summaryrefslogtreecommitdiffstats
path: root/include
diff options
context:
space:
mode:
authorAlan Stern <stern@rowland.harvard.edu>2006-03-27 01:16:30 -0800
committerLinus Torvalds <torvalds@g5.osdl.org>2006-03-27 08:44:50 -0800
commite041c683412d5bf44dc2b109053e3b837b71742d (patch)
tree9d271066ef379da0c0fb3b8cb4137abd5d2ebba0 /include
parent76b81e2b0e2241accebcc68e126bc5ab958661b9 (diff)
downloadop-kernel-dev-e041c683412d5bf44dc2b109053e3b837b71742d.zip
op-kernel-dev-e041c683412d5bf44dc2b109053e3b837b71742d.tar.gz
[PATCH] Notifier chain update: API changes
The kernel's implementation of notifier chains is unsafe. There is no protection against entries being added to or removed from a chain while the chain is in use. The issues were discussed in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2 We noticed that notifier chains in the kernel fall into two basic usage classes: "Blocking" chains are always called from a process context and the callout routines are allowed to sleep; "Atomic" chains can be called from an atomic context and the callout routines are not allowed to sleep. We decided to codify this distinction and make it part of the API. Therefore this set of patches introduces three new, parallel APIs: one for blocking notifiers, one for atomic notifiers, and one for "raw" notifiers (which is really just the old API under a new name). New kinds of data structures are used for the heads of the chains, and new routines are defined for registration, unregistration, and calling a chain. The three APIs are explained in include/linux/notifier.h and their implementation is in kernel/sys.c. With atomic and blocking chains, the implementation guarantees that the chain links will not be corrupted and that chain callers will not get messed up by entries being added or removed. For raw chains the implementation provides no guarantees at all; users of this API must provide their own protections. (The idea was that situations may come up where the assumptions of the atomic and blocking APIs are not appropriate, so it should be possible for users to handle these things in their own way.) There are some limitations, which should not be too hard to live with. For atomic/blocking chains, registration and unregistration must always be done in a process context since the chain is protected by a mutex/rwsem. Also, a callout routine for a non-raw chain must not try to register or unregister entries on its own chain. (This did happen in a couple of places and the code had to be changed to avoid it.) Since atomic chains may be called from within an NMI handler, they cannot use spinlocks for synchronization. Instead we use RCU. The overhead falls almost entirely in the unregister routine, which is okay since unregistration is much less frequent that calling a chain. Here is the list of chains that we adjusted and their classifications. None of them use the raw API, so for the moment it is only a placeholder. ATOMIC CHAINS ------------- arch/i386/kernel/traps.c: i386die_chain arch/ia64/kernel/traps.c: ia64die_chain arch/powerpc/kernel/traps.c: powerpc_die_chain arch/sparc64/kernel/traps.c: sparc64die_chain arch/x86_64/kernel/traps.c: die_chain drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list kernel/panic.c: panic_notifier_list kernel/profile.c: task_free_notifier net/bluetooth/hci_core.c: hci_notifier net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain net/ipv6/addrconf.c: inet6addr_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_chain net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain net/netlink/af_netlink.c: netlink_chain BLOCKING CHAINS --------------- arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain arch/s390/kernel/process.c: idle_chain arch/x86_64/kernel/process.c idle_notifier drivers/base/memory.c: memory_chain drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list drivers/macintosh/adb.c: adb_client_list drivers/macintosh/via-pmu.c sleep_notifier_list drivers/macintosh/via-pmu68k.c sleep_notifier_list drivers/macintosh/windfarm_core.c wf_client_list drivers/usb/core/notify.c usb_notifier_list drivers/video/fbmem.c fb_notifier_list kernel/cpu.c cpu_chain kernel/module.c module_notify_list kernel/profile.c munmap_notifier kernel/profile.c task_exit_notifier kernel/sys.c reboot_notifier_list net/core/dev.c netdev_chain net/decnet/dn_dev.c: dnaddr_chain net/ipv4/devinet.c: inetaddr_chain It's possible that some of these classifications are wrong. If they are, please let us know or submit a patch to fix them. Note that any chain that gets called very frequently should be atomic, because the rwsem read-locking used for blocking chains is very likely to incur cache misses on SMP systems. (However, if the chain's callout routines may sleep then the chain cannot be atomic.) The patch set was written by Alan Stern and Chandra Seetharaman, incorporating material written by Keith Owens and suggestions from Paul McKenney and Andrew Morton. [jes@sgi.com: restructure the notifier chain initialization macros] Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com> Signed-off-by: Jes Sorensen <jes@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'include')
-rw-r--r--include/asm-i386/kdebug.h10
-rw-r--r--include/asm-ia64/kdebug.h4
-rw-r--r--include/asm-powerpc/kdebug.h12
-rw-r--r--include/asm-sparc64/kdebug.h11
-rw-r--r--include/asm-x86_64/kdebug.h23
-rw-r--r--include/linux/adb.h2
-rw-r--r--include/linux/kernel.h2
-rw-r--r--include/linux/memory.h1
-rw-r--r--include/linux/netfilter_ipv4/ip_conntrack.h17
-rw-r--r--include/linux/notifier.h96
-rw-r--r--include/net/netfilter/nf_conntrack.h17
11 files changed, 134 insertions, 61 deletions
diff --git a/include/asm-i386/kdebug.h b/include/asm-i386/kdebug.h
index 316138e..96d0828 100644
--- a/include/asm-i386/kdebug.h
+++ b/include/asm-i386/kdebug.h
@@ -17,11 +17,9 @@ struct die_args {
int signr;
};
-/* Note - you should never unregister because that can race with NMIs.
- If you really want to do it first unregister - then synchronize_sched - then free.
- */
-int register_die_notifier(struct notifier_block *nb);
-extern struct notifier_block *i386die_chain;
+extern int register_die_notifier(struct notifier_block *);
+extern int unregister_die_notifier(struct notifier_block *);
+extern struct atomic_notifier_head i386die_chain;
/* Grossly misnamed. */
@@ -51,7 +49,7 @@ static inline int notify_die(enum die_val val, const char *str,
.trapnr = trap,
.signr = sig
};
- return notifier_call_chain(&i386die_chain, val, &args);
+ return atomic_notifier_call_chain(&i386die_chain, val, &args);
}
#endif
diff --git a/include/asm-ia64/kdebug.h b/include/asm-ia64/kdebug.h
index 8b01a08..218c458 100644
--- a/include/asm-ia64/kdebug.h
+++ b/include/asm-ia64/kdebug.h
@@ -40,7 +40,7 @@ struct die_args {
extern int register_die_notifier(struct notifier_block *);
extern int unregister_die_notifier(struct notifier_block *);
-extern struct notifier_block *ia64die_chain;
+extern struct atomic_notifier_head ia64die_chain;
enum die_val {
DIE_BREAK = 1,
@@ -81,7 +81,7 @@ static inline int notify_die(enum die_val val, char *str, struct pt_regs *regs,
.signr = sig
};
- return notifier_call_chain(&ia64die_chain, val, &args);
+ return atomic_notifier_call_chain(&ia64die_chain, val, &args);
}
#endif
diff --git a/include/asm-powerpc/kdebug.h b/include/asm-powerpc/kdebug.h
index 7c16265..c01786a 100644
--- a/include/asm-powerpc/kdebug.h
+++ b/include/asm-powerpc/kdebug.h
@@ -16,13 +16,9 @@ struct die_args {
int signr;
};
-/*
- Note - you should never unregister because that can race with NMIs.
- If you really want to do it first unregister - then synchronize_sched -
- then free.
