Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Merge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fix from Ingo Molnar:
"Update the <linux/swait.h> documentation to discourage their use"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/swait: Document it clearly that the swait facilities are special and shouldn't be used
+16
-11
+16
-11
include/linux/swait.h
+16
-11
include/linux/swait.h
···
9
9
/*
10
10
* Simple wait queues
11
11
*
12
-
* While these are very similar to the other/complex wait queues (wait.h) the
13
-
* most important difference is that the simple waitqueue allows for
14
-
* deterministic behaviour -- IOW it has strictly bounded IRQ and lock hold
15
-
* times.
12
+
* While these are very similar to regular wait queues (wait.h) the most
13
+
* important difference is that the simple waitqueue allows for deterministic
14
+
* behaviour -- IOW it has strictly bounded IRQ and lock hold times.
16
15
*
17
-
* In order to make this so, we had to drop a fair number of features of the
18
-
* other waitqueue code; notably:
16
+
* Mainly, this is accomplished by two things. Firstly not allowing swake_up_all
17
+
* from IRQ disabled, and dropping the lock upon every wakeup, giving a higher
18
+
* priority task a chance to run.
19
+
*
20
+
* Secondly, we had to drop a fair number of features of the other waitqueue
21
+
* code; notably:
19
22
*
20
23
* - mixing INTERRUPTIBLE and UNINTERRUPTIBLE sleeps on the same waitqueue;
21
24
* all wakeups are TASK_NORMAL in order to avoid O(n) lookups for the right
···
27
24
* - the exclusive mode; because this requires preserving the list order
28
25
* and this is hard.
29
26
*
30
-
* - custom wake functions; because you cannot give any guarantees about
31
-
* random code.
27
+
* - custom wake callback functions; because you cannot give any guarantees
28
+
* about random code. This also allows swait to be used in RT, such that
29
+
* raw spinlock can be used for the swait queue head.
32
30
*
33
-
* As a side effect of this; the data structures are slimmer.
34
-
*
35
-
* One would recommend using this wait queue where possible.
31
+
* As a side effect of these; the data structures are slimmer albeit more ad-hoc.
32
+
* For all the above, note that simple wait queues should _only_ be used under
33
+
* very specific realtime constraints -- it is best to stick with the regular
34
+
* wait queues in most cases.
36
35
*/
37
36
38
37
struct task_struct;