From: "J. R. Okajima" <hooanon05g@gmail.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: kbuild test robot <lkp@intel.com>,
kbuild-all@01.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/3] lockdep: new annotation lock_downgrade()
Date: Fri, 03 Feb 2017 06:05:31 +0900 [thread overview]
Message-ID: <2978.1486069531@jrobl> (raw)
In-Reply-To: <20170202184540.GZ6515@twins.programming.kicks-ass.net>
Peter Zijlstra:
> > >> kernel/locking/rwsem.c:126:2: error: implicit declaration of functi=
on 'lock_downgrade' [-Werror=3Dimplicit-function-declaration]
> > lock_downgrade(&sem->dep_map, _RET_IP_);
> > ^~~~~~~~~~~~~~
:::
> This is what you need !LOCKDEP stubs for ;-)
Ok, here is the update.
Just one line added.
J. R. Okajima
commit 6874cbfb3c4f757efecbeb800bfd4db1050698f6
Author: J. R. Okajima <hooanon05g@gmail.com>
Date: Thu Feb 2 23:41:38 2017 +0900
lockdep: new annotation lock_downgrade()
=
The commit
f831948 2016-11-30 locking/lockdep: Provide a type check for lock_is_=
held
didn't fully support rwsem. Here downgrade_write() supports the added =
type.
=
Originally-written-by: Peter Zijlstra <peterz@infradead.org>
See-also: http://marc.info/?l=3Dlinux-kernel&m=3D148581164003149&w=3D2
Signed-off-by: J. R. Okajima <hooanon05g@gmail.com>
diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
index 0345cbf..ba75d06 100644
--- a/include/linux/lockdep.h
+++ b/include/linux/lockdep.h
@@ -361,6 +361,8 @@ static inline void lock_set_subclass(struct lockdep_ma=
p *lock,
lock_set_class(lock, lock->name, lock->key, subclass, ip);
}
=
+extern void lock_downgrade(struct lockdep_map *lock, unsigned long ip);
+
extern void lockdep_set_current_reclaim_state(gfp_t gfp_mask);
extern void lockdep_clear_current_reclaim_state(void);
extern void lockdep_trace_alloc(gfp_t mask);
@@ -413,6 +415,7 @@ static inline void lockdep_on(void)
=
# define lock_acquire(l, s, t, r, c, n, i) do { } while (0)
# define lock_release(l, n, i) do { } while (0)
+# define lock_downgrade(l, i) do { } while (0)
# define lock_set_class(l, n, k, s, i) do { } while (0)
# define lock_set_subclass(l, s, i) do { } while (0)
# define lockdep_set_current_reclaim_state(g) do { } while (0)
diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c
index 7dc8f8e..6a4a740 100644
--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3523,6 +3523,44 @@ __lock_set_class(struct lockdep_map *lock, const ch=
ar *name,
return 1;
}
=
+static int __lock_downgrade(struct lockdep_map *lock, unsigned long ip)
+{
+ struct task_struct *curr =3D current;
+ struct held_lock *hlock;
+ unsigned int depth;
+ int i;
+
+ depth =3D curr->lockdep_depth;
+ /*
+ * This function is about (re)setting the class of a held lock,
+ * yet we're not actually holding any locks. Naughty user!
+ */
+ if (DEBUG_LOCKS_WARN_ON(!depth))
+ return 0;
+
+ hlock =3D find_held_lock(curr, lock, depth, &i);
+ if (!hlock)
+ return print_unlock_imbalance_bug(curr, lock, ip);
+
+ curr->lockdep_depth =3D i;
+ curr->curr_chain_key =3D hlock->prev_chain_key;
+
+ WARN(hlock->read, "downgrading a read lock");
+ hlock->read =3D 1;
+ hlock->acquire_ip =3D ip;
+
+ if (validate_held_lock(curr, depth, i))
+ return 0;
+
+ /*
+ * I took it apart and put it back together again, except now I have
+ * these 'spare' parts.. where shall I put them.
+ */
+ if (DEBUG_LOCKS_WARN_ON(curr->lockdep_depth !=3D depth))
+ return 0;
+ return 1;
+}
+
/*
* Remove the lock to the list of currently held locks - this gets
* called on mutex_unlock()/spin_unlock*() (or on a failed
@@ -3749,6 +3787,23 @@ void lock_set_class(struct lockdep_map *lock, const=
char *name,
}
EXPORT_SYMBOL_GPL(lock_set_class);
=
+void lock_downgrade(struct lockdep_map *lock, unsigned long ip)
+{
+ unsigned long flags;
+
+ if (unlikely(current->lockdep_recursion))
+ return;
+
+ raw_local_irq_save(flags);
+ current->lockdep_recursion =3D 1;
+ check_flags(flags);
+ if (__lock_downgrade(lock, ip))
+ check_chain_key(current);
+ current->lockdep_recursion =3D 0;
+ raw_local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(lock_downgrade);
+
/*
* We are not always called with irqs disabled - do that here,
* and also avoid lockdep recursion:
diff --git a/kernel/locking/rwsem.c b/kernel/locking/rwsem.c
index 45ba475..31db3ef 100644
--- a/kernel/locking/rwsem.c
+++ b/kernel/locking/rwsem.c
@@ -123,10 +123,8 @@ EXPORT_SYMBOL(up_write);
*/
void downgrade_write(struct rw_semaphore *sem)
{
- /*
- * lockdep: a downgraded write will live on as a write
- * dependency.
- */
+ lock_downgrade(&sem->dep_map, _RET_IP_);
+
rwsem_set_reader_owned(sem);
__downgrade_write(sem);
}
next prev parent reply other threads:[~2017-02-02 21:05 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-01-30 21:25 Q: lockdep_assert_held_read() after downgrade_write() J. R. Okajima
2017-01-30 21:30 ` Jens Axboe
2017-01-31 10:36 ` Peter Zijlstra
2017-01-31 11:25 ` Peter Zijlstra
2017-01-31 14:23 ` Waiman Long
2017-01-31 14:26 ` Peter Zijlstra
2017-01-31 15:40 ` J. R. Okajima
2017-01-31 16:32 ` Peter Zijlstra
2017-02-02 16:33 ` J. R. Okajima
2017-02-02 16:38 ` [PATCH 1/3] lockdep: consolidate by new find_held_lock() J. R. Okajima
2017-02-02 16:38 ` [PATCH 2/3] lockdep: consolidate by new validate_held_lock() J. R. Okajima
2017-02-14 12:09 ` Peter Zijlstra
2017-03-16 11:25 ` [tip:locking/core] locking/lockdep: Factor out the validate_held_lock() helper function tip-bot for J. R. Okajima
2017-02-02 16:38 ` [PATCH 3/3] lockdep: new annotation lock_downgrade() J. R. Okajima
2017-02-02 17:59 ` kbuild test robot
2017-02-02 18:45 ` Peter Zijlstra
2017-02-02 21:05 ` J. R. Okajima [this message]
2017-02-14 12:11 ` Peter Zijlstra
2017-03-16 11:25 ` [tip:locking/core] locking/lockdep: Add new check to lock_downgrade() tip-bot for J. R. Okajima
2017-03-16 11:24 ` [tip:locking/core] locking/lockdep: Factor out the find_held_lock() helper function tip-bot for J. R. Okajima
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=2978.1486069531@jrobl \
--to=hooanon05g@gmail.com \
--cc=kbuild-all@01.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lkp@intel.com \
--cc=peterz@infradead.org \
/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).