Techwizz/kernel-fxtec-pro1x
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Documentation: rw_lock lessons learned

Browse source 
In recent months, two different network projects erroneously
strayed down the rw_lock path.  Update the Documentation
based upon comments by Eric Dumazet and Paul E. McKenney in
those threads.

Further updates await somebody else with more expertise.

Changes:
  - Merged with extensive content by Stephen Hemminger.
  - Fix one of the comments by Linus Torvalds.

Signed-off-by: William.Allen.Simpson@gmail.com
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This commit is contained in:
William Allen Simpson 2009-12-13 15:12:46 -05:00 committed by Linus Torvalds
parent f40542532e
commit fb0bbb92d4
1 changed files with 84 additions and 100 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

180
Documentation/spinlocks.txt
View file

@ -1,87 +1,20 @@
SPIN_LOCK_UNLOCKED and RW_LOCK_UNLOCKED defeat lockdep state tracking and
are hence deprecated.
Lesson 1: Spin locks
Please use DEFINE_SPINLOCK()/DEFINE_RWLOCK() or
__SPIN_LOCK_UNLOCKED()/__RW_LOCK_UNLOCKED() as appropriate for static
initialization.
Most of the time, you can simply turn:
static spinlock_t xxx_lock = SPIN_LOCK_UNLOCKED;
into:
The most basic primitive for locking is spinlock.
static DEFINE_SPINLOCK(xxx_lock);
Static structure member variables go from:
struct foo bar {
.lock = SPIN_LOCK_UNLOCKED;
};
to:
struct foo bar {
.lock = __SPIN_LOCK_UNLOCKED(bar.lock);
};
Declaration of static rw_locks undergo a similar transformation.
Dynamic initialization, when necessary, may be performed as
demonstrated below.
spinlock_t xxx_lock;
rwlock_t xxx_rw_lock;
static int __init xxx_init(void)
{
spin_lock_init(&xxx_lock);
rwlock_init(&xxx_rw_lock);
...
}
module_init(xxx_init);
The following discussion is still valid, however, with the dynamic
initialization of spinlocks or with DEFINE_SPINLOCK, etc., used
instead of SPIN_LOCK_UNLOCKED.
-----------------------
On Fri, 2 Jan 1998, Doug Ledford wrote:
>
> I'm working on making the aic7xxx driver more SMP friendly (as well as
> importing the latest FreeBSD sequencer code to have 7895 support) and wanted
> to get some info from you. The goal here is to make the various routines
> SMP safe as well as UP safe during interrupts and other manipulating
> routines. So far, I've added a spin_lock variable to things like my queue
> structs. Now, from what I recall, there are some spin lock functions I can
> use to lock these spin locks from other use as opposed to a (nasty)
> save_flags(); cli(); stuff; restore_flags(); construct. Where do I find
> these routines and go about making use of them? Do they only lock on a
> per-processor basis or can they also lock say an interrupt routine from
> mucking with a queue if the queue routine was manipulating it when the
> interrupt occurred, or should I still use a cli(); based construct on that
> one?
See <asm/spinlock.h>. The basic version is:
spinlock_t xxx_lock = SPIN_LOCK_UNLOCKED;
unsigned long flags;
spin_lock_irqsave(&xxx_lock, flags);
... critical section here ..
spin_unlock_irqrestore(&xxx_lock, flags);
and the above is always safe. It will disable interrupts _locally_, but the
The above is always safe. It will disable interrupts _locally_, but the
spinlock itself will guarantee the global lock, so it will guarantee that
there is only one thread-of-control within the region(s) protected by that
lock.
Note that it works well even under UP - the above sequence under UP
essentially is just the same as doing a
lock. This works well even under UP. The above sequence under UP
essentially is just the same as doing
unsigned long flags;
@ -91,15 +24,13 @@ essentially is just the same as doing a
so the code does _not_ need to worry about UP vs SMP issues: the spinlocks
work correctly under both (and spinlocks are actually more efficient on
architectures that allow doing the "save_flags + cli" in one go because I
don't export that interface normally).
architectures that allow doing the "save_flags + cli" in one operation).
NOTE NOTE NOTE! The reason the spinlock is so much faster than a global
interrupt lock under SMP is exactly because it disables interrupts only on
the local CPU. The spin-lock is safe only when you _also_ use the lock
itself to do locking across CPU's, which implies that EVERYTHING that
touches a shared variable has to agree about the spinlock they want to
use.
NOTE! Implications of spin_locks for memory are further described in:
Documentation/memory-barriers.txt
(5) LOCK operations.
(6) UNLOCK operations.
The above is usually pretty simple (you usually need and want only one
