git: 6c324fb645e6 - stable/13 - linux(4): Add a dedicated statat() implementation

From: Dmitry Chagin <dchagin_at_FreeBSD.org>
Date: Thu, 29 Jun 2023 08:20:26 UTC
The branch stable/13 has been updated by dchagin:

URL: https://cgit.FreeBSD.org/src/commit/?id=6c324fb645e64e0090e16a5846eafdeb723f4105

commit 6c324fb645e64e0090e16a5846eafdeb723f4105
Author:     Dmitry Chagin <dchagin@FreeBSD.org>
AuthorDate: 2023-04-28 08:55:04 +0000
Commit:     Dmitry Chagin <dchagin@FreeBSD.org>
CommitDate: 2023-06-29 08:15:45 +0000

    linux(4): Add a dedicated statat() implementation
    
    Get rid of calling Linux stat translation hook and specific to Linux
    handling of non-vnode dirfd from kern_statat(),
    
    Reviewed by:            kib, mjg
    Differential revision:  https://reviews.freebsd.org/D35474
    
    (cherry picked from commit cb858340dcbf214cc4c4d78dbb741620d7b3a252)
---
 sys/compat/cloudabi/cloudabi_file.c   |  2 +-
 sys/compat/freebsd32/freebsd32_misc.c | 13 ++++++-------
 sys/compat/linux/linux_file.c         |  8 ++++----
 sys/compat/linux/linux_stats.c        | 34 ++++++++++++++++++++++++++++++++--
 sys/kern/vfs_mountroot.c              |  2 +-
 sys/kern/vfs_syscalls.c               | 34 +++++++++++-----------------------
 sys/sys/syscallsubr.h                 |  4 ++--
 7 files changed, 57 insertions(+), 40 deletions(-)

diff --git a/sys/compat/cloudabi/cloudabi_file.c b/sys/compat/cloudabi/cloudabi_file.c
index 83987b79aa0c..1a1def879977 100644
--- a/sys/compat/cloudabi/cloudabi_file.c
+++ b/sys/compat/cloudabi/cloudabi_file.c
@@ -659,7 +659,7 @@ cloudabi_sys_file_stat_get(struct thread *td,
 
 	error = kern_statat(td,
 	    (uap->fd.flags & CLOUDABI_LOOKUP_SYMLINK_FOLLOW) != 0 ? 0 :
-	    AT_SYMLINK_NOFOLLOW, uap->fd.fd, path, UIO_SYSSPACE, &sb, NULL);
+	    AT_SYMLINK_NOFOLLOW, uap->fd.fd, path, UIO_SYSSPACE, &sb);
 	cloudabi_freestr(path);
 	if (error != 0)
 		return (error);
diff --git a/sys/compat/freebsd32/freebsd32_misc.c b/sys/compat/freebsd32/freebsd32_misc.c
index 70eece0311f4..9fd24a91080f 100644
--- a/sys/compat/freebsd32/freebsd32_misc.c
+++ b/sys/compat/freebsd32/freebsd32_misc.c
@@ -2251,8 +2251,7 @@ ofreebsd32_stat(struct thread *td, struct ofreebsd32_stat_args *uap)
 	struct ostat32 sb32;
 	int error;
 
-	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE,
-	    &sb, NULL);
+	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb);
 	if (error)
 		return (error);
 	copy_ostat(&sb, &sb32);
@@ -2301,7 +2300,7 @@ freebsd32_fstatat(struct thread *td, struct freebsd32_fstatat_args *uap)
 	int error;
 
 	error = kern_statat(td, uap->flag, uap->fd, uap->path, UIO_USERSPACE,
-	    &ub, NULL);
+	    &ub);
 	if (error)
 		return (error);
 	copy_stat(&ub, &ub32);
@@ -2318,7 +2317,7 @@ ofreebsd32_lstat(struct thread *td, struct ofreebsd32_lstat_args *uap)
 	int error;
 
 	error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error)
 		return (error);
 	copy_ostat(&sb, &sb32);
@@ -2437,7 +2436,7 @@ freebsd11_freebsd32_stat(struct thread *td,
 	int error;
 
