From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <06d201c0d47a$f06d2fc0$e8b7c6d4@SOMA> From: "Boyd Roberts" To: <9fans@cse.psu.edu> References: <20010504050636.542.qmail@infidel.intergate.ca> Subject: Re: [9fans] tcpiput() MIME-Version: 1.0 Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: 7bit Date: Fri, 4 May 2001 11:16:44 +0200 Topicbox-Message-UUID: 9a2c43ce-eac9-11e9-9e20-41e7f4b1d025 From: "Taj Khattra" > > i think /sys/src/9/ip/tcp.c:1256 in the mar 27 '01 pkg should be > > if((h->tcpcksum[0] || h->tcpcksum[1]) && > > instead of > > if((h->tcpcksum[0] || h->tcpcksum[0]) && > sounds reasonable IIRC. reminds me when sun turned off udp checksums and nfs requests would survive the ethernet crc but would be passed along corrupt which the udp checksum probably would have caught.