", "int", "bool"); +SDT_PROBE_DEFINE3(vfs, fplookup, lookup, done, "struct nameidata *", "int", + "enum cache_fpl_status"); SDT_PROBE_DECLARE(vfs, namei, lookup, entry); SDT_PROBE_DECLARE(vfs, namei, lookup, return); @@ -6420,15 +6421,11 @@ out: cache_fpl_smr_assert_not_entered(&fpl); cache_fpl_assert_status(&fpl); *status = fpl.status; - if (SDT_PROBES_ENABLED()) { - SDT_PROBE3(vfs, fplookup, lookup, done, ndp, fpl.line, fpl.status); - if (fpl.status == CACHE_FPL_STATUS_HANDLED) - SDT_PROBE4(vfs, namei, lookup, return, error, ndp->ni_vp, true, - ndp); - } - + SDT_PROBE3(vfs, fplookup, lookup, done, ndp, fpl.line, fpl.status); if (__predict_true(fpl.status == CACHE_FPL_STATUS_HANDLED)) { MPASS(error != CACHE_FPL_FAILED); + SDT_PROBE4(vfs, namei, lookup, return, error, ndp->ni_vp, true, + ndp); if (error != 0) { cache_fpl_cleanup_cnp(fpl.cnp); MPASS(fpl.dvp == NULL); diff --git a/sys/kern/vfs_inotify.c b/sys/kern/vfs_inotify.c index 41e73bb41a49..d3cd0d1f9832 100644 --- a/sys/kern/vfs_inotify.c +++ b/sys/kern/vfs_inotify.c @@ -371,7 +371,7 @@ inotify_unlink_watch_locked(struct inotify_softc *sc, struct inotify_watch *watc TAILQ_REMOVE(&vp->v_pollinfo->vpi_inotify, watch, vlink); if (TAILQ_EMPTY(&vp->v_pollinfo->vpi_inotify)) - vn_irflag_unset_locked(vp, VIRF_INOTIFY); + vn_irflag_unset(vp, VIRF_INOTIFY); } /* @@ -675,7 +675,8 @@ vn_inotify(struct vnode *vp, struct vnode *dvp, struct componentname *cnp, struct vattr va; int error; - error = VOP_GETATTR(vp, &va, cnp->cn_cred); + error = VOP_GETATTR(vp, &va, + cnp->cn_cred); if (error == 0 && va.va_nlink != 0) selfevent = 0; } @@ -760,9 +761,11 @@ vn_inotify_add_watch(struct vnode *vp, struct inotify_softc *sc, uint32_t mask, * directory if it's specified as a vnode. */ vrefact(vp); + VOP_UNLOCK(vp); NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, dp->d_name, vp); error = namei(&nd); + vn_lock(vp, LK_SHARED | LK_RETRY); if (error != 0) break; vn_irflag_set_cond(nd.ni_vp, VIRF_INOTIFY_PARENT); diff --git a/sys/kern/vfs_syscalls.c b/sys/kern/vfs_syscalls.c index d880733cbfe7..c71e0d9ee569 100644 --- a/sys/kern/vfs_syscalls.c +++ b/sys/kern/vfs_syscalls.c @@ -4314,10 +4314,6 @@ kern_getdirentries(struct thread *td, int fd, char *buf, size_t count, vp = fp->f_vnode; foffset = foffset_lock(fp, 0); unionread: - if (vp->v_type != VDIR) { - error = EINVAL; - goto fail; - } if (__predict_false((vp->v_vflag & VV_UNLINKED) != 0)) { error = ENOENT; goto fail; @@ -4330,6 +4326,19 @@ unionread: auio.uio_segflg = bufseg; auio.uio_td = td; vn_lock(vp, LK_SHARED | LK_RETRY); + /* + * We want to return ENOTDIR for anything that is not VDIR, but + * not for VBAD, and we can't check for VBAD while the vnode is + * unlocked. + */ + if (vp->v_type != VDIR) { + if (vp->v_type == VBAD) + error = EBADF; + else + error = ENOTDIR; + VOP_UNLOCK(vp); + goto fail; + } AUDIT_ARG_VNODE1(vp); loff = auio.uio_offset = foffset; #ifdef MAC |