spinlock for most things - using more than one spinlock can make things a
@ -125,15 +56,19 @@ can't stomp on each other).
The above is a problem mainly if you end up mixing code - for example the
routines in ll_rw_block() tend to use cli/sti to protect the atomicity of
their actions, and if a driver uses spinlocks instead then you should
think about issues like the above..
think about issues like the above.
This is really the only really hard part about spinlocks: once you start
using spinlocks they tend to expand to areas you might not have noticed
before, because you have to make sure the spinlocks correctly protect the
shared data structures _everywhere_ they are used. The spinlocks are most
easily added to places that are completely independent of other code (ie
internal driver data structures that nobody else ever touches, for
example).
easily added to places that are completely independent of other code (for
example, internal driver data structures that nobody else ever touches).
NOTE! The spin-lock is safe only when you _also_ use the lock itself
to do locking across CPU's, which implies that EVERYTHING that
touches a shared variable has to agree about the spinlock they want
to use.
----
@ -141,14 +76,18 @@ Lesson 2: reader-writer spinlocks.
If your data accesses have a very natural pattern where you usually tend
to mostly read from the shared variables, the reader-writer locks
(rw_lock) versions of the spinlocks are often nicer. They allow multiple
(rw_lock) versions of the spinlocks are sometimes useful. They allow multiple
readers to be in the same critical region at once, but if somebody wants
to change the variables it has to get an exclusive write lock. The
routines look the same as above:
to change the variables it has to get an exclusive write lock.
NOTE! reader-writer locks require more atomic memory operations than
simple spinlocks. Unless the reader critical section is long, you
are better off just using spinlocks.
The routines look the same as above:
rwlock_t xxx_lock = RW_LOCK_UNLOCKED;
unsigned long flags;
read_lock_irqsave(&xxx_lock, flags);
@ -159,18 +98,21 @@ routines look the same as above:
.. read and write exclusive access to the info ...
write_unlock_irqrestore(&xxx_lock, flags);
The above kind of lock is useful for complex data structures like linked
lists etc, especially when you know that most of the work is to just
traverse the list searching for entries without changing the list itself,
for example. Then you can use the read lock for that kind of list
traversal, which allows many concurrent readers. Anything that _changes_
the list will have to get the write lock.
The above kind of lock may be useful for complex data structures like
linked lists, especially searching for entries without changing the list
itself. The read lock allows many concurrent readers. Anything that
_changes_ the list will have to get the write lock.
Note: you cannot "upgrade" a read-lock to a write-lock, so if you at _any_
NOTE! RCU is better for list traversal, but requires careful
attention to design detail (see Documentation/RCU/listRCU.txt).
Also, you cannot "upgrade" a read-lock to a write-lock, so if you at _any_
time need to do any changes (even if you don't do it every time), you have
to get the write-lock at the very beginning. I could fairly easily add a
primitive to create a "upgradeable" read-lock, but it hasn't been an issue
yet. Tell me if you'd want one.
to get the write-lock at the very beginning.
NOTE! We are working hard to remove reader-writer spinlocks in most
cases, so please don't add a new one without consensus. (Instead, see
Documentation/RCU/rcu.txt for complete information.)
----
@ -233,4 +175,46 @@ indeed), while write-locks need to protect themselves against interrupts.
Linus
----
Reference information:
For dynamic initialization, use spin_lock_init() or rwlock_init() as
appropriate:
spinlock_t xxx_lock;
rwlock_t xxx_rw_lock;
static int __init xxx_init(void)
{
spin_lock_init(&xxx_lock);
rwlock_init(&xxx_rw_lock);
...
}
module_init(xxx_init);
For static initialization, use DEFINE_SPINLOCK() / DEFINE_RWLOCK() or
__SPIN_LOCK_UNLOCKED() / __RW_LOCK_UNLOCKED() as appropriate.
SPIN_LOCK_UNLOCKED and RW_LOCK_UNLOCKED are deprecated. These interfere
with lockdep state tracking.
Most of the time, you can simply turn:
static spinlock_t xxx_lock = SPIN_LOCK_UNLOCKED;
into:
static DEFINE_SPINLOCK(xxx_lock);
Static structure member variables go from:
struct foo bar {
.lock = SPIN_LOCK_UNLOCKED;
};
to:
struct foo bar {
.lock = __SPIN_LOCK_UNLOCKED(bar.lock);
};
Declaration of static rw_locks undergo a similar transformation.