 	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE,
-	    &sb, NULL);
+	    &sb);
 	if (error != 0)
 		return (error);
 	error = freebsd11_cvtstat32(&sb, &sb32);
@@ -2472,7 +2471,7 @@ freebsd11_freebsd32_fstatat(struct thread *td,
 	int error;
 
 	error = kern_statat(td, uap->flag, uap->fd, uap->path, UIO_USERSPACE,
-	    &sb, NULL);
+	    &sb);
 	if (error != 0)
 		return (error);
 	error = freebsd11_cvtstat32(&sb, &sb32);
@@ -2490,7 +2489,7 @@ freebsd11_freebsd32_lstat(struct thread *td,
 	int error;
 
 	error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	error = freebsd11_cvtstat32(&sb, &sb32);
diff --git a/sys/compat/linux/linux_file.c b/sys/compat/linux/linux_file.c
index db4b1151bc60..249318a2e619 100644
--- a/sys/compat/linux/linux_file.c
+++ b/sys/compat/linux/linux_file.c
@@ -732,7 +732,7 @@ linux_unlink(struct thread *td, struct linux_unlink_args *args)
 		if (error == EPERM) {
 			/* Introduce POSIX noncompliant behaviour of Linux */
 			if (kern_statat(td, 0, AT_FDCWD, args->path,
-			    UIO_USERSPACE, &st, NULL) == 0) {
+			    UIO_USERSPACE, &st) == 0) {
 				if (S_ISDIR(st.st_mode))
 					error = EISDIR;
 			}
@@ -742,8 +742,8 @@ linux_unlink(struct thread *td, struct linux_unlink_args *args)
 		error = kern_funlinkat(td, AT_FDCWD, path, FD_NONE, UIO_SYSSPACE, 0, 0);
 		if (error == EPERM) {
 			/* Introduce POSIX noncompliant behaviour of Linux */
-			if (kern_statat(td, 0, AT_FDCWD, path, UIO_SYSSPACE, &st,
-			    NULL) == 0) {
+			if (kern_statat(td, 0, AT_FDCWD, path, UIO_SYSSPACE,
+			    &st) == 0) {
 				if (S_ISDIR(st.st_mode))
 					error = EISDIR;
 			}
@@ -769,7 +769,7 @@ linux_unlinkat_impl(struct thread *td, enum uio_seg pathseg, const char *path,
 	if (error == EPERM && !(args->flag & LINUX_AT_REMOVEDIR)) {
 		/* Introduce POSIX noncompliant behaviour of Linux */
 		if (kern_statat(td, AT_SYMLINK_NOFOLLOW, dfd, path,
-		    pathseg, &st, NULL) == 0 && S_ISDIR(st.st_mode))
+		    pathseg, &st) == 0 && S_ISDIR(st.st_mode))
 			error = EISDIR;
 	}
 	return (error);
diff --git a/sys/compat/linux/linux_stats.c b/sys/compat/linux/linux_stats.c
index 58c378d6bc08..934fc6e5e740 100644
--- a/sys/compat/linux/linux_stats.c
+++ b/sys/compat/linux/linux_stats.c
@@ -29,17 +29,23 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
+#include "opt_ktrace.h"
+
 #include <sys/param.h>
 #include <sys/capsicum.h>
 #include <sys/dirent.h>
 #include <sys/lock.h>
 #include <sys/malloc.h>
 #include <sys/mutex.h>
+#include <sys/namei.h>
 #include <sys/proc.h>
 #include <sys/stat.h>
 #include <sys/syscallsubr.h>
 #include <sys/tty.h>
 #include <sys/vnode.h>
+#ifdef KTRACE
+#include <sys/ktrace.h>
+#endif
 
