diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2019-12-15 11:36:12 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2019-12-15 11:36:12 -0800 |
commit | 2e6d304515ba9936d85265ad93dddc4c13c17d06 (patch) | |
tree | 0aeaf6e255deb791f02e919e670f563e5505c84f | |
parent | 510c9788991c58827373bca719d8cffa4d65f846 (diff) | |
parent | 8243186f0cc7c57cf9d6a110cd7315c44e3e0be8 (diff) |
Merge branch 'remove-ksys-mount-dup' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/linux
Pull ksys_mount() and ksys_dup() removal from Dominik Brodowski:
"This small series replaces all in-kernel calls to the
userspace-focused ksys_mount() and ksys_dup() with calls to
kernel-centric functions:
For each replacement of ksys_mount() with do_mount(), one needs to
verify that the first and third parameter (char *dev_name, char *type)
are strings allocated in kernelspace and that the fifth parameter
(void *data) is either NULL or refers to a full page (only occurence
in init/do_mounts.c::do_mount_root()). The second and fourth
parameters (char *dir_name, unsigned long flags) are passed by
ksys_mount() to do_mount() unchanged, and therefore do not require
particular care.
Moreover, instead of pretending to be userspace, the opening of
/dev/console as stdin/stdout/stderr can be implemented using in-kernel
functions as well. Thereby, ksys_dup() can be removed for good"
[ This doesn't get rid of the special "kernel init runs with KERNEL_DS"
case, but it at least removes _some_ of the users of "treat kernel
pointers as user pointers for our magical init sequence".
One day we'll hopefully be rid of it all, and can initialize our
init_thread addr_limit to USER_DS. - Linus ]
* 'remove-ksys-mount-dup' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/linux:
fs: remove ksys_dup()
init: unify opening /dev/console as stdin/stdout/stderr
init: use do_mount() instead of ksys_mount()
initrd: use do_mount() instead of ksys_mount()
devtmpfs: use do_mount() instead of ksys_mount()
-rw-r--r-- | drivers/base/devtmpfs.c | 6 | ||||
-rw-r--r-- | fs/file.c | 7 | ||||
-rw-r--r-- | fs/namespace.c | 10 | ||||
-rw-r--r-- | include/linux/device.h | 4 | ||||
-rw-r--r-- | include/linux/initrd.h | 2 | ||||
-rw-r--r-- | include/linux/syscalls.h | 3 | ||||
-rw-r--r-- | init/do_mounts.c | 30 | ||||
-rw-r--r-- | init/do_mounts_initrd.c | 11 | ||||
-rw-r--r-- | init/main.c | 31 |
9 files changed, 63 insertions, 41 deletions
diff --git a/drivers/base/devtmpfs.c b/drivers/base/devtmpfs.c index 30d0523014e0..6cdbf1531238 100644 --- a/drivers/base/devtmpfs.c +++ b/drivers/base/devtmpfs.c @@ -359,7 +359,7 @@ static int handle_remove(const char *nodename, struct device *dev) * If configured, or requested by the commandline, devtmpfs will be * auto-mounted after the kernel mounted the root filesystem. */ -int devtmpfs_mount(const char *mntdir) +int devtmpfs_mount(void) { int err; @@ -369,7 +369,7 @@ int devtmpfs_mount(const char *mntdir) if (!thread) return 0; - err = ksys_mount("devtmpfs", mntdir, "devtmpfs", MS_SILENT, NULL); + err = do_mount("devtmpfs", "dev", "devtmpfs", MS_SILENT, NULL); if (err) printk(KERN_INFO "devtmpfs: error mounting %i\n", err); else @@ -394,7 +394,7 @@ static int devtmpfsd(void *p) *err = ksys_unshare(CLONE_NEWNS); if (*err) goto out; - *err = ksys_mount("devtmpfs", "/", "devtmpfs", MS_SILENT, NULL); + *err = do_mount("devtmpfs", "/", "devtmpfs", MS_SILENT, NULL); if (*err) goto out; ksys_chdir("/.."); /* will traverse into overmounted root */ diff --git a/fs/file.c b/fs/file.c index 3da91a112bab..2f4fcf985079 100644 --- a/fs/file.c +++ b/fs/file.c @@ -960,7 +960,7 @@ SYSCALL_DEFINE2(dup2, unsigned int, oldfd, unsigned int, newfd) return ksys_dup3(oldfd, newfd, 0); } -int ksys_dup(unsigned int fildes) +SYSCALL_DEFINE1(dup, unsigned int, fildes) { int ret = -EBADF; struct file *file = fget_raw(fildes); @@ -975,11 +975,6 @@ int ksys_dup(unsigned int fildes) return ret; } -SYSCALL_DEFINE1(dup, unsigned int, fildes) -{ - return ksys_dup(fildes); -} - int f_dupfd(unsigned int from, struct file *file, unsigned flags) { int err; diff --git a/fs/namespace.c b/fs/namespace.c index 2fd0c8bcb8c1..be601d3a8008 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -3325,8 +3325,8 @@ struct dentry *mount_subtree(struct vfsmount *m, const char *name) } EXPORT_SYMBOL(mount_subtree); -int ksys_mount(const char __user *dev_name, const char __user *dir_name, - const char __user *type, unsigned long flags, void __user *data) +SYSCALL_DEFINE5(mount, char __user *, dev_name, char __user *, dir_name, + char __user *, type, unsigned long, flags, void __user *, data) { int ret; char *kernel_type; @@ -3359,12 +3359,6 @@ out_type: return ret; } -SYSCALL_DEFINE5(mount, char __user *, dev_name, char __user *, dir_name, - char __user *, type, unsigned long, flags, void __user *, data) -{ - return ksys_mount(dev_name, dir_name, type, flags, data); -} - /* * Create a kernel mount representation for a new, prepared superblock * (specified by fs_fd) and attach to an open_tree-like file descriptor. diff --git a/include/linux/device.h b/include/linux/device.h index e226030c1df3..96ff76731e93 100644 --- a/include/linux/device.h +++ b/include/linux/device.h @@ -1666,11 +1666,11 @@ extern bool kill_device(struct device *dev); #ifdef CONFIG_DEVTMPFS extern int devtmpfs_create_node(struct device *dev); extern int devtmpfs_delete_node(struct device *dev); -extern int devtmpfs_mount(const char *mntdir); +extern int devtmpfs_mount(void); #else static inline int devtmpfs_create_node(struct device *dev) { return 0; } static inline int devtmpfs_delete_node(struct device *dev) { return 0; } -static inline int devtmpfs_mount(const char *mountpoint) { return 0; } +static inline int devtmpfs_mount(void) { return 0; } #endif /* drivers/base/power/shutdown.c */ diff --git a/include/linux/initrd.h b/include/linux/initrd.h index d77fe34fb00a..aa5914355728 100644 --- a/include/linux/initrd.h +++ b/include/linux/initrd.h @@ -28,3 +28,5 @@ extern unsigned int real_root_dev; extern char __initramfs_start[]; extern unsigned long __initramfs_size; + +void console_on_rootfs(void); diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index d0391cc2dae9..2960dedcfde8 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -1231,10 +1231,7 @@ asmlinkage long sys_ni_syscall(void); * the ksys_xyzyyz() functions prototyped below. */ -int ksys_mount(const char __user *dev_name, const char __user *dir_name, - const char __user *type, unsigned long flags, void __user *data); int ksys_umount(char __user *name, int flags); -int ksys_dup(unsigned int fildes); int ksys_chroot(const char __user *filename); ssize_t ksys_write(unsigned int fd, const char __user *buf, size_t count); int ksys_chdir(const char __user *filename); diff --git a/init/do_mounts.c b/init/do_mounts.c index af9cda887a23..f55cbd9cb818 100644 --- a/init/do_mounts.c +++ b/init/do_mounts.c @@ -387,12 +387,25 @@ static void __init get_fs_names(char *page) *s = '\0'; } -static int __init do_mount_root(char *name, char *fs, int flags, void *data) +static int __init do_mount_root(const char *name, const char *fs, + const int flags, const void *data) { struct super_block *s; - int err = ksys_mount(name, "/root", fs, flags, data); - if (err) - return err; + char *data_page; + struct page *p; + int ret; + + /* do_mount() requires a full page as fifth argument */ + p = alloc_page(GFP_KERNEL); + if (!p) + return -ENOMEM; + + data_page = page_address(p); + strncpy(data_page, data, PAGE_SIZE - 1); + + ret = do_mount(name, "/root", fs, flags, data_page); + if (ret) + goto out; ksys_chdir("/root"); s = current->fs->pwd.dentry->d_sb; @@ -402,7 +415,10 @@ static int __init do_mount_root(char *name, char *fs, int flags, void *data) s->s_type->name, sb_rdonly(s) ? " readonly" : "", MAJOR(ROOT_DEV), MINOR(ROOT_DEV)); - return 0; + +out: + put_page(p); + return ret; } void __init mount_block_root(char *name, int flags) @@ -670,8 +686,8 @@ void __init prepare_namespace(void) mount_root(); out: - devtmpfs_mount("dev"); - ksys_mount(".", "/", NULL, MS_MOVE, NULL); + devtmpfs_mount(); + do_mount(".", "/", NULL, MS_MOVE, NULL); ksys_chroot("."); } diff --git a/init/do_mounts_initrd.c b/init/do_mounts_initrd.c index a9c6cc56f505..dab8b1151b56 100644 --- a/init/do_mounts_initrd.c +++ b/init/do_mounts_initrd.c @@ -48,13 +48,10 @@ early_param("initrd", early_initrd); static int init_linuxrc(struct subprocess_info *info, struct cred *new) { ksys_unshare(CLONE_FS | CLONE_FILES); - /* stdin/stdout/stderr for /linuxrc */ - ksys_open("/dev/console", O_RDWR, 0); - ksys_dup(0); - ksys_dup(0); + console_on_rootfs(); /* move initrd over / and chdir/chroot in initrd root */ ksys_chdir("/root"); - ksys_mount(".", "/", NULL, MS_MOVE, NULL); + do_mount(".", "/", NULL, MS_MOVE, NULL); ksys_chroot("."); ksys_setsid(); return 0; @@ -89,7 +86,7 @@ static void __init handle_initrd(void) current->flags &= ~PF_FREEZER_SKIP; /* move initrd to rootfs' /old */ - ksys_mount("..", ".", NULL, MS_MOVE, NULL); + do_mount("..", ".", NULL, MS_MOVE, NULL); /* switch root and cwd back to / of rootfs */ ksys_chroot(".."); @@ -103,7 +100,7 @@ static void __init handle_initrd(void) mount_root(); printk(KERN_NOTICE "Trying to move old root to /initrd ... "); - error = ksys_mount("/old", "/root/initrd", NULL, MS_MOVE, NULL); + error = do_mount("/old", "/root/initrd", NULL, MS_MOVE, NULL); if (!error) printk("okay\n"); else { diff --git a/init/main.c b/init/main.c index 91f6ebb30ef0..ec3a1463ac69 100644 --- a/init/main.c +++ b/init/main.c @@ -93,6 +93,7 @@ #include <linux/rodata_test.h> #include <linux/jump_label.h> #include <linux/mem_encrypt.h> +#include <linux/file.h> #include <asm/io.h> #include <asm/bugs.h> @@ -1155,6 +1156,30 @@ static int __ref kernel_init(void *unused) "See Linux Documentation/admin-guide/init.rst for guidance."); } +void console_on_rootfs(void) +{ + struct file *file; + unsigned int i; + + /* Open /dev/console in kernelspace, this should never fail */ + file = filp_open("/dev/console", O_RDWR, 0); + if (!file) + goto err_out; + + /* create stdin/stdout/stderr, this should never fail */ + for (i = 0; i < 3; i++) { + if (f_dupfd(i, file, 0) != i) + goto err_out; + } + + return; + +err_out: + /* no panic -- this might not be fatal */ + pr_err("Warning: unable to open an initial console.\n"); + return; +} + static noinline void __init kernel_init_freeable(void) { /* @@ -1190,12 +1215,8 @@ static noinline void __init kernel_init_freeable(void) do_basic_setup(); - /* Open the /dev/console on the rootfs, this should never fail */ - if (ksys_open((const char __user *) "/dev/console", O_RDWR, 0) < 0) - pr_err("Warning: unable to open an initial console.\n"); + console_on_rootfs(); - (void) ksys_dup(0); - (void) ksys_dup(0); /* * check if there is an early userspace init. If yes, let it do all * the work |