Revision bbcd53c960713507ae764bf81970651b5577b95a authored by David Hildenbrand on 07 May 2021, 01:05:55 UTC, committed by Linus Torvalds on 07 May 2021, 07:26:34 UTC
Patch series "drivers/char: remove /dev/kmem for good".

Exploring /dev/kmem and /dev/mem in the context of memory hot(un)plug and
memory ballooning, I started questioning the existence of /dev/kmem.

Comparing it with the /proc/kcore implementation, it does not seem to be
able to deal with things like

a) Pages unmapped from the direct mapping (e.g., to be used by secretmem)
  -> kern_addr_valid(). virt_addr_valid() is not sufficient.

b) Special cases like gart aperture memory that is not to be touched
  -> mem_pfn_is_ram()

Unless I am missing something, it's at least broken in some cases and might
fault/crash the machine.

Looks like its existence has been questioned before in 2005 and 2010 [1],
after ~11 additional years, it might make sense to revive the discussion.

CONFIG_DEVKMEM is only enabled in a single defconfig (on purpose or by
mistake?).  All distributions disable it: in Ubuntu it has been disabled
for more than 10 years, in Debian since 2.6.31, in Fedora at least
starting with FC3, in RHEL starting with RHEL4, in SUSE starting from
15sp2, and OpenSUSE has it disabled as well.

1) /dev/kmem was popular for rootkits [2] before it got disabled
   basically everywhere. Ubuntu documents [3] "There is no modern user of
   /dev/kmem any more beyond attackers using it to load kernel rootkits.".
   RHEL documents in a BZ [5] "it served no practical purpose other than to
   serve as a potential security problem or to enable binary module drivers
   to access structures/functions they shouldn't be touching"

2) /proc/kcore is a decent interface to have a controlled way to read
   kernel memory for debugging puposes. (will need some extensions to
   deal with memory offlining/unplug, memory ballooning, and poisoned
   pages, though)

3) It might be useful for corner case debugging [1]. KDB/KGDB might be a
   better fit, especially, to write random memory; harder to shoot
   yourself into the foot.

4) "Kernel Memory Editor" [4] hasn't seen any updates since 2000 and seems
   to be incompatible with 64bit [1]. For educational purposes,
   /proc/kcore might be used to monitor value updates -- or older
   kernels can be used.

5) It's broken on arm64, and therefore, completely disabled there.

Looks like it's essentially unused and has been replaced by better
suited interfaces for individual tasks (/proc/kcore, KDB/KGDB). Let's
just remove it.

[1] https://lwn.net/Articles/147901/
[2] https://www.linuxjournal.com/article/10505
[3] https://wiki.ubuntu.com/Security/Features#A.2Fdev.2Fkmem_disabled
[4] https://sourceforge.net/projects/kme/
[5] https://bugzilla.redhat.com/show_bug.cgi?id=154796

Link: https://lkml.kernel.org/r/20210324102351.6932-1-david@redhat.com
Link: https://lkml.kernel.org/r/20210324102351.6932-2-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Alexander A. Klimov" <grandmaster@al2klimov.de>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Alexandre Belloni <alexandre.belloni@bootlin.com>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Andrey Zhizhikin <andrey.zhizhikin@leica-geosystems.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Brian Cain <bcain@codeaurora.org>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Chris Zankel <chris@zankel.net>
Cc: Corentin Labbe <clabbe@baylibre.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Cc: Greentime Hu <green.hu@gmail.com>
Cc: Gregory Clement <gregory.clement@bootlin.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Helge Deller <deller@gmx.de>
Cc: Hillf Danton <hdanton@sina.com>
Cc: huang ying <huang.ying.caritas@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: James Troup <james.troup@canonical.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kairui Song <kasong@redhat.com>
Cc: Krzysztof Kozlowski <krzk@kernel.org>
Cc: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Cc: Liviu Dudau <liviu.dudau@arm.com>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Niklas Schnelle <schnelle@linux.ibm.com>
Cc: Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>
Cc: openrisc@lists.librecores.org
Cc: Palmer Dabbelt <palmerdabbelt@google.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: "Pavel Machek (CIP)" <pavel@denx.de>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Cc: Pierre Morel <pmorel@linux.ibm.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Robert Richter <rric@kernel.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Cc: sparclinux@vger.kernel.org
Cc: Stafford Horne <shorne@gmail.com>
Cc: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Cc: Theodore Dubois <tblodt@icloud.com>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: William Cohen <wcohen@redhat.com>
Cc: Xiaoming Ni <nixiaoming@huawei.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent cb152a1
Raw File
dir.c
// SPDX-License-Identifier: GPL-2.0
/*
 *	fs/bfs/dir.c
 *	BFS directory operations.
 *	Copyright (C) 1999-2018  Tigran Aivazian <aivazian.tigran@gmail.com>
 *  Made endianness-clean by Andrew Stribblehill <ads@wompom.org> 2005
 */