- */
-int register_die_notifier(struct notifier_block *nb);
-extern struct notifier_block *powerpc_die_chain;
+extern int register_die_notifier(struct notifier_block *);
+extern int unregister_die_notifier(struct notifier_block *);
+extern struct atomic_notifier_head powerpc_die_chain;
/* Grossly misnamed. */
enum die_val {
@@ -37,7 +33,7 @@ enum die_val {
static inline int notify_die(enum die_val val,char *str,struct pt_regs *regs,long err,int trap, int sig)
{
struct die_args args = { .regs=regs, .str=str, .err=err, .trapnr=trap,.signr=sig };
- return notifier_call_chain(&powerpc_die_chain, val, &args);
+ return atomic_notifier_call_chain(&powerpc_die_chain, val, &args);
}
#endif /* __KERNEL__ */
diff --git a/include/asm-sparc64/kdebug.h b/include/asm-sparc64/kdebug.h
index 6321f5a..4040d12 100644
--- a/include/asm-sparc64/kdebug.h
+++ b/include/asm-sparc64/kdebug.h
@@ -15,12 +15,9 @@ struct die_args {
int signr;
};
-/* Note - you should never unregister because that can race with NMIs.
- * If you really want to do it first unregister - then synchronize_sched
- * - then free.
- */
-int register_die_notifier(struct notifier_block *nb);
-extern struct notifier_block *sparc64die_chain;
+extern int register_die_notifier(struct notifier_block *);
+extern int unregister_die_notifier(struct notifier_block *);
+extern struct atomic_notifier_head sparc64die_chain;
extern void bad_trap(struct pt_regs *, long);
@@ -46,7 +43,7 @@ static inline int notify_die(enum die_val val,char *str, struct pt_regs *regs,
.trapnr = trap,
.signr = sig };
- return notifier_call_chain(&sparc64die_chain, val, &args);
+ return atomic_notifier_call_chain(&sparc64die_chain, val, &args);
}
#endif
diff --git a/include/asm-x86_64/kdebug.h b/include/asm-x86_64/kdebug.h
index b9ed4c0..cf79563 100644
--- a/include/asm-x86_64/kdebug.h
+++ b/include/asm-x86_64/kdebug.h
@@ -5,21 +5,20 @@
struct pt_regs;
-struct die_args {
+struct die_args {
struct pt_regs *regs;
const char *str;
- long err;
+ long err;
int trapnr;
int signr;
-};
+};
+
+extern int register_die_notifier(struct notifier_block *);
+extern int unregister_die_notifier(struct notifier_block *);
+extern struct atomic_notifier_head die_chain;
-/* Note - you should never unregister because that can race with NMIs.
- If you really want to do it first unregister - then synchronize_sched - then free.
- */
-int register_die_notifier(struct notifier_block *nb);
-extern struct notifier_block *die_chain;
/* Grossly misnamed. */
-enum die_val {
+enum die_val {
DIE_OOPS = 1,
DIE_INT3,
DIE_DEBUG,
@@ -33,8 +32,8 @@ enum die_val {
DIE_CALL,
DIE_NMI_IPI,
DIE_PAGE_FAULT,
-};
-
+};
+
static inline int notify_die(enum die_val val, const char *str,
struct pt_regs *regs, long err, int trap, int sig)
{
@@ -45,7 +44,7 @@ static inline int notify_die(enum die_val val, const char *str,
.trapnr = trap,
.signr = sig
};
- return notifier_call_chain(&die_chain, val, &args);
+ return atomic_notifier_call_chain(&die_chain, val, &args);
}
extern int printk_address(unsigned long address);
diff --git a/include/linux/adb.h b/include/linux/adb.h
index e9fdc63..b7305b1 100644
--- a/include/linux/adb.h
+++ b/include/linux/adb.h
@@ -85,7 +85,7 @@ enum adb_message {
ADB_MSG_POST_RESET /* Called after resetting the bus (re-do init & register) */
};
extern struct adb_driver *adb_controller;
-extern struct notifier_block *adb_client_list;
+extern struct blocking_notifier_head adb_client_list;
int adb_request(struct adb_request *req, void (*done)(struct adb_request *),
int flags, int nbytes, ...);
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 03d6cfa..a3720f9 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -87,7 +87,7 @@ extern int cond_resched(void);
(__x < 0) ? -__x : __x; \
})
-extern struct notifier_block *panic_notifier_list;
+extern struct atomic_notifier_head panic_notifier_list;
extern long (*panic_blink)(long time);
NORET_TYPE void panic(const char * fmt, ...)
