Date: Mon, 24 Dec 2007 11:38:44 GMT From: Robert Watson <rwatson@FreeBSD.org> To: Perforce Change Reviews <perforce@freebsd.org> Subject: PERFORCE change 131515 for review Message-ID: <200712241138.lBOBcifm084185@repoman.freebsd.org>
next in thread | raw e-mail | index | archive | help
http://perforce.freebsd.org/chv.cgi?CH=131515 Change 131515 by rwatson@rwatson_cinnamon on 2007/12/24 11:38:24 Update comments. Affected files ... .. //depot/projects/zcopybpf/src/contrib/libpcap/pcap-bpf.c#14 edit Differences ... ==== //depot/projects/zcopybpf/src/contrib/libpcap/pcap-bpf.c#14 (text+ko) ==== @@ -143,8 +143,11 @@ #ifdef BIOCGETBUFMODE /* - * Simple version of pcap_next_zbuf() -- do the shared memory part, but no - * blocking. + * Zero-copy BPF buffer routines to check for and acknowledge BPF data in + * shared memory buffers. + * + * pcap_next_zbuf_shm(): Check for a newly available shared memory buffer, + * and set up p->buffer and cc to reflect one if available. */ static int pcap_next_zbuf_shm(pcap_t *p, u_int *cc) @@ -178,12 +181,11 @@ } /* - * Selection routine for zero-copy BPF: identify the next completed buffer, - * if any. Try shared memory first, and if that doesn't work, make a system - * call, which may dislodge a buffer. - * - * Return (1) if the buffer is found, (0) if a retry is required, and (-1) if - * there is an unrecoverable error. + * pcap_next_zbuf() -- Similar to pcap_next_zbuf_shm(), except wait using + * select() for data or a timeout, and possibly force rotation of the buffer + * in the event we time out or are in immediate mode. Invoke the shared + * memory check before doing system calls in order to avoid doing avoidable + * work. */ static int pcap_next_zbuf(pcap_t *p, u_int *cc)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200712241138.lBOBcifm084185>