Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1213456
| From | "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem |
| Date | 2015-08-26 02:30 +0200 |
| Message-ID | <q1wLf-69E-3@gated-at.bofh.it> (permalink) |
| References | <pZYBY-2dA-11@gated-at.bofh.it> <q0jZM-7yI-5@gated-at.bofh.it> <q120P-3p7-9@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
On Mon, Aug 24, 2015 at 05:34:31PM +0200, Oleg Nesterov wrote:
> On 08/22, Paul E. McKenney wrote:
> >
> > Queued for testing, thank you, Oleg!
>
> Thanks Paul!
I cannot really test it, but thus far it has at least not broken anything
else.
> > Right now, this is mostly relying on 0day and -next testing. Any thoughts
> > for a useful torture test for this?
>
> Right now I do not have any idea how to write the meaningful test for
> rcu_sync... Perhaps something like
>
> struct rcu_sync_struct rss;
> spinlock_t lock;
>
> int A, B;
>
> void read(void)
> {
> rcu_read_lock();
>
> bool need_lock = !rcu_sync_is_idle(&rss);
> if (need_lock)
> spin_lock(&lock);
>
> BUG_ON(A != B);
>
> if (need_lock)
> spin_unlock(&lock);
>
> rcu_read_unlock();
> }
>
> void modify(void)
> {
> rcu_sync_enter(&rss);
>
> spin_lock(&lock);
> A++; B++;
> spin_unlock(&lock);
>
> rcu_sync_exit(&rss);
> }
>
> makes sense... I'll try to think.
That looks like a promising start. There would need to be some sleep
time in modify() between rcu_sync_exit() and rcu_sync_enter().
> > One approach would be to treat it
> > like a reader-writer lock. Other thoughts?
>
> I booted the kernel with the additional patch below, and nothing bad has
> happened, it continues to print
>
> Writes: Total: 2 Max/Min: 0/0 Fail: 0
> Reads : Total: 2 Max/Min: 0/0 Fail: 0
>
> However, I do not know what this code actually does, so currently I have
> no idea if this test makes any sense for percpu_rw_semaphore.
Actually, unless I am really confused, that does not look good...
I would expect something like this, from a run with rwsem_lock:
[ 16.336057] Writes: Total: 473 Max/Min: 0/0 Fail: 0
[ 16.337615] Reads : Total: 219 Max/Min: 0/0 Fail: 0
[ 31.338152] Writes: Total: 959 Max/Min: 0/0 Fail: 0
[ 31.339114] Reads : Total: 437 Max/Min: 0/0 Fail: 0
[ 46.340167] Writes: Total: 1365 Max/Min: 0/0 Fail: 0
[ 46.341952] Reads : Total: 653 Max/Min: 0/0 Fail: 0
[ 61.343027] Writes: Total: 1795 Max/Min: 0/0 Fail: 0
[ 61.343968] Reads : Total: 865 Max/Min: 0/0 Fail: 0
[ 76.344034] Writes: Total: 2220 Max/Min: 0/0 Fail: 0
[ 76.345243] Reads : Total: 1071 Max/Min: 0/0 Fail: 0
The "Total" should increase for writes and for reads -- if you are
just seeing "Total: 2" over and over, that indicates that either
the torture test or rcu_sync got stuck somewhere.
Thanx, Paul
> Oleg.
