qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] seqlock: fix seqlock_write_unlock_impl function
@ 2020-01-29 14:49 Luc Michel
  2020-01-29 15:49 ` Philippe Mathieu-Daudé
  2020-01-30  0:27 ` Paolo Bonzini
  0 siblings, 2 replies; 3+ messages in thread
From: Luc Michel @ 2020-01-29 14:49 UTC (permalink / raw)
  To: qemu-devel, Paolo Bonzini; +Cc: Luc Michel

The seqlock write unlock function was incorrectly calling
seqlock_write_begin() instead of seqlock_write_end(), and was releasing
the lock before incrementing the sequence. This could lead to a race
condition and a corrupted sequence number becoming odd even though the
lock is not held.

Signed-off-by: Luc Michel <luc.michel@greensocs.com>
---
 include/qemu/seqlock.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/qemu/seqlock.h b/include/qemu/seqlock.h
index fd408b7ec5..8b6b4ee4bb 100644
--- a/include/qemu/seqlock.h
+++ b/include/qemu/seqlock.h
@@ -53,15 +53,15 @@ static inline void seqlock_write_lock_impl(QemuSeqLock *sl, QemuLockable *lock)
     seqlock_write_begin(sl);
 }
 #define seqlock_write_lock(sl, lock) \
     seqlock_write_lock_impl(sl, QEMU_MAKE_LOCKABLE(lock))
 
-/* Lock out other writers and update the count.  */
+/* Update the count and release the lock.  */
 static inline void seqlock_write_unlock_impl(QemuSeqLock *sl, QemuLockable *lock)
 {
+    seqlock_write_end(sl);
     qemu_lockable_unlock(lock);
-    seqlock_write_begin(sl);
 }
 #define seqlock_write_unlock(sl, lock) \
     seqlock_write_unlock_impl(sl, QEMU_MAKE_LOCKABLE(lock))
 
 
-- 
2.25.0



^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] seqlock: fix seqlock_write_unlock_impl function
  2020-01-29 14:49 [PATCH] seqlock: fix seqlock_write_unlock_impl function Luc Michel
@ 2020-01-29 15:49 ` Philippe Mathieu-Daudé
  2020-01-30  0:27 ` Paolo Bonzini
  1 sibling, 0 replies; 3+ messages in thread
From: Philippe Mathieu-Daudé @ 2020-01-29 15:49 UTC (permalink / raw)
  To: Luc Michel, qemu-devel, Paolo Bonzini

On 1/29/20 3:49 PM, Luc Michel wrote:
> The seqlock write unlock function was incorrectly calling
> seqlock_write_begin() instead of seqlock_write_end(), and was releasing
> the lock before incrementing the sequence. This could lead to a race
> condition and a corrupted sequence number becoming odd even though the
> lock is not held.

I'm surprised it took 18 months to figure this out.

Fixes: 988fcafc730
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> Signed-off-by: Luc Michel <luc.michel@greensocs.com>
> ---
>   include/qemu/seqlock.h | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/include/qemu/seqlock.h b/include/qemu/seqlock.h
> index fd408b7ec5..8b6b4ee4bb 100644
> --- a/include/qemu/seqlock.h
> +++ b/include/qemu/seqlock.h
> @@ -53,15 +53,15 @@ static inline void seqlock_write_lock_impl(QemuSeqLock *sl, QemuLockable *lock)
>       seqlock_write_begin(sl);
>   }
>   #define seqlock_write_lock(sl, lock) \
>       seqlock_write_lock_impl(sl, QEMU_MAKE_LOCKABLE(lock))
>   
> -/* Lock out other writers and update the count.  */
> +/* Update the count and release the lock.  */
>   static inline void seqlock_write_unlock_impl(QemuSeqLock *sl, QemuLockable *lock)
>   {
> +    seqlock_write_end(sl);
>       qemu_lockable_unlock(lock);
> -    seqlock_write_begin(sl);
>   }
>   #define seqlock_write_unlock(sl, lock) \
>       seqlock_write_unlock_impl(sl, QEMU_MAKE_LOCKABLE(lock))
>   
>   
> 



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] seqlock: fix seqlock_write_unlock_impl function
  2020-01-29 14:49 [PATCH] seqlock: fix seqlock_write_unlock_impl function Luc Michel
  2020-01-29 15:49 ` Philippe Mathieu-Daudé
@ 2020-01-30  0:27 ` Paolo Bonzini
  1 sibling, 0 replies; 3+ messages in thread
From: Paolo Bonzini @ 2020-01-30  0:27 UTC (permalink / raw)
  To: Luc Michel, qemu-devel

On 29/01/20 15:49, Luc Michel wrote:
> The seqlock write unlock function was incorrectly calling
> seqlock_write_begin() instead of seqlock_write_end(), and was releasing
> the lock before incrementing the sequence. This could lead to a race
> condition and a corrupted sequence number becoming odd even though the
> lock is not held.
> 
> Signed-off-by: Luc Michel <luc.michel@greensocs.com>
> ---
>  include/qemu/seqlock.h | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/include/qemu/seqlock.h b/include/qemu/seqlock.h
> index fd408b7ec5..8b6b4ee4bb 100644
> --- a/include/qemu/seqlock.h
> +++ b/include/qemu/seqlock.h
> @@ -53,15 +53,15 @@ static inline void seqlock_write_lock_impl(QemuSeqLock *sl, QemuLockable *lock)
>      seqlock_write_begin(sl);
>  }
>  #define seqlock_write_lock(sl, lock) \
>      seqlock_write_lock_impl(sl, QEMU_MAKE_LOCKABLE(lock))
>  
> -/* Lock out other writers and update the count.  */
> +/* Update the count and release the lock.  */
>  static inline void seqlock_write_unlock_impl(QemuSeqLock *sl, QemuLockable *lock)
>  {
> +    seqlock_write_end(sl);
>      qemu_lockable_unlock(lock);
> -    seqlock_write_begin(sl);
>  }
>  #define seqlock_write_unlock(sl, lock) \
>      seqlock_write_unlock_impl(sl, QEMU_MAKE_LOCKABLE(lock))
>  
>  
> 

Queued, thanks very much.

Paolo



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2020-01-30  0:28 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-01-29 14:49 [PATCH] seqlock: fix seqlock_write_unlock_impl function Luc Michel
2020-01-29 15:49 ` Philippe Mathieu-Daudé
2020-01-30  0:27 ` Paolo Bonzini

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).