[openbeosnetteam] Re: Missing stuff
- From: Oliver Tappe <openbeos@xxxxxxxxxxxxxxx>
- To: openbeosnetteam@xxxxxxxxxxxxx
- Date: Fri, 21 Jul 2006 16:24:38 +0200
Hi there,
On 2006-07-21 at 12:35:07 [+0200], Axel Dörfler <axeld@xxxxxxxxxxxxxxxx>
wrote:
>
> if you'll find the time to work on something in the next few days, be
> aware that there are some bits missing that you might need or want
> (mostly from bone_util):
> - a generic FIFO would be nice for collecting incoming data
> - timer functionality
> - a non-net_buffer based checksum routine
> - many things in the buffer handling you will need (like cloning,
> splitting, ...)
>
> I'll fill those in as needed, but you are allowed to do that, too, of
> course ;-)
> For raw sockets, that I will work on next, I'll need buffer cloning and
> FIFOs.
> I'll start with implementing buffer cloning, and I probably won't get
> to much more this weekend.
> (if I do, I'll let you know)
I have just grabbed UDP as a starting point for myself and when I looked at
the existing structures I have noticed that there seems to be no notion of
protocol control blocks (PCBs). While UDP can certainly get away with the
data provided by net_socket, I suppose TCP surely will need to extend the
per-socket (well, per-connection) info in order to keep connection state.
Should we introduce some kind of protocol_state structure and a pointer to
such a beast in net_socket (the pointer would usually be empty for many
protocols, but could be used for TCP)?
Or is there some other way this data can be kept that I am just unable to see?
cheers,
Oliver
- Follow-Ups:
- [openbeosnetteam] Re: Missing stuff
- From: Axel Dörfler
- References:
- [openbeosnetteam] Missing stuff
- From: Axel Dörfler
Other related posts:
- » [openbeosnetteam] Missing stuff
- » [openbeosnetteam] Re: Missing stuff
- » [openbeosnetteam] Re: Missing stuff
- [openbeosnetteam] Re: Missing stuff
- From: Axel Dörfler
- [openbeosnetteam] Missing stuff
- From: Axel Dörfler