__attribute__ ((NORET_AND format (printf, 1, 2)));
diff --git a/include/linux/memory.h b/include/linux/memory.h
index e251dc4..8f04143 100644
--- a/include/linux/memory.h
+++ b/include/linux/memory.h
@@ -77,7 +77,6 @@ extern int remove_memory_block(unsigned long, struct mem_section *, int);
#define CONFIG_MEM_BLOCK_SIZE (PAGES_PER_SECTION<<PAGE_SHIFT)
-struct notifier_block;
#endif /* CONFIG_MEMORY_HOTPLUG */
diff --git a/include/linux/netfilter_ipv4/ip_conntrack.h b/include/linux/netfilter_ipv4/ip_conntrack.h
index f32d75c..d54d7b2 100644
--- a/include/linux/netfilter_ipv4/ip_conntrack.h
+++ b/include/linux/netfilter_ipv4/ip_conntrack.h
@@ -308,29 +308,30 @@ DECLARE_PER_CPU(struct ip_conntrack_ecache, ip_conntrack_ecache);
#define CONNTRACK_ECACHE(x) (__get_cpu_var(ip_conntrack_ecache).x)
-extern struct notifier_block *ip_conntrack_chain;
-extern struct notifier_block *ip_conntrack_expect_chain;
+extern struct atomic_notifier_head ip_conntrack_chain;
+extern struct atomic_notifier_head ip_conntrack_expect_chain;
static inline int ip_conntrack_register_notifier(struct notifier_block *nb)
{
- return notifier_chain_register(&ip_conntrack_chain, nb);
+ return atomic_notifier_chain_register(&ip_conntrack_chain, nb);
}
static inline int ip_conntrack_unregister_notifier(struct notifier_block *nb)
{
- return notifier_chain_unregister(&ip_conntrack_chain, nb);
+ return atomic_notifier_chain_unregister(&ip_conntrack_chain, nb);
}
static inline int
ip_conntrack_expect_register_notifier(struct notifier_block *nb)
{
- return notifier_chain_register(&ip_conntrack_expect_chain, nb);
+ return atomic_notifier_chain_register(&ip_conntrack_expect_chain, nb);
}
static inline int
ip_conntrack_expect_unregister_notifier(struct notifier_block *nb)
{
- return notifier_chain_unregister(&ip_conntrack_expect_chain, nb);
+ return atomic_notifier_chain_unregister(&ip_conntrack_expect_chain,
+ nb);
}
extern void ip_ct_deliver_cached_events(const struct ip_conntrack *ct);
@@ -355,14 +356,14 @@ static inline void ip_conntrack_event(enum ip_conntrack_events event,
struct ip_conntrack *ct)
{
if (is_confirmed(ct) && !is_dying(ct))
- notifier_call_chain(&ip_conntrack_chain, event, ct);
+ atomic_notifier_call_chain(&ip_conntrack_chain, event, ct);
}
static inline void
ip_conntrack_expect_event(enum ip_conntrack_expect_events event,
struct ip_conntrack_expect *exp)
{
- notifier_call_chain(&ip_conntrack_expect_chain, event, exp);
+ atomic_notifier_call_chain(&ip_conntrack_expect_chain, event, exp);
}
#else /* CONFIG_IP_NF_CONNTRACK_EVENTS */
static inline void ip_conntrack_event_cache(enum ip_conntrack_events event,
diff --git a/include/linux/notifier.h b/include/linux/notifier.h
index 5937dd60..51dbab9 100644
--- a/include/linux/notifier.h
+++ b/include/linux/notifier.h
@@ -10,25 +10,107 @@
#ifndef _LINUX_NOTIFIER_H
#define _LINUX_NOTIFIER_H
#include <linux/errno.h>
+#include <linux/mutex.h>
+#include <linux/rwsem.h>
-struct notifier_block
-{
- int (*notifier_call)(struct notifier_block *self, unsigned long, void *);
+/*
+ * Notifier chains are of three types:
+ *
+ * Atomic notifier chains: Chain callbacks run in interrupt/atomic
+ * context. Callouts are not allowed to block.
