[PATCH 2/3] ipc: merge ipc_rcu and kern_ipc_perm

From: Manfred Spraul
Date: Mon May 15 2017 - 13:19:58 EST


ipc has two management structures that exist for every id:
- struct kern_ipc_perm, it contains e.g. the permissions.
- struct ipc_rcu, it contains the rcu head for rcu handling and
the refcount.

The patch merges both structures.
As a bonus, we may save one cacheline, because both structures are
cacheline aligned.
In addition, it reduces the number of casts, instead most codepaths can
use container_of.

Signed-off-by: Manfred Spraul <manfred@xxxxxxxxxxxxxxxx>
---
include/linux/ipc.h | 3 +++
ipc/msg.c | 22 ++++++++++++++--------
ipc/sem.c | 35 ++++++++++++++++++++---------------
ipc/shm.c | 21 ++++++++++++++-------
ipc/util.c | 33 +++++++++++++++------------------
ipc/util.h | 18 +++++++-----------
6 files changed, 73 insertions(+), 59 deletions(-)

diff --git a/include/linux/ipc.h b/include/linux/ipc.h
index 71fd92d..5591f055 100644
--- a/include/linux/ipc.h
+++ b/include/linux/ipc.h
@@ -20,6 +20,9 @@ struct kern_ipc_perm {
umode_t mode;
unsigned long seq;
void *security;
+
+ struct rcu_head rcu;
+ atomic_t refcount;
} ____cacheline_aligned_in_smp;

#endif /* _LINUX_IPC_H */
diff --git a/ipc/msg.c b/ipc/msg.c
index 104926d..e9785c4 100644
--- a/ipc/msg.c
+++ b/ipc/msg.c
@@ -97,8 +97,8 @@ static inline void msg_rmid(struct ipc_namespace *ns, struct msg_queue *s)

static void msg_rcu_free(struct rcu_head *head)
{
- struct ipc_rcu *p = container_of(head, struct ipc_rcu, rcu);
- struct msg_queue *msq = ipc_rcu_to_struct(p);
+ struct kern_ipc_perm *p = container_of(head, struct kern_ipc_perm, rcu);
+ struct msg_queue *msq = container_of(p, struct msg_queue, q_perm);

security_msg_queue_free(msq);
ipc_rcu_free(head);
@@ -118,7 +118,13 @@ static int newque(struct ipc_namespace *ns, struct ipc_params *params)
key_t key = params->key;
int msgflg = params->flg;

- msq = ipc_rcu_alloc(sizeof(*msq));
+ if (offsetof(struct msg_queue, q_perm) != 0) {
+ pr_err("Invalid struct sem_perm, failing msgget().\n");
+ return -ENOMEM;
+ }
+
+ msq = container_of(ipc_rcu_alloc(sizeof(*msq)), struct msg_queue,
+ q_perm);
if (!msq)
return -ENOMEM;

@@ -128,7 +134,7 @@ static int newque(struct ipc_namespace *ns, struct ipc_params *params)
msq->q_perm.security = NULL;
retval = security_msg_queue_alloc(msq);
if (retval) {
- ipc_rcu_putref(msq, ipc_rcu_free);
+ ipc_rcu_putref(&msq->q_perm, ipc_rcu_free);
return retval;
}

@@ -144,7 +150,7 @@ static int newque(struct ipc_namespace *ns, struct ipc_params *params)
/* ipc_addid() locks msq upon success. */
id = ipc_addid(&msg_ids(ns), &msq->q_perm, ns->msg_ctlmni);
if (id < 0) {
- ipc_rcu_putref(msq, msg_rcu_free);
+ ipc_rcu_putref(&msq->q_perm, msg_rcu_free);
return id;
}

@@ -249,7 +255,7 @@ static void freeque(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
free_msg(msg);
}
atomic_sub(msq->q_cbytes, &ns->msg_bytes);
- ipc_rcu_putref(msq, msg_rcu_free);
+ ipc_rcu_putref(&msq->q_perm, msg_rcu_free);
}

