Googling around about LWIP I got some suggestions. So as an experiment I changed
#define MEM_SIZE 4000
to
#define MEM_SIZE 0x2000
So from 4000 to 8192 - and now my problem linux client task does not stop the linux server task from getting data and completing. This means the linux client goes as fast as my network and script can go (presumably putting a load on the picow). My picow driven client running every 10 seconds can complete.
OK, so I can solve one problem by changing a magic number. The problem is there are lots of little knobs to change things inside LWIP. None of my searches gave any clarity to how to calculate a needed value for any of them.
My preferred method is to run a slower mode when resources are depleted, not just break. Ok, enough whining (or do you Brits call it whinging?). I know rpi just includes LWIP, and is not responsible for LWIP foibles.
Now, not happy having a success I tried starting a second linux client task, this one breaks in a new (non ERR_MEM) way.
#define MEM_SIZE 4000
to
#define MEM_SIZE 0x2000
So from 4000 to 8192 - and now my problem linux client task does not stop the linux server task from getting data and completing. This means the linux client goes as fast as my network and script can go (presumably putting a load on the picow). My picow driven client running every 10 seconds can complete.
OK, so I can solve one problem by changing a magic number. The problem is there are lots of little knobs to change things inside LWIP. None of my searches gave any clarity to how to calculate a needed value for any of them.
My preferred method is to run a slower mode when resources are depleted, not just break. Ok, enough whining (or do you Brits call it whinging?). I know rpi just includes LWIP, and is not responsible for LWIP foibles.
Now, not happy having a success I tried starting a second linux client task, this one breaks in a new (non ERR_MEM) way.
Statistics: Posted by nospamcalfee — Sat Jul 12, 2025 11:50 pm