> ---
>
> diff --git a/kernel/locking/locktorture.c b/kernel/locking/locktorture.c
> index ec8cce2..62561ec 100644
> --- a/kernel/locking/locktorture.c
> +++ b/kernel/locking/locktorture.c
> @@ -55,7 +55,7 @@ torture_param(int, stutter, 5, "Number of jiffies to run/halt test, 0=disable");
> torture_param(bool, verbose, true,
> "Enable verbose debugging printk()s");
>
> -static char *torture_type = "spin_lock";
> +static char *torture_type = "rwsem_lock";
> module_param(torture_type, charp, 0444);
> MODULE_PARM_DESC(torture_type,
> "Type of lock to torture (spin_lock, spin_lock_irq, mutex_lock, ...)");
> @@ -361,10 +361,12 @@ static struct lock_torture_ops mutex_lock_ops = {
> .name = "mutex_lock"
> };
>
> -static DECLARE_RWSEM(torture_rwsem);
> -static int torture_rwsem_down_write(void) __acquires(torture_rwsem)
> +#include <linux/percpu-rwsem.h>
> +static struct percpu_rw_semaphore pcpu_rwsem;
> +
> +static int torture_rwsem_down_write(void) __acquires(pcpu_rwsem)
> {
> - down_write(&torture_rwsem);
> + percpu_down_write(&pcpu_rwsem);
> return 0;
> }
>
> @@ -384,14 +386,14 @@ static void torture_rwsem_write_delay(struct torture_random_state *trsp)
> #endif
> }
>
> -static void torture_rwsem_up_write(void) __releases(torture_rwsem)
> +static void torture_rwsem_up_write(void) __releases(pcpu_rwsem)
> {
> - up_write(&torture_rwsem);
> + percpu_up_write(&pcpu_rwsem);
> }
>
> -static int torture_rwsem_down_read(void) __acquires(torture_rwsem)
> +static int torture_rwsem_down_read(void) __acquires(pcpu_rwsem)
> {
> - down_read(&torture_rwsem);
> + percpu_down_read(&pcpu_rwsem);
> return 0;
> }
>
> @@ -411,9 +413,9 @@ static void torture_rwsem_read_delay(struct torture_random_state *trsp)
> #endif
> }
>
> -static void torture_rwsem_up_read(void) __releases(torture_rwsem)
> +static void torture_rwsem_up_read(void) __releases(pcpu_rwsem)
> {
> - up_read(&torture_rwsem);
> + percpu_up_read(&pcpu_rwsem);
> }
>
> static struct lock_torture_ops rwsem_lock_ops = {
> @@ -645,6 +647,11 @@ static int __init lock_torture_init(void)
> &rwsem_lock_ops,
> };
>
> + /*
> + * TODO: DECLARE_PERCPU_RWSEM(). The patch already exists.
> + */
> + BUG_ON(percpu_init_rwsem(&pcpu_rwsem));
> +
> if (!torture_init_begin(torture_type, verbose, &torture_runnable))
> return -EBUSY;
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 3/8] rcusync: Add the CONFIG_PROVE_RCU checks Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 6/8] percpu-rwsem: change it to rely on rss_sync infrastructure Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 5/8] percpu-rwsem: make percpu_free_rwsem() after kzalloc() safe Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 8/8] percpu-rwsem: cleanup the lockdep annotations in percpu_down_read() Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 4/8] rcusync: Introduce rcu_sync_dtor() Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 7/8] percpu-rwsem: fix the comments outdated by rcu_sync Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 2/8] rcusync: Introduce struct rcu_sync_ops Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
[PATCH v2 1/8] rcu: Create rcu_sync infrastructure Oleg Nesterov <oleg@redhat.com> - 2015-08-21 19:50 +0200
Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2015-08-22 18:40 +0200
Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem Oleg Nesterov <oleg@redhat.com> - 2015-08-24 17:40 +0200
parse_args() is too unforgivable? Oleg Nesterov <oleg@redhat.com> - 2015-08-24 20:40 +0200
Re: parse_args() is too unforgivable? Rusty Russell <rusty@rustcorp.com.au> - 2015-08-25 03:40 +0200
[PATCH 0/1] params: don't ignore the rest of cmdline if parse_one() fails Oleg Nesterov <oleg@redhat.com> - 2015-08-25 17:30 +0200
[PATCH 1/1] params: don't ignore the rest of cmdline if parse_one() fails Oleg Nesterov <oleg@redhat.com> - 2015-08-25 17:30 +0200
Re: [PATCH 1/1] params: don't ignore the rest of cmdline if parse_one() fails Rusty Russell <rusty@rustcorp.com.au> - 2015-08-26 03:20 +0200
Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2015-08-26 02:30 +0200
Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem Oleg Nesterov <oleg@redhat.com> - 2015-08-26 14:20 +0200
Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem Oleg Nesterov <oleg@redhat.com> - 2015-08-26 15:00 +0200
Re: [PATCH v2 0/8] Add rcu_sync infrastructure to avoid _expedited() in percpu-rwsem "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2015-08-26 16:40 +0200
csiph-web