Re: [PATCH 02/19] fs: don't take the i_lock in inode_inc_iversion

[NeilBrown]

On Wed, Dec 13 2017, Jeff Layton wrote:

> On Wed, 2017-12-13 at 09:20 -0500, Jeff Layton wrote:
>> From: Jeff Layton <jlayton@xxxxxxxxxx>
>> 
>> The rationale for taking the i_lock when incrementing this value is
>> lost in antiquity. The readers of the field don't take it (at least
>> not universally), so my assumption is that it was only done here to
>> serialize incrementors.
>> 
>> If that is indeed the case, then we can drop the i_lock from this
>> codepath and treat it as a atomic64_t for the purposes of
>> incrementing it. This allows us to use inode_inc_iversion without
>> any danger of lock inversion.
>> 
>> Note that the read side is not fetched atomically with this change.
>> The assumption here is that that is not a critical issue since the
>> i_version is not fully synchronized with anything else anyway.
>> 
>> Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
>> ---
>>  include/linux/fs.h | 6 +++---
>>  1 file changed, 3 insertions(+), 3 deletions(-)
>> 
>> diff --git a/include/linux/fs.h b/include/linux/fs.h
>> index 5001e77342fd..c234fac4bb77 100644
>> --- a/include/linux/fs.h
>> +++ b/include/linux/fs.h
>> @@ -2136,9 +2136,9 @@ inode_set_iversion_queried(struct inode *inode, const u64 new)
>>  static inline bool
>>  inode_maybe_inc_iversion(struct inode *inode, bool force)
>>  {
>> -	spin_lock(&inode->i_lock);
>> -	inode->i_version++;
>> -	spin_unlock(&inode->i_lock);
>> +	atomic64_t *ivp = (atomic64_t *)&inode->i_version;
>> +
>> +	atomic64_inc(ivp);
>>  	return true;
>>  }
>>  
>
> FWIW, I'm not sure this patch is strictly necessary as an interim step.
>
> Adding the i_lock into the all of the places where we currently just do
> inode->i_version++ without properly auditing all of them gave me pause
> though.
>
> In any case, the last patch in the series cleans this nastiness up.

Yes, I thought "nastiness" too, and was happy to see it cleaned up.

I would have guessed that the purpose of the spinlock was to avoid the
risk for torn-reads/writes on 32bit platforms that cannot access a 64bit
value atomically.  In either case, using atomic64_t is the right thing
to do.

Thanks,
NeilBrown

Attachment: signature.asc
Description: PGP signature