Revision c9d398fa237882ea07167e23bcfc5e6847066518 authored by Naoya Horiguchi on 31 March 2017, 22:11:55 UTC, committed by Linus Torvalds on 01 April 2017, 00:13:30 UTC
I found the race condition which triggers the following bug when move_pages() and soft offline are called on a single hugetlb page concurrently. Soft offlining page 0x119400 at 0x700000000000 BUG: unable to handle kernel paging request at ffffea0011943820 IP: follow_huge_pmd+0x143/0x190 PGD 7ffd2067 PUD 7ffd1067 PMD 0 [61163.582052] Oops: 0000 [#1] SMP Modules linked in: binfmt_misc ppdev virtio_balloon parport_pc pcspkr i2c_piix4 parport i2c_core acpi_cpufreq ip_tables xfs libcrc32c ata_generic pata_acpi virtio_blk 8139too crc32c_intel ata_piix serio_raw libata virtio_pci 8139cp virtio_ring virtio mii floppy dm_mirror dm_region_hash dm_log dm_mod [last unloaded: cap_check] CPU: 0 PID: 22573 Comm: iterate_numa_mo Tainted: P OE 4.11.0-rc2-mm1+ #2 Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011 RIP: 0010:follow_huge_pmd+0x143/0x190 RSP: 0018:ffffc90004bdbcd0 EFLAGS: 00010202 RAX: 0000000465003e80 RBX: ffffea0004e34d30 RCX: 00003ffffffff000 RDX: 0000000011943800 RSI: 0000000000080001 RDI: 0000000465003e80 RBP: ffffc90004bdbd18 R08: 0000000000000000 R09: ffff880138d34000 R10: ffffea0004650000 R11: 0000000000c363b0 R12: ffffea0011943800 R13: ffff8801b8d34000 R14: ffffea0000000000 R15: 000077ff80000000 FS: 00007fc977710740(0000) GS:ffff88007dc00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffea0011943820 CR3: 000000007a746000 CR4: 00000000001406f0 Call Trace: follow_page_mask+0x270/0x550 SYSC_move_pages+0x4ea/0x8f0 SyS_move_pages+0xe/0x10 do_syscall_64+0x67/0x180 entry_SYSCALL64_slow_path+0x25/0x25 RIP: 0033:0x7fc976e03949 RSP: 002b:00007ffe72221d88 EFLAGS: 00000246 ORIG_RAX: 0000000000000117 RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fc976e03949 RDX: 0000000000c22390 RSI: 0000000000001400 RDI: 0000000000005827 RBP: 00007ffe72221e00 R08: 0000000000c2c3a0 R09: 0000000000000004 R10: 0000000000c363b0 R11: 0000000000000246 R12: 0000000000400650 R13: 00007ffe72221ee0 R14: 0000000000000000 R15: 0000000000000000 Code: 81 e4 ff ff 1f 00 48 21 c2 49 c1 ec 0c 48 c1 ea 0c 4c 01 e2 49 bc 00 00 00 00 00 ea ff ff 48 c1 e2 06 49 01 d4 f6 45 bc 04 74 90 <49> 8b 7c 24 20 40 f6 c7 01 75 2b 4c 89 e7 8b 47 1c 85 c0 7e 2a RIP: follow_huge_pmd+0x143/0x190 RSP: ffffc90004bdbcd0 CR2: ffffea0011943820 ---[ end trace e4f81353a2d23232 ]--- Kernel panic - not syncing: Fatal exception Kernel Offset: disabled This bug is triggered when pmd_present() returns true for non-present hugetlb, so fixing the present check in follow_huge_pmd() prevents it. Using pmd_present() to determine present/non-present for hugetlb is not correct, because pmd_present() checks multiple bits (not only _PAGE_PRESENT) for historical reason and it can misjudge hugetlb state. Fixes: e66f17ff7177 ("mm/hugetlb: take page table lock in follow_huge_pmd()") Link: http://lkml.kernel.org/r/1490149898-20231-1-git-send-email-n-horiguchi@ah.jp.nec.com Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Acked-by: Hillf Danton <hillf.zj@alibaba-inc.com> Cc: Hugh Dickins <hughd@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Cc: Mike Kravetz <mike.kravetz@oracle.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com> Cc: <stable@vger.kernel.org> [4.0+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent 0cefabd
audit.h
/* audit -- definition of audit_context structure and supporting types
*
* Copyright 2003-2004 Red Hat, Inc.
