Re: [lwip-devel] Best method to allocate space for external packets

[Jonathan Larmour]

Grubb, Jared wrote:
> Just like I am proposing for user-defined pools, we could create
> user-defined pbuf pools. This would require a few changes, but the total
> size of the changes would not be that large.

That's the thing. It would only be needed for some hardware, so if it can 
be done in a way which doesn't affect existing footprint for current 
drivers, then that would definitely be good.

Although there are quite a few systems out there that use PCI windows, or 
have alignment constraints for DMA (at a minimum alignment to a data cache 
line) so I think this improvement could be quite widely applicable.

> We have a lot of
> flexibility with the memp_std.h file, and in the latest version I posted
> a few days ago, there is a separate LWIP_PBUF_POOL macro to simplify the
> declaration of the two pbuf pools used in lwIP right now.

It's quite different indeed for the code in memp.c when the location of the 
pools themselves can be different.

> I can create a mock-up of what this would require in pbuf.c later today
> for you all to review, if you think this would be useful.

If you think that would be quicker than saying what you're thinking of then 
ok :-).

> I feel it would be useful because currently we say that PBUF_POOL is
> useful for, among other things, hardware receive buffers, but the
> PBUF_POOL is used a bit in lwIP too (SNMP and PPP come to mind).
> Allowing these to be separate pools would be useful. To preserve
> backwards compatibility, we can make it so there is only one pbuf
> mempool.

PBUF_POOL could be used by two different ethernet drivers as well too, so 
yes, some way to have different pools would be nice.

Jifl

> -----Original Message-----
> From: address@hidden
> [mailto:address@hidden On Behalf Of
> Jonathan Larmour
> Sent: Thursday, August 16, 2007 4:34 PM
> To: lwip-devel
> Subject: Re: [lwip-devel] Best method to allocate space for external
> packets
>
> Thomas Taranowski wrote:
>
> I don't think we can try and fit anything into the existing memp
> allocator 
> system, since it's inherent in the design of that that all the pools are
>
> contiguous and sequential.
>
> The simplification here is that _all_ pool pbufs would then be allocated
>
> like that. This is fine most of the time but could be more complicated
> if 
> there are multiple devices, not all of which need this system. Should we
>
> be worried about that? I don't know but from my thinking about it, it 
> would be a very hard problem to solve without breaking abstractions
> and/or 
> lots of APIs. Maybe we should leave it to the port and driver.
>
> Thoughts?
>
>
> _______________________________________________
> lwip-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/lwip-devel
>
>
> _______________________________________________
> lwip-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/lwip-devel


--
eCosCentric Limited      http://www.eCosCentric.com/     The eCos experts
Barnwell House, Barnwell Drive, Cambridge, UK.       Tel: +44 1223 245571
Registered in England and Wales: Reg No 4422071.
------["The best things in life aren't things."]------      Opinions==mine