Just a thought: do you know that you are getting ENOBUFS from the 
driver?  Most computers today can generate output way faster than the 
devices can manage, so it's possible that you are getting the failure 
from actually taking up all the mbufs in the system.  What does 'netstat 
-m' show?  Don't forget that there really isn't any backpressure in this 
case, either from UDP or from the driver (AFAIK).

Does the socket layer even check with sowriteable() in the case of UDP?  
Last I looked, data on datagram sockets went straight to the protocol 
layer, bypassing all (socket) buffering completely.

Regards,

Justin

On Tuesday, May 14, 2002, at 05:18 PM, sudog wrote:

> On Tuesday 14 May 2002 14:20, Jonathan Stone wrote:
>>
>> You could use setsockopt() juggle the low watermark once you decide
>> you've sent "too much".
>>
>> See man 4 setsockopt, SO_SNDLOWAT, SO_RCVLOWAT.
>
> Looks like this didn't work. The following:
>
> getsockopt(mysock, SOL_SOCKET, SO_SNDLOWAT, (void *)&counter, (socklen_t
> *)&counter2);
> printf("getsockopt counter returned: %ld\n", counter);
>
> Shows that the default amount is 2048. I tried setting it to 4096
> (successfully, if the getsockopt() was any indication) and I tried 
> shrinking
> the size of the packet I was sending with sendto() but I was still 
> flooded
> with ENOBUFS.
>
> So, off to spelunk inside the kernel. Looks like there's a relevant 
> macro..
>
> in: sys/socketvar.h
>
> #define sowriteable(so) \
>         ((sbspace(&(so)->so_snd) >= (so)->so_snd.sb_lowat && \
>             (((so)->so_state&SS_ISCONNECTED) || \
>               ((so)->so_proto->pr_flags&PR_CONNREQUIRED)==0)) || \
>          ((so)->so_state & SS_CANTSENDMORE) || \
>          (so)->so_error)
>
> This is the only place I can find that references the low watermark, 
> and it
> looks like it's geared for streamed sockets.
>
> Further, it looks like that stuff is only called by soo_poll, which I 
> can't
> seem to find anywhere but attached to the poll(2) command. Let's see..
>
> I converted it to use poll(2) but the same result happened.
> So is the low water mark only used with streaming sockets then?
>
> Thanks for your help; I'll look at this some more than I get home.
>
> -Marc Tooley
>
>
--
Justin C. Walker, Curmudgeon-At-Large  *
Institute for General Semantics        |   When LuteFisk is outlawed
                                        |   Only outlaws will have
                                        |       LuteFisk
*--------------------------------------*-------------------------------*