May 2024 - Wine-Devel - list.winehq.org

[PATCH v5 08/28] ntsync: Introduce NTSYNC_IOC_EVENT_RESET.

by Elizabeth Figura

This corresponds to the NT syscall NtResetEvent(). This sets the event to the unsignaled state, and returns its previous state. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 24 ++++++++++++++++++++++++ include/uapi/linux/ntsync.h | 1 + 2 files changed, 25 insertions(+) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index 97d9b6047fbe..b070ceccc3af 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -559,6 +559,28 @@ static int ntsync_event_set(struct ntsync_obj *event, void __user *argp) return 0; } +static int ntsync_event_reset(struct ntsync_obj *event, void __user *argp) +{ + struct ntsync_device *dev = event->dev; + __u32 prev_state; + bool all; + + if (event->type != NTSYNC_TYPE_EVENT) + return -EINVAL; + + all = ntsync_lock_obj(dev, event); + + prev_state = event->u.event.signaled; + event->u.event.signaled = false; + + ntsync_unlock_obj(dev, event, all); + + if (put_user(prev_state, (__u32 __user *)argp)) + return -EFAULT; + + return 0; +} + static int ntsync_obj_release(struct inode *inode, struct file *file) { struct ntsync_obj *obj = file->private_data; @@ -584,6 +606,8 @@ static long ntsync_obj_ioctl(struct file *file, unsigned int cmd, return ntsync_mutex_kill(obj, argp); case NTSYNC_IOC_EVENT_SET: return ntsync_event_set(obj, argp); + case NTSYNC_IOC_EVENT_RESET: + return ntsync_event_reset(obj, argp); default: return -ENOIOCTLCMD; } diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index 36d903521bbe..7fdf79729b20 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -52,5 +52,6 @@ struct ntsync_wait_args { #define NTSYNC_IOC_MUTEX_UNLOCK _IOWR('N', 0x85, struct ntsync_mutex_args) #define NTSYNC_IOC_MUTEX_KILL _IOW ('N', 0x86, __u32) #define NTSYNC_IOC_EVENT_SET _IOR ('N', 0x88, __u32) +#define NTSYNC_IOC_EVENT_RESET _IOR ('N', 0x89, __u32) #endif -- 2.43.0

May 19, 2024

[PATCH v5 07/28] ntsync: Introduce NTSYNC_IOC_EVENT_SET.

by Elizabeth Figura

This corresponds to the NT syscall NtSetEvent(). This sets the event to the signaled state, and returns its previous state. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 27 +++++++++++++++++++++++++++ include/uapi/linux/ntsync.h | 1 + 2 files changed, 28 insertions(+) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index 2bce03187c17..97d9b6047fbe 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -534,6 +534,31 @@ static int ntsync_mutex_kill(struct ntsync_obj *mutex, void __user *argp) return ret; } +static int ntsync_event_set(struct ntsync_obj *event, void __user *argp) +{ + struct ntsync_device *dev = event->dev; + __u32 prev_state; + bool all; + + if (event->type != NTSYNC_TYPE_EVENT) + return -EINVAL; + + all = ntsync_lock_obj(dev, event); + + prev_state = event->u.event.signaled; + event->u.event.signaled = true; + if (all) + try_wake_all_obj(dev, event); + try_wake_any_event(event); + + ntsync_unlock_obj(dev, event, all); + + if (put_user(prev_state, (__u32 __user *)argp)) + return -EFAULT; + + return 0; +} + static int ntsync_obj_release(struct inode *inode, struct file *file) { struct ntsync_obj *obj = file->private_data; @@ -557,6 +582,8 @@ static long ntsync_obj_ioctl(struct file *file, unsigned int cmd, return ntsync_mutex_unlock(obj, argp); case NTSYNC_IOC_MUTEX_KILL: return ntsync_mutex_kill(obj, argp); + case NTSYNC_IOC_EVENT_SET: + return ntsync_event_set(obj, argp); default: return -ENOIOCTLCMD; } diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index 4c0c4271c7de..36d903521bbe 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -51,5 +51,6 @@ struct ntsync_wait_args { #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) #define NTSYNC_IOC_MUTEX_UNLOCK _IOWR('N', 0x85, struct ntsync_mutex_args) #define NTSYNC_IOC_MUTEX_KILL _IOW ('N', 0x86, __u32) +#define NTSYNC_IOC_EVENT_SET _IOR ('N', 0x88, __u32) #endif -- 2.43.0

May 19, 2024

[PATCH v5 06/28] ntsync: Introduce NTSYNC_IOC_CREATE_EVENT.

by Elizabeth Figura

This correspond to the NT syscall NtCreateEvent(). An NT event holds a single bit of state denoting whether it is signaled or unsignaled. There are two types of events: manual-reset and automatic-reset. When an automatic-reset event is acquired via a wait function, its state is reset to unsignaled. Manual-reset events are not affected by wait functions. Whether the event is manual-reset, and its initial state, are specified at creation time. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 62 +++++++++++++++++++++++++++++++++++++ include/uapi/linux/ntsync.h | 7 +++++ 2 files changed, 69 insertions(+) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index 5aaf9dad76b6..2bce03187c17 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -26,6 +26,7 @@ enum ntsync_type { NTSYNC_TYPE_SEM, NTSYNC_TYPE_MUTEX, + NTSYNC_TYPE_EVENT, }; /* @@ -61,6 +62,10 @@ struct ntsync_obj { pid_t owner; bool ownerdead; } mutex; + struct { + bool manual; + bool signaled; + } event; } u; /* @@ -233,6 +238,8 @@ static bool is_signaled(struct ntsync_obj *obj, __u32 owner) if (obj->u.mutex.owner && obj->u.mutex.owner != owner) return false; return obj->u.mutex.count < UINT_MAX; + case NTSYNC_TYPE_EVENT: + return obj->u.event.signaled; } WARN(1, "bad object type %#x\n", obj->type); @@ -283,6 +290,10 @@ static void try_wake_all(struct ntsync_device *dev, struct ntsync_q *q, obj->u.mutex.count++; obj->u.mutex.owner = q->owner; break; + case NTSYNC_TYPE_EVENT: + if (!obj->u.event.manual) + obj->u.event.signaled = false; + break; } } wake_up_process(q->task); @@ -353,6 +364,28 @@ static void try_wake_any_mutex(struct ntsync_obj *mutex) } } +static void try_wake_any_event(struct ntsync_obj *event) +{ + struct ntsync_q_entry *entry; + + ntsync_assert_held(event); + lockdep_assert(event->type == NTSYNC_TYPE_EVENT); + + list_for_each_entry(entry, &event->any_waiters, node) { + struct ntsync_q *q = entry->q; + int signaled = -1; + + if (!event->u.event.signaled) + break; + + if (atomic_try_cmpxchg(&q->signaled, &signaled, entry->index)) { + if (!event->u.event.manual) + event->u.event.signaled = false; + wake_up_process(q->task); + } + } +} + /* * Actually change the semaphore state, returning -EOVERFLOW if it is made * invalid. @@ -629,6 +662,30 @@ static int ntsync_create_mutex(struct ntsync_device *dev, void __user *argp) return put_user(fd, &user_args->mutex); } +static int ntsync_create_event(struct ntsync_device *dev, void __user *argp) +{ + struct ntsync_event_args __user *user_args = argp; + struct ntsync_event_args args; + struct ntsync_obj *event; + int fd; + + if (copy_from_user(&args, argp, sizeof(args))) + return -EFAULT; + + event = ntsync_alloc_obj(dev, NTSYNC_TYPE_EVENT); + if (!event) + return -ENOMEM; + event->u.event.manual = args.manual; + event->u.event.signaled = args.signaled; + fd = ntsync_obj_get_fd(event); + if (fd < 0) { + kfree(event); + return fd; + } + + return put_user(fd, &user_args->event); +} + static struct ntsync_obj *get_obj(struct ntsync_device *dev, int fd) { struct file *file = fget(fd); @@ -759,6 +816,9 @@ static void try_wake_any_obj(struct ntsync_obj *obj) case NTSYNC_TYPE_MUTEX: try_wake_any_mutex(obj); break; + case NTSYNC_TYPE_EVENT: + try_wake_any_event(obj); + break; } } @@ -948,6 +1008,8 @@ static long ntsync_char_ioctl(struct file *file, unsigned int cmd, void __user *argp = (void __user *)parm; switch (cmd) { + case NTSYNC_IOC_CREATE_EVENT: + return ntsync_create_event(dev, argp); case NTSYNC_IOC_CREATE_MUTEX: return ntsync_create_mutex(dev, argp); case NTSYNC_IOC_CREATE_SEM: diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index d7996180c1d2..4c0c4271c7de 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -22,6 +22,12 @@ struct ntsync_mutex_args { __u32 count; }; +struct ntsync_event_args { + __u32 event; + __u32 manual; + __u32 signaled; +}; + #define NTSYNC_WAIT_REALTIME 0x1 struct ntsync_wait_args { @@ -40,6 +46,7 @@ struct ntsync_wait_args { #define NTSYNC_IOC_WAIT_ANY _IOWR('N', 0x82, struct ntsync_wait_args) #define NTSYNC_IOC_WAIT_ALL _IOWR('N', 0x83, struct ntsync_wait_args) #define NTSYNC_IOC_CREATE_MUTEX _IOWR('N', 0x84, struct ntsync_sem_args) +#define NTSYNC_IOC_CREATE_EVENT _IOWR('N', 0x87, struct ntsync_event_args) #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) #define NTSYNC_IOC_MUTEX_UNLOCK _IOWR('N', 0x85, struct ntsync_mutex_args) -- 2.43.0

