From: Ingo Molnar <mingo@elte.hu>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>,
"Tantilov, Emil S" <emil.s.tantilov@intel.com>,
"Kirsher, Jeffrey T" <jeffrey.t.kirsher@intel.com>,
netdev <netdev@vger.kernel.org>,
David Miller <davem@davemloft.net>,
"Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@intel.com>,
"Duyck, Alexander H" <alexander.h.duyck@intel.com>,
Eric Dumazet <dada1@cosmosbay.com>
Subject: [patch] locking, percpu counters: introduce separate lock classes
Date: Mon, 29 Dec 2008 13:44:44 +0100 [thread overview]
Message-ID: <20081229124444.GA20306@elte.hu> (raw)
In-Reply-To: <20081229123819.GA18321@elte.hu>
* Ingo Molnar <mingo@elte.hu> wrote:
> hm, even with the revert i got the splat below. So some other commits
> are causing this too?
in any case, i picked up and tidied up Peter's annotation patch - see it
below. Chances are that this plus the revert will do the trick - and
that's what i'm testing now in tip/master.
Ingo
------------------------>
>From ea319518ba3de282c13ae1cf4bf2215c5e03e67e Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
Date: Fri, 26 Dec 2008 15:08:55 +0100
Subject: [PATCH] locking, percpu counters: introduce separate lock classes
Impact: fix lockdep false positives
Classify percpu_counter instances similar to regular lock objects --
that is, per instantiation site.
The networking code has increased its use of percpu_counters, which
leads to false positives if they are treated as a single class.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
include/linux/percpu_counter.h | 14 ++++++++++----
lib/percpu_counter.c | 18 ++++--------------
lib/proportions.c | 6 +++---
mm/backing-dev.c | 2 +-
4 files changed, 18 insertions(+), 22 deletions(-)
diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h
index 9007ccd..96bdde3 100644
--- a/include/linux/percpu_counter.h
+++ b/include/linux/percpu_counter.h
@@ -30,8 +30,16 @@ struct percpu_counter {
#define FBC_BATCH (NR_CPUS*4)
#endif
-int percpu_counter_init(struct percpu_counter *fbc, s64 amount);
-int percpu_counter_init_irq(struct percpu_counter *fbc, s64 amount);
+int __percpu_counter_init(struct percpu_counter *fbc, s64 amount,
+ struct lock_class_key *key);
+
+#define percpu_counter_init(fbc, value) \
+ ({ \
+ static struct lock_class_key __key; \
+ \
+ __percpu_counter_init(fbc, value, &__key); \
+ })
+
void percpu_counter_destroy(struct percpu_counter *fbc);
void percpu_counter_set(struct percpu_counter *fbc, s64 amount);
void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch);
@@ -85,8 +93,6 @@ static inline int percpu_counter_init(struct percpu_counter *fbc, s64 amount)
return 0;
}
-#define percpu_counter_init_irq percpu_counter_init
-
static inline void percpu_counter_destroy(struct percpu_counter *fbc)
{
}
diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c
index a866389..c7fe2e4 100644
--- a/lib/percpu_counter.c
+++ b/lib/percpu_counter.c
@@ -71,11 +71,11 @@ s64 __percpu_counter_sum(struct percpu_counter *fbc)
}
EXPORT_SYMBOL(__percpu_counter_sum);
-static struct lock_class_key percpu_counter_irqsafe;
-
-int percpu_counter_init(struct percpu_counter *fbc, s64 amount)
+int __percpu_counter_init(struct percpu_counter *fbc, s64 amount,
+ struct lock_class_key *key)
{
spin_lock_init(&fbc->lock);
+ lockdep_set_class(&fbc->lock, key);
fbc->count = amount;
fbc->counters = alloc_percpu(s32);
