ESP8266 Community Forum

- Fri Jun 12, 2015 9:45 am #20271 Priviet,

You got further with SDK 1.1.1 than me.
My firmware didn't come so far and didn't startup user_init.

I am quite sure that I can locate WDT errors in general if I find some time but until then:

I think I remember changes in the printing of wdt errors in the SDK.
If you run the SDK 1.1.1 version with 74880 baud or maybe 115200 baud the wdt reset message
may contain additional info like epc1 etc.
That may tell us the last known location before watchdog feed got cut off.
Using the mapfile/objdump it should be straightforward to identify the code.
That may help.

Regarding SDK 1.0.1: Good work.
But pure binaries are less than this https://github.com/tuanpmt/nodemcu-official-sdk.
But as you know finding your own way is fun, so I will try it myself, too.

It would be nice if you publish your versions to compare and learn from.

Cal

Get your modules here:

https://www.banggood.com/custlink/KvGGhGF4wG

- Sat Jun 13, 2015 2:17 am #20318

I am quite sure that I can locate WDT errors in general if I find some time but until then:

I think I remember changes in the printing of wdt errors in the SDK.

In the SDK 1.1.x WDT algorithm changed (+timer) and the CPU vectors transferred to IRAM base.

call_user_start:
l32r a2, _stext
wsr.vecbase a2
call0 call_user_start_local

I did not change the code official NodeMCU.
Used only changes SDK library. NodeMCU can not work with the new SDK. Code NodeMCU is outdated.

- Sat Jun 13, 2015 2:45 am #20320 I know that sdk 1.1.1 changes the vecbase. So does the RTOS sdk and so do I.
I don't believe that your failing nodemcu image you named 1.1.1 has a changed vecbase until you give us
a way to proof ourselves.
Either document your actions or provide elf image and bins of what you named 1.1.1.
More stuff less fluff, please.
I see a lot of expertise in you disassemblings you made public.
I don't understand why you make a big thing out of simple things.

Cal

Get your modules here:

https://www.banggood.com/custlink/KvGGhGF4wG

- Sat Jun 13, 2015 3:13 am #20323

I know that sdk 1.1.1 changes the vecbase. So does the RTOS sdk and so do I.
I don't believe that your failing nodemcu image you named 1.1.1 has a changed vecbase until you give us
a way to proof ourselves.
Either document your actions or provide elf image and bins of what you named 1.1.1.
More stuff less fluff, please.
I see a lot of expertise in you disassemblings you made public.
I don't understand why you make a big thing out of simple things.

Cal

You do not have the required permissions to view the files attached to this post.