Re: [PATCH v2] target/file: add support of direct and async I/O

2018-05-08 Thread Mike Christie
On 05/08/2018 12:47 AM, Martin K. Petersen wrote: > > Bryant, > >> Can you review this patch and pull it into scsi since Nicholas has >> been out for awhile? >> >> I have tested this patch and really like the performance enhancements >> as a result of it. > > No problem queuing it up if I get an

Re: [PATCH v2] target/file: add support of direct and async I/O

2018-05-07 Thread Martin K. Petersen
Bryant, > Can you review this patch and pull it into scsi since Nicholas has > been out for awhile? > > I have tested this patch and really like the performance enhancements > as a result of it. No problem queuing it up if I get an ACK from Christoph or Mike. -- Martin K. Petersen Oracle

Re: [PATCH v2] target/file: add support of direct and async I/O

2018-05-07 Thread Bryant G. Ly
Hi Martin, Can you review this patch and pull it into scsi since Nicholas has been out for awhile? I have tested this patch and really like the performance enhancements as a result of it. Thanks, Bryant On 4/19/18 1:29 PM, Andrei Vagin wrote: > Hello Nicholas, > > What do you think about th

Re: [PATCH v2] target/file: add support of direct and async I/O

2018-04-19 Thread Andrei Vagin
Hello Nicholas, What do you think about this patch? Thanks, Andrei On Wed, Mar 21, 2018 at 11:55:02PM -0700, Andrei Vagin wrote: > There are two advantages: > * Direct I/O allows to avoid the write-back cache, so it reduces > affects to other processes in the system. > * Async I/O allows to ha

Re: [PATCH v2] target/file: add support of direct and async I/O

2018-03-22 Thread Andrei Vagin
On Thu, Mar 22, 2018 at 10:34:34AM -0700, Christoph Hellwig wrote: > > > > DIF (PI) emulation doesn't work when a target uses async I/O, because > > DIF metadata is saved in a separate file, and it is another non-trivial > > task how to synchronize writing in two files, so that a following read >

Re: [PATCH v2] target/file: add support of direct and async I/O

2018-03-22 Thread Bryant G. Ly
On 3/22/18 12:34 PM, Christoph Hellwig wrote: >> DIF (PI) emulation doesn't work when a target uses async I/O, because >> DIF metadata is saved in a separate file, and it is another non-trivial >> task how to synchronize writing in two files, so that a following read >> operation always returns a

Re: [PATCH v2] target/file: add support of direct and async I/O

2018-03-22 Thread Christoph Hellwig
> > DIF (PI) emulation doesn't work when a target uses async I/O, because > DIF metadata is saved in a separate file, and it is another non-trivial > task how to synchronize writing in two files, so that a following read > operation always returns a consisten metadata for a specified block. As sa

[PATCH v2] target/file: add support of direct and async I/O

2018-03-21 Thread Andrei Vagin
There are two advantages: * Direct I/O allows to avoid the write-back cache, so it reduces affects to other processes in the system. * Async I/O allows to handle a few commands concurrently. DIO + AIO shows a better perfomance for random write operations: Mode: O_DSYNC Async: 1 $ ./fio --bs=4K