Replies: 5 comments 1 reply
-
Not complete but see https://github.com/mikee47/esp_rgbww_firmware/tree/feature/sming5 |
Beta Was this translation helpful? Give feedback.
-
ok, wow, that's clearly not what I have expected! Thank you! I really don't know what to say. Am I right in thinking that the OTA update will have to be rewritten for the esp32(c3) SOCs? pj |
Beta Was this translation helpful? Give feedback.
-
The signed/unsigned can be converted to a warning by building with |
Beta Was this translation helpful? Give feedback.
-
so, I did fiddle some more with the partition layout (the original fw used two spiffs partitions, I don't think that's necesary, but I'll keep it for now). This is the debug output:
many seconds later, this is the next Network message:
Looking at Wireshark, the esp ACKs the get request, reporting a window of 5179 - which seems reasonable and fits the fact that small files are being served, large ones fail. After that, I get three keepalive requests from the browser that are ACK'd. After that, the server resets the connection. No data is ever received by the client. Any idea would be helpful. btw: the actual tcp code uses "debug_tcp_d" rather than "debug_d" - is there a build switch to get that activated? |
Beta Was this translation helpful? Give feedback.
-
ventured a bit into gdb. Not sure how valid this information is, due to gdb stub ram useage as well as the connection having long timed out as I'm stepping through this, but:
Line TcpConnection.cpp calls getAvialableWriteSize() and that seems to go from 2670 to 0 over four calls. I have added a debug_i to look at "available" and built without GDB and got this:
So it looks like filling up the write buffer and not freeing it (either due to a missing free or due to not sending bytes, which seems more in line with the fact that I'm not receiving bytes) is what is happening. [edit] checking tcp_write return always comes back "0 ERR_OK" but once the availableWriteSize is down to 0, obviously, tcp_write is not called anymore as there is no available buffer. [edit 2] I've tried to enable TCP_DEBUG und MEM_DEBUG in lwip, but building resets both lwipopts.h and user_config.h in /home/pjakobs/devel/esp_rgbww_firmware/Sming/Sming/Arch/Esp8266/Components/esp-open-lwip/esp-open-lwip/include/ - not sure why that would be the case. |
Beta Was this translation helpful? Give feedback.
-
dear sming community,
I have inherited a firmware project built on a rather old version of sming and am struggeling porting it forward to the current sming version.
On the way there, I have written a very rudimentary hardware pwm support fort for sming, but going through the process of understanding the differences throughout the versions is really tough as I have no previous experience with sming.
If anyone has the time and interest to help out, I'd be greatly appreciative.
here's the project: https://github.com/verybadsoldier/esp_rgbww_firmware
thanks,
pj
Beta Was this translation helpful? Give feedback.
All reactions