/*
@@ -688,7 +694,7 @@ long do_msgsnd(int msqid, long mtype, void __user *mtext,
/* enqueue the sender and prepare to block */
ss_add(msq, &s, msgsz);

- if (!ipc_rcu_getref(msq)) {
+ if (!ipc_rcu_getref(&msq->q_perm)) {
err = -EIDRM;
goto out_unlock0;
}
@@ -700,7 +706,7 @@ long do_msgsnd(int msqid, long mtype, void __user *mtext,
rcu_read_lock();
ipc_lock_object(&msq->q_perm);

- ipc_rcu_putref(msq, msg_rcu_free);
+ ipc_rcu_putref(&msq->q_perm, msg_rcu_free);
/* raced with RMID? */
if (!ipc_valid_object(&msq->q_perm)) {
err = -EIDRM;
diff --git a/ipc/sem.c b/ipc/sem.c
index fff8337..18241c1 100644
--- a/ipc/sem.c
+++ b/ipc/sem.c
@@ -260,8 +260,8 @@ static void merge_queues(struct sem_array *sma)

static void sem_rcu_free(struct rcu_head *head)
{
- struct ipc_rcu *p = container_of(head, struct ipc_rcu, rcu);
- struct sem_array *sma = ipc_rcu_to_struct(p);
+ struct kern_ipc_perm *p = container_of(head, struct kern_ipc_perm, rcu);
+ struct sem_array *sma = container_of(p, struct sem_array, sem_perm);

security_sem_free(sma);
ipc_rcu_free(head);
@@ -438,7 +438,7 @@ static inline struct sem_array *sem_obtain_object_check(struct ipc_namespace *ns
static inline void sem_lock_and_putref(struct sem_array *sma)
{
sem_lock(sma, NULL, -1);
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
}

static inline void sem_rmid(struct ipc_namespace *ns, struct sem_array *s)
@@ -469,8 +469,13 @@ static int newary(struct ipc_namespace *ns, struct ipc_params *params)
if (ns->used_sems + nsems > ns->sc_semmns)
return -ENOSPC;

+ if (offsetof(struct sem_array, sem_perm) != 0) {
+ pr_err("Invalid struct sem_perm, failing semget().\n");
+ return -ENOMEM;
+ }
+
size = sizeof(*sma) + nsems * sizeof(sma->sems[0]);
- sma = ipc_rcu_alloc(size);
+ sma = container_of(ipc_rcu_alloc(size), struct sem_array, sem_perm);
if (!sma)
return -ENOMEM;

@@ -482,7 +487,7 @@ static int newary(struct ipc_namespace *ns, struct ipc_params *params)
sma->sem_perm.security = NULL;
retval = security_sem_alloc(sma);
if (retval) {
- ipc_rcu_putref(sma, ipc_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, ipc_rcu_free);
return retval;
}

@@ -502,7 +507,7 @@ static int newary(struct ipc_namespace *ns, struct ipc_params *params)

id = ipc_addid(&sem_ids(ns), &sma->sem_perm, ns->sc_semmni);
if (id < 0) {
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
return id;
}
ns->used_sems += nsems;
@@ -1122,7 +1127,7 @@ static void freeary(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)

wake_up_q(&wake_q);
ns->used_sems -= sma->sem_nsems;
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
}

static unsigned long copy_semid_to_user(void __user *buf, struct semid64_ds *in, int version)
@@ -1362,7 +1367,7 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
goto out_unlock;
}
if (nsems > SEMMSL_FAST) {
- if (!ipc_rcu_getref(sma)) {
+ if (!ipc_rcu_getref(&sma->sem_perm)) {
err = -EIDRM;
goto out_unlock;
}
@@ -1370,7 +1375,7 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
rcu_read_unlock();
sem_io = ipc_alloc(sizeof(ushort)*nsems);
if (sem_io == NULL) {
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
return -ENOMEM;
}

@@ -1395,7 +1400,7 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
int i;
struct sem_undo *un;

