Date: Fri, 8 Jun 2007 10:54:12 -0400 (EDT) From: Andrew Gallatin <gallatin@cs.duke.edu> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/net pfil.h Message-ID: <18025.28075.163371.116715@grasshopper.cs.duke.edu> In-Reply-To: <20070608135015.GP89017@FreeBSD.org> References: <200706081243.l58ChQnr045086@repoman.freebsd.org> <20070608135015.GP89017@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Gleb Smirnoff writes: > Nice! I guess you can tell how much performance did we lose before? Thanks. I was running packet blasting tests to mxge(4). At ~1.3Mpps, where my old, slow machine was cpu bound, hwpmc showed an rw lock (presumably from pfil_run_hooks), and pfil_run_hooks() as the top cpu users behind my driver's rx handler, ip_input, and udp_input. I got very busy, and never had a chance to re-run the tests with the fix in place. > Do you plan to MFC this? I honestly hadn't checked to see if the bug was in -stable. If it is, then I'll MFC it next week (if nobody beats me to it). Drew
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?18025.28075.163371.116715>