Re: [lwip-users] lwIP hangs on transfers with many short packets

[multiplex]

Tyrel Newton wrote:
> 
> 
>> One last thing. If we take Xilinx' lastezd and greatezd SDK 12.1 - we get
>> 1.3.2 as well. Would that help, IYO?
> 
> I would provide more insight on this, but I'm currently using ISE 11.4
> with the emaclite (10/100 mac) peripheral, FreeRTOS, and the latest CVS
> head for lwIP. As I mentioned previously, I also re-wrote the lwIP driver
> provided by Xilinx because their driver was rather terrible and
> inefficient.
> 
> As I have been unimpressed with the application-level code provided by
> Xilinx, I'm betting their lwIP temac driver is not the best. The temac
> peripheral has a lot of customizable runtime settings and I'm betting the
> lwIP driver does not set them correctly or doesn't provide hooks to set
> them at all. For example, perhaps you are running out of some resource at
> the lwIP temac driver level, such as queue space for outgoing packets or
> buffer descriptors for the scatter-gather DMA stuff.
> 
> I highly doubt moving to 12.1 and 1.3.2 will fix your problem. Its very
> likely at the lwIP temac driver level, which Xilinx probably didn't update
> between releases (the 12.1 release heavily focused on support for the
> Spartan/Virtex-6 FPGAs).
> 
> Tyrel
> _______________________________________________
> lwip-users mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/lwip-users
> 
> 

Unfortunately, IIRC, emaclite does not support backtoback packets. Since we
are sitting right behind a packet processor, we do expect this to happen -
and we have to process non-TCP as well so cannot afford to drop packets at
high rates. Has this restriction changed? Would you provide your driver,
should we ask?

-- 
View this message in context: 
http://old.nabble.com/lwIP-hangs-on-transfers-with-many-short-packets-tp28607115p28703682.html
Sent from the lwip-users mailing list archive at Nabble.com.