- if (!ipc_rcu_getref(sma)) {
+ if (!ipc_rcu_getref(&sma->sem_perm)) {
err = -EIDRM;
goto out_rcu_wakeup;
}
@@ -1404,20 +1409,20 @@ static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
if (nsems > SEMMSL_FAST) {
sem_io = ipc_alloc(sizeof(ushort)*nsems);
if (sem_io == NULL) {
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
return -ENOMEM;
}
}

if (copy_from_user(sem_io, p, nsems*sizeof(ushort))) {
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
err = -EFAULT;
goto out_free;
}

for (i = 0; i < nsems; i++) {
if (sem_io[i] > SEMVMX) {
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
err = -ERANGE;
goto out_free;
}
@@ -1699,7 +1704,7 @@ static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
}

nsems = sma->sem_nsems;
- if (!ipc_rcu_getref(sma)) {
+ if (!ipc_rcu_getref(&sma->sem_perm)) {
rcu_read_unlock();
un = ERR_PTR(-EIDRM);
goto out;
@@ -1709,7 +1714,7 @@ static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
/* step 2: allocate new undo structure */
new = kzalloc(sizeof(struct sem_undo) + sizeof(short)*nsems, GFP_KERNEL);
if (!new) {
- ipc_rcu_putref(sma, sem_rcu_free);
+ ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
return ERR_PTR(-ENOMEM);
}

diff --git a/ipc/shm.c b/ipc/shm.c
index 34c4344..cec6df1 100644
--- a/ipc/shm.c
+++ b/ipc/shm.c
@@ -174,9 +174,10 @@ static inline void shm_lock_by_ptr(struct shmid_kernel *ipcp)

static void shm_rcu_free(struct rcu_head *head)
{
- struct ipc_rcu *p = container_of(head, struct ipc_rcu, rcu);
- struct shmid_kernel *shp = ipc_rcu_to_struct(p);
-
+ struct kern_ipc_perm *ptr = container_of(head, struct kern_ipc_perm,
+ rcu);
+ struct shmid_kernel *shp = container_of(ptr, struct shmid_kernel,
+ shm_perm);
security_shm_free(shp);
ipc_rcu_free(head);
}
@@ -241,7 +242,7 @@ static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
user_shm_unlock(i_size_read(file_inode(shm_file)),
shp->mlock_user);
fput(shm_file);
- ipc_rcu_putref(shp, shm_rcu_free);
+ ipc_rcu_putref(&shp->shm_perm, shm_rcu_free);
}

/*
@@ -542,7 +543,13 @@ static int newseg(struct ipc_namespace *ns, struct ipc_params *params)
ns->shm_tot + numpages > ns->shm_ctlall)
return -ENOSPC;

- shp = ipc_rcu_alloc(sizeof(*shp));
+ if (offsetof(struct shmid_kernel, shm_perm) != 0) {
+ pr_err("Invalid struct sem_perm, failing msgget().\n");
+ return -ENOMEM;
+ }
+
+ shp = container_of(ipc_rcu_alloc(sizeof(*shp)), struct shmid_kernel,
+ shm_perm);
if (!shp)
return -ENOMEM;

@@ -553,7 +560,7 @@ static int newseg(struct ipc_namespace *ns, struct ipc_params *params)
shp->shm_perm.security = NULL;
error = security_shm_alloc(shp);
if (error) {
- ipc_rcu_putref(shp, ipc_rcu_free);
+ ipc_rcu_putref(&shp->shm_perm, ipc_rcu_free);
return error;
}

@@ -624,7 +631,7 @@ static int newseg(struct ipc_namespace *ns, struct ipc_params *params)
user_shm_unlock(size, shp->mlock_user);
fput(file);
no_file:
- ipc_rcu_putref(shp, shm_rcu_free);
+ ipc_rcu_putref(&shp->shm_perm, shm_rcu_free);
return error;
}

diff --git a/ipc/util.c b/ipc/util.c
index caec7b1..9dcc08b 100644
--- a/ipc/util.c
+++ b/ipc/util.c
@@ -418,46 +418,43 @@ void ipc_free(void *ptr)
}

