Messages in this thread | | | Date | Wed, 16 Dec 1998 10:37:38 +0100 (CET) | From | MOLNAR Ingo <> | Subject | Re: PATCH: Raw device IO for 2.1.131 |
| |
On Tue, 15 Dec 1998, David S. Miller wrote:
> Just one situation I want clarified, do you think we'll be able to fit > into our framework a way to (for example) get at the user's > application level headers during a write (ie. http header, then the > data)?
i can see 3 solutions:
2) if we are lucky: ie. the headers are predictable (ipv6, video) or the hardware plays well too, we can splice the data flow before the DMA is done, thus we can feed both user-visible buffers and 'content' buffers. (or in the case of vide, where frame size is predictable, the mapping can be static) This is not always possible.
1) we are smart: we put some _simple_ dataflow-changing functionality into the kernel. The ability to parse very simple and secure headers. Push anything more complex to user-space: it's going to cost time anyway, it's complex ...
(this functionality might be in the card as well...)
the problem is with the VM again: the VM has page-alignment and granularity issues, making 'data-flow splice' harder.
-- mingo
- 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/
| |