Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-26 Thread Qu Wenruo
On 07/26/2016 06:57 PM, Filipe Manana wrote: On Tue, Jul 26, 2016 at 11:04 AM, Qu Wenruo wrote: At 07/26/2016 05:28 PM, Filipe Manana wrote: On Tue, Jul 26, 2016 at 1:57 AM, Qu Wenruo wrote: At 07/25/2016 09:48 PM, Filipe Manana

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-26 Thread Filipe Manana
On Tue, Jul 26, 2016 at 11:04 AM, Qu Wenruo wrote: > > > At 07/26/2016 05:28 PM, Filipe Manana wrote: >> >> On Tue, Jul 26, 2016 at 1:57 AM, Qu Wenruo >> wrote: >>> >>> >>> >>> At 07/25/2016 09:48 PM, Filipe Manana wrote: On Mon,

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-26 Thread Qu Wenruo
Some typo fixes: On 07/26/2016 06:04 PM, Qu Wenruo wrote: At 07/26/2016 05:28 PM, Filipe Manana wrote: On Tue, Jul 26, 2016 at 1:57 AM, Qu Wenruo wrote: At 07/25/2016 09:48 PM, Filipe Manana wrote: On Mon, Jul 25, 2016 at 8:19 AM, Qu Wenruo

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-26 Thread Qu Wenruo
At 07/26/2016 05:28 PM, Filipe Manana wrote: On Tue, Jul 26, 2016 at 1:57 AM, Qu Wenruo wrote: At 07/25/2016 09:48 PM, Filipe Manana wrote: On Mon, Jul 25, 2016 at 8:19 AM, Qu Wenruo wrote: This patch will disable clone detection of

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-26 Thread Filipe Manana
On Tue, Jul 26, 2016 at 1:57 AM, Qu Wenruo wrote: > > > At 07/25/2016 09:48 PM, Filipe Manana wrote: >> >> On Mon, Jul 25, 2016 at 8:19 AM, Qu Wenruo >> wrote: >>> >>> This patch will disable clone detection of send. >>> >>> The main problem of

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-25 Thread Qu Wenruo
At 07/25/2016 11:37 PM, David Sterba wrote: On Mon, Jul 25, 2016 at 03:19:59PM +0800, Qu Wenruo wrote: This patch will disable clone detection of send. Making that unconditional is not the right way. We do have the send operation flags in place so if you really think there's a need for

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-25 Thread Qu Wenruo
At 07/25/2016 09:48 PM, Filipe Manana wrote: On Mon, Jul 25, 2016 at 8:19 AM, Qu Wenruo wrote: This patch will disable clone detection of send. The main problem of clone detetion in send is its memory usage and long execution time. The clone detection is done by

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-25 Thread David Sterba
On Mon, Jul 25, 2016 at 03:19:59PM +0800, Qu Wenruo wrote: > This patch will disable clone detection of send. Making that unconditional is not the right way. We do have the send operation flags in place so if you really think there's a need for disabling the clones, let's add a flag for that

Re: [PATCH RFC] btrfs: send: Disable clone detection

2016-07-25 Thread Filipe Manana
On Mon, Jul 25, 2016 at 8:19 AM, Qu Wenruo wrote: > This patch will disable clone detection of send. > > The main problem of clone detetion in send is its memory usage and long > execution time. > > The clone detection is done by iterating all backrefs and adding backref

[PATCH RFC] btrfs: send: Disable clone detection

2016-07-25 Thread Qu Wenruo
This patch will disable clone detection of send. The main problem of clone detetion in send is its memory usage and long execution time. The clone detection is done by iterating all backrefs and adding backref whose root is the source. However iterating all backrefs is already quite a bad idea,