/**
- * ipc_rcu_alloc - allocate ipc and rcu space
+ * ipc_rcu_alloc - allocate ipc space
* @size: size desired
*
- * Allocate memory for the rcu header structure + the object.
- * Returns the pointer to the object or NULL upon failure.
+ * Allocate memory for an ipc object.
+ * The first member must be struct kern_ipc_perm.
*/
-void *ipc_rcu_alloc(int size)
+struct kern_ipc_perm *ipc_rcu_alloc(int size)
{
/*
* We prepend the allocation with the rcu struct
*/
- struct ipc_rcu *out = ipc_alloc(sizeof(struct ipc_rcu) + size);
+ struct kern_ipc_perm *out = ipc_alloc(size);
if (unlikely(!out))
return NULL;
atomic_set(&out->refcount, 1);
- return out + 1;
+ return out;
}

-int ipc_rcu_getref(void *ptr)
+int ipc_rcu_getref(struct kern_ipc_perm *ptr)
{
- struct ipc_rcu *p = ((struct ipc_rcu *)ptr) - 1;
-
- return atomic_inc_not_zero(&p->refcount);
+ return atomic_inc_not_zero(&ptr->refcount);
}

-void ipc_rcu_putref(void *ptr, void (*func)(struct rcu_head *head))
+void ipc_rcu_putref(struct kern_ipc_perm *ptr,
+ void (*func)(struct rcu_head *head))
{
- struct ipc_rcu *p = ((struct ipc_rcu *)ptr) - 1;
-
- if (!atomic_dec_and_test(&p->refcount))
+ if (!atomic_dec_and_test(&ptr->refcount))
return;

- call_rcu(&p->rcu, func);
+ call_rcu(&ptr->rcu, func);
}

-void ipc_rcu_free(struct rcu_head *head)
+void ipc_rcu_free(struct rcu_head *h)
{
- struct ipc_rcu *p = container_of(head, struct ipc_rcu, rcu);
+ struct kern_ipc_perm *ptr = container_of(h, struct kern_ipc_perm, rcu);

- kvfree(p);
+ kvfree(ptr);
}

/**
diff --git a/ipc/util.h b/ipc/util.h
index 60ddccc..09d0f91 100644
--- a/ipc/util.h
+++ b/ipc/util.h
@@ -47,13 +47,6 @@ static inline void msg_exit_ns(struct ipc_namespace *ns) { }
static inline void shm_exit_ns(struct ipc_namespace *ns) { }
#endif

-struct ipc_rcu {
- struct rcu_head rcu;
- atomic_t refcount;
-} ____cacheline_aligned_in_smp;
-
-#define ipc_rcu_to_struct(p) ((void *)(p+1))
-
/*
* Structure that holds the parameters needed by the ipc operations
* (see after)
@@ -125,11 +118,14 @@ void ipc_free(void *ptr);
* Objects are reference counted, they start with reference count 1.
* getref increases the refcount, the putref call that reduces the recount
* to 0 schedules the rcu destruction. Caller must guarantee locking.
+ *
+ * struct kern_ipc_perm must be the first member in the allocated structure.
*/
-void *ipc_rcu_alloc(int size);
-int ipc_rcu_getref(void *ptr);
-void ipc_rcu_putref(void *ptr, void (*func)(struct rcu_head *head));
-void ipc_rcu_free(struct rcu_head *head);
+struct kern_ipc_perm *ipc_rcu_alloc(int size);
+int ipc_rcu_getref(struct kern_ipc_perm *ptr);
+void ipc_rcu_putref(struct kern_ipc_perm *ptr,
+ void (*func)(struct rcu_head *head));
+void ipc_rcu_free(struct rcu_head *h);

struct kern_ipc_perm *ipc_lock(struct ipc_ids *, int);
struct kern_ipc_perm *ipc_obtain_object_idr(struct ipc_ids *ids, int id);
--
2.9.3