Date: Fri, 19 Feb 2021 13:48:22 +0000 From: "Scheffenegger, Richard" <Richard.Scheffenegger@netapp.com> To: "re@freebsd.org" <re@freebsd.org>, Glen Barber <gjb@freebsd.org> Cc: "transport@freebsd.org" <transport@freebsd.org> Subject: approval for D28780 / 853fd7a2e39802e46bd3d6476529796ac22412d9 Message-ID: <SN4PR0601MB3728F2C95010B94345239B6186849@SN4PR0601MB3728.namprd06.prod.outlook.com>
next in thread | raw e-mail | index | archive | help
Hi Glen, I received feedback around PRR which shows up as a performance degradation. I would like to MFC this commit in about one week into STABLE/13, and ideal= ly to releng/13.0 around the same time (or should there be at least 3 days = linger time in STABLE, after one week in HEAD)? https://reviews.freebsd.org/D28780 Under some circumstances, PRR may end up with a fully collapsed cwnd when finalizing the loss recovery. Also, in case PRR is working against an empty scoreboard, a possibilty for an integer overflow exists. Performing signed arithmetics to address a possible sudden jump of cwnd. Richard Scheffenegger
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?SN4PR0601MB3728F2C95010B94345239B6186849>