May 19, 2024

[PATCH v5 05/28] ntsync: Introduce NTSYNC_IOC_MUTEX_KILL.

by Elizabeth Figura

This does not correspond to any NT syscall. Rather, when a thread dies, it should be called by the NT emulator for each mutex, with the TID of the dying thread. NT mutexes are robust (in the pthread sense). When an NT thread dies, any mutexes it owned are immediately released. Acquisition of those mutexes by other threads will return a special value indicating that the mutex was abandoned, like EOWNERDEAD returned from pthread_mutex_lock(), and EOWNERDEAD is indeed used here for that purpose. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 61 +++++++++++++++++++++++++++++++++++-- include/uapi/linux/ntsync.h | 1 + 2 files changed, 60 insertions(+), 2 deletions(-) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index f00af9b15164..5aaf9dad76b6 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -59,6 +59,7 @@ struct ntsync_obj { struct { __u32 count; pid_t owner; + bool ownerdead; } mutex; } u; @@ -107,6 +108,7 @@ struct ntsync_q { atomic_t signaled; bool all; + bool ownerdead; __u32 count; struct ntsync_q_entry entries[]; }; @@ -275,6 +277,9 @@ static void try_wake_all(struct ntsync_device *dev, struct ntsync_q *q, obj->u.sem.count--; break; case NTSYNC_TYPE_MUTEX: + if (obj->u.mutex.ownerdead) + q->ownerdead = true; + obj->u.mutex.ownerdead = false; obj->u.mutex.count++; obj->u.mutex.owner = q->owner; break; @@ -338,6 +343,9 @@ static void try_wake_any_mutex(struct ntsync_obj *mutex) continue; if (atomic_try_cmpxchg(&q->signaled, &signaled, entry->index)) { + if (mutex->u.mutex.ownerdead) + q->ownerdead = true; + mutex->u.mutex.ownerdead = false; mutex->u.mutex.count++; mutex->u.mutex.owner = q->owner; wake_up_process(q->task); @@ -447,6 +455,52 @@ static int ntsync_mutex_unlock(struct ntsync_obj *mutex, void __user *argp) return ret; } +/* + * Actually change the mutex state to mark its owner as dead, + * returning -EPERM if not the owner. + */ +static int kill_mutex_state(struct ntsync_obj *mutex, __u32 owner) +{ + ntsync_assert_held(mutex); + + if (mutex->u.mutex.owner != owner) + return -EPERM; + + mutex->u.mutex.ownerdead = true; + mutex->u.mutex.owner = 0; + mutex->u.mutex.count = 0; + return 0; +} + +static int ntsync_mutex_kill(struct ntsync_obj *mutex, void __user *argp) +{ + struct ntsync_device *dev = mutex->dev; + __u32 owner; + bool all; + int ret; + + if (get_user(owner, (__u32 __user *)argp)) + return -EFAULT; + if (!owner) + return -EINVAL; + + if (mutex->type != NTSYNC_TYPE_MUTEX) + return -EINVAL; + + all = ntsync_lock_obj(dev, mutex); + + ret = kill_mutex_state(mutex, owner); + if (!ret) { + if (all) + try_wake_all_obj(dev, mutex); + try_wake_any_mutex(mutex); + } + + ntsync_unlock_obj(dev, mutex, all); + + return ret; +} + static int ntsync_obj_release(struct inode *inode, struct file *file) { struct ntsync_obj *obj = file->private_data; @@ -468,6 +522,8 @@ static long ntsync_obj_ioctl(struct file *file, unsigned int cmd, return ntsync_sem_post(obj, argp); case NTSYNC_IOC_MUTEX_UNLOCK: return ntsync_mutex_unlock(obj, argp); + case NTSYNC_IOC_MUTEX_KILL: + return ntsync_mutex_kill(obj, argp); default: return -ENOIOCTLCMD; } @@ -659,6 +715,7 @@ static int setup_wait(struct ntsync_device *dev, q->owner = args->owner; atomic_set(&q->signaled, -1); q->all = all; + q->ownerdead = false; q->count = count; for (i = 0; i < count; i++) { @@ -767,7 +824,7 @@ static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) struct ntsync_wait_args __user *user_args = argp; /* even if we caught a signal, we need to communicate success */ - ret = 0; + ret = q->ownerdead ? -EOWNERDEAD : 0; if (put_user(signaled, &user_args->index)) ret = -EFAULT; @@ -848,7 +905,7 @@ static int ntsync_wait_all(struct ntsync_device *dev, void __user *argp) struct ntsync_wait_args __user *user_args = argp; /* even if we caught a signal, we need to communicate success */ - ret = 0; + ret = q->ownerdead ? -EOWNERDEAD : 0; if (put_user(signaled, &user_args->index)) ret = -EFAULT; diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index a633db34f284..d7996180c1d2 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -43,5 +43,6 @@ struct ntsync_wait_args { #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) #define NTSYNC_IOC_MUTEX_UNLOCK _IOWR('N', 0x85, struct ntsync_mutex_args) +#define NTSYNC_IOC_MUTEX_KILL _IOW ('N', 0x86, __u32) #endif -- 2.43.0

May 19, 2024

[PATCH v5 04/28] ntsync: Introduce NTSYNC_IOC_MUTEX_UNLOCK.

