RE: [PATCH V2 4/5] RDMA/iser: support iWARP devices

["Steve Wise" <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>]

> -----Original Message-----
> From: Or Gerlitz [mailto:ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org]
> Sent: Wednesday, July 01, 2015 2:40 AM
> To: Steve Wise
> Cc: 'Hefty, Sean'; 'Jason Gunthorpe'; 'Sagi Grimberg'; dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org; roid-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org; sagig-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org; linux-
> rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> Subject: Re: [PATCH V2 4/5] RDMA/iser: support iWARP devices
> 
> On 6/30/2015 9:42 PM, Steve Wise wrote:
> >>>> I prefer to decouple the iSER changes with this core work.
> >>>> Jason/Sean... thoughts?  I could do the iSER w/o patch 3, and the
> >>>> follow up with a series that includes our final solution on
> >>>> transport independent memory registration and change all the TI
> >>>> kernel users (iser and nfsrdma) along with it.
> >>> There has a been a big push lately to drop the strange transport
> >>> specific stuff - if you are comitted to seeing the access flag clean
> >>> up series through then I don't see a problem with using whatever order
> >>> you like.
> >> I agree.
> > Ok, then I'll do this:
> >
> > 1) series with iser changes to enable iwarp including mlx/ipath/qib patches to set max_sge_rd
> > 2) series on transport independent access flags / memory reg - updating the kernel ULPs too.
> not following... OTOH for iser to work over iwarp too we **only** (your
> #1) need to patch some IB/RoCE drivers to properly set their advertized
> max_sge_rd but no core changes but OTOH we **also** need  (your #2) an
> IB core patch/series with independent access flags for porting (say)
> NFSoRDMA to use the same code for multiple transports? explain...
> 
>

First, I'll submit iSER patches that enable iWARP in the same manner that NFSRDMA does.  Namely, have iSER look at the transport
type and set the MR access flags accordingly.  That is a very simple set of changes.  I submitted this as the first series earlier
in the week.  The main comment was to fix the rdma drivers to set max_sge_rd and use that to adjust the read sge depth.  

Next, I'll submit the transport independent core changes to remove the ULPS from having to know about transport access_flag
differences.  Included in this series will be changes to iSER and NFSRDMA to make use of the new services.

Steve.

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html