* Copyright 2005 Hewlett-Packard Development Company, L.P.
* Copyright 2005 IBM Corporation
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#include <linux/fs.h>
#include <linux/audit.h>
#include <linux/skbuff.h>
#include <uapi/linux/mqueue.h>
#include <linux/tty.h>
/* AUDIT_NAMES is the number of slots we reserve in the audit_context
* for saving names from getname(). If we get more names we will allocate
* a name dynamically and also add those to the list anchored by names_list. */
#define AUDIT_NAMES 5
/* At task start time, the audit_state is set in the audit_context using
a per-task filter. At syscall entry, the audit_state is augmented by
the syscall filter. */
enum audit_state {
AUDIT_DISABLED, /* Do not create per-task audit_context.
* No syscall-specific audit records can
* be generated. */
AUDIT_BUILD_CONTEXT, /* Create the per-task audit_context,
* and fill it in at syscall
* entry time. This makes a full
* syscall record available if some
* other part of the kernel decides it
* should be recorded. */
AUDIT_RECORD_CONTEXT /* Create the per-task audit_context,
* always fill it in at syscall entry
* time, and always write out the audit
* record at syscall exit time. */
};
/* Rule lists */
struct audit_watch;
struct audit_fsnotify_mark;
struct audit_tree;
struct audit_chunk;
struct audit_entry {
struct list_head list;
struct rcu_head rcu;
struct audit_krule rule;
};
struct audit_cap_data {
kernel_cap_t permitted;
kernel_cap_t inheritable;
union {
unsigned int fE; /* effective bit of file cap */
kernel_cap_t effective; /* effective set of process */
};
};
/* When fs/namei.c:getname() is called, we store the pointer in name and bump
* the refcnt in the associated filename struct.
*
* Further, in fs/namei.c:path_lookup() we store the inode and device.
*/
struct audit_names {
struct list_head list; /* audit_context->names_list */
struct filename *name;
int name_len; /* number of chars to log */
bool hidden; /* don't log this record */
unsigned long ino;
dev_t dev;
umode_t mode;
kuid_t uid;
kgid_t gid;
dev_t rdev;
u32 osid;
struct audit_cap_data fcap;
unsigned int fcap_ver;
unsigned char type; /* record type */
/*
* This was an allocated audit_names and not from the array of
* names allocated in the task audit context. Thus this name
* should be freed on syscall exit.
*/
bool should_free;
};
struct audit_proctitle {
int len; /* length of the cmdline field. */
char *value; /* the cmdline field */
};
/* The per-task audit context. */
struct audit_context {
int dummy; /* must be the first element */
int in_syscall; /* 1 if task is in a syscall */
enum audit_state state, current_state;
unsigned int serial; /* serial number for record */
int major; /* syscall number */
struct timespec ctime; /* time of syscall entry */
unsigned long argv[4]; /* syscall arguments */
long return_code;/* syscall return code */
u64 prio;
int return_valid; /* return code is valid */
/*
* The names_list is the list of all audit_names collected during this
* syscall. The first AUDIT_NAMES entries in the names_list will
* actually be from the preallocated_names array for performance
* reasons. Except during allocation they should never be referenced
* through the preallocated_names array and should only be found/used
* by running the names_list.