#include <linux/time.h>
#include <linux/string.h>
#include <linux/fs.h>
#include <linux/buffer_head.h>
#include <linux/sched.h>
#include "bfs.h"

#undef DEBUG

#ifdef DEBUG
#define dprintf(x...)	printf(x)
#else
#define dprintf(x...)
#endif

static int bfs_add_entry(struct inode *dir, const struct qstr *child, int ino);
static struct buffer_head *bfs_find_entry(struct inode *dir,
				const struct qstr *child,
				struct bfs_dirent **res_dir);

static int bfs_readdir(struct file *f, struct dir_context *ctx)
{
	struct inode *dir = file_inode(f);
	struct buffer_head *bh;
	struct bfs_dirent *de;
	unsigned int offset;
	int block;

	if (ctx->pos & (BFS_DIRENT_SIZE - 1)) {
		printf("Bad f_pos=%08lx for %s:%08lx\n",
					(unsigned long)ctx->pos,
					dir->i_sb->s_id, dir->i_ino);
		return -EINVAL;
	}

	while (ctx->pos < dir->i_size) {
		offset = ctx->pos & (BFS_BSIZE - 1);
		block = BFS_I(dir)->i_sblock + (ctx->pos >> BFS_BSIZE_BITS);
		bh = sb_bread(dir->i_sb, block);
		if (!bh) {
			ctx->pos += BFS_BSIZE - offset;
			continue;
		}
		do {
			de = (struct bfs_dirent *)(bh->b_data + offset);
			if (de->ino) {
				int size = strnlen(de->name, BFS_NAMELEN);
				if (!dir_emit(ctx, de->name, size,
						le16_to_cpu(de->ino),
						DT_UNKNOWN)) {
					brelse(bh);
					return 0;
				}
			}
			offset += BFS_DIRENT_SIZE;
			ctx->pos += BFS_DIRENT_SIZE;
		} while ((offset < BFS_BSIZE) && (ctx->pos < dir->i_size));
		brelse(bh);
	}
	return 0;
}

const struct file_operations bfs_dir_operations = {
	.read		= generic_read_dir,
	.iterate_shared	= bfs_readdir,
	.fsync		= generic_file_fsync,
	.llseek		= generic_file_llseek,
};

static int bfs_create(struct user_namespace *mnt_userns, struct inode *dir,
		      struct dentry *dentry, umode_t mode, bool excl)
{
	int err;
	struct inode *inode;
	struct super_block *s = dir->i_sb;
	struct bfs_sb_info *info = BFS_SB(s);
	unsigned long ino;

	inode = new_inode(s);
	if (!inode)
		return -ENOMEM;
	mutex_lock(&info->bfs_lock);
	ino = find_first_zero_bit(info->si_imap, info->si_lasti + 1);
	if (ino > info->si_lasti) {
		mutex_unlock(&info->bfs_lock);
		iput(inode);
		return -ENOSPC;
	}
	set_bit(ino, info->si_imap);
	info->si_freei--;
	inode_init_owner(&init_user_ns, inode, dir, mode);
	inode->i_mtime = inode->i_atime = inode->i_ctime = current_time(inode);
	inode->i_blocks = 0;
	inode->i_op = &bfs_file_inops;
	inode->i_fop = &bfs_file_operations;
	inode->i_mapping->a_ops = &bfs_aops;
	inode->i_ino = ino;
	BFS_I(inode)->i_dsk_ino = ino;
	BFS_I(inode)->i_sblock = 0;
	BFS_I(inode)->i_eblock = 0;
	insert_inode_hash(inode);
        mark_inode_dirty(inode);
	bfs_dump_imap("create", s);

