[PATCH v1 1/2] IB/iser: set block queue_virt_boundary

2015-10-11 Thread Sagi Grimberg
By limiting the sg lists that we are allowing to meet and 4k virtual boundary, we can completely remove the entire bounce buffering logic. Signed-off-by: Sagi Grimberg Reviewed-by: Christoph Hellwig Cc: Jens Axboe Cc: Martin K. Petersen

Re: [PATCH v1 1/2] IB/iser: set block queue_virt_boundary

2015-10-11 Thread Or Gerlitz
On Sun, Oct 11, 2015 at 6:35 PM, Sagi Grimberg wrote: > By limiting the sg lists that we are allowing to meet > and 4k virtual boundary, we can completely remove the > entire bounce buffering logic. "limiting the sg lists that we are allowing to meet and 4k virtual boundary"

Re: [PATCH v1 1/2] IB/iser: set block queue_virt_boundary

2015-10-11 Thread Bart Van Assche
On 10/11/15 12:04, Or Gerlitz wrote: If the 4X assertion made here is correct, why not keeping the current BB logic to come into play for such devices? I know that without the BB code our driver looks much nicer and elegant, but life is sometimes more complex... thoughts? Hello Or, Whether or

Re: [PATCH v1 1/2] IB/iser: set block queue_virt_boundary

2015-10-11 Thread Or Gerlitz
On Mon, Oct 12, 2015 at 7:34 AM, Sagi Grimberg wrote: >> What happens now when an app wants to use 1K, 2K, 3K IOs? they can >> only use 1/4, 1/2 or 3/4 of the available memory for these >> transactions? > What? I'm not sure what you are talking about. what available

Re: [PATCH v1 1/2] IB/iser: set block queue_virt_boundary

2015-10-11 Thread Sagi Grimberg
What happens now when an app wants to use 1K, 2K, 3K IOs? they can only use 1/4, 1/2 or 3/4 of the available memory for these transactions? What? I'm not sure what you are talking about. what available memory? and we are going to fix it with devices like mlx5 over the new API you're