Date: Tue, 16 Aug 2016 06:21:14 -0700 From: Randall Stewart <rrs@netflix.com> To: Slawa Olhovchenkov <slw@zxy.spb.ru> Cc: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r304218 - head/sys/netinet Message-ID: <16561701-B1C6-4BE3-B9BA-3535F564620F@netflix.com> In-Reply-To: <20160816131805.GK22212@zxy.spb.ru> References: <201608161240.u7GCeuWS082118@repo.freebsd.org> <20160816131805.GK22212@zxy.spb.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
In theory it *could* be MFC=E2=80=99d to stable-10 and 11 but I am not = sure we want to do that. I am told by Drew that it does improve performance since in stable-10 you are = getting the INFO_WLOCK() but I am not sure if folks want it MFC=E2=80=99d=E2=80=A6 One thing that this code leads us towards is we *in theory* could move = the lock acquisition to the timer code itself (I think).. we would have to make sure that the = callout functions did do the unlock since thats part of the lock-dance with reference=E2=80=A6 but = its theoretically possible :-) R > On Aug 16, 2016, at 6:18 AM, Slawa Olhovchenkov <slw@zxy.spb.ru> = wrote: >=20 > On Tue, Aug 16, 2016 at 12:40:56PM +0000, Randall Stewart wrote: >=20 >> Author: rrs >> Date: Tue Aug 16 12:40:56 2016 >> New Revision: 304218 >> URL: https://svnweb.freebsd.org/changeset/base/304218 >>=20 >> Log: >> This cleans up the timer code in TCP and also makes it so we do not >> take the INFO lock *unless* we are really going to delete the TCB. >>=20 >> Differential Revision: D7136 >=20 > Is this related to stable/10? -------- Randall Stewart rrs@netflix.com 803-317-4952
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?16561701-B1C6-4BE3-B9BA-3535F564620F>