 #ifdef COMPAT_LINUX32
 #include <machine/../linux32/linux.h>
@@ -57,9 +63,33 @@ static int
 linux_kern_statat(struct thread *td, int flag, int fd, const char *path,
     enum uio_seg pathseg, struct stat *sbp)
 {
+	struct nameidata nd;
+	int error;
+
+	if ((flag & ~(AT_SYMLINK_NOFOLLOW | AT_RESOLVE_BENEATH |
+	    AT_EMPTY_PATH)) != 0)
+		return (EINVAL);
+
+	NDINIT_ATRIGHTS(&nd, LOOKUP, at2cnpflags(flag, AT_RESOLVE_BENEATH |
+	    AT_SYMLINK_NOFOLLOW | AT_EMPTY_PATH) | LOCKSHARED | LOCKLEAF |
+	    AUDITVNODE1, pathseg, path, fd, &cap_fstat_rights, td);
 
-	return (kern_statat(td, flag, fd, path, pathseg, sbp,
-	    translate_vnhook_major_minor));
+	if ((error = namei(&nd)) != 0) {
+		if (error == ENOTDIR &&
+		    (nd.ni_resflags & NIRES_EMPTYPATH) != 0)
+			error = kern_fstat(td, fd, sbp);
+		return (error);
+	}
+	error = VOP_STAT(nd.ni_vp, sbp, td->td_ucred, NOCRED, td);
+	if (error == 0)
+		translate_vnhook_major_minor(nd.ni_vp, sbp);
+	NDFREE_PNBUF(&nd);
+	vput(nd.ni_vp);
+#ifdef KTRACE
+	if (KTRPOINT(td, KTR_STRUCT))
+		ktrstat_error(sbp, error);
+#endif
+	return (error);
 }
 
 #ifdef LINUX_LEGACY_SYSCALLS
diff --git a/sys/kern/vfs_mountroot.c b/sys/kern/vfs_mountroot.c
index 1886d3b7aca3..8a4b67836206 100644
--- a/sys/kern/vfs_mountroot.c
+++ b/sys/kern/vfs_mountroot.c
@@ -595,7 +595,7 @@ parse_dir_md(char **conf)
 	free(tok, M_TEMP);
 
 	/* Get file status. */
-	error = kern_statat(td, 0, AT_FDCWD, path, UIO_SYSSPACE, &sb, NULL);
+	error = kern_statat(td, 0, AT_FDCWD, path, UIO_SYSSPACE, &sb);
 	if (error)
 		goto out;
 
diff --git a/sys/kern/vfs_syscalls.c b/sys/kern/vfs_syscalls.c
index 3b3947b2ccb7..89f45e898d8e 100644
--- a/sys/kern/vfs_syscalls.c
+++ b/sys/kern/vfs_syscalls.c
@@ -109,7 +109,7 @@ static int kern_readlink_vp(struct vnode *vp, char *buf, enum uio_seg bufseg,
 static int kern_linkat_vp(struct thread *td, struct vnode *vp, int fd,
     const char *path, enum uio_seg segflag);
 
-static uint64_t
+uint64_t
 at2cnpflags(u_int at_flags, u_int mask)
 {
 	u_int64_t res;
@@ -2187,8 +2187,7 @@ ostat(struct thread *td, struct ostat_args *uap)
 	struct ostat osb;
 	int error;
 
-	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE,
-	    &sb, NULL);
+	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	cvtstat(&sb, &osb);
@@ -2212,7 +2211,7 @@ olstat(struct thread *td, struct olstat_args *uap)
 	int error;
 
 	error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	cvtstat(&sb, &osb);
@@ -2330,8 +2329,7 @@ freebsd11_stat(struct thread *td, struct freebsd11_stat_args* uap)
 	struct freebsd11_stat osb;
 	int error;
 
