Re: [PATCH] Btrfs: don't take the chunk_mutex/dev_list mutex in statfs V2

2014-11-14 Thread David Sterba
On Thu, Nov 06, 2014 at 09:35:17AM -0500, Josef Bacik wrote: Yeah df if you change alloc_start in the middle of it is going to be slightly racey, which is fine since I'm going to kill alloc_start soon anyway. Go on, seems that there was never a real usecase for that. According to the

Re: [PATCH] Btrfs: don't take the chunk_mutex/dev_list mutex in statfs V2

2014-11-06 Thread Josef Bacik
On 11/04/2014 05:26 AM, Miao Xie wrote: On Mon, 3 Nov 2014 08:56:50 -0500, Josef Bacik wrote: Our gluster boxes get several thousand statfs() calls per second, which begins to suck hardcore with all of the lock contention on the chunk mutex and dev list mutex. We don't really need to hold

Re: [PATCH] Btrfs: don't take the chunk_mutex/dev_list mutex in statfs V2

2014-11-04 Thread Miao Xie
On Mon, 3 Nov 2014 08:56:50 -0500, Josef Bacik wrote: Our gluster boxes get several thousand statfs() calls per second, which begins to suck hardcore with all of the lock contention on the chunk mutex and dev list mutex. We don't really need to hold these things, if we have transient