+ * Blocking notifier chains: Chain callbacks run in process context.
+ * Callouts are allowed to block.
+ * Raw notifier chains: There are no restrictions on callbacks,
+ * registration, or unregistration. All locking and protection
+ * must be provided by the caller.
+ *
+ * atomic_notifier_chain_register() may be called from an atomic context,
+ * but blocking_notifier_chain_register() must be called from a process
+ * context. Ditto for the corresponding _unregister() routines.
+ *
+ * atomic_notifier_chain_unregister() and blocking_notifier_chain_unregister()
+ * _must not_ be called from within the call chain.
+ */
+
+struct notifier_block {
+ int (*notifier_call)(struct notifier_block *, unsigned long, void *);
struct notifier_block *next;
int priority;
};
+struct atomic_notifier_head {
+ spinlock_t lock;
+ struct notifier_block *head;
+};
+
+struct blocking_notifier_head {
+ struct rw_semaphore rwsem;
+ struct notifier_block *head;
+};
+
+struct raw_notifier_head {
+ struct notifier_block *head;
+};
+
+#define ATOMIC_INIT_NOTIFIER_HEAD(name) do { \
+ spin_lock_init(&(name)->lock); \
+ (name)->head = NULL; \
+ } while (0)
+#define BLOCKING_INIT_NOTIFIER_HEAD(name) do { \
+ init_rwsem(&(name)->rwsem); \
+ (name)->head = NULL; \
+ } while (0)
+#define RAW_INIT_NOTIFIER_HEAD(name) do { \
+ (name)->head = NULL; \
+ } while (0)
+
+#define ATOMIC_NOTIFIER_INIT(name) { \
+ .lock = SPIN_LOCK_UNLOCKED, \
+ .head = NULL }
+#define BLOCKING_NOTIFIER_INIT(name) { \
+ .rwsem = __RWSEM_INITIALIZER((name).rwsem), \
+ .head = NULL }
+#define RAW_NOTIFIER_INIT(name) { \
+ .head = NULL }
+
+#define ATOMIC_NOTIFIER_HEAD(name) \
+ struct atomic_notifier_head name = \
+ ATOMIC_NOTIFIER_INIT(name)
+#define BLOCKING_NOTIFIER_HEAD(name) \
+ struct blocking_notifier_head name = \
+ BLOCKING_NOTIFIER_INIT(name)
+#define RAW_NOTIFIER_HEAD(name) \
+ struct raw_notifier_head name = \
+ RAW_NOTIFIER_INIT(name)
#ifdef __KERNEL__
-extern int notifier_chain_register(struct notifier_block **list, struct notifier_block *n);
-extern int notifier_chain_unregister(struct notifier_block **nl, struct notifier_block *n);
-extern int notifier_call_chain(struct notifier_block **n, unsigned long val, void *v);
+extern int atomic_notifier_chain_register(struct atomic_notifier_head *,
+ struct notifier_block *);
+extern int blocking_notifier_chain_register(struct blocking_notifier_head *,
+ struct notifier_block *);
+extern int raw_notifier_chain_register(struct raw_notifier_head *,
+ struct notifier_block *);
+
+extern int atomic_notifier_chain_unregister(struct atomic_notifier_head *,
+ struct notifier_block *);
+extern int blocking_notifier_chain_unregister(struct blocking_notifier_head *,
+ struct notifier_block *);
+extern int raw_notifier_chain_unregister(struct raw_notifier_head *,
+ struct notifier_block *);
+