by Elizabeth Figura

This corresponds to the NT syscall NtReleaseMutant(). This syscall decrements the mutex's recursion count by one, and returns the previous value. If the mutex is not owned by the current task, the function instead fails and returns -EPERM. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 53 +++++++++++++++++++++++++++++++++++++ include/uapi/linux/ntsync.h | 1 + 2 files changed, 54 insertions(+) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index cfe802c79d7d..f00af9b15164 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -396,6 +396,57 @@ static int ntsync_sem_post(struct ntsync_obj *sem, void __user *argp) return ret; } +/* + * Actually change the mutex state, returning -EPERM if not the owner. + */ +static int unlock_mutex_state(struct ntsync_obj *mutex, + const struct ntsync_mutex_args *args) +{ + ntsync_assert_held(mutex); + + if (mutex->u.mutex.owner != args->owner) + return -EPERM; + + if (!--mutex->u.mutex.count) + mutex->u.mutex.owner = 0; + return 0; +} + +static int ntsync_mutex_unlock(struct ntsync_obj *mutex, void __user *argp) +{ + struct ntsync_mutex_args __user *user_args = argp; + struct ntsync_device *dev = mutex->dev; + struct ntsync_mutex_args args; + __u32 prev_count; + bool all; + int ret; + + if (copy_from_user(&args, argp, sizeof(args))) + return -EFAULT; + if (!args.owner) + return -EINVAL; + + if (mutex->type != NTSYNC_TYPE_MUTEX) + return -EINVAL; + + all = ntsync_lock_obj(dev, mutex); + + prev_count = mutex->u.mutex.count; + ret = unlock_mutex_state(mutex, &args); + if (!ret) { + if (all) + try_wake_all_obj(dev, mutex); + try_wake_any_mutex(mutex); + } + + ntsync_unlock_obj(dev, mutex, all); + + if (!ret && put_user(prev_count, &user_args->count)) + ret = -EFAULT; + + return ret; +} + static int ntsync_obj_release(struct inode *inode, struct file *file) { struct ntsync_obj *obj = file->private_data; @@ -415,6 +466,8 @@ static long ntsync_obj_ioctl(struct file *file, unsigned int cmd, switch (cmd) { case NTSYNC_IOC_SEM_POST: return ntsync_sem_post(obj, argp); + case NTSYNC_IOC_MUTEX_UNLOCK: + return ntsync_mutex_unlock(obj, argp); default: return -ENOIOCTLCMD; } diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index d5e5a2fbcb4d..a633db34f284 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -42,5 +42,6 @@ struct ntsync_wait_args { #define NTSYNC_IOC_CREATE_MUTEX _IOWR('N', 0x84, struct ntsync_sem_args) #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) +#define NTSYNC_IOC_MUTEX_UNLOCK _IOWR('N', 0x85, struct ntsync_mutex_args) #endif -- 2.43.0

May 19, 2024

[PATCH v5 03/28] ntsync: Introduce NTSYNC_IOC_CREATE_MUTEX.

by Elizabeth Figura

This corresponds to the NT syscall NtCreateMutant(). An NT mutex is recursive, with a 32-bit recursion counter. When acquired via NtWaitForMultipleObjects(), the recursion counter is incremented by one. The OS records the thread which acquired it. The OS records the thread which acquired it. However, in order to keep this driver self-contained, the owning thread ID is managed by user-space, and passed as a parameter to all relevant ioctls. The initial owner and recursion count, if any, are specified when the mutex is created. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 77 +++++++++++++++++++++++++++++++++++-- include/uapi/linux/ntsync.h | 10 ++++- 2 files changed, 83 insertions(+), 4 deletions(-) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index a2f2dfadc3ee..cfe802c79d7d 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -25,6 +25,7 @@ enum ntsync_type { NTSYNC_TYPE_SEM, + NTSYNC_TYPE_MUTEX, }; /* @@ -55,6 +56,10 @@ struct ntsync_obj { __u32 count; __u32 max; } sem; + struct { + __u32 count; + pid_t owner; + } mutex; } u; /* @@ -92,6 +97,7 @@ struct ntsync_q_entry { struct ntsync_q { struct task_struct *task; + __u32 owner; /* * Protected via atomic_try_cmpxchg(). Only the thread that wins the @@ -214,13 +220,17 @@ static void ntsync_unlock_obj(struct ntsync_device *dev, struct ntsync_obj *obj, ((lockdep_is_held(&(obj)->dev->wait_all_lock) != LOCK_STATE_NOT_HELD) && \ (obj)->dev_locked)) -static bool is_signaled(struct ntsync_obj *obj) +static bool is_signaled(struct ntsync_obj *obj, __u32 owner) { ntsync_assert_held(obj); switch (obj->type) { case NTSYNC_TYPE_SEM: return !!obj->u.sem.count; + case NTSYNC_TYPE_MUTEX: + if (obj->u.mutex.owner && obj->u.mutex.owner != owner) + return false; + return obj->u.mutex.count < UINT_MAX; } WARN(1, "bad object type %#x\n", obj->type); @@ -250,7 +260,7 @@ static void try_wake_all(struct ntsync_device *dev, struct ntsync_q *q, } for (i = 0; i < count; i++) { - if (!is_signaled(q->entries[i].obj)) { + if (!is_signaled(q->entries[i].obj, q->owner)) { can_wake = false; break; } @@ -264,6 +274,10 @@ static void try_wake_all(struct ntsync_device *dev, struct ntsync_q *q, case NTSYNC_TYPE_SEM: obj->u.sem.count--; break; + case NTSYNC_TYPE_MUTEX: + obj->u.mutex.count++; + obj->u.mutex.owner = q->owner; + break; } } wake_up_process(q->task); @@ -307,6 +321,30 @@ static void try_wake_any_sem(struct ntsync_obj *sem) } } +static void try_wake_any_mutex(struct ntsync_obj *mutex) +{ + struct ntsync_q_entry *entry; + + ntsync_assert_held(mutex); + lockdep_assert(mutex->type == NTSYNC_TYPE_MUTEX); + + list_for_each_entry(entry, &mutex->any_waiters, node) { + struct ntsync_q *q = entry->q; + int signaled = -1; + + if (mutex->u.mutex.count == UINT_MAX) + break; + if (mutex->u.mutex.owner && mutex->u.mutex.owner != q->owner) + continue; + + if (atomic_try_cmpxchg(&q->signaled, &signaled, entry->index)) { + mutex->u.mutex.count++; + mutex->u.mutex.owner = q->owner; + wake_up_process(q->task); + } + } +} + /* * Actually change the semaphore state, returning -EOVERFLOW if it is made * invalid. @@ -455,6 +493,33 @@ static int ntsync_create_sem(struct ntsync_device *dev, void __user *argp) return put_user(fd, &user_args->sem); } +static int ntsync_create_mutex(struct ntsync_device *dev, void __user *argp) +{ + struct ntsync_mutex_args __user *user_args = argp; + struct ntsync_mutex_args args; + struct ntsync_obj *mutex; + int fd; + + if (copy_from_user(&args, argp, sizeof(args))) + return -EFAULT; + + if (!args.owner != !args.count) + return -EINVAL; + + mutex = ntsync_alloc_obj(dev, NTSYNC_TYPE_MUTEX); + if (!mutex) + return -ENOMEM; + mutex->u.mutex.count = args.count; + mutex->u.mutex.owner = args.owner; + fd = ntsync_obj_get_fd(mutex); + if (fd < 0) { + kfree(mutex); + return fd; + } + + return put_user(fd, &user_args->mutex); +} + static struct ntsync_obj *get_obj(struct ntsync_device *dev, int fd) { struct file *file = fget(fd); @@ -524,7 +589,7 @@ static int setup_wait(struct ntsync_device *dev, struct ntsync_q *q; __u32 i, j; - if (args->pad[0] || args->pad[1] || args->pad[2] || (args->flags & ~NTSYNC_WAIT_REALTIME)) + if (args->pad[0] || args->pad[1] || (args->flags & ~NTSYNC_WAIT_REALTIME)) return -EINVAL; if (args->count > NTSYNC_MAX_WAIT_COUNT) @@ -538,6 +603,7 @@ static int setup_wait(struct ntsync_device *dev, if (!q) return -ENOMEM; q->task = current; + q->owner = args->owner; atomic_set(&q->signaled, -1); q->all = all; q->count = count; @@ -580,6 +646,9 @@ static void try_wake_any_obj(struct ntsync_obj *obj) case NTSYNC_TYPE_SEM: try_wake_any_sem(obj); break; + case NTSYNC_TYPE_MUTEX: + try_wake_any_mutex(obj); + break; } } @@ -769,6 +838,8 @@ static long ntsync_char_ioctl(struct file *file, unsigned int cmd, void __user *argp = (void __user *)parm; switch (cmd) { + case NTSYNC_IOC_CREATE_MUTEX: + return ntsync_create_mutex(dev, argp); case NTSYNC_IOC_CREATE_SEM: return ntsync_create_sem(dev, argp); case NTSYNC_IOC_WAIT_ALL: diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index addf187b1573..d5e5a2fbcb4d 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -16,6 +16,12 @@ struct ntsync_sem_args { __u32 max; }; +struct ntsync_mutex_args { + __u32 mutex; + __u32 owner; + __u32 count; +}; + #define NTSYNC_WAIT_REALTIME 0x1 struct ntsync_wait_args { @@ -24,7 +30,8 @@ struct ntsync_wait_args { __u32 count; __u32 index; __u32 flags; - __u32 pad[3]; + __u32 owner; + __u32 pad[2]; }; #define NTSYNC_MAX_WAIT_COUNT 64 @@ -32,6 +39,7 @@ struct ntsync_wait_args { #define NTSYNC_IOC_CREATE_SEM _IOWR('N', 0x80, struct ntsync_sem_args) #define NTSYNC_IOC_WAIT_ANY _IOWR('N', 0x82, struct ntsync_wait_args) #define NTSYNC_IOC_WAIT_ALL _IOWR('N', 0x83, struct ntsync_wait_args) +#define NTSYNC_IOC_CREATE_MUTEX _IOWR('N', 0x84, struct ntsync_sem_args) #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) -- 2.43.0