*/
struct audit_names preallocated_names[AUDIT_NAMES];
int name_count; /* total records in names_list */
struct list_head names_list; /* struct audit_names->list anchor */
char *filterkey; /* key for rule that triggered record */
struct path pwd;
struct audit_aux_data *aux;
struct audit_aux_data *aux_pids;
struct sockaddr_storage *sockaddr;
size_t sockaddr_len;
/* Save things to print about task_struct */
pid_t pid, ppid;
kuid_t uid, euid, suid, fsuid;
kgid_t gid, egid, sgid, fsgid;
unsigned long personality;
int arch;
pid_t target_pid;
kuid_t target_auid;
kuid_t target_uid;
unsigned int target_sessionid;
u32 target_sid;
char target_comm[TASK_COMM_LEN];
struct audit_tree_refs *trees, *first_trees;
struct list_head killed_trees;
int tree_count;
int type;
union {
struct {
int nargs;
long args[6];
} socketcall;
struct {
kuid_t uid;
kgid_t gid;
umode_t mode;
u32 osid;
int has_perm;
uid_t perm_uid;
gid_t perm_gid;
umode_t perm_mode;
unsigned long qbytes;
} ipc;
struct {
mqd_t mqdes;
struct mq_attr mqstat;
} mq_getsetattr;
struct {
mqd_t mqdes;
int sigev_signo;
} mq_notify;
struct {
mqd_t mqdes;
size_t msg_len;
unsigned int msg_prio;
struct timespec abs_timeout;
} mq_sendrecv;
struct {
int oflag;
umode_t mode;
struct mq_attr attr;
} mq_open;
struct {
pid_t pid;
struct audit_cap_data cap;
} capset;
struct {
int fd;
int flags;
} mmap;
struct {
int argc;
} execve;
struct {
char *name;
} module;
};
int fds[2];
struct audit_proctitle proctitle;
};
extern u32 audit_ever_enabled;
extern void audit_copy_inode(struct audit_names *name,
const struct dentry *dentry,
struct inode *inode);
extern void audit_log_cap(struct audit_buffer *ab, char *prefix,
kernel_cap_t *cap);
extern void audit_log_name(struct audit_context *context,
struct audit_names *n, const struct path *path,
int record_num, int *call_panic);
extern int auditd_test_task(const struct task_struct *task);
#define AUDIT_INODE_BUCKETS 32
extern struct list_head audit_inode_hash[AUDIT_INODE_BUCKETS];
static inline int audit_hash_ino(u32 ino)
{
return (ino & (AUDIT_INODE_BUCKETS-1));
}
/* Indicates that audit should log the full pathname. */
#define AUDIT_NAME_FULL -1
extern int audit_match_class(int class, unsigned syscall);
extern int audit_comparator(const u32 left, const u32 op, const u32 right);
extern int audit_uid_comparator(kuid_t left, u32 op, kuid_t right);
extern int audit_gid_comparator(kgid_t left, u32 op, kgid_t right);
extern int parent_len(const char *path);
extern int audit_compare_dname_path(const char *dname, const char *path, int plen);
extern struct sk_buff *audit_make_reply(__u32 portid, int seq, int type,
int done, int multi,
const void *payload, int size);
extern void audit_panic(const char *message);
struct audit_netlink_list {
__u32 portid;
struct net *net;
struct sk_buff_head q;
};
int audit_send_list(void *);
extern int selinux_audit_rule_update(void);
extern struct mutex audit_filter_mutex;
extern int audit_del_rule(struct audit_entry *);
extern void audit_free_rule_rcu(struct rcu_head *);
extern struct list_head audit_filter_list[];
extern struct audit_entry *audit_dupe_rule(struct audit_krule *old);
extern void audit_log_d_path_exe(struct audit_buffer *ab,
struct mm_struct *mm);
extern struct tty_struct *audit_get_tty(struct task_struct *tsk);
extern void audit_put_tty(struct tty_struct *tty);
/* audit watch functions */
#ifdef CONFIG_AUDIT_WATCH
