Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-28 Thread Or Gerlitz
On Tue, Apr 12, 2016 at 5:13 PM, Christoph Hellwig wrote: > So that we don't overflow the number of MR segments allocated because > we have to split on SGL segment into multiple MR segments. > > Signed-off-by: Christoph Hellwig nit, but please fix IB/iser: set [...] --> IB/iser: Set max segmen

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-13 Thread Christoph Hellwig
On Tue, Apr 12, 2016 at 11:43:09AM -0700, Christoph Hellwig wrote: > > I think this means that there is a mismatch between the current block layer > > limits and what NVMe / RDMA drivers need ... > > If we tell the block layer that we can only handle page sized comments > using max_segent_size it

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-13 Thread Sagi Grimberg
Acked-by: Sagi Grimberg -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-13 Thread Sagi Grimberg
In iser we sorta rely on 4k pages so we avoid PAGE_SIZE but rather set SIZE_4K for these sort of things (like we did in the virt_boundary). So you still want only 4k segments even on PPC where the PAGE_SIZE is 16k? Yes, iSER has the "no-gaps" constraint (like nvme) and some applications in t

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread Christoph Hellwig
On Tue, Apr 12, 2016 at 11:13:52AM -0700, Bart Van Assche wrote: > >That's what NVMe does, but I don't think it's a good idea. Because > >of the unaligned start into the page this means you have to set the limit > >to one lower than the actual hardware limit. > > I think this means that there is

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread Bart Van Assche
On 04/12/2016 09:51 AM, Christoph Hellwig wrote: On Tue, Apr 12, 2016 at 08:34:03AM -0700, Bart Van Assche wrote: ib_sg_to_pages() can handle segments that are larger than mr->page_size. The interface handles it fine, but we'll still end up using a segment per (MR) page. Have you considered

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread Christoph Hellwig
On Tue, Apr 12, 2016 at 08:34:03AM -0700, Bart Van Assche wrote: > Hello Christoph, > > ib_sg_to_pages() can handle segments that are larger than mr->page_size. The interface handles it fine, but we'll still end up using a segment per (MR) page. > Have you considered to set queue_limits.max_hw_s

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread Bart Van Assche
On 04/12/2016 07:13 AM, Christoph Hellwig wrote: So that we don't overflow the number of MR segments allocated because we have to split on SGL segment into multiple MR segments. Signed-off-by: Christoph Hellwig --- drivers/infiniband/ulp/iser/iscsi_iser.c | 1 + 1 file changed, 1 insertion(+

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread Christoph Hellwig
On Tue, Apr 12, 2016 at 07:27:36AM -0700, James Bottomley wrote: > > In iser we sorta rely on 4k pages so we avoid > > PAGE_SIZE but rather set SIZE_4K for these sort > > of things (like we did in the virt_boundary). > > So you still want only 4k segments even on PPC where the PAGE_SIZE is > 16k?

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread James Bottomley
On Tue, 2016-04-12 at 13:48 +0300, Sagi Grimberg wrote: > > So that we don't overflow the number of MR segments allocated > > because > > we have to split on SGL segment into multiple MR segments. > > > > Signed-off-by: Christoph Hellwig > > --- > > drivers/infiniband/ulp/iser/iscsi_iser.c | 1

Re: [PATCH 2/2] IB/iser: set max_segment_size

2016-04-12 Thread Sagi Grimberg
So that we don't overflow the number of MR segments allocated because we have to split on SGL segment into multiple MR segments. Signed-off-by: Christoph Hellwig --- drivers/infiniband/ulp/iser/iscsi_iser.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/infiniband/ulp/iser/iscs