Messages in this thread | | | Subject | Re: PATCH: Raw device IO for 2.1.131 | From | Jes Sorensen <> | Date | 14 Dec 1998 14:10:08 +0100 |
| |
>>>>> "David" == David S Miller <davem@dm.cobaltmicro.com> writes:
David> From: Jes Sorensen <Jes.Sorensen@cern.ch> Date: 14 Dec 1998 David> 11:10:14 +0100
> SGI managed to get zero copy right for write() on a socket > and it works great, I'd love having Linux do the same.
David> Here is where I disagree with you, sendfile() is the right David> answer to pushing data to the wire, SGI's solution is not right David> and adds complexity to the kernel where it doesn't need to David> exist.
Provided that data is read directly from disk and put on the wire, yes. For the case where you have data coming from other sources than disk, this is not the solution. We have people using lotsa Linux machines to take data from the detectors on the accelerator here, using home made data aquisition cards, and spit it out on the wire using TCP. Zero Copy on write() would be a big win here with the data rates we are talking (some are talking 50-100MB/sec or more per source if I remember correctly).
Jes
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |