Date: Thu, 13 Feb 2014 14:34:01 -0500 From: Garrett Wollman <wollman@bimajority.org> To: Scott Long <scott4long@yahoo.com> Cc: FreeBSD Net <freebsd-net@freebsd.org> Subject: Re: Use of contiguous physical memory in cxgbe driver Message-ID: <21245.7721.789554.761901@hergotha.csail.mit.edu> In-Reply-To: <5CAE71A4-EA1D-481D-AFBA-3738F8E66087@yahoo.com> References: <21216.22944.314697.179039@hergotha.csail.mit.edu> <201402111348.52135.jhb@freebsd.org> <CAJ-VmonCdNQPUCQwm0OhqQ3Kt_7x6-g-JwGVZQfzWTgrDYfmqw@mail.gmail.com> <201402121446.19278.jhb@freebsd.org> <21244.20212.423983.960018@hergotha.csail.mit.edu> <20140213075651.GY34851@funkthat.com> <21245.1163.754141.154430@hergotha.csail.mit.edu> <5CAE71A4-EA1D-481D-AFBA-3738F8E66087@yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Thu, 13 Feb 2014 10:05:18 -0800, Scott Long <scott4long@yahoo.com> said: > So what you’re saying is that all of the other memory allocations that go along with > moving data through a socket wind up fragmenting the free memory pool to the point > where it becomes impossible to find 3 contig pages for a 9k jumbo RX frame. I should mention that this behavior is particularly problematic when only some clients are even sending jumbo frames. Demanding three physically contiguous pages when 70% of the packets received are not more than 1514 bytes is particularly wasteful, although hopefully this is mitigated with LRO in the common case. (I don't know if LRO even works with NFS. How would I check?) -GAWollman
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?21245.7721.789554.761901>