if (!fbc->counters)
@@ -87,17 +87,7 @@ int percpu_counter_init(struct percpu_counter *fbc, s64 amount)
#endif
return 0;
}
-EXPORT_SYMBOL(percpu_counter_init);
-
-int percpu_counter_init_irq(struct percpu_counter *fbc, s64 amount)
-{
- int err;
-
- err = percpu_counter_init(fbc, amount);
- if (!err)
- lockdep_set_class(&fbc->lock, &percpu_counter_irqsafe);
- return err;
-}
+EXPORT_SYMBOL(__percpu_counter_init);
void percpu_counter_destroy(struct percpu_counter *fbc)
{
diff --git a/lib/proportions.c b/lib/proportions.c
index 4f387a6..7367f2b 100644
--- a/lib/proportions.c
+++ b/lib/proportions.c
@@ -83,11 +83,11 @@ int prop_descriptor_init(struct prop_descriptor *pd, int shift)
pd->index = 0;
pd->pg[0].shift = shift;
mutex_init(&pd->mutex);
- err = percpu_counter_init_irq(&pd->pg[0].events, 0);
+ err = percpu_counter_init(&pd->pg[0].events, 0);
if (err)
goto out;
- err = percpu_counter_init_irq(&pd->pg[1].events, 0);
+ err = percpu_counter_init(&pd->pg[1].events, 0);
if (err)
percpu_counter_destroy(&pd->pg[0].events);
@@ -191,7 +191,7 @@ int prop_local_init_percpu(struct prop_local_percpu *pl)
spin_lock_init(&pl->lock);
pl->shift = 0;
pl->period = 0;
- return percpu_counter_init_irq(&pl->events, 0);
+ return percpu_counter_init(&pl->events, 0);
}
void prop_local_destroy_percpu(struct prop_local_percpu *pl)
diff --git a/mm/backing-dev.c b/mm/backing-dev.c
index f2e574d..f3b1258 100644
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -220,7 +220,7 @@ int bdi_init(struct backing_dev_info *bdi)
bdi->max_prop_frac = PROP_FRAC_BASE;
for (i = 0; i < NR_BDI_STAT_ITEMS; i++) {
- err = percpu_counter_init_irq(&bdi->bdi_stat[i], 0);
+ err = percpu_counter_init(&bdi->bdi_stat[i], 0);
if (err)
goto err;
}
next prev parent reply other threads:[~2008-12-29 12:45 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-12-25 10:25 unsafe locks seen with netperf on net-2.6.29 tree Jeff Kirsher
2008-12-25 11:26 ` Herbert Xu
2008-12-26 14:08 ` Peter Zijlstra
2008-12-27 19:38 ` Tantilov, Emil S
2008-12-27 20:38 ` Peter Zijlstra
2008-12-28 0:54 ` Tantilov, Emil S
2008-12-29 10:02 ` Peter Zijlstra
2008-12-29 10:07 ` Herbert Xu
2008-12-29 10:16 ` Peter Zijlstra
2008-12-29 10:22 ` Herbert Xu
2008-12-29 10:31 ` Herbert Xu
2008-12-29 10:37 ` Herbert Xu
2008-12-29 11:28 ` Ingo Molnar
2008-12-29 11:31 ` Ingo Molnar
2008-12-29 11:49 ` Herbert Xu
2008-12-29 11:58 ` Ingo Molnar
2008-12-29 12:01 ` Herbert Xu
2008-12-29 12:16 ` Ingo Molnar
2008-12-29 12:38 ` Ingo Molnar
2008-12-29 12:44 ` Ingo Molnar [this message]
2008-12-29 14:14 ` [patch] locking, percpu counters: introduce separate lock classes Ingo Molnar
2008-12-30 3:58 ` Herbert Xu
2008-12-30 6:05 ` Ingo Molnar
2008-12-30 6:39 ` David Miller
2008-12-30 6:56 ` Ingo Molnar
2008-12-30 7:04 ` David Miller
2008-12-30 7:21 ` Ingo Molnar
2008-12-29 12:49 ` unsafe locks seen with netperf on net-2.6.29 tree Herbert Xu
2008-12-29 12:55 ` Ingo Molnar
2008-12-29 9:57 ` Ingo Molnar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20081229124444.GA20306@elte.hu \
--to=mingo@elte.hu \
--cc=a.p.zijlstra@chello.nl \
--cc=alexander.h.duyck@intel.com \
--cc=dada1@cosmosbay.com \
--cc=davem@davemloft.net \
--cc=emil.s.tantilov@intel.com \
--cc=herbert@gondor.apana.org.au \
--cc=jeffrey.t.kirsher@intel.com \
--cc=netdev@vger.kernel.org \
--cc=peter.p.waskiewicz.jr@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).