Hello,
I'm using
#define LWIP_NETIF_TX_SINGLE_PBUF 1
with
#define TCP_OVERSIZE TCP_MSS
which is usually working quite well with STABLE-2_0_3_RELEASE.
However, I sometimes encounter a fragmented pbuf in the linkouput function,
wich breaks thinks with the link layer api.
There ar
On ven., oct. 20, 2017 at 09:32:14 -0300, Sergio R. Caprile wrote:
> I can comment on the ESP8266 side.
> It does not use 1.4.1 nor 2.0.3 but a custom (heavily modified) git head
> grabbed some time between 1.4.1 and 2.0 when there was a probable 1.5 to
> come, as that is what they advertise on the
> The whole point of LWIP_NETIF_TX_SINGLE_PBUF was that the stack tries to do
> its best.
> A driver should still be prepared to get chained pbufs and copy them into one
> piece.
>
> LWIP_NETIF_TX_SINGLE_PBUF should be a performance optmization to e.g. prevent
> copying
> data into a chained pb
On Fri, Oct 20, 2017 at 03:02:45PM +0200, Dirk Ziegelmeier wrote:
> Simon tried to explain you this:
>
> https://git.savannah.gnu.org/cgit/lwip.git/tree/src/include/lwip/opt.h#n1593
>
> Dirk
Thanks!
These infos are not yet in stable, so I'll have a try with master.
> The whole point of LWIP_NETIF_TX_SINGLE_PBUF was that the stack tries to do
> its best.
> A driver should still be prepared to get chained pbufs and copy them into one
> piece.
>
> LWIP_NETIF_TX_SINGLE_PBUF should be a performance optmization to e.g. prevent
> copying
> data into a chained pb
[...]
> and be aware that the ESP8266 TCP/IP is a
> proprietary patched stack, forked some time in the 1.4.1 days of yore,
> perhaps older. So... your mileage may (will) vary.
True. FWIW, esp8266/Arduino has lwIP-2.0.3 stable and unpatched if that
could help debugging. Upgrading solved a numb
Hello,
I'm in the process of integrating IPv6 in our user stack.
I'm using lwip-2.1.0RC1, and thanks for that (especially partial SACK since
last stable that saved my day this morning).
Beeing a complete beginner with IPv6, I started with the basics: ping.
I set up a link-local address by ha
Have you tried with LWIP_TCP_SACK_OUT option enabled ?
This option alone solves quite a number of issues on some low-mem devices.
To use it you would need to upgrade to lwIP-2.1.2.
I must take this opportunity to thank all lwIP devs for their work,
this is much appreciated (I'm speaking from esp
On Sat, Aug 03, 2019 at 11:55:20AM +0200, goldsi...@gmx.de wrote:
> Am 03.08.2019 um 00:48 schrieb mav:
> > Could anybody please recomend good and proven lwIP NAT solution? I've found
> > this one https://github.com/ajaybhargav/lwip_nat but it is randomly
> > crashing for me. Any suggestions?
>
