Date: Fri, 14 Apr 1995 19:43:37 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> To: julian@TFS.COM (Julian Elischer) Cc: freebsd-hackers@FreeBSD.org Subject: Just how fast can we go... (was: Re: SCSI target) Message-ID: <199504150243.TAA02395@gndrsh.aac.dev.com> In-Reply-To: <m0rzxO2-0003vnC@TFS.COM> from "Julian Elischer" at Apr 14, 95 07:15:02 pm
next in thread | previous in thread | raw e-mail | index | archive | help
[cc: trimmed to hackers, seems the best place, since we are talking
about both scsi and 100MB/sec ether :-)]
>
> >
> > >
> > Well, with 100MB/S ethernet support now being a reality TCP/IP over
> > SCSI only has an advantage for *wide* scsi controllers.
> weeeellll, no, there are advandages in being able to transfer
> 128KB of scatter-gather data in one hit with NO
> cpu intervention..... :)
A few comments form the if_de.c driver:
* The above used to be true (the stupid restriction is still true)
* but we gone to directly DMA'ing into MBUFs because with 100Mb
* cards the copying is just too much of a hit.
#define TULIP_RXDESCS 16
#define TULIP_TXDESCS 128
/*
* First find out how many and which different pages
* the mbuf data occupies. Then check to see if we
* have enough descriptor space in our transmit ring
* to actually send it.
*/
segcnt = tulip_txsegment(sc, m, addrvec,
min(ri->ri_max - 1, TULIP_MAX_TXSEG));
The 21040/21140 chips are bus master just like a scsi controller, and
can infact do some very large scatter-gather's in there own right.
--
Rod Grimes rgrimes@gndrsh.aac.dev.com
Accurate Automation Company Custom computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504150243.TAA02395>