1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
|
diff --git a/fs/eventfd.c b/fs/eventfd.c
index 8aa0ea8c55e8..6cdb1b854341 100644
--- a/fs/eventfd.c
+++ b/fs/eventfd.c
@@ -24,18 +24,32 @@
#include <linux/seq_file.h>
#include <linux/idr.h>
+#define EPOLLALLMASK64 ((__force __u64)EPOLLALLMASK)
+
static DEFINE_IDA(eventfd_ida);
struct eventfd_ctx {
struct kref kref;
wait_queue_head_t wqh;
/*
- * Every time that a write(2) is performed on an eventfd, the
- * value of the __u64 being written is added to "count" and a
- * wakeup is performed on "wqh". A read(2) will return the "count"
- * value to userspace, and will reset "count" to zero. The kernel
- * side eventfd_signal() also, adds to the "count" counter and
- * issue a wakeup.
+ * If the EFD_VPOLL flag was NOT set at eventfd creation:
+ * Every time that a write(2) is performed on an eventfd, the
+ * value of the __u64 being written is added to "count" and a
+ * wakeup is performed on "wqh". A read(2) will return the "count"
+ * value to userspace, and will reset "count" to zero (or decrement
+ * "count" by 1 if the flag EFD_SEMAPHORE has been set). The kernel
+ * side eventfd_signal() also, adds to the "count" counter and
+ * issue a wakeup.
+ *
+ * If the EFD_VPOLL flag was set at eventfd creation:
+ * count is the set of pending EPOLL events.
+ * read(2) returns the current value of count.
+ * The argument of write(2) is an 8-byte integer:
+ * it is an or-composition of a control command (EFD_VPOLL_ADDEVENTS,
+ * EFD_VPOLL_DELEVENTS or EFD_VPOLL_MODEVENTS) and the bitmap of
+ * events to be added, deleted to the current set of pending events.
+ * (i.e. which bits of "count" must be set or reset).
+ * EFD_VPOLL_MODEVENTS redefines the set of pending events.
*/
__u64 count;
unsigned int flags;
@@ -295,6 +309,78 @@ static ssize_t eventfd_write(struct file *file, const char __user *buf, size_t c
return res;
}
+static __poll_t eventfd_vpoll_poll(struct file *file, poll_table *wait)
+{
+ struct eventfd_ctx *ctx = file->private_data;
+ __poll_t events = 0;
+ u64 count;
+
+ poll_wait(file, &ctx->wqh, wait);
+
+ count = READ_ONCE(ctx->count);
+
+ events = (((__force __poll_t)count) & EPOLLALLMASK);
+
+ return events;
+}
+
+static ssize_t eventfd_vpoll_read(struct file *file, char __user *buf,
+ size_t count, loff_t *ppos)
+{
+ struct eventfd_ctx *ctx = file->private_data;
+ ssize_t res;
+ __u64 ucnt = 0;
+
+ if (count < sizeof(ucnt))
+ return -EINVAL;
+ res = sizeof(ucnt);
+ ucnt = READ_ONCE(ctx->count);
+ if (put_user(ucnt, (__u64 __user *)buf))
+ return -EFAULT;
+
+ return res;
+}
+
+static ssize_t eventfd_vpoll_write(struct file *file, const char __user *buf,
+ size_t count, loff_t *ppos)
+{
+ struct eventfd_ctx *ctx = file->private_data;
+ ssize_t res;
+ __u64 ucnt;
+ __u32 events;
+
+ if (count < sizeof(ucnt))
+ return -EINVAL;
+ if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
+ return -EFAULT;
+ spin_lock_irq(&ctx->wqh.lock);
+
+ events = ucnt & EPOLLALLMASK64;
+ res = sizeof(ucnt);
+ switch (ucnt & ~EPOLLALLMASK64) {
+ case EFD_VPOLL_ADDEVENTS:
+ ctx->count |= events;
+ break;
+ case EFD_VPOLL_DELEVENTS:
+ ctx->count &= ~(events);
+ break;
+ case EFD_VPOLL_MODEVENTS:
+ ctx->count = (ctx->count & ~EPOLLALLMASK64) | events;
+ break;
+ default:
+ res = -EINVAL;
+ }
+
+ /* wake up waiting threads */
+ if (res >= 0 && waitqueue_active(&ctx->wqh))
+ wake_up_locked_poll(&ctx->wqh, ((__force __poll_t)ctx->count) & EPOLLALLMASK);
+
+ spin_unlock_irq(&ctx->wqh.lock);
+
+ return res;
+
+}
+
#ifdef CONFIG_PROC_FS
static void eventfd_show_fdinfo(struct seq_file *m, struct file *f)
{
@@ -319,6 +405,17 @@ static const struct file_operations eventfd_fops = {
.llseek = noop_llseek,
};
+static const struct file_operations eventfd_vpoll_fops = {
+#ifdef CONFIG_PROC_FS
+ .show_fdinfo = eventfd_show_fdinfo,
+#endif
+ .release = eventfd_release,
+ .poll = eventfd_vpoll_poll,
+ .read = eventfd_vpoll_read,
+ .write = eventfd_vpoll_write,
+ .llseek = noop_llseek,
+};
+
/**
* eventfd_fget - Acquire a reference of an eventfd file descriptor.
* @fd: [in] Eventfd file descriptor.
@@ -391,6 +488,7 @@ EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
static int do_eventfd(unsigned int count, int flags)
{
struct eventfd_ctx *ctx;
+ const struct file_operations *fops = &eventfd_fops;
int fd;
/* Check the EFD_* constants for consistency. */
@@ -410,7 +508,11 @@ static int do_eventfd(unsigned int count, int flags)
ctx->flags = flags;
ctx->id = ida_simple_get(&eventfd_ida, 0, 0, GFP_KERNEL);
- fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx,
+ if (flags & EFD_VPOLL) {
+ fops = &eventfd_vpoll_fops;
+ ctx->count &= EPOLLALLMASK64;
+ }
+ fd = anon_inode_getfd("[eventfd]", fops, ctx,
O_RDWR | (flags & EFD_SHARED_FCNTL_FLAGS));
if (fd < 0)
eventfd_free_ctx(ctx);
diff --git a/include/linux/eventfd.h b/include/linux/eventfd.h
index ffcc7724ca21..5b1e6ef56651 100644
--- a/include/linux/eventfd.h
+++ b/include/linux/eventfd.h
@@ -21,11 +21,16 @@
* shared O_* flags.
*/
#define EFD_SEMAPHORE (1 << 0)
+#define EFD_VPOLL (1 << 1)
#define EFD_CLOEXEC O_CLOEXEC
#define EFD_NONBLOCK O_NONBLOCK
#define EFD_SHARED_FCNTL_FLAGS (O_CLOEXEC | O_NONBLOCK)
-#define EFD_FLAGS_SET (EFD_SHARED_FCNTL_FLAGS | EFD_SEMAPHORE)
+#define EFD_FLAGS_SET (EFD_SHARED_FCNTL_FLAGS | EFD_SEMAPHORE | EFD_VPOLL)
+
+#define EFD_VPOLL_ADDEVENTS (1ULL << 32)
+#define EFD_VPOLL_DELEVENTS (2ULL << 32)
+#define EFD_VPOLL_MODEVENTS (3ULL << 32)
struct eventfd_ctx;
struct file;
diff --git a/include/uapi/linux/eventpoll.h b/include/uapi/linux/eventpoll.h
index 8a3432d0f0dc..814de6d869c7 100644
--- a/include/uapi/linux/eventpoll.h
+++ b/include/uapi/linux/eventpoll.h
@@ -41,6 +41,8 @@
#define EPOLLMSG (__force __poll_t)0x00000400
#define EPOLLRDHUP (__force __poll_t)0x00002000
+#define EPOLLALLMASK ((__force __poll_t)0x0fffffff)
+
/* Set exclusive wakeup mode for the target file descriptor */
#define EPOLLEXCLUSIVE ((__force __poll_t)(1U << 28))
|
