From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Gunthorpe Subject: Re: IB: increase DMA max_segment_size on Mellanox hardware Date: Tue, 18 Jan 2011 11:01:29 -0700 Message-ID: <20110118180129.GH995@obsidianresearch.com> References: <1295230184.2574.26.camel@obelisk.thedillows.org> <4D342510.80701@voltaire.com> <1295318783.3051.81.camel@obelisk.thedillows.org> <4D353E5E.2020500@voltaire.com> <1295353524.3051.121.camel@obelisk.thedillows.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1295353524.3051.121.camel-1q1vX8mYZiGLUyTwlgNVppKKF0rrzTr+@public.gmane.org> Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: David Dillow Cc: Or Gerlitz , linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Roland Dreier , Eli Cohen List-Id: linux-rdma@vger.kernel.org On Tue, Jan 18, 2011 at 07:25:24AM -0500, David Dillow wrote: > > Also when working with direct I/O from user space and/or under file-system, > > did you really see many BIOs that can be merged? I was under the impression, > > that (specifically after some time the system is active) for the most case, > > I get totally scattered SGs, whose pages can't be coalesced at all. > > I would expect that to be a common case, but there are systems out there > that this is not an issue. They typically allocate buffers early on when > they can be contiguous, and just keep reusing those. It has become really easy to use hugetlb now, you can just mmap anonymous memory with MAP_HUGETLB and you get 2M pages. If an app is using O_DIRECT it seem worth making an attempt to use huge pages.. IIRC databases are already doing this. Jason -- 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