	err = bfs_add_entry(dir, &dentry->d_name, inode->i_ino);
	if (err) {
		inode_dec_link_count(inode);
		mutex_unlock(&info->bfs_lock);
		iput(inode);
		return err;
	}
	mutex_unlock(&info->bfs_lock);
	d_instantiate(dentry, inode);
	return 0;
}

static struct dentry *bfs_lookup(struct inode *dir, struct dentry *dentry,
						unsigned int flags)
{
	struct inode *inode = NULL;
	struct buffer_head *bh;
	struct bfs_dirent *de;
	struct bfs_sb_info *info = BFS_SB(dir->i_sb);

	if (dentry->d_name.len > BFS_NAMELEN)
		return ERR_PTR(-ENAMETOOLONG);

	mutex_lock(&info->bfs_lock);
	bh = bfs_find_entry(dir, &dentry->d_name, &de);
	if (bh) {
		unsigned long ino = (unsigned long)le16_to_cpu(de->ino);
		brelse(bh);
		inode = bfs_iget(dir->i_sb, ino);
	}
	mutex_unlock(&info->bfs_lock);
	return d_splice_alias(inode, dentry);
}

static int bfs_link(struct dentry *old, struct inode *dir,
						struct dentry *new)
{
	struct inode *inode = d_inode(old);
	struct bfs_sb_info *info = BFS_SB(inode->i_sb);
	int err;

	mutex_lock(&info->bfs_lock);
	err = bfs_add_entry(dir, &new->d_name, inode->i_ino);
	if (err) {
		mutex_unlock(&info->bfs_lock);
		return err;
	}
	inc_nlink(inode);
	inode->i_ctime = current_time(inode);
	mark_inode_dirty(inode);
	ihold(inode);
	d_instantiate(new, inode);
	mutex_unlock(&info->bfs_lock);
	return 0;
}

static int bfs_unlink(struct inode *dir, struct dentry *dentry)
{
	int error = -ENOENT;
	struct inode *inode = d_inode(dentry);
	struct buffer_head *bh;
	struct bfs_dirent *de;
	struct bfs_sb_info *info = BFS_SB(inode->i_sb);

	mutex_lock(&info->bfs_lock);
	bh = bfs_find_entry(dir, &dentry->d_name, &de);
	if (!bh || (le16_to_cpu(de->ino) != inode->i_ino))
		goto out_brelse;

	if (!inode->i_nlink) {
		printf("unlinking non-existent file %s:%lu (nlink=%d)\n",
					inode->i_sb->s_id, inode->i_ino,
					inode->i_nlink);
		set_nlink(inode, 1);
	}
	de->ino = 0;
	mark_buffer_dirty_inode(bh, dir);
	dir->i_ctime = dir->i_mtime = current_time(dir);
	mark_inode_dirty(dir);
	inode->i_ctime = dir->i_ctime;
	inode_dec_link_count(inode);
	error = 0;

out_brelse:
	brelse(bh);
	mutex_unlock(&info->bfs_lock);
	return error;
}

static int bfs_rename(struct user_namespace *mnt_userns, struct inode *old_dir,
		      struct dentry *old_dentry, struct inode *new_dir,
		      struct dentry *new_dentry, unsigned int flags)
{
	struct inode *old_inode, *new_inode;
	struct buffer_head *old_bh, *new_bh;
	struct bfs_dirent *old_de, *new_de;
	struct bfs_sb_info *info;
	int error = -ENOENT;

	if (flags & ~RENAME_NOREPLACE)
		return -EINVAL;

	old_bh = new_bh = NULL;
	old_inode = d_inode(old_dentry);
	if (S_ISDIR(old_inode->i_mode))
		return -EINVAL;

	info = BFS_SB(old_inode->i_sb);

	mutex_lock(&info->bfs_lock);
	old_bh = bfs_find_entry(old_dir, &old_dentry->d_name, &old_de);

	if (!old_bh || (le16_to_cpu(old_de->ino) != old_inode->i_ino))
		goto end_rename;

	error = -EPERM;
	new_inode = d_inode(new_dentry);
	new_bh = bfs_find_entry(new_dir, &new_dentry->d_name, &new_de);

	if (new_bh && !new_inode) {
		brelse(new_bh);
		new_bh = NULL;
	}
	if (!new_bh) {
		error = bfs_add_entry(new_dir, &new_dentry->d_name,
					old_inode->i_ino);
		if (error)
			goto end_rename;
	}
	old_de->ino = 0;
	old_dir->i_ctime = old_dir->i_mtime = current_time(old_dir);
	mark_inode_dirty(old_dir);
	if (new_inode) {
		new_inode->i_ctime = current_time(new_inode);
		inode_dec_link_count(new_inode);
	}
	mark_buffer_dirty_inode(old_bh, old_dir);
	error = 0;

end_rename:
	mutex_unlock(&info->bfs_lock);
	brelse(old_bh);
	brelse(new_bh);
	return error;
}

const struct inode_operations bfs_dir_inops = {
	.create			= bfs_create,
	.lookup			= bfs_lookup,
	.link			= bfs_link,
	.unlink			= bfs_unlink,
	.rename			= bfs_rename,
};

static int bfs_add_entry(struct inode *dir, const struct qstr *child, int ino)
{
	const unsigned char *name = child->name;
	int namelen = child->len;
	struct buffer_head *bh;
	struct bfs_dirent *de;
	int block, sblock, eblock, off, pos;
	int i;

	dprintf("name=%s, namelen=%d\n", name, namelen);

	if (!namelen)
		return -ENOENT;
	if (namelen > BFS_NAMELEN)
		return -ENAMETOOLONG;

	sblock = BFS_I(dir)->i_sblock;
	eblock = BFS_I(dir)->i_eblock;
	for (block = sblock; block <= eblock; block++) {
		bh = sb_bread(dir->i_sb, block);
		if (!bh)
			return -EIO;
		for (off = 0; off < BFS_BSIZE; off += BFS_DIRENT_SIZE) {
			de = (struct bfs_dirent *)(bh->b_data + off);
			if (!de->ino) {
				pos = (block - sblock) * BFS_BSIZE + off;
				if (pos >= dir->i_size) {
					dir->i_size += BFS_DIRENT_SIZE;
					dir->i_ctime = current_time(dir);
				}
				dir->i_mtime = current_time(dir);
				mark_inode_dirty(dir);
				de->ino = cpu_to_le16((u16)ino);
				for (i = 0; i < BFS_NAMELEN; i++)
					de->name[i] =
						(i < namelen) ? name[i] : 0;
				mark_buffer_dirty_inode(bh, dir);
				brelse(bh);
				return 0;
			}
		}
		brelse(bh);
	}
	return -ENOSPC;
}

static inline int bfs_namecmp(int len, const unsigned char *name,
							const char *buffer)
{
	if ((len < BFS_NAMELEN) && buffer[len])
		return 0;
	return !memcmp(name, buffer, len);
}

static struct buffer_head *bfs_find_entry(struct inode *dir,
			const struct qstr *child,
			struct bfs_dirent **res_dir)
{
	unsigned long block = 0, offset = 0;
	struct buffer_head *bh = NULL;
	struct bfs_dirent *de;
	const unsigned char *name = child->name;
	int namelen = child->len;

	*res_dir = NULL;
	if (namelen > BFS_NAMELEN)
		return NULL;

	while (block * BFS_BSIZE + offset < dir->i_size) {
		if (!bh) {
			bh = sb_bread(dir->i_sb, BFS_I(dir)->i_sblock + block);
			if (!bh) {
				block++;
				continue;
			}
		}
		de = (struct bfs_dirent *)(bh->b_data + offset);
		offset += BFS_DIRENT_SIZE;
		if (le16_to_cpu(de->ino) &&
				bfs_namecmp(namelen, name, de->name)) {
			*res_dir = de;
			return bh;
		}
		if (offset < bh->b_size)
			continue;
		brelse(bh);
		bh = NULL;
		offset = 0;
		block++;
	}
	brelse(bh);
	return NULL;
}
back to top