-	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE,
-	    &sb, NULL);
+	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	error = freebsd11_cvtstat(&sb, &osb);
@@ -2348,7 +2346,7 @@ freebsd11_lstat(struct thread *td, struct freebsd11_lstat_args* uap)
 	int error;
 
 	error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	error = freebsd11_cvtstat(&sb, &osb);
@@ -2385,7 +2383,7 @@ freebsd11_fstatat(struct thread *td, struct freebsd11_fstatat_args* uap)
 	int error;
 
 	error = kern_statat(td, uap->flag, uap->fd, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	error = freebsd11_cvtstat(&sb, &osb);
@@ -2413,7 +2411,7 @@ sys_fstatat(struct thread *td, struct fstatat_args *uap)
 	int error;
 
 	error = kern_statat(td, uap->flag, uap->fd, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error == 0)
 		error = copyout(&sb, uap->buf, sizeof (sb));
 	return (error);
@@ -2421,8 +2419,7 @@ sys_fstatat(struct thread *td, struct fstatat_args *uap)
 
 int
 kern_statat(struct thread *td, int flag, int fd, const char *path,
-    enum uio_seg pathseg, struct stat *sbp,
-    void (*hook)(struct vnode *vp, struct stat *sbp))
+    enum uio_seg pathseg, struct stat *sbp)
 {
 	struct nameidata nd;
 	int error;
@@ -2435,17 +2432,9 @@ kern_statat(struct thread *td, int flag, int fd, const char *path,
 	    AT_SYMLINK_NOFOLLOW | AT_EMPTY_PATH) | LOCKSHARED | LOCKLEAF |
 	    AUDITVNODE1, pathseg, path, fd, &cap_fstat_rights, td);
 
-	if ((error = namei(&nd)) != 0) {
-		if (error == ENOTDIR &&
-		    (nd.ni_resflags & NIRES_EMPTYPATH) != 0)
-			error = kern_fstat(td, fd, sbp);
+	if ((error = namei(&nd)) != 0)
 		return (error);
-	}
 	error = VOP_STAT(nd.ni_vp, sbp, td->td_ucred, NOCRED, td);
-	if (error == 0) {
-		if (__predict_false(hook != NULL))
-			hook(nd.ni_vp, sbp);
-	}
 	NDFREE_NOTHING(&nd);
 	vput(nd.ni_vp);
 #ifdef __STAT_TIME_T_EXT
@@ -2501,8 +2490,7 @@ freebsd11_nstat(struct thread *td, struct freebsd11_nstat_args *uap)
 	struct nstat nsb;
 	int error;
 
-	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE,
-	    &sb, NULL);
+	error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	freebsd11_cvtnstat(&sb, &nsb);
@@ -2526,7 +2514,7 @@ freebsd11_nlstat(struct thread *td, struct freebsd11_nlstat_args *uap)
 	int error;
 
 	error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path,
-	    UIO_USERSPACE, &sb, NULL);
+	    UIO_USERSPACE, &sb);
 	if (error != 0)
 		return (error);
 	freebsd11_cvtnstat(&sb, &nsb);
diff --git a/sys/sys/syscallsubr.h b/sys/sys/syscallsubr.h
index ae17a10d01da..fb7bbe58a63a 100644
--- a/sys/sys/syscallsubr.h
+++ b/sys/sys/syscallsubr.h
@@ -80,6 +80,7 @@ struct mmap_req {
 	mmap_check_fp_fn	mr_check_fp_fn;
 };
 
+uint64_t at2cnpflags(u_int at_flags, u_int mask);
 int	kern___getcwd(struct thread *td, char *buf, enum uio_seg bufseg,
 	    size_t buflen, size_t path_max);
 int	kern_accept(struct thread *td, int s, struct sockaddr **name,
@@ -312,8 +313,7 @@ int	kern_sigqueue(struct thread *td, pid_t pid, int signum,
 	    union sigval *value);
 int	kern_socket(struct thread *td, int domain, int type, int protocol);
 int	kern_statat(struct thread *td, int flag, int fd, const char *path,
-	    enum uio_seg pathseg, struct stat *sbp,
-	    void (*hook)(struct vnode *vp, struct stat *sbp));
+	    enum uio_seg pathseg, struct stat *sbp);
 int	kern_specialfd(struct thread *td, int type, void *arg);
 int	kern_statfs(struct thread *td, const char *path, enum uio_seg pathseg,
 	    struct statfs *buf);