May 19, 2024

[PATCH v5 02/28] ntsync: Introduce NTSYNC_IOC_WAIT_ALL.

by Elizabeth Figura

This is similar to NTSYNC_IOC_WAIT_ANY, but waits until all of the objects are simultaneously signaled, and then acquires all of them as a single atomic operation. Because acquisition of multiple objects is atomic, some complex locking is required. We cannot simply spin-lock multiple objects simultaneously, as that may disable preëmption for a problematically long time. Instead, modifying any object which may be involved in a wait-all operation takes a device-wide sleeping mutex, "wait_all_lock", instead of the normal object spinlock. Because wait-for-all is a rare operation, in order to optimize wait-for-any, this lock is only taken when necessary. "all_hint" is used to mark objects which are involved in a wait-for-all operation, and if an object is not, only its spinlock is taken. The locking scheme used here was written by Peter Zijlstra. Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 334 ++++++++++++++++++++++++++++++++++-- include/uapi/linux/ntsync.h | 1 + 2 files changed, 322 insertions(+), 13 deletions(-) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index d5864891caf0..a2f2dfadc3ee 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -13,6 +13,7 @@ #include <linux/ktime.h> #include <linux/miscdevice.h> #include <linux/module.h> +#include <linux/mutex.h> #include <linux/overflow.h> #include <linux/sched.h> #include <linux/sched/signal.h> @@ -41,6 +42,7 @@ enum ntsync_type { struct ntsync_obj { spinlock_t lock; + int dev_locked; enum ntsync_type type; @@ -55,7 +57,30 @@ struct ntsync_obj { } sem; } u; + /* + * any_waiters is protected by the object lock, but all_waiters is + * protected by the device wait_all_lock. + */ struct list_head any_waiters; + struct list_head all_waiters; + + /* + * Hint describing how many tasks are queued on this object in a + * wait-all operation. + * + * Any time we do a wake, we may need to wake "all" waiters as well as + * "any" waiters. In order to atomically wake "all" waiters, we must + * lock all of the objects, and that means grabbing the wait_all_lock + * below (and, due to lock ordering rules, before locking this object). + * However, wait-all is a rare operation, and grabbing the wait-all + * lock for every wake would create unnecessary contention. + * Therefore we first check whether all_hint is zero, and, if it is, + * we skip trying to wake "all" waiters. + * + * Since wait requests must originate from user-space threads, we're + * limited here by PID_MAX_LIMIT, so there's no risk of overflow. + */ + atomic_t all_hint; }; struct ntsync_q_entry { @@ -75,19 +100,198 @@ struct ntsync_q { */ atomic_t signaled; + bool all; __u32 count; struct ntsync_q_entry entries[]; }; struct ntsync_device { + /* + * Wait-all operations must atomically grab all objects, and be totally + * ordered with respect to each other and wait-any operations. + * If one thread is trying to acquire several objects, another thread + * cannot touch the object at the same time. + * + * This device-wide lock is used to serialize wait-for-all + * operations, and operations on an object that is involved in a + * wait-for-all. + */ + struct mutex wait_all_lock; + struct file *file; }; +/* + * Single objects are locked using obj->lock. + * + * Multiple objects are 'locked' while holding dev->wait_all_lock. + * In this case however, individual objects are not locked by holding + * obj->lock, but by setting obj->dev_locked. + * + * This means that in order to lock a single object, the sequence is slightly + * more complicated than usual. Specifically it needs to check obj->dev_locked + * after acquiring obj->lock, if set, it needs to drop the lock and acquire + * dev->wait_all_lock in order to serialize against the multi-object operation. + */ + +static void dev_lock_obj(struct ntsync_device *dev, struct ntsync_obj *obj) +{ + lockdep_assert_held(&dev->wait_all_lock); + lockdep_assert(obj->dev == dev); + spin_lock(&obj->lock); + /* + * By setting obj->dev_locked inside obj->lock, it is ensured that + * anyone holding obj->lock must see the value. + */ + obj->dev_locked = 1; + spin_unlock(&obj->lock); +} + +static void dev_unlock_obj(struct ntsync_device *dev, struct ntsync_obj *obj) +{ + lockdep_assert_held(&dev->wait_all_lock); + lockdep_assert(obj->dev == dev); + spin_lock(&obj->lock); + obj->dev_locked = 0; + spin_unlock(&obj->lock); +} + +static void obj_lock(struct ntsync_obj *obj) +{ + struct ntsync_device *dev = obj->dev; + + for (;;) { + spin_lock(&obj->lock); + if (likely(!obj->dev_locked)) + break; + + spin_unlock(&obj->lock); + mutex_lock(&dev->wait_all_lock); + spin_lock(&obj->lock); + /* + * obj->dev_locked should be set and released under the same + * wait_all_lock section, since we now own this lock, it should + * be clear. + */ + lockdep_assert(!obj->dev_locked); + spin_unlock(&obj->lock); + mutex_unlock(&dev->wait_all_lock); + } +} + +static void obj_unlock(struct ntsync_obj *obj) +{ + spin_unlock(&obj->lock); +} + +static bool ntsync_lock_obj(struct ntsync_device *dev, struct ntsync_obj *obj) +{ + bool all; + + obj_lock(obj); + all = atomic_read(&obj->all_hint); + if (unlikely(all)) { + obj_unlock(obj); + mutex_lock(&dev->wait_all_lock); + dev_lock_obj(dev, obj); + } + + return all; +} + +static void ntsync_unlock_obj(struct ntsync_device *dev, struct ntsync_obj *obj, bool all) +{ + if (all) { + dev_unlock_obj(dev, obj); + mutex_unlock(&dev->wait_all_lock); + } else { + obj_unlock(obj); + } +} + +#define ntsync_assert_held(obj) \ + lockdep_assert((lockdep_is_held(&(obj)->lock) != LOCK_STATE_NOT_HELD) || \ + ((lockdep_is_held(&(obj)->dev->wait_all_lock) != LOCK_STATE_NOT_HELD) && \ + (obj)->dev_locked)) + +static bool is_signaled(struct ntsync_obj *obj) +{ + ntsync_assert_held(obj); + + switch (obj->type) { + case NTSYNC_TYPE_SEM: + return !!obj->u.sem.count; + } + + WARN(1, "bad object type %#x\n", obj->type); + return false; +} + +/* + * "locked_obj" is an optional pointer to an object which is already locked and + * should not be locked again. This is necessary so that changing an object's + * state and waking it can be a single atomic operation. + */ +static void try_wake_all(struct ntsync_device *dev, struct ntsync_q *q, + struct ntsync_obj *locked_obj) +{ + __u32 count = q->count; + bool can_wake = true; + int signaled = -1; + __u32 i; + + lockdep_assert_held(&dev->wait_all_lock); + if (locked_obj) + lockdep_assert(locked_obj->dev_locked); + + for (i = 0; i < count; i++) { + if (q->entries[i].obj != locked_obj) + dev_lock_obj(dev, q->entries[i].obj); + } + + for (i = 0; i < count; i++) { + if (!is_signaled(q->entries[i].obj)) { + can_wake = false; + break; + } + } + + if (can_wake && atomic_try_cmpxchg(&q->signaled, &signaled, 0)) { + for (i = 0; i < count; i++) { + struct ntsync_obj *obj = q->entries[i].obj; + + switch (obj->type) { + case NTSYNC_TYPE_SEM: + obj->u.sem.count--; + break; + } + } + wake_up_process(q->task); + } + + for (i = 0; i < count; i++) { + if (q->entries[i].obj != locked_obj) + dev_unlock_obj(dev, q->entries[i].obj); + } +} + +static void try_wake_all_obj(struct ntsync_device *dev, struct ntsync_obj *obj) +{ + struct ntsync_q_entry *entry; + + lockdep_assert_held(&dev->wait_all_lock); + lockdep_assert(obj->dev_locked); + + list_for_each_entry(entry, &obj->all_waiters, node) + try_wake_all(dev, entry->q, obj); +} + static void try_wake_any_sem(struct ntsync_obj *sem) { struct ntsync_q_entry *entry; - lockdep_assert_held(&sem->lock); + ntsync_assert_held(sem); + lockdep_assert(sem->type == NTSYNC_TYPE_SEM); list_for_each_entry(entry, &sem->any_waiters, node) { struct ntsync_q *q = entry->q; @@ -111,7 +315,7 @@ static int post_sem_state(struct ntsync_obj *sem, __u32 count) { __u32 sum; - lockdep_assert_held(&sem->lock); + ntsync_assert_held(sem); if (check_add_overflow(sem->u.sem.count, count, &sum) || sum > sem->u.sem.max) @@ -123,9 +327,11 @@ static int post_sem_state(struct ntsync_obj *sem, __u32 count) static int ntsync_sem_post(struct ntsync_obj *sem, void __user *argp) { + struct ntsync_device *dev = sem->dev; __u32 __user *user_args = argp; __u32 prev_count; __u32 args; + bool all; int ret; if (copy_from_user(&args, argp, sizeof(args))) @@ -134,14 +340,17 @@ static int ntsync_sem_post(struct ntsync_obj *sem, void __user *argp) if (sem->type != NTSYNC_TYPE_SEM) return -EINVAL; - spin_lock(&sem->lock); + all = ntsync_lock_obj(dev, sem); prev_count = sem->u.sem.count; ret = post_sem_state(sem, args); - if (!ret) + if (!ret) { + if (all) + try_wake_all_obj(dev, sem); try_wake_any_sem(sem); + } - spin_unlock(&sem->lock); + ntsync_unlock_obj(dev, sem, all); if (!ret && put_user(prev_count, user_args)) ret = -EFAULT; @@ -194,6 +403,8 @@ static struct ntsync_obj *ntsync_alloc_obj(struct ntsync_device *dev, get_file(dev->file); spin_lock_init(&obj->lock); INIT_LIST_HEAD(&obj->any_waiters); + INIT_LIST_HEAD(&obj->all_waiters); + atomic_set(&obj->all_hint, 0); return obj; } @@ -305,7 +516,7 @@ static int ntsync_schedule(const struct ntsync_q *q, const struct ntsync_wait_ar * Allocate and initialize the ntsync_q structure, but do not queue us yet. */ static int setup_wait(struct ntsync_device *dev, - const struct ntsync_wait_args *args, + const struct ntsync_wait_args *args, bool all, struct ntsync_q **ret_q) { const __u32 count = args->count; @@ -328,6 +539,7 @@ static int setup_wait(struct ntsync_device *dev, return -ENOMEM; q->task = current; atomic_set(&q->signaled, -1); + q->all = all; q->count = count; for (i = 0; i < count; i++) { @@ -337,6 +549,16 @@ static int setup_wait(struct ntsync_device *dev, if (!obj) goto err; + if (all) { + /* Check that the objects are all distinct. */ + for (j = 0; j < i; j++) { + if (obj == q->entries[j].obj) { + put_obj(obj); + goto err; + } + } + } + entry->obj = obj; entry->q = q; entry->index = i; @@ -366,13 +588,14 @@ static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) struct ntsync_wait_args args; struct ntsync_q *q; int signaled; + bool all; __u32 i; int ret; if (copy_from_user(&args, argp, sizeof(args))) return -EFAULT; - ret = setup_wait(dev, &args, &q); + ret = setup_wait(dev, &args, false, &q); if (ret < 0) return ret; @@ -382,9 +605,9 @@ static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) struct ntsync_q_entry *entry = &q->entries[i]; struct ntsync_obj *obj = entry->obj; - spin_lock(&obj->lock); + all = ntsync_lock_obj(dev, obj); list_add_tail(&entry->node, &obj->any_waiters); - spin_unlock(&obj->lock); + ntsync_unlock_obj(dev, obj, all); } /* check if we are already signaled */ @@ -395,9 +618,9 @@ static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) if (atomic_read(&q->signaled) != -1) break; - spin_lock(&obj->lock); + all = ntsync_lock_obj(dev, obj); try_wake_any_obj(obj); - spin_unlock(&obj->lock); + ntsync_unlock_obj(dev, obj, all); } /* sleep */ @@ -410,9 +633,9 @@ static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) struct ntsync_q_entry *entry = &q->entries[i]; struct ntsync_obj *obj = entry->obj; - spin_lock(&obj->lock); + all = ntsync_lock_obj(dev, obj); list_del(&entry->node); - spin_unlock(&obj->lock); + ntsync_unlock_obj(dev, obj, all); put_obj(obj); } @@ -434,6 +657,87 @@ static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) return ret; } +static int ntsync_wait_all(struct ntsync_device *dev, void __user *argp) +{ + struct ntsync_wait_args args; + struct ntsync_q *q; + int signaled; + __u32 i; + int ret; + + if (copy_from_user(&args, argp, sizeof(args))) + return -EFAULT; + + ret = setup_wait(dev, &args, true, &q); + if (ret < 0) + return ret; + + /* queue ourselves */ + + mutex_lock(&dev->wait_all_lock); + + for (i = 0; i < args.count; i++) { + struct ntsync_q_entry *entry = &q->entries[i]; + struct ntsync_obj *obj = entry->obj; + + atomic_inc(&obj->all_hint); + + /* + * obj->all_waiters is protected by dev->wait_all_lock rather + * than obj->lock, so there is no need to acquire obj->lock + * here. + */ + list_add_tail(&entry->node, &obj->all_waiters); + } + + /* check if we are already signaled */ + + try_wake_all(dev, q, NULL); + + mutex_unlock(&dev->wait_all_lock); + + /* sleep */ + + ret = ntsync_schedule(q, &args); + + /* and finally, unqueue */ + + mutex_lock(&dev->wait_all_lock); + + for (i = 0; i < args.count; i++) { + struct ntsync_q_entry *entry = &q->entries[i]; + struct ntsync_obj *obj = entry->obj; + + /* + * obj->all_waiters is protected by dev->wait_all_lock rather + * than obj->lock, so there is no need to acquire it here. + */ + list_del(&entry->node); + + atomic_dec(&obj->all_hint); + + put_obj(obj); + } + + mutex_unlock(&dev->wait_all_lock); + + signaled = atomic_read(&q->signaled); + if (signaled != -1) { + struct ntsync_wait_args __user *user_args = argp; + + /* even if we caught a signal, we need to communicate success */ + ret = 0; + + if (put_user(signaled, &user_args->index)) + ret = -EFAULT; + } else if (!ret) { + ret = -ETIMEDOUT; + } + + kfree(q); + return ret; +} + static int ntsync_char_open(struct inode *inode, struct file *file) { struct ntsync_device *dev; @@ -442,6 +746,8 @@ static int ntsync_char_open(struct inode *inode, struct file *file) if (!dev) return -ENOMEM; + mutex_init(&dev->wait_all_lock); + file->private_data = dev; dev->file = file; return nonseekable_open(inode, file); @@ -465,6 +771,8 @@ static long ntsync_char_ioctl(struct file *file, unsigned int cmd, switch (cmd) { case NTSYNC_IOC_CREATE_SEM: return ntsync_create_sem(dev, argp); + case NTSYNC_IOC_WAIT_ALL: + return ntsync_wait_all(dev, argp); case NTSYNC_IOC_WAIT_ANY: return ntsync_wait_any(dev, argp); default: diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index edc12c7a10dc..addf187b1573 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -31,6 +31,7 @@ struct ntsync_wait_args { #define NTSYNC_IOC_CREATE_SEM _IOWR('N', 0x80, struct ntsync_sem_args) #define NTSYNC_IOC_WAIT_ANY _IOWR('N', 0x82, struct ntsync_wait_args) +#define NTSYNC_IOC_WAIT_ALL _IOWR('N', 0x83, struct ntsync_wait_args) #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) -- 2.43.0

May 19, 2024

[PATCH v5 01/28] ntsync: Introduce NTSYNC_IOC_WAIT_ANY.

by Elizabeth Figura

This corresponds to part of the functionality of the NT syscall NtWaitForMultipleObjects(). Specifically, it implements the behaviour where the third argument (wait_any) is TRUE, and it does not handle alertable waits. Those features have been split out into separate patches to ease review. This patch therefore implements the wait/wake infrastructure which comprises the core of ntsync's functionality. NTSYNC_IOC_WAIT_ANY is a vectored wait function similar to poll(). Unlike poll(), it "consumes" objects when they are signaled. For semaphores, this means decreasing one from the internal counter. At most one object can be consumed by this function. This wait/wake model is fundamentally different from that used anywhere else in the kernel, and for that reason ntsync does not use any existing infrastructure, such as futexes, kernel mutexes or semaphores, or wait_event(). Up to 64 objects can be waited on at once. As soon as one is signaled, the object with the lowest index is consumed, and that index is returned via the "index" field. A timeout is supported. The timeout is passed as a u64 nanosecond value, which represents absolute time measured against either the MONOTONIC or REALTIME clock (controlled by the flags argument). If U64_MAX is passed, the ioctl waits indefinitely. This ioctl validates that all objects belong to the relevant device. This is not necessary for any technical reason related to NTSYNC_IOC_WAIT_ANY, but will be necessary for NTSYNC_IOC_WAIT_ALL introduced in the following patch. Some padding fields are added for alignment and for fields which will be added in future patches (split out to ease review). Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- drivers/misc/ntsync.c | 245 ++++++++++++++++++++++++++++++++++++ include/uapi/linux/ntsync.h | 14 +++ 2 files changed, 259 insertions(+) diff --git a/drivers/misc/ntsync.c b/drivers/misc/ntsync.c index 3c2f743c58b0..d5864891caf0 100644 --- a/drivers/misc/ntsync.c +++ b/drivers/misc/ntsync.c @@ -6,11 +6,16 @@ */ #include <linux/anon_inodes.h> +#include <linux/atomic.h> #include <linux/file.h> #include <linux/fs.h> +#include <linux/hrtimer.h> +#include <linux/ktime.h> #include <linux/miscdevice.h> #include <linux/module.h> #include <linux/overflow.h> +#include <linux/sched.h> +#include <linux/sched/signal.h> #include <linux/slab.h> #include <linux/spinlock.h> #include <uapi/linux/ntsync.h> @@ -30,6 +35,8 @@ enum ntsync_type { * * Both rely on struct file for reference counting. Individual * ntsync_obj objects take a reference to the device when created. + * Wait operations take a reference to each object being waited on for + * the duration of the wait. */ struct ntsync_obj { @@ -47,12 +54,55 @@ struct ntsync_obj { __u32 max; } sem; } u; + + struct list_head any_waiters; +}; + +struct ntsync_q_entry { + struct list_head node; + struct ntsync_q *q; + struct ntsync_obj *obj; + __u32 index; +}; + +struct ntsync_q { + struct task_struct *task; + + /* + * Protected via atomic_try_cmpxchg(). Only the thread that wins the + * compare-and-swap may actually change object states and wake this + * task. + */ + atomic_t signaled; + + __u32 count; + struct ntsync_q_entry entries[]; }; struct ntsync_device { struct file *file; }; +static void try_wake_any_sem(struct ntsync_obj *sem) +{ + struct ntsync_q_entry *entry; + + lockdep_assert_held(&sem->lock); + + list_for_each_entry(entry, &sem->any_waiters, node) { + struct ntsync_q *q = entry->q; + int signaled = -1; + + if (!sem->u.sem.count) + break; + + if (atomic_try_cmpxchg(&q->signaled, &signaled, entry->index)) { + sem->u.sem.count--; + wake_up_process(q->task); + } + } +} + /* * Actually change the semaphore state, returning -EOVERFLOW if it is made * invalid. @@ -88,6 +138,8 @@ static int ntsync_sem_post(struct ntsync_obj *sem, void __user *argp) prev_count = sem->u.sem.count; ret = post_sem_state(sem, args); + if (!ret) + try_wake_any_sem(sem); spin_unlock(&sem->lock); @@ -141,6 +193,7 @@ static struct ntsync_obj *ntsync_alloc_obj(struct ntsync_device *dev, obj->dev = dev; get_file(dev->file); spin_lock_init(&obj->lock); + INIT_LIST_HEAD(&obj->any_waiters); return obj; } @@ -191,6 +244,196 @@ static int ntsync_create_sem(struct ntsync_device *dev, void __user *argp) return put_user(fd, &user_args->sem); } +static struct ntsync_obj *get_obj(struct ntsync_device *dev, int fd) +{ + struct file *file = fget(fd); + struct ntsync_obj *obj; + + if (!file) + return NULL; + + if (file->f_op != &ntsync_obj_fops) { + fput(file); + return NULL; + } + + obj = file->private_data; + if (obj->dev != dev) { + fput(file); + return NULL; + } + + return obj; +} + +static void put_obj(struct ntsync_obj *obj) +{ + fput(obj->file); +} + +static int ntsync_schedule(const struct ntsync_q *q, const struct ntsync_wait_args *args) +{ + ktime_t timeout = ns_to_ktime(args->timeout); + clockid_t clock = CLOCK_MONOTONIC; + ktime_t *timeout_ptr; + int ret = 0; + + timeout_ptr = (args->timeout == U64_MAX ? NULL : &timeout); + + if (args->flags & NTSYNC_WAIT_REALTIME) + clock = CLOCK_REALTIME; + + do { + if (signal_pending(current)) { + ret = -ERESTARTSYS; + break; + } + + set_current_state(TASK_INTERRUPTIBLE); + if (atomic_read(&q->signaled) != -1) { + ret = 0; + break; + } + ret = schedule_hrtimeout_range_clock(timeout_ptr, 0, HRTIMER_MODE_ABS, clock); + } while (ret < 0); + __set_current_state(TASK_RUNNING); + + return ret; +} + +/* + * Allocate and initialize the ntsync_q structure, but do not queue us yet. + */ +static int setup_wait(struct ntsync_device *dev, + const struct ntsync_wait_args *args, + struct ntsync_q **ret_q) +{ + const __u32 count = args->count; + int fds[NTSYNC_MAX_WAIT_COUNT]; + struct ntsync_q *q; + __u32 i, j; + + if (args->pad[0] || args->pad[1] || args->pad[2] || (args->flags & ~NTSYNC_WAIT_REALTIME)) + return -EINVAL; + + if (args->count > NTSYNC_MAX_WAIT_COUNT) + return -EINVAL; + + if (copy_from_user(fds, u64_to_user_ptr(args->objs), + array_size(count, sizeof(*fds)))) + return -EFAULT; + + q = kmalloc(struct_size(q, entries, count), GFP_KERNEL); + if (!q) + return -ENOMEM; + q->task = current; + atomic_set(&q->signaled, -1); + q->count = count; + + for (i = 0; i < count; i++) { + struct ntsync_q_entry *entry = &q->entries[i]; + struct ntsync_obj *obj = get_obj(dev, fds[i]); + + if (!obj) + goto err; + + entry->obj = obj; + entry->q = q; + entry->index = i; + } + + *ret_q = q; + return 0; + +err: + for (j = 0; j < i; j++) + put_obj(q->entries[j].obj); + kfree(q); + return -EINVAL; +} + +static void try_wake_any_obj(struct ntsync_obj *obj) +{ + switch (obj->type) { + case NTSYNC_TYPE_SEM: + try_wake_any_sem(obj); + break; + } +} + +static int ntsync_wait_any(struct ntsync_device *dev, void __user *argp) +{ + struct ntsync_wait_args args; + struct ntsync_q *q; + int signaled; + __u32 i; + int ret; + + if (copy_from_user(&args, argp, sizeof(args))) + return -EFAULT; + + ret = setup_wait(dev, &args, &q); + if (ret < 0) + return ret; + + /* queue ourselves */ + + for (i = 0; i < args.count; i++) { + struct ntsync_q_entry *entry = &q->entries[i]; + struct ntsync_obj *obj = entry->obj; + + spin_lock(&obj->lock); + list_add_tail(&entry->node, &obj->any_waiters); + spin_unlock(&obj->lock); + } + + /* check if we are already signaled */ + + for (i = 0; i < args.count; i++) { + struct ntsync_obj *obj = q->entries[i].obj; + + if (atomic_read(&q->signaled) != -1) + break; + + spin_lock(&obj->lock); + try_wake_any_obj(obj); + spin_unlock(&obj->lock); + } + + /* sleep */ + + ret = ntsync_schedule(q, &args); + + /* and finally, unqueue */ + + for (i = 0; i < args.count; i++) { + struct ntsync_q_entry *entry = &q->entries[i]; + struct ntsync_obj *obj = entry->obj; + + spin_lock(&obj->lock); + list_del(&entry->node); + spin_unlock(&obj->lock); + + put_obj(obj); + } + + signaled = atomic_read(&q->signaled); + if (signaled != -1) { + struct ntsync_wait_args __user *user_args = argp; + + /* even if we caught a signal, we need to communicate success */ + ret = 0; + + if (put_user(signaled, &user_args->index)) + ret = -EFAULT; + } else if (!ret) { + ret = -ETIMEDOUT; + } + + kfree(q); + return ret; +} + static int ntsync_char_open(struct inode *inode, struct file *file) { struct ntsync_device *dev; @@ -222,6 +465,8 @@ static long ntsync_char_ioctl(struct file *file, unsigned int cmd, switch (cmd) { case NTSYNC_IOC_CREATE_SEM: return ntsync_create_sem(dev, argp); + case NTSYNC_IOC_WAIT_ANY: + return ntsync_wait_any(dev, argp); default: return -ENOIOCTLCMD; } diff --git a/include/uapi/linux/ntsync.h b/include/uapi/linux/ntsync.h index dcfa38fdc93c..edc12c7a10dc 100644 --- a/include/uapi/linux/ntsync.h +++ b/include/uapi/linux/ntsync.h @@ -16,7 +16,21 @@ struct ntsync_sem_args { __u32 max; }; +#define NTSYNC_WAIT_REALTIME 0x1 + +struct ntsync_wait_args { + __u64 timeout; + __u64 objs; + __u32 count; + __u32 index; + __u32 flags; + __u32 pad[3]; +}; + +#define NTSYNC_MAX_WAIT_COUNT 64 + #define NTSYNC_IOC_CREATE_SEM _IOWR('N', 0x80, struct ntsync_sem_args) +#define NTSYNC_IOC_WAIT_ANY _IOWR('N', 0x82, struct ntsync_wait_args) #define NTSYNC_IOC_SEM_POST _IOWR('N', 0x81, __u32) -- 2.43.0

May 19, 2024

[PATCH v5 00/28] NT synchronization primitive driver

by Elizabeth Figura

This patch series implements a new char misc driver, /dev/ntsync, which is used to implement Windows NT synchronization primitives. NT synchronization primitives are unique in that the wait functions both are vectored, operate on multiple types of object with different behaviour (mutex, semaphore, event), and affect the state of the objects they wait on. This model is not compatible with existing kernel synchronization objects or interfaces, and therefore the ntsync driver implements its own wait queues and locking. This patch series is rebased against the "char-misc-next" branch of gregkh/char-misc.git. == Background == The Wine project emulates the Windows API in user space. One particular part of that API, namely the NT synchronization primitives, have historically been implemented via RPC to a dedicated "kernel" process. However, more recent applications use these APIs more strenuously, and the overhead of RPC has become a bottleneck. The NT synchronization APIs are too complex to implement on top of existing primitives without sacrificing correctness. Certain operations, such as NtPulseEvent() or the "wait-for-all" mode of NtWaitForMultipleObjects(), require direct control over the underlying wait queue, and implementing a wait queue sufficiently robust for Wine in user space is not possible. This proposed driver, therefore, implements the problematic interfaces directly in the Linux kernel. This driver was presented at Linux Plumbers Conference 2023. For those further interested in the history of synchronization in Wine and past attempts to solve this problem in user space, a recording of the presentation can be viewed here: https://www.youtube.com/watch?v=NjU4nyWyhU8 == Performance == The performance measurements described below are copied from earlier versions of the patch set. While some of the code has changed, I do not currently anticipate that it has changed drastically enough to affect those measurements. The gain in performance varies wildly depending on the application in question and the user's hardware. For some games NT synchronization is not a bottleneck and no change can be observed, but for others frame rate improvements of 50 to 150 percent are not atypical. The following table lists frame rate measurements from a variety of games on a variety of hardware, taken by users Dmitry Skvortsov, FuzzyQuils, OnMars, and myself: Game Upstream ntsync improvement =========================================================================== Anger Foot 69 99 43% Call of Juarez 99.8 224.1 125% Dirt 3 110.6 860.7 678% Forza Horizon 5 108 160 48% Lara Croft: Temple of Osiris 141 326 131% Metro 2033 164.4 199.2 21% Resident Evil 2 26 77 196% The Crew 26 51 96% Tiny Tina's Wonderlands 130 360 177% Total War Saga: Troy 109 146 34% =========================================================================== == Patches == The intended semantics of the patches are broadly intended to match those of the corresponding Windows functions. For those not already familiar with the Windows functions (or their undocumented behaviour), patch 27/28 provides a detailed specification, and individual patches also include a brief description of the API they are implementing. The patches making use of this driver in Wine can be retrieved or browsed here: https://repo.or.cz/wine/zf.git/shortlog/refs/heads/ntsync5 == Previous versions == Changes from v4: * Rework wait-all locking code to avoid taking more than one spinlock at a time, and also to fix a race where the wait-all lock would not be not correctly taken. The new locking mechanism involves taking a simple spinlock for normal "any" waits, and taking a device-wide mutex for "all" waits or when locking any object that is involved in an "all" wait. The mechanism was written by Peter Zijlstra. * Try to reword or clarify various parts of the documentation (patch 27), per Peter Zijlstra. * I did not rename NTSYNC_IOC_SEM_POST to RELEASE (like NT) although this was suggested by Peter Zijlstra, mostly because it's not clear to me that renaming an already committed ioctl would be fine. The API committed isn't actually usable yet, though, so if altering it would be fine on those grounds, I can revise this series to rename the function accordingly. * Similarly, I did not change the create ioctls to return the fd directly, although this was suggested and would be a bit simpler and cleaner, because NTSYNC_IOC_CREATE_SEM already exists upstream and returns the fd through a struct. I can make this change in the next revision if that'd be preferable. I also still would appreciate a clarification on the advice in [1]. [1] https://docs.kernel.org/driver-api/ioctl.html#return-code * Link to v4: https://lore.kernel.org/lkml/20240416010837.333694-1-zfigura(a)codeweavers.… * Link to v3: https://lore.kernel.org/lkml/20240329000621.148791-1-zfigura(a)codeweavers.… * Link to v2: https://lore.kernel.org/lkml/20240219223833.95710-1-zfigura(a)codeweavers.c… * Link to v1: https://lore.kernel.org/lkml/20240214233645.9273-1-zfigura(a)codeweavers.co… * Link to RFC v2: https://lore.kernel.org/lkml/20240131021356.10322-1-zfigura(a)codeweavers.c… * Link to RFC v1: https://lore.kernel.org/lkml/20240124004028.16826-1-zfigura(a)codeweavers.c… Elizabeth Figura (28): ntsync: Introduce NTSYNC_IOC_WAIT_ANY. ntsync: Introduce NTSYNC_IOC_WAIT_ALL. ntsync: Introduce NTSYNC_IOC_CREATE_MUTEX. ntsync: Introduce NTSYNC_IOC_MUTEX_UNLOCK. ntsync: Introduce NTSYNC_IOC_MUTEX_KILL. ntsync: Introduce NTSYNC_IOC_CREATE_EVENT. ntsync: Introduce NTSYNC_IOC_EVENT_SET. ntsync: Introduce NTSYNC_IOC_EVENT_RESET. ntsync: Introduce NTSYNC_IOC_EVENT_PULSE. ntsync: Introduce NTSYNC_IOC_SEM_READ. ntsync: Introduce NTSYNC_IOC_MUTEX_READ. ntsync: Introduce NTSYNC_IOC_EVENT_READ. ntsync: Introduce alertable waits. selftests: ntsync: Add some tests for semaphore state. selftests: ntsync: Add some tests for mutex state. selftests: ntsync: Add some tests for NTSYNC_IOC_WAIT_ANY. selftests: ntsync: Add some tests for NTSYNC_IOC_WAIT_ALL. selftests: ntsync: Add some tests for wakeup signaling with WINESYNC_IOC_WAIT_ANY. selftests: ntsync: Add some tests for wakeup signaling with WINESYNC_IOC_WAIT_ALL. selftests: ntsync: Add some tests for manual-reset event state. selftests: ntsync: Add some tests for auto-reset event state. selftests: ntsync: Add some tests for wakeup signaling with events. selftests: ntsync: Add tests for alertable waits. selftests: ntsync: Add some tests for wakeup signaling via alerts. selftests: ntsync: Add a stress test for contended waits. maintainers: Add an entry for ntsync. docs: ntsync: Add documentation for the ntsync uAPI. ntsync: No longer depend on BROKEN. Documentation/userspace-api/index.rst | 1 + Documentation/userspace-api/ntsync.rst | 398 +++++ MAINTAINERS | 9 + drivers/misc/Kconfig | 1 - drivers/misc/ntsync.c | 989 +++++++++++- include/uapi/linux/ntsync.h | 39 + tools/testing/selftests/Makefile | 1 + .../selftests/drivers/ntsync/.gitignore | 1 + .../testing/selftests/drivers/ntsync/Makefile | 7 + tools/testing/selftests/drivers/ntsync/config | 1 + .../testing/selftests/drivers/ntsync/ntsync.c | 1407 +++++++++++++++++ 11 files changed, 2850 insertions(+), 4 deletions(-) create mode 100644 Documentation/userspace-api/ntsync.rst create mode 100644 tools/testing/selftests/drivers/ntsync/.gitignore create mode 100644 tools/testing/selftests/drivers/ntsync/Makefile create mode 100644 tools/testing/selftests/drivers/ntsync/config create mode 100644 tools/testing/selftests/drivers/ntsync/ntsync.c base-commit: f5b335dc025cfee90957efa90dc72fada0d5abb4 -- 2.43.0

May 19, 2024

Wine staging 9.9 release

by Alistair Leslie-Hughes

Binary packages for various distributions will be available from: https://www.winehq.org/download Summary since last release * Rebased to current wine 9.9 (427 patches are applied to wine vanilla) Upstreamed (Either directly from staging or fixed with a similar patch). * None Removed (No longer required). * None Added: * None Updated: * vkd3d-latest * ntdll-Junction_Points Where can you help * Run Steam/Battle.net/GOG/UPlay/Epic * Test your favorite game. * Test your favorite applications. * Improve staging patches and get them accepted upstream. * Suggest patches to be included in staging. As always, if you find a bug, please report it via https://bugs.winehq.org Best Regards Alistair.

May 18, 2024