LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Waiman Long <longman@redhat.com>
To: Ingo Molnar <mingo@redhat.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	Davidlohr Bueso <dave@stgolabs.net>,
	"Theodore Y. Ts'o" <tytso@mit.edu>,
	Oleg Nesterov <oleg@redhat.com>,
	Amir Goldstein <amir73il@gmail.com>, Jan Kara <jack@suse.cz>,
	Matthew Wilcox <willy@infradead.org>,
	Waiman Long <longman@redhat.com>
Subject: [PATCH v3 2/2] locking/percpu-rwsem: Annotate rwsem ownership transfer by setting RWSEM_OWNER_UNKNOWN
Date: Tue, 15 May 2018 13:38:04 -0400	[thread overview]
Message-ID: <1526405884-4860-3-git-send-email-longman@redhat.com> (raw)
In-Reply-To: <1526405884-4860-1-git-send-email-longman@redhat.com>

The filesystem freezing code needs to transfer ownership of a rwsem
embedded in a percpu-rwsem from the task that does the freezing to
another one that does the thawing by calling percpu_rwsem_release()
after freezing and percpu_rwsem_acquire() before thawing.

However, the new rwsem debug code runs afoul with this scheme by warning
that the task that releases the rwsem isn't the one that acquires it.

[   20.302978] ------------[ cut here ]------------
[   20.305016] DEBUG_LOCKS_WARN_ON(sem->owner != get_current())
[   20.305029] WARNING: CPU: 1 PID: 1401 at
/home/amir/build/src/linux/kernel/locking/rwsem.c:133 up_write+0x59/0x79
[   20.311252] CPU: 1 PID: 1401 Comm: fsfreeze Not tainted 4.17.0-rc3-xfstests-00049-g39e47bf59eb3 #3276
[   20.314808] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014
[   20.318403] RIP: 0010:up_write+0x59/0x79
[   20.320928] RSP: 0018:ffffc90000717e48 EFLAGS: 00010286
[   20.322955] RAX: 0000000000000030 RBX: ffff880078f1c680 RCX: ffff880078e42200
[   20.325665] RDX: ffffffff810cc9c1 RSI: 0000000000000001 RDI: 0000000000000202
[   20.328844] RBP: ffffc90000717e80 R08: 0000000000000001 R09: 0000000000000001
[   20.332340] R10: ffffc90000717c58 R11: ffffffff836807ad R12: ffff880078f1c388
[   20.335095] R13: ffff880078a8b980 R14: 0000000000000000 R15: 00000000fffffff7
[   20.338009] FS:  00007fb61ca42700(0000) GS:ffff88007f400000(0000) knlGS:0000000000000000
[   20.341423] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   20.343772] CR2: 00007fb61c559b30 CR3: 0000000078da6000 CR4: 00000000000006e0
[   20.346463] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   20.349201] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   20.351960] Call Trace:
[   20.352911]  percpu_up_write+0x1f/0x28
[   20.354344]  thaw_super_locked+0xdf/0x120
[   20.355944]  do_vfs_ioctl+0x270/0x5f1
[   20.357390]  ? __se_sys_newfstat+0x2e/0x39
[   20.358969]  ? entry_SYSCALL_64_after_hwframe+0x59/0xbe
[   20.360991]  ksys_ioctl+0x52/0x71
[   20.362384]  __x64_sys_ioctl+0x16/0x19
[   20.363702]  do_syscall_64+0x5d/0x167
[   20.365099]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

To work properly with the rwsem debug code, we need to annotate that the
rwsem ownership is unknown during the tranfer period until a brave soul
comes forward to acquire the ownership. During that period, optimistic
spinning will be disabled.