extern void audit_put_watch(struct audit_watch *watch);
extern void audit_get_watch(struct audit_watch *watch);
extern int audit_to_watch(struct audit_krule *krule, char *path, int len, u32 op);
extern int audit_add_watch(struct audit_krule *krule, struct list_head **list);
extern void audit_remove_watch_rule(struct audit_krule *krule);
extern char *audit_watch_path(struct audit_watch *watch);
extern int audit_watch_compare(struct audit_watch *watch, unsigned long ino, dev_t dev);
extern struct audit_fsnotify_mark *audit_alloc_mark(struct audit_krule *krule, char *pathname, int len);
extern char *audit_mark_path(struct audit_fsnotify_mark *mark);
extern void audit_remove_mark(struct audit_fsnotify_mark *audit_mark);
extern void audit_remove_mark_rule(struct audit_krule *krule);
extern int audit_mark_compare(struct audit_fsnotify_mark *mark, unsigned long ino, dev_t dev);
extern int audit_dupe_exe(struct audit_krule *new, struct audit_krule *old);
extern int audit_exe_compare(struct task_struct *tsk, struct audit_fsnotify_mark *mark);
#else
#define audit_put_watch(w) {}
#define audit_get_watch(w) {}
#define audit_to_watch(k, p, l, o) (-EINVAL)
#define audit_add_watch(k, l) (-EINVAL)
#define audit_remove_watch_rule(k) BUG()
#define audit_watch_path(w) ""
#define audit_watch_compare(w, i, d) 0
#define audit_alloc_mark(k, p, l) (ERR_PTR(-EINVAL))
#define audit_mark_path(m) ""
#define audit_remove_mark(m)
#define audit_remove_mark_rule(k)
#define audit_mark_compare(m, i, d) 0
#define audit_exe_compare(t, m) (-EINVAL)
#define audit_dupe_exe(n, o) (-EINVAL)
#endif /* CONFIG_AUDIT_WATCH */
#ifdef CONFIG_AUDIT_TREE
extern struct audit_chunk *audit_tree_lookup(const struct inode *);
extern void audit_put_chunk(struct audit_chunk *);
extern bool audit_tree_match(struct audit_chunk *, struct audit_tree *);
extern int audit_make_tree(struct audit_krule *, char *, u32);
extern int audit_add_tree_rule(struct audit_krule *);
extern int audit_remove_tree_rule(struct audit_krule *);
extern void audit_trim_trees(void);
extern int audit_tag_tree(char *old, char *new);
extern const char *audit_tree_path(struct audit_tree *);
extern void audit_put_tree(struct audit_tree *);
extern void audit_kill_trees(struct list_head *);
#else
#define audit_remove_tree_rule(rule) BUG()
#define audit_add_tree_rule(rule) -EINVAL
#define audit_make_tree(rule, str, op) -EINVAL
#define audit_trim_trees() (void)0
#define audit_put_tree(tree) (void)0
#define audit_tag_tree(old, new) -EINVAL
#define audit_tree_path(rule) "" /* never called */
#define audit_kill_trees(list) BUG()
#endif
extern char *audit_unpack_string(void **, size_t *, size_t);
extern pid_t audit_sig_pid;
extern kuid_t audit_sig_uid;
extern u32 audit_sig_sid;
extern int audit_filter(int msgtype, unsigned int listtype);
#ifdef CONFIG_AUDITSYSCALL
extern int __audit_signal_info(int sig, struct task_struct *t);
static inline int audit_signal_info(int sig, struct task_struct *t)
{
if (auditd_test_task(t) || (audit_signals && !audit_dummy_context()))
return __audit_signal_info(sig, t);
return 0;
}
extern void audit_filter_inodes(struct task_struct *, struct audit_context *);
extern struct list_head *audit_killed_trees(void);
#else
#define audit_signal_info(s,t) AUDIT_DISABLED
#define audit_filter_inodes(t,c) AUDIT_DISABLED
#endif
extern struct mutex audit_cmd_mutex;
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...