Date: Wed, 7 Feb 2001 10:13:10 -0800 (PST) From: Luigi Rizzo <rizzo@aciri.org> To: wes@softweyr.com (Wes Peters) Cc: bright@wintelcom.net, rizzo@aciri.org, net@FreeBSD.ORG Subject: Re: send problem on udp socket... Message-ID: <200102071813.f17IDA975689@iguana.aciri.org> In-Reply-To: <3A818B31.9B12C8B3@softweyr.com> from Wes Peters at "Feb 7, 2001 10:51:45 am"
next in thread | previous in thread | raw e-mail | index | archive | help
> > ENOBUFS == ESYSADMINNEEDSTORAISENMBCLUSTERS > > Or perhaps ENOBUFS == E_SYSTEM_NEEDS_TO_RAISE_NMBCLUSTERS_ALL_ON_ITS_OWN? it is not an NMBCLUSTERS problem, it is just the device queue which is filling up, and this is a perfectly normal and desired behaviour. One would just want that to be handled as in the case of writes to a disk or the like -- i.e. wait until the subsystem (in this case the network device) is ready. For disks and TCP the synchronization is achieved through the socket buffer and (presumably for disks as well, i am no expert on filesystems) explicit acks on write ops. For UDP sockets things are unfortunately a bit more complex as there is no ack nor socket buffers. Plus there might be around a ton of software which relies on the current behaviour... cheers luigi ----------------------------------+----------------------------------------- Luigi RIZZO, luigi@iet.unipi.it . ACIRI/ICSI (on leave from Univ. di Pisa) http://www.iet.unipi.it/~luigi/ . 1947 Center St, Berkeley CA 94704 Phone: (510) 666 2927 ----------------------------------+----------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200102071813.f17IDA975689>