From owner-svn-src-head@FreeBSD.ORG Fri Feb 4 17:49:55 2011 Return-Path: Delivered-To: svn-src-head@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 98867106566B; Fri, 4 Feb 2011 17:49:55 +0000 (UTC) (envelope-from yongari@FreeBSD.org) Received: from svn.freebsd.org (svn.freebsd.org [IPv6:2001:4f8:fff6::2c]) by mx1.freebsd.org (Postfix) with ESMTP id 6D9578FC08; Fri, 4 Feb 2011 17:49:55 +0000 (UTC) Received: from svn.freebsd.org (localhost [127.0.0.1]) by svn.freebsd.org (8.14.3/8.14.3) with ESMTP id p14HntfA092692; Fri, 4 Feb 2011 17:49:55 GMT (envelope-from yongari@svn.freebsd.org) Received: (from yongari@localhost) by svn.freebsd.org (8.14.3/8.14.3/Submit) id p14Hntjo092690; Fri, 4 Feb 2011 17:49:55 GMT (envelope-from yongari@svn.freebsd.org) Message-Id: <201102041749.p14Hntjo092690@svn.freebsd.org> From: Pyun YongHyeon Date: Fri, 4 Feb 2011 17:49:55 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org X-SVN-Group: head MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Cc: Subject: svn commit: r218289 - head/sys/dev/re X-BeenThere: svn-src-head@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: SVN commit messages for the src tree for head/-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 04 Feb 2011 17:49:55 -0000 Author: yongari Date: Fri Feb 4 17:49:55 2011 New Revision: 218289 URL: http://svn.freebsd.org/changeset/base/218289 Log: Disable TX IP checksum offloading for RTL8168C controllers. The controller in question generates frames with bad IP checksum value if packets contain IP options. For instance, packets generated by ping(8) with record route option have wrong IP checksum value. The controller correctly computes checksum for normal TCP/UDP packets though. There are two known RTL8168/8111C variants in market and the issue I observed happened on RL_HWREV_8168C_SPIN2. I'm not sure RL_HWREV_8168C also has the same issue but it would be better to assume it has the same issue since they shall share same core. RTL8102E which is supposed to be released at the time of RTL8168/8111C announcement does not have the issue. Tested by: Konstantin V. Krotov ( kkv <> insysnet dot ru ) Modified: head/sys/dev/re/if_re.c Modified: head/sys/dev/re/if_re.c ============================================================================== --- head/sys/dev/re/if_re.c Fri Feb 4 17:26:44 2011 (r218288) +++ head/sys/dev/re/if_re.c Fri Feb 4 17:49:55 2011 (r218289) @@ -1527,7 +1527,16 @@ re_attach(device_t dev) ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; ifp->if_ioctl = re_ioctl; ifp->if_start = re_start; - ifp->if_hwassist = RE_CSUM_FEATURES | CSUM_TSO; + /* + * RTL8168/8111C generates wrong IP checksummed frame if the + * packet has IP options so disable TX IP checksum offloading. + */ + if (sc->rl_hwrev->rl_rev == RL_HWREV_8168C || + sc->rl_hwrev->rl_rev == RL_HWREV_8168C_SPIN2) + ifp->if_hwassist = CSUM_TCP | CSUM_UDP; + else + ifp->if_hwassist = CSUM_IP | CSUM_TCP | CSUM_UDP; + ifp->if_hwassist |= CSUM_TSO; ifp->if_capabilities = IFCAP_HWCSUM | IFCAP_TSO4; ifp->if_capenable = ifp->if_capabilities; ifp->if_init = re_init; @@ -3209,6 +3218,7 @@ re_ioctl(struct ifnet *ifp, u_long comma struct rl_softc *sc = ifp->if_softc; struct ifreq *ifr = (struct ifreq *) data; struct mii_data *mii; + uint32_t rev; int error = 0; switch (command) { @@ -3294,9 +3304,14 @@ re_ioctl(struct ifnet *ifp, u_long comma if ((mask & IFCAP_TXCSUM) != 0 && (ifp->if_capabilities & IFCAP_TXCSUM) != 0) { ifp->if_capenable ^= IFCAP_TXCSUM; - if ((ifp->if_capenable & IFCAP_TXCSUM) != 0) - ifp->if_hwassist |= RE_CSUM_FEATURES; - else + if ((ifp->if_capenable & IFCAP_TXCSUM) != 0) { + rev = sc->rl_hwrev->rl_rev; + if (rev == RL_HWREV_8168C || + rev == RL_HWREV_8168C_SPIN2) + ifp->if_hwassist |= CSUM_TCP | CSUM_UDP; + else + ifp->if_hwassist |= RE_CSUM_FEATURES; + } else ifp->if_hwassist &= ~RE_CSUM_FEATURES; reinit = 1; }