+extern int atomic_notifier_call_chain(struct atomic_notifier_head *,
+ unsigned long val, void *v);
+extern int blocking_notifier_call_chain(struct blocking_notifier_head *,
+ unsigned long val, void *v);
+extern int raw_notifier_call_chain(struct raw_notifier_head *,
+ unsigned long val, void *v);
#define NOTIFY_DONE 0x0000 /* Don't care */
#define NOTIFY_OK 0x0001 /* Suits me */
#define NOTIFY_STOP_MASK 0x8000 /* Don't call further */
-#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002) /* Bad/Veto action */
+#define NOTIFY_BAD (NOTIFY_STOP_MASK|0x0002)
+ /* Bad/Veto action */
/*
* Clean way to return from the notifier and stop further calls.
*/
diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h
index b6f0905..916013c 100644
--- a/include/net/netfilter/nf_conntrack.h
+++ b/include/net/netfilter/nf_conntrack.h
@@ -300,29 +300,30 @@ DECLARE_PER_CPU(struct nf_conntrack_ecache, nf_conntrack_ecache);
#define CONNTRACK_ECACHE(x) (__get_cpu_var(nf_conntrack_ecache).x)
-extern struct notifier_block *nf_conntrack_chain;
-extern struct notifier_block *nf_conntrack_expect_chain;
+extern struct atomic_notifier_head nf_conntrack_chain;
+extern struct atomic_notifier_head nf_conntrack_expect_chain;
static inline int nf_conntrack_register_notifier(struct notifier_block *nb)
{
- return notifier_chain_register(&nf_conntrack_chain, nb);
+ return atomic_notifier_chain_register(&nf_conntrack_chain, nb);
}
static inline int nf_conntrack_unregister_notifier(struct notifier_block *nb)
{
- return notifier_chain_unregister(&nf_conntrack_chain, nb);
+ return atomic_notifier_chain_unregister(&nf_conntrack_chain, nb);
}
static inline int
nf_conntrack_expect_register_notifier(struct notifier_block *nb)
{
- return notifier_chain_register(&nf_conntrack_expect_chain, nb);
+ return atomic_notifier_chain_register(&nf_conntrack_expect_chain, nb);
}
static inline int
nf_conntrack_expect_unregister_notifier(struct notifier_block *nb)
{
- return notifier_chain_unregister(&nf_conntrack_expect_chain, nb);
+ return atomic_notifier_chain_unregister(&nf_conntrack_expect_chain,
+ nb);
}
extern void nf_ct_deliver_cached_events(const struct nf_conn *ct);
@@ -347,14 +348,14 @@ static inline void nf_conntrack_event(enum ip_conntrack_events event,
struct nf_conn *ct)
{
if (nf_ct_is_confirmed(ct) && !nf_ct_is_dying(ct))
- notifier_call_chain(&nf_conntrack_chain, event, ct);
+ atomic_notifier_call_chain(&nf_conntrack_chain, event, ct);
}
static inline void
nf_conntrack_expect_event(enum ip_conntrack_expect_events event,
struct nf_conntrack_expect *exp)
{
- notifier_call_chain(&nf_conntrack_expect_chain, event, exp);
+ atomic_notifier_call_chain(&nf_conntrack_expect_chain, event, exp);
}
#else /* CONFIG_NF_CONNTRACK_EVENTS */
static inline void nf_conntrack_event_cache(enum ip_conntrack_events event,
OpenPOWER on IntegriCloud