Signed-off-by: Waiman Long <longman@redhat.com>
---
 include/linux/percpu-rwsem.h | 6 +++++-
 include/linux/rwsem.h        | 5 +++++
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/include/linux/percpu-rwsem.h b/include/linux/percpu-rwsem.h
index b1f37a8..79b99d6 100644
--- a/include/linux/percpu-rwsem.h
+++ b/include/linux/percpu-rwsem.h
@@ -133,7 +133,7 @@ static inline void percpu_rwsem_release(struct percpu_rw_semaphore *sem,
 	lock_release(&sem->rw_sem.dep_map, 1, ip);
 #ifdef CONFIG_RWSEM_SPIN_ON_OWNER
 	if (!read)
-		sem->rw_sem.owner = NULL;
+		sem->rw_sem.owner = RWSEM_OWNER_UNKNOWN;
 #endif
 }
 
@@ -141,6 +141,10 @@ static inline void percpu_rwsem_acquire(struct percpu_rw_semaphore *sem,
 					bool read, unsigned long ip)
 {
 	lock_acquire(&sem->rw_sem.dep_map, 0, 1, read, 1, NULL, ip);
+#ifdef CONFIG_RWSEM_SPIN_ON_OWNER
+	if (!read)
+		sem->rw_sem.owner = current;
+#endif
 }
 
 #endif
diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h
index 56707d5..fc2ed27 100644
--- a/include/linux/rwsem.h
+++ b/include/linux/rwsem.h
@@ -44,6 +44,11 @@ struct rw_semaphore {
 #endif
 };
 
+/*
+ * Owner value to indicate the rwsem's owner is not currently known.
+ */
+#define RWSEM_OWNER_UNKNOWN	((struct task_struct *)-1)
+
 extern struct rw_semaphore *rwsem_down_read_failed(struct rw_semaphore *sem);
 extern struct rw_semaphore *rwsem_down_read_failed_killable(struct rw_semaphore *sem);
 extern struct rw_semaphore *rwsem_down_write_failed(struct rw_semaphore *sem);
-- 
1.8.3.1

  parent reply	other threads:[~2018-05-15 17:39 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-15 17:38 [PATCH v3 0/2] locking/rwsem: Fix DEBUG_RWSEM warning from thaw_super() Waiman Long
2018-05-15 17:38 ` [PATCH v3 1/2] locking/rwsem: Add a new RWSEM_ANONYMOUSLY_OWNED flag Waiman Long
2018-05-15 17:46   ` Peter Zijlstra
2018-05-15 17:48     ` Waiman Long
2018-05-15 17:48   ` Peter Zijlstra
2018-05-15 17:52     ` Waiman Long
2018-05-15 17:55       ` Peter Zijlstra
2018-05-15 17:56         ` Waiman Long
2018-05-15 18:02   ` Peter Zijlstra
2018-05-15 17:38 ` Waiman Long [this message]
2018-05-15 17:58   ` [PATCH v3 2/2] locking/percpu-rwsem: Annotate rwsem ownership transfer by setting RWSEM_OWNER_UNKNOWN Peter Zijlstra
2018-05-15 18:02     ` Waiman Long
2018-05-15 18:05       ` Amir Goldstein
2018-05-15 18:45         ` Waiman Long
2018-05-15 18:05       ` Peter Zijlstra
2018-05-15 18:35         ` Waiman Long
2018-05-15 18:02     ` Matthew Wilcox
2018-05-15 18:07       ` Peter Zijlstra
2018-05-15 18:45       ` Waiman Long
2018-05-15 21:21         ` Matthew Wilcox
2018-05-15 21:30           ` Waiman Long

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=1526405884-4860-3-git-send-email-longman@redhat.com \
    --to=longman@redhat.com \
    --cc=amir73il@gmail.com \
    --cc=dave@stgolabs.net \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=oleg@redhat.com \
    --cc=peterz@infradead.org \
    --cc=tglx@linutronix.de \
    --cc=tytso@mit.edu \
    --cc=willy@infradead.org \
    --subject='Re: [PATCH v3 2/2] locking/percpu-rwsem: Annotate rwsem ownership transfer by setting RWSEM_OWNER_UNKNOWN' \
    /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

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).