Date: Thu, 29 Mar 2001 12:13:15 -0800 From: Mike Ruhl <mruhl@network-alchemy.com> To: Jim King <jim@jimking.net> Cc: Nate Williams <nate@yogotech.com>, Paul Marquis <pmarquis@pobox.com>, Brian Matthews <blm@actzero.com>, Allen Landsidel <all@biosys.net>, freebsd-stable@FreeBSD.ORG Subject: Re: Threads vs. blocking sockets Message-ID: <3AC3975B.C2F624AF@network-alchemy.com> References: <F0D64494733BD411BB9A00D0B74A0264021C9E@cpe-24-221-167-196.ca.sprintbbd.net> <15043.35980.669828.971544@nomad.yogotech.com> <00dd01c0b886$d8510250$524c8486@jking> <3AC390FD.6C03BF8F@pobox.com> <00f201c0b889$c4ff4d30$524c8486@jking> <15043.37950.941991.251263@nomad.yogotech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------9F0D9F61BA9412F5C0A087B4
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Just two other comments on this topic...
Note: a send of 0 (zero) bytes is just as valid as a send of > 0
bytes.
Also, depending on how the socket was created, send(2) will behave
differently.
s = socket(domain, SOCK_STREAM, 0);
Does completly different things than:
s = socket(domain, SOCK_DGRAM, 0);
When doing send()s on sockets, make sure you understand the underlying
protocol before you decide that something works or does not work.
Mike
Nate Williams wrote:
>
> > None of the documented values for errno as set by send(2) indicate "only
> > part of the data was sent", at least by my reading.
>
> That's because it's not an error to send part of the data.
>
> > How do I tell how many bytes were sent?
>
> The call returns the number of characters (bytes) sent. You do know how
> much data you were sending, don't you?
>
> Here's a clue for you, since you apparently aren't getting one
> otherwise. This code probably won't even compile due to const
> poisoning, but hopefully you'll get the idea.
>
> boolean
> mySocketSend(int s, const void *msg, size_t len, int flags)
> {
> size_t totalSent;
> void *data;
>
> totalSent = 0;
> data = msg;
> do
> {
> size_t sent;
> sent = send(s, data, len - sentBytes, flags);
> if (sent < 0)
> {
> /* XXX - Handle the error here */
> return false;
> }
> else if (sent > 0)
> {
> totalSent += sent;
> data += sent;
> }
> } while (totalSent < len);
>
> /* Success, sent all the bytes */
> return true;
> }
>
> >
> > Still sounds broken to me.
> >
> > Jim
> >
> > "Paul Marquis" <pmarquis@pobox.com> wrote:
> >
> > > Under the manual entry for send(2), the RETURN VALUES section states:
> > >
> > > The call returns the number of characters sent, or -1 if an
> > > error occured.
> > >
> > > So if you request to send 100 bytes, the OS may only be able to sent
> > > half those and you need to check how many you sent an resend the part
> > > that wasn't sent.
> > >
> > > Nate is right.
> > >
> > > Jim King wrote:
> > > > Nate Williams wrote:
> > > > > Again, all threading libraries I've used (not just on FreeBSD)
> > *require*
> > > > > the user to check that when sending/receiving data, the caller must
> > make
> > > > > sure that all the expected data has been sent/received.
> > > >
> > > > The man page for send(2) doesn't mention this. It sounds broken to me.
> >
> >
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-stable" in the body of the message
--------------9F0D9F61BA9412F5C0A087B4
Content-Type: text/x-vcard; charset=us-ascii;
name="mruhl.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Mike Ruhl
Content-Disposition: attachment;
filename="mruhl.vcf"
begin:vcard
n:Ruhl;Michael J
tel;work:(831) 440-6472
x-mozilla-html:FALSE
org:Nokia
adr:;;1538 Pacific Avenue;Santa Cruz;CA.;95060;U.S.A.
version:2.1
email;internet:mruhl@cips.nokia.com
title:Tall Blond Guy
x-mozilla-cpt:;0
fn:Michael J Ruhl
end:vcard
--------------9F0D9F61BA9412F5C0A087B4--
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3AC3975B.C2F624AF>