On Wed, Jun 13, 2012 at 09:51:03AM -0400, Josef Bacik wrote:
> On Wed, Jun 13, 2012 at 03:49:07PM +0200, Stefan Behrens wrote:
> > The last sentence of chapter 2 of Documentation/CodingStyle is quite
> > unambiguous. Here is the full quote of that chapter:
> >
> > Chapter 2: Breaki
On Wed, Jun 13, 2012 at 03:49:07PM +0200, Stefan Behrens wrote:
> On Wed, 13 Jun 2012 09:14:27 -0400, Josef Bacik wrote:
> > On Wed, Jun 13, 2012 at 12:35:26AM +0200, David Sterba wrote:
> >> On Tue, Jun 12, 2012 at 03:50:41PM -0400, Josef Bacik wrote:
> >>> @@ -4694,8 +4716,11 @@ int btrfs_init_de
On Wed, 13 Jun 2012 09:14:27 -0400, Josef Bacik wrote:
> On Wed, Jun 13, 2012 at 12:35:26AM +0200, David Sterba wrote:
>> On Tue, Jun 12, 2012 at 03:50:41PM -0400, Josef Bacik wrote:
>>> @@ -4694,8 +4716,11 @@ int btrfs_init_dev_stats(struct btrfs_fs_info
>>> *fs_info)
>>> key.offset =
On Wed, Jun 13, 2012 at 12:35:26AM +0200, David Sterba wrote:
> On Tue, Jun 12, 2012 at 03:50:41PM -0400, Josef Bacik wrote:
> > +++ b/fs/btrfs/check-integrity.c
> > @@ -93,6 +93,7 @@
> > #include "print-tree.h"
> > #include "locking.h"
> > #include "check-integrity.h"
> > +#include "rcu-string.
On Wed, 13 Jun 2012 00:35:26 +0200, David Sterba wrote:
> On Tue, Jun 12, 2012 at 03:50:41PM -0400, Josef Bacik wrote:
>> +++ b/fs/btrfs/check-integrity.c
>> @@ -93,6 +93,7 @@
>> #include "print-tree.h"
>> #include "locking.h"
>> #include "check-integrity.h"
>> +#include "rcu-string.h"
>>
>>
On Tue, Jun 12, 2012 at 03:50:41PM -0400, Josef Bacik wrote:
> +++ b/fs/btrfs/check-integrity.c
> @@ -93,6 +93,7 @@
> #include "print-tree.h"
> #include "locking.h"
> #include "check-integrity.h"
> +#include "rcu-string.h"
>
> #define BTRFSIC_BLOCK_HASHTABLE_SIZE 0x1
> #define BTRFSIC_BL
Al pointed out that we can just toss out the old name on a device and add a
new one arbitrarily, so anybody who uses device->name in printk could
possibly use free'd memory. Instead of adding locking around all of this he
suggested doing it with RCU, so I've introduced a struct rcu_string that
doe