1 On Fri, 2 Jan 1998, Doug Ledford wrote:
3 > I'm working on making the aic7xxx driver more SMP friendly (as well as
4 > importing the latest FreeBSD sequencer code to have 7895 support) and wanted
5 > to get some info from you. The goal here is to make the various routines
6 > SMP safe as well as UP safe during interrupts and other manipulating
7 > routines. So far, I've added a spin_lock variable to things like my queue
8 > structs. Now, from what I recall, there are some spin lock functions I can
9 > use to lock these spin locks from other use as opposed to a (nasty)
10 > save_flags(); cli(); stuff; restore_flags(); construct. Where do I find
11 > these routines and go about making use of them? Do they only lock on a
12 > per-processor basis or can they also lock say an interrupt routine from
13 > mucking with a queue if the queue routine was manipulating it when the
14 > interrupt occurred, or should I still use a cli(); based construct on that
17 See <asm/spinlock.h>. The basic version is:
19 spinlock_t xxx_lock = SPIN_LOCK_UNLOCKED;
24 spin_lock_irqsave(&xxx_lock, flags);
25 ... critical section here ..
26 spin_unlock_irqrestore(&xxx_lock, flags);
28 and the above is always safe. It will disable interrupts _locally_, but the
29 spinlock itself will guarantee the global lock, so it will guarantee that
30 there is only one thread-of-control within the region(s) protected by that
33 Note that it works well even under UP - the above sequence under UP
34 essentially is just the same as doing a
38 save_flags(flags); cli();
39 ... critical section ...
42 so the code does _not_ need to worry about UP vs SMP issues: the spinlocks
43 work correctly under both (and spinlocks are actually more efficient on
44 architectures that allow doing the "save_flags + cli" in one go because I
45 don't export that interface normally).
47 NOTE NOTE NOTE! The reason the spinlock is so much faster than a global
48 interrupt lock under SMP is exactly because it disables interrupts only on
49 the local CPU. The spin-lock is safe only when you _also_ use the lock
50 itself to do locking across CPU's, which implies that EVERYTHING that
51 touches a shared variable has to agree about the spinlock they want to
54 The above is usually pretty simple (you usually need and want only one
55 spinlock for most things - using more than one spinlock can make things a
56 lot more complex and even slower and is usually worth it only for
57 sequences that you _know_ need to be split up: avoid it at all cost if you
58 aren't sure). HOWEVER, it _does_ mean that if you have some code that does
61 .. critical section ..
64 and another sequence that does
66 spin_lock_irqsave(flags);
67 .. critical section ..
68 spin_unlock_irqrestore(flags);
70 then they are NOT mutually exclusive, and the critical regions can happen
71 at the same time on two different CPU's. That's fine per se, but the
72 critical regions had better be critical for different things (ie they
73 can't stomp on each other).
75 The above is a problem mainly if you end up mixing code - for example the
76 routines in ll_rw_block() tend to use cli/sti to protect the atomicity of
77 their actions, and if a driver uses spinlocks instead then you should
78 think about issues like the above..
80 This is really the only really hard part about spinlocks: once you start
81 using spinlocks they tend to expand to areas you might not have noticed
82 before, because you have to make sure the spinlocks correctly protect the
83 shared data structures _everywhere_ they are used. The spinlocks are most
84 easily added to places that are completely independent of other code (ie
85 internal driver data structures that nobody else ever touches, for
90 Lesson 2: reader-writer spinlocks.
92 If your data accesses have a very natural pattern where you usually tend
93 to mostly read from the shared variables, the reader-writer locks
94 (rw_lock) versions of the spinlocks are often nicer. They allow multiple
95 readers to be in the same critical region at once, but if somebody wants
96 to change the variables it has to get an exclusive write lock. The
97 routines look the same as above:
99 rwlock_t xxx_lock = RW_LOCK_UNLOCKED;
104 read_lock_irqsave(&xxx_lock, flags);
105 .. critical section that only reads the info ...
106 read_unlock_irqrestore(&xxx_lock, flags);
108 write_lock_irqsave(&xxx_lock, flags);
109 .. read and write exclusive access to the info ...
110 write_unlock_irqrestore(&xxx_lock, flags);
112 The above kind of lock is useful for complex data structures like linked
113 lists etc, especially when you know that most of the work is to just
114 traverse the list searching for entries without changing the list itself,
115 for example. Then you can use the read lock for that kind of list
116 traversal, which allows many concurrent readers. Anything that _changes_
117 the list will have to get the write lock.
119 Note: you cannot "upgrade" a read-lock to a write-lock, so if you at _any_
120 time need to do any changes (even if you don't do it every time), you have
121 to get the write-lock at the very beginning. I could fairly easily add a
122 primitive to create a "upgradeable" read-lock, but it hasn't been an issue
123 yet. Tell me if you'd want one.
127 Lesson 3: spinlocks revisited.
129 The single spin-lock primitives above are by no means the only ones. They
130 are the most safe ones, and the ones that work under all circumstances,
131 but partly _because_ they are safe they are also fairly slow. They are
132 much faster than a generic global cli/sti pair, but slower than they'd
133 need to be, because they do have to disable interrupts (which is just a
134 single instruction on a x86, but it's an expensive one - and on other
135 architectures it can be worse).
137 If you have a case where you have to protect a data structure across
138 several CPU's and you want to use spinlocks you can potentially use
139 cheaper versions of the spinlocks. IFF you know that the spinlocks are
140 never used in interrupt handlers, you can use the non-irq versions:
146 (and the equivalent read-write versions too, of course). The spinlock will
147 guarantee the same kind of exclusive access, and it will be much faster.
148 This is useful if you know that the data in question is only ever
149 manipulated from a "process context", ie no interrupts involved.
151 The reasons you mustn't use these versions if you have interrupts that
152 play with the spinlock is that you can get deadlocks:
156 <- interrupt comes in:
159 where an interrupt tries to lock an already locked variable. This is ok if
160 the other interrupt happens on another CPU, but it is _not_ ok if the
161 interrupt happens on the same CPU that already holds the lock, because the
162 lock will obviously never be released (because the interrupt is waiting
163 for the lock, and the lock-holder is interrupted by the interrupt and will
164 not continue until the interrupt has been processed).
166 (This is also the reason why the irq-versions of the spinlocks only need
167 to disable the _local_ interrupts - it's ok to use spinlocks in interrupts
168 on other CPU's, because an interrupt on another CPU doesn't interrupt the
169 CPU that holds the lock, so the lock-holder can continue and eventually
172 Note that you can be clever with read-write locks and interrupts. For
173 example, if you know that the interrupt only ever gets a read-lock, then
174 you can use a non-irq version of read locks everywhere - because they
175 don't block on each other (and thus there is no dead-lock wrt interrupts.
176 But when you do the write-lock, you have to use the irq-safe version.
178 For an example of being clever with rw-locks, see the "waitqueue_lock"
179 handling in kernel/sched.c - nothing ever _changes_ a wait-queue from
180 within an interrupt, they only read the queue in order to know whom to
181 wake up. So read-locks are safe (which is good: they are very common
182 indeed), while write-locks need to protect themselves against interrupts.