Date: Sun, 03 Feb 2019 04:54:00 +0000 From: bugzilla-noreply@freebsd.org To: net@FreeBSD.org Subject: [Bug 235031] [em] em0: poor NFS performance, strange behavior Message-ID: <bug-235031-7501-HgGGMgaD1H@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-235031-7501@https.bugs.freebsd.org/bugzilla/> References: <bug-235031-7501@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D235031 --- Comment #28 from Eugene Grosbein <eugen@freebsd.org> --- (In reply to Rick Macklem from comment #27) Hardware-assisted TSO itself may be not useful in case of 1Gbps or less linerate for multiple reasons: overwhelming horsepower of modern (and even pretty old modern) amd64 CPUs, NFS mounts over UDP, extensive use of ipfw nat/libalias for routed jails/guest virtual machines as libalias is not compartible with TSO etc. So when not dealing with 40Gbps or faster hardwar= e, I choose to disable TSO per-interface and not fight against corresponding bug= s. As for NIC rings, even pretty old Intel Gigabit have tunnables to setup big enough rings to not bother of this issue. Again, even old Intel Gigabit NICs have multiple interrupt tunnables to set= up right timings for local traffic profile. As for testing different client mount option sets, I'll try it later and respond. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-